The Defining Moment: The Great Depression and the American Economy in the Twentieth Century 9780226066912

Citation preview

The Defining Moment

A National Bureau of Economic Research Project Report

The Defining Moment The Great Depression and the American Economy in the Twentieth Century

Edited by

Michael D. Bordo, Claudia Goldin, and Eugene N. White

The University of Chicago Press

Chicago and London

MICHAEL D. BORDOis professor of economics at Rutgers University and director of the Center for Monetary and Financial History at Rutgers. He is also a research associate of the National Bureau of Economic Research. CLAUDIA GOLDINis professor of economics at Harvard University and director of the Development of the American Economy program at the National Bureau of Economic Research. EUGENEN. WHITEis professor of economics at Rutgers University and a research associate of the National Bureau of Economic Research.

The University of Chicago Press, Chicago 60637 The University of Chicago Press, Ltd., London 0 1998 by the National Bureau of Economic Research All rights reserved. Published 1998 Printed in the United States of America 07060504030201009998 1 2 3 4 5 ISBN: 0-226-06589-8 (cloth)

Library of Congress Cataloging-in-Publication Data The defining moment : the Great Depression and the American economy in the twentieth century I edited by Michael D. Bordo, Claudia Goldin, and Eugene N. White. p. cm. - (National Bureau of Economic Research project report) Includes bibliographical references and index. ISBN 0-226-06589-8 (alk. paper) 1. Depressions-1929-United States. 2. United States-Economic conditions. 3. United States-Economic policy. I. Bordo, Michael D. 11. Goldin, Claudia. 111. White, Eugene Nelson, 1952- . IV. Series. HB3717 1929 .B673 1998 338.5’42-DC21 97-2975 1 CIP

@ The paper used in this publication meets the minimum requirements of the American National Standard for Information Sciences-Permanence of Paper for Printed Library Materials, ANSI 239.48-1992.

National Bureau of Economic Research Officers

John H. Biggs, chairman Carl F. Christ, vice-chairman Martin Feldstein, president and chief executive ojjicer Gerald A. Polansky, treasurer

Sam Parker, director ofjinance and corporate secretary Susan Colligan, assistant corporate secretary Deborah Mankiw, assistant corporate secretary

Directors at Large Peter C. Aldrich Elizabeth E. Bailey John H. Biggs Andrew Brimmer Carl F. Christ Don R. Conlan Kathleen B. Cooper Jean A. Crockett

George C. Eads Martin Feldstein George Hatsopoulos Karen N. Horn Lawrence R. Klein Leo Melamed Merton H. Miller Michael H. Moskow

Directors by University Appointment George Akerlof, California, Berkeley Jagdish Bhagwati, Columbia William C. Brainard, Yale Glen G. Cain, Wisconsin Franklin Fisher, Massachusetts Institute of Technology Saul H. Hymans, Michigan Marjorie B. McElroy, Duke

Robert T. Parry Peter G. Peterson Richard N. Rosett Bert Seidman Kathleen P. Utgoff Marina v. N. Whitman John 0.Wilson

Joel Mokyr, Northwestern Andrew Postlewaite, Pennsylvania Nathan Rosenberg, Stanford Harold T. Shapiro, Princeton Craig Swan, Minnesota David B. Yoffie, Harvard Arnold Zellner, Chicago

Directors by Appointment of Other Organizations Marcel Boyer, Canadian Economics Association Mark Drabenstott, American Agricultural Economics Association William C. Dunkelberg, National Association of Business Economists Richard A. Easterlin, Economic History Association Gail D. Fosler, The Conference Board A. Ronald Gallant, American Statistical Association

Robert S. Hamada, American Finance Association Rudolph A. Oswald, American Federation of Labor and Congress of Industrial Organizations Gerald A. Polansky, Americun Institute of Certified Public Accountants John J. Siegfried, American Economic Association Josh S. Weston,Committeefor Economic Development

Directors Emeriti Moses Abramovitz George T. Conklin, Jr Thomas D. Flynn

Franklin A. Lindsay Paul W. McCracken Geoffrey H. Moore

James J. O’Leary George B. Roberts Eli Shapiro

Relation of the Directors to the Work and Publications of the National Bureau of Economic Research 1. The object of the National Bureau of Economic Research is to ascertain and to present to the public important economic facts and their interpretation in a scientific and impartial manner. The Board of Directors is charged with the responsibility of ensuring that the work of the National Bureau is carried on in strict conformity with this object. 2. The President of the National Bureau shall submit to the Board of Directors, or to its Executive Committee, for their formal adoption all specific proposals for research to be instituted. 3. No research report shall be published by the National Bureau until the President has sent each member of the Board a notice that a manuscript is recommended for publication and that in the President’s opinion it is suitable for publication in accordance with the principles of the National Bureau. Such notification will include an abstract or summary of the manuscript’s content and a response form for use by those Directors who desire a copy of the manuscript for review. Each manuscript shall contain a summary drawing attention to the nature and treatment of the problem studied, the character of the data and their utilization in the report, and the main conclusions reached. 4. For each manuscript so submitted, a special committee of the Directors (including Directors Emeriti) shall be appointed by majority agreement of the President and Vice Presidents (or by the Executive Committee in case of inability to decide on the part of the President and Vice Presidents), consisting of three Directors selected as nearly as may be one from each general division of the Board. The names of the special manuscript committee shall be stated to each Director when notice of the proposed publication is submitted to him. It shall be the duty of each member of the special manuscript committee to read the manuscript. If each member of the manuscript committee signifies his approval within thirty days of the transmittal of the manuscript, the report may be published. If at the end of that period any member of the manuscript committee withholds his approval, the President shall then notify each member of the Board, requesting approval or disapproval of publication, and thirty days additional shall be granted for this purpose. The manuscript shall then not be published unless at least a majority of the entire Board who shall have voted on the proposal within the time fixed for the receipt of votes shall have approved. 5 . No manuscript may be published, though approved by each member of the special manuscript committee, until forty-five days have elapsed from the transmittal of the report in manuscript form. The interval is allowed for the receipt of any memorandum of dissent or reservation, together with a brief statement of his reasons, that any member may wish to express; and such memorandum of dissent or reservation shall be published with the manuscript if he so desires. Publication does not, however, imply that each member of the Board has read the manuscript, or that either members of the Board in general or the special committee have passed on its validity in every detail. 6. Publications of the National Bureau issued for informational purposes concerning the work of the Bureau and its staff, or issued to inform the public of activities of Bureau staff, and volumes issued as a result of various conferences involving the National Bureau shall contain a specific disclaimer noting that such publication has not passed through the normal review procedures required in this resolution. The Executive Committee of the Board is charged with review of all such publications from time to time to ensure that they do not take on the character of formal research reports of the National Bureau, requiring formal Board approval. 7. Unless otherwise determined by the Board or exempted by the terms of paragraph 6 , a copy of this resolution shall be printed in each National Bureau publication. (Resolution adopted October 25, 1926, as revised through September 30, 1974)

Contents

Acknowledgments List of Abbreviations Time Line

ix xi xv

The Defining Moment Hypothesis: 1 The Editors’ Introduction Michael D. Bordo, Claudia Goldin, and Eugene N. White I. THEBIRTHOF ACTIVIST MACROECONOMIC POLICY

1. Was the Great Depression a Watershed for American Monetary Policy? Charles W. Calomiris and David C. Wheelock

23

2. Fiscal Policy in the Shadow of the Great Depression 67 J. Bradford De Long

3. The Legacy of Deposit Insurance: The Growth, Spread, and Cost of Insuring Financial Intermediaries Eugene N. White

87

11. EXPANDING GOVERNMENT

4. By Way of Analogy: The Expansion of the Federal Government in the 1930s 125 Hugh Rockoff vii

viii

Contents 5. The Impact of the New Deal on

American Federalism John Joseph Wallis and Wallace E. Oates

155

6. The Great Depression and the Regulating State:

Federal Government Regulation of Agriculture, 1884-1970 Gary D. Libecap

181

111. INSURING HOUSEHOLDS AND WORKERS

7. A Distinctive System: Origins and Impact of U.S. Unemployment Compensation

221

Katherine Baicker, Claudia Goldin, and Lawrence E Katz

8. Spurts in Union Growth: Defining Moments and Social Processes Richard B. Freeman 9. The Genesis and Evolution of Social Security Jeffrey A. Miron and David N. Weil

265 297

IV. INTERNATIONAL PERSPECTIVES 10. From Smoot-Hawley to Reciprocal Trade Agreements: Changing the Course of U.S. Trade Policy in the 1930s Douglas A. Irwin 11. The Great Depression as a Watershed: International Capital Mobility over the Long Run Maurice Obstfeld and Alan M. Taylor

325

353

12. Implications of the Great Depression for

the Development of the International Monetary System Michael D. Bordo and Barry Eichengreen

403

Contributors

455

Name Index

457

Subject Index

465

Acknowledgments

The preconference, held 29 March 1996 in Cambridge, Mass., at the NBER, and the conference, held 11-12 October 1996 at Kiawah Island, S.C., were sponsored by the National Science foundation (SBR95-11481) and the National Bureau of Economic Research. The coeditors thank both these organizations for their generous financial support. The coeditors also thank the discussants, Anna J. Schwartz (Calomiris and Wheelock), N. Gregory Mankiw (De Long), Lawrence J. White (White), Stanley L. Engerman (Rockoff), Robert A. Margo (Wallis and Oates), Sam Peltzman (Libecap), Bruce D. Meyer (Baicker, Goldin, and Katz), Alan B. Krueger (Freeman), Dora L. Costa (Miron and Weil), Robert E. Baldwin (Irwin), Lance E. Davis (Obstfeld and Taylor), James M. Boughton (Bordo and Eichengreen), and panelists Anna J. Schwartz and Peter Temin.

ix

This Page Intentionally Left Blank

Abbreviations

Editors’ Note: We have included in this list of abbreviations only those that bear directly on the subject of the volume. Therefore, commonly used abbreviations (e.g., GNP, CPI, PAC) are not listed.

AAA ACSS ACUC ACW ADC AFDC AFL AMEX ASCS BIS BLS

ccc CEA CED CES CFA

CIO CPS DIDMCA EEC EMS EPU xi

Agricultural Adjustment Administration (or Act) Advisory Council on Social Security Advisory Council on Unemployment Compensation Amalgamated Clothing Workers aid to dependent children aid to familes with dependent children American Federation of Labor American Stock Exchange Agricultural Stabilization and Conservation Service Bank for International Settlements Bureau of Labor Statistics Commodity Credit Corporation (Note: Civilian Conservation Corps is not abbreviated) Council of Economic Advisers Committee on Economic Development Committee on Economic Security Communautg Financikre Africaine (African Financial Community) Congress of Industrial Organizations Current Population Survey Depository Institutions Deregulation and Monetary Control Act European Economic Community European Monetary System European Payments Union

xii

Abbreviations

ERAA ESA FDIC FDICIA FERA FHLB FHLBB FOMC FSLIC GATT IBEW ICU ILGWU IMF IT0 MFN NAFTA NASD NCUA NCUSIF NEA NIPA NIRA NIRA NLRB NPEB NRA NYSE OAA OAI OASI

OECD PBGC PIK PPP RTAA SDR SEC SIPA SIPC SSA SSI SWOC TVA

Emergency Relief Appropriations Act Economic Security Act Federal Deposit Insurance Corporation Federal Deposit Insurance Corporation Improvement Act Federal Emergency Relief Administration Federal Home Loan Bank Federal Home Loan Bank Board Federal Open Market Committee Federal Savings and Loan Insurance Corporation General Agreement on Tariffs and Trade International Brotherhood of Electrical Workers International Clearing Union (never created) International Ladies’ Garment Workers’ Union International Monetary Fund International Trade Organization most favored nation (trade policy) North American Free Trade Association National Association of Securities Dealers National Credit Union Administration National Credit Union Share Insurance Fund National Education Association national income and product accounts National Industrial Recovery Act National Industrial Relations Act (in Freeman chapter only) National Labor Relations Board National Production and Employment Budget National Recovery Administration New York Stock Exchange old-age assistance old-age insurance old-age and survivors insurance Organization for Economic Cooperation and Development Pension Benefit Guaranty Corporation payments in kind purchasing power parity Reciprocal Trade Agreements Act special drawing right Securities and Exchange Commission Securities Investor Protection Act Securities Investor Protection Corporation Social Security Act supplemental security income Steelworkers Organizing Committee Tennessee Valley Authority

xiii

UAW UE UFCW UI UIP UMW USDA USTC

wc

WPA

Abbreviations United Auto Workers United Electrical Workers United Food and Commercial Workers unemployment insurance uncovered interest parity United Mine Workers U.S. Department of Agriculture U.S. Tariff Corporation workers’ compensation Works Progress (later Projects) Administration

This Page Intentionally Left Blank

Time Line

Editors’ Note: The time line includes only those events and dates mentioned

in the volume. 1914-1 8 1925 1929 1930 1931 1932 1933

April August October June December March-June September February November January-March March March-June March April May

June

xv

World War I Gold exchange standard began Great Depression began Stock market crash Smoot-Hawley Tariff Act First banking crisis Second banking crisis Britain left the gold standard Reconstruction Finance Corporation Glass-Steagall Act of 1932 Roosevelt elected president for first term Third banking crisis Roosevelt sworn in as president First Hundred Days of the New Deal Emergency Banking Act Civilian Conservation Corps and Reforestation Relief Act United States left the gold standard Federal Emergency Relief Act Agricultural Adjustment Act of 1933 Tennessee Valley Authority Act Federal Securities Act National Industrial Recovery Act National Employment System Act

xvi

TimeLine

October November 1934

January June

1935

January April

July August 1936

January February June

1937

September February April August September

Banking Act of 1933 or “Glass-SteagallAct,” set up Federal Deposit Insurance Corporation Commodity Credit Corporation, by executive order Civil Works Administration, by executive order Gold Reserve Act United States devalued the dollar (from $20.67 to $35 per ounce) Reciprocal Trade Agreements Act Securities and Exchange Commission National Housing Act, created Federal Savings and Loan Insurance Corporation Silver Purchase Act Committee on National Security issued its report Roosevelt’s State of the Union Address Emergency Relief and Appropriations Act (created the Works Progress Administration and the Rural Electrification Administration, both in May 1935) Schechter Poultry Corporation et al. v. United States, National Recovery Administration ruled unconstitutional National Labor Relations (Wagner) Act, set up National Labor Relations Board Banking Act of 1935 Social Security Act United States v. Butlei; declared key provisions of the Agricultural Adjustment Act unconstitutional Soil Conservation Act and Domestic Allotment Act replaced the Agricultural Adjustment Act Walsh-Healey Public Contracts Act Robinson-Patman Act France abandoned gold convertibility Court-packing bill sent to Congress National Labor Relations Board v. Jones and Laughlin, upheld National Labor Relations Board Miller-Tydings Act Farm Security Administration

xvii

1938 1939 1941 1944 1945 1946 1947 1948 1950 1951 1952 1956 1958 1965 1966 1970 1971 1973 1974 1980 1989

Time Line

February June August September December

Agricultural Adjustment Act of 1938 Fair Labor Standards Act Social Security Act of 1939 World War I1 began United States entered World War I1 Bretton Woods Agreement World War I1 ended; Anglo-American loan; cold war began International Monetary Fund began operations Employment Act of 1946 Formation of the General Agreements on Tariffs and Trade Taft-Hartley Act, modified 1935 Wagner Act Marshall Plan began Federal Deposit Insurance Act of 1950 Treasury-Federal Reserve Accord Marshall Plan ended National Defense Highway Act European countries achieved current account convertibility U S . dollar entered crisis zone; France began converting dollars into gold Financial Institutions Supervisory and Insurance Act of 1966 Securities Investor Protection Act National Credit Union Share Insurance Fund United States closed gold window Managed float began Employee Retirement Income Security Act Depository Institutions Deregulation and Monetary Control Act Financial Institutions Reform, Recovery and Enforcement Act; Federal Savings and Loan Insurance Corporation eliminated

This Page Intentionally Left Blank

The Defining Moment Hypothesis: The Editors’ Introduction Michael D. Bordo, Claudia Goldin, and Eugene N. White

There is a widely held belief that the Great Depression was the “defining moment” in the development of the American economy. According to this view, the severity and length of the depression altered the basic rules, institutions, and attitudes governing the economy. Bolstering this perception is well-known evidence that the growth of govemment as a share of GNP accelerated in the 193Os, although some of the increase was due to the collapse of GNP (fig. 1). Still, government as a share of GNP remained high even as the economy began to recover. Not only did the relative size of government expand, but the relationship between the federal govemment and state and local governments was also irrevocably altered in the mid1930s.A striking change occurred around 1935 when the federal government, as a fraction of all government expenditures, grew largely at the expense of the localities. This change was independent of the growth of defense, international relations, and debt servicing and is apparent if intergovernmental grants are attributed to either the granting or receiving government (fig. 2 ) . Added to these time-series facts about the growth of government is that the most important social programs today originated in the 1930s. Social security, old-age assistance, welfare, and unemployment insurance were all part of the same bill passed in 1935. Regulation of agriculture, banking, and finance was vastly expanded, and organized labor was finally awarded its “bill of rights,” all as part of the New Deal. The notion that the government was responsible for the health of the American economy was codified, at the end of World War Michael D. Bordo is professor of economics at Rutgers University and director of the Center for Monetary and Financial History at Rutgers. He is also a research associate of the National Bureau of Economic Research. Claudia Goldin is professor of economics at Harvard University and director of the Development of the American Economy program at the National Bureau of Economic Research. Eugene N. White is professor of economics at Rutgers University and a research associate of the National Bureau of Economic Research.

2

Michael D. Bordo, Claudia Goldin, and Eugene N. White

50 c

0

3

-0

2

5

(Government Expenditures)/GNP 40

(TI

K

0

z

(TI

tn

E

v)

c

0 a,

ul m c

30

20

c

a,

g

a

10 w-

0

LJ

,

1900

1910

1920

1930

1940

1950

1960

,

/

1970

,

I

.

1980

,

1990

Year

Fig. 1 Government (national, state, and local) as a share of GNP, 1902-84

Source: Higgs (1987, table 2.1). Notes: Government expenditures include all spending by the three levels of government, whereas government purchases of goods and services exclude transfer payments. In all cases, intergovernmental grants are excluded to avoid double-counting.Thus, the growth in government, as a share of GNF’, since the late 1950s has almost all come from an increase in transfer payments such as social security. The National Income and Product Account revisions, which extend back to the late 1950s, make government expenditure data for the most recent period noncomparable with historical data and lower the share of GNF’ that is government by reducing expenditures by state and local governments.

11, in the Employment Act of 1946. There is a very good case to be made, or so it would appear, for the idea that the 1930s was the “defining moment” in twentieth-century U.S. economic history.

The “De6ning Moment” Hypothesis The argument that the Great Depression was a watershed considers the protracted economic crisis as inducing fundamental change in the relationship between government and the private sector. But many of the innovations embraced in the 1930s-most of which were part of the Roosevelt administration’s New Deal-had been under consideration for some time, at both the state and national levels. Some important reforms had been passed decades earlier in other industrialized countries. It could thus be argued that change, already proceeding, was simply accelerated by the economic collapse.’ In 1. Hughes (1977, 146), e.g., views the 1930s as a continuation of previous trends regarding government and the economy, possibly with some acceleration, but not a break.

A

*2 3

70

-

se + Debt Financing)

60-

CI .-

-0 C

0

Q X

W

rc

5040

-

0

Q) 0) 3 0 -

(0

+-

c

8 t

n

2010 -

0"

'

1900

'

'

1910

'

'

'

1920

'

1930

'

'

'

'

1940

'

1950

'

'

1960

'

1970

I980

1990

2000

Year 70

3

L

Localities 60

3

c .W ,=

W

Federal - (Defense + Debt Financing)

\\

50

0. X .k-

0

W

c

8

a

40

30 20 10

0

,

1900

1910

1920

.

I

1930

,

t

1940

,

t

1950

.

/

1960

.

I

1970

.

I

1980

.

I

1990

,

I

2000

Year Fig. 2 Share of government expenditures by (A)granting and (B) receiving government, 1902-92

Source: Wallis and Oates (chap. 5 in this volume, figs. 5.1 and 5.2). Notes: Shares for all three levels of government sum (vertically) to 100 percent for each year. Expenditures on national defense, international relations, and interest on the government debt have been excluded from federal expenditures. In panel A, all intergovernmental grants have been attributed to the granting ( i t . , revenue-raising) government. In panel B, all intergovernmental grants have been attributed to the receiving ( i t . , spending) government. Thus, the "granting" graph makes the federal government appear relatively larger than does the "receiving" graph.

4

Michael D. Bordo, Claudia Goldin, and Eugene N. White

some cases, the acceleration may have had real effects in altering the institutional details of legislation, features that remained in place for decades to come. But many aspects of the New Deal have not endured, and one question is why some became permanent fixtures whereas others did not. Some scholars point to World War I as paving the way for many of the economic changes of the 1930s. For example, certain New Deal regulatory policies may have been accepted with greater alacrity because the government takeover of the railroads in World War I fostered the belief that government could succeed when private enterprise did not. The income tax, inaugurated in 1913, was so greatly increased in 1917, as was the taxation of the very rich, that the new system has been termed “the most significant domestic initiative to emerge from the war” (Brownlee 1996,48).Agricultural price supports originated in World War I, although their function was to be altered by the New Deal. The War Finance Corporation was the model for the Reconstruction Finance Corporation, set up by President Hoover in 1932 and then used extensively by President Roosevelt in support of the New Deal. And wage and price controls during World War I, it has been claimed, set the stage for the New Deal’s ill-fated National Industrial Recovery Act. But if World War I provided the opening wedge for many New Deal programs, World War I1 may have cemented them in place, solidifying the notion that big government was crucial to the health of the economy and democracy. Figure 1 even appears to suggest that the cold war may have been the real factor that advanced government purchases as a fraction of GNP. The contributors to this volume are aware that the defining moment hypothesis is a complicated one. The answer may involve understanding what the New Deal would have been in the absence of World War I, as well as what the New Deal’s legacy would have been without World War I1 and the cold war.2 The New Deal was not one but at least two programs. The “second” New Deal gained from the lessons of the “first.” Had the Great Depression ended, say, before 1935, the legacy of the New Deal would have been far different. Parts of the National Industrial Recovery Act, struck down by the Supreme Court in 1935, lived a second life in subsequent legislation. Sections were resurrected by the National Labor Relations (Wagner) Act in 1935, the Robinson-Patman Act of 1936 (which made price discrimination illegal), the Miller-Tydings Act of 1937 (which exempted resale price maintenance contracts from antitrust laws), and the Fair Labor Standards Act in 1938 (which set hours and wage standards). The legislation of the post-1934 New Deal was often drafted to pass the test of a Supreme Court that had already nullified many of Roosevelt’s key programs. This special crafting altered the form of many institutional details. Most difficult of all the aspects of the defining moment hypothesis is how the Great Depression affected the perceptions of the American people about the role of government. We often read that American faith in institutions is 2. On the notion that almost all crises increase the size of government, see Higgs (1987).

5

The Defining Moment Hypothesis: The Editors’ Introduction

today at a historic low point. Yet, government expenditures as a fraction of GNP have never been greater (fig. l).3Although there is marked distrust, there is also, today, a general acceptance of some governmental role in public goods provision, social insurance, regulation, the completion of various markets, the internalization of externalities, and even outright redistribution. Did the Great Depression alter the public’s view concerning the functions of government, particularly those at the national level? It would certainly appear that it did. Even though New Deal programs were politicized, the 1930smay have been a turning point in the extrication of government from control by the political parties. Prior to the 1930s the effect that government had on the public was often felt through the impact political parties had on people’s lives. In the nation’s cities, for example, government was identified with political machines and operated through the favors of city bosses. Patronage appointments at the state level and, before the federal civil service, at the national level were other routes through which the public received government largesse. The Union Army pension, the largest single social program prior to the New Deal, was clearly identified with the Republican Party. With the New Deal, however, social programs began to exist independently of political involvement and party affiliation. Whatever the resolution of these difficult issues, one thing does seem clear. The increase in the size and scope of business and the greater proportion of Americans living in cities and working as wage earners gave rise to the notion that government could function better if it, too, were larger in size and more centralized in scope. Workers, no longer largely self-employed, were less personally responsible for the hardships that befell them, such as unemployment. More complex technology at home and in the workplace gave government an expanded regulatory role, and the greater density of population heightened the need for government to contain various “spillovers.” But because local and state governments could not effectively regulate or tax mobile capital (and labor), the national government was destined to grow. States and localities required a coordinating mechanism, a role that could best be served by the national government. The “call” for increased government and for greater scope of government built up gradually throughout the nineteenth ~ e n t u r yImportant .~ debates continue to rage concerning whether the call for increased government came from 3. Although the last date in fig. 1 is 1984 (see note to fig. 1 concerning the reasons why), recent data (not consistent with pre-1960 data) do not show a decline from the early 1980s. 4.See, e.g., Hofstadter (1955) on the populist and progressive roots of the New Deal. Whereas the New Deal was a reaction to economic collapse, populism and progressivism were reforms directed at the bigness of enterprise and the corruption of state and local governments. Populism and progressivism were liberal in the sense that they were concerned with the protection of individual rights, but Hofstadter reminds his readers that reform was often reactionary and ambiguous. Populism, in particular, embraced many conservative elements such as nativism. Patterson (1969), however, argues that progressivism was essentially moribund on the eve of the Great Depression. Whatever the states did accomplish by 1933 in reaction to the economic crisis, and it was precious little, served as a reminder of how much had changed since the Progressive Era and how important the national government had become.

6

Michael D. Bordo, Claudia Goldin, and Eugene N. White

the right or the left, from business or labor, from big business or small busin e ~ sFor . ~ us, the key issue is whether, in the midst of this long-run process of economic change, the depression altered the form and function of government and, if so, through what route. To examine the legacy of the Great Depression at the end of the twentieth century, a conference was organized and held in October 1996. Each of the 12 conference papers posed the question whether the depression decade of the 1930s was a defining moment for various aspects of economic policy and for particular sectors of the economy. Although the manner in which the topics were selected could have influenced the conclusions, we do not believe that to be the case. We began by identifying the four major sections of the volume: macroeconomic policy, government, social insurance and labor, and the world economy. Under each heading we chose topics considered to be among the most significant and for which we could identify a scholar to produce the chapter. As a group, we were never tied to a specific view on the subject (as our rather diverse contributions will attest). The majority of the papers found support for the idea of a sharp break with the past, although some major changes represented accelerations of earlier developments. Several common themes emerge from the papers that explain how the depression influenced the U.S. economy’s subsequent development. First, skepticism about the efficacy of government intervention withered as the public adopted the attitude that the government could “get the job done” if the free market did not. Business, however, was not always as enthusiastic. Although there is evidence that parts of the intellectual community were already disposed toward a bigger role for government, the depression convinced many doubters. Second, many innovations introduced by the New Deal were forms of social insurance. These new programs survived and often grew far beyond the intentions of their progenitors because they created groups of readily identifiable and organized beneficiaries while costs were diffused among the general public. Third, the character of federalism moved from “coordinate”to “COoperative” with extensive intergovernmental grants, giving greater influence to centralized government. Last, the conduct of economic policy-both domestic and international-changed to give more weight to employment targets and less to a stable price level and exchange rate. These developments became key features of the economy, shaping the course of its growth over the remainder of the century.

Origins and Persistence of the Depression The papers in this volume are all concerned with the legacy of the Great Depression, not its immediate impact or its causes. But the changes it induced 5 . For two novel interpretations, see Koko (1967) on the conservative origins of government regulation and Libecap (1992) on the role of rent-seeking small business in the passage of the Meat Inspection Act and the Sherman Antitrust Act.

7

The Defining Moment Hypothesis: The Editors’ Introduction

10,000 5,000 3,000

2,000 1,000

500 300 200 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 Fig. 3 Real GNP (billion 1990 dollars), 1900-1995 Sources: Nominal GNP and GNP deflator: 1900-1938, Balke and Gordon (1986, 781-83); 193958, U.S. Council of Economic Advisors (1991, tables B-1 and B-4); 1959-95, U.S. Council of Economic Advisers (1997, tables B-4 and B-24). Notes: Real GNP is graphed in (natural) log form. The GNP deflator is spliced so that 1990 = 100. Vertical lines delineate the period of the Great Depression.

were consequences, in no small part, of its magnitude and perceived origins. It is, therefore, important to recall the dimensions of the depression that earned the appellation “great” as well as the literature on its origins. In terms of its impact on economic performance, the depression was a disaster without equal in the twentieth century. The contraction phase of the depression, extending from August 1929 to March 1933, saw the most severe decline in key economic aggregates in the annals of U.S. business cycle history. Real GNP fell by more than one-third (fig. 3), as did the price level (fig. 4). Industrial production declined by more than 50 percent. UQemployment rose to 25 percent by 1933 (fig. 5 ) . Not only was the descent precipitous, but the recovery from the business cycle trough was slow. The economy did not regain its 1929 GNP level until 1939. Interpretations of the causes of the Great Depression have shifted radically over time. Businessmen, and especially bankers and financiers, were initially blamed for the collapse. Congressional hearings on the stock market crash of 1929 and the banking crises sought to identify specific practices and accountable individuals. The Securities Act of 1933, the Securities Exchange Act of 1934, and the Banking Acts of 1933, 1934, and 1935 endeavored to forbid reckless financial activities and to constrain institutions and markets to be more prudent. For some New Dealers, the operation of individual markets failed. Government intervention was needed to assist agriculture through the Agricul-

Michael D. Bordo, Claudia Goldin, and Eugene N. White

8

200 100

5 ........................... 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 1 1 1 , 1 " 1 , 1 , ' 1 ' , " " " 1 1 , 1 1 1 1 ' , ' ' , , ~ , , 1 , , , ' ' , , ~ ' 1 ~ ' ~ ' i

Fig. 4 Consumer price index, 1900-1995 (log scale; 1990 = 100)

Sources: 1990-1970, U.S. Bureau of the Census (1975, series E-135); 1971-95, U.S. Council of Economic Advisers (1996, table B-56). Notes: The series have been transformed so that 1990 = 100. The consumer price index is graphed in (natural) log form. Vertical lines delineate the period of the Great Depression.

tural Adjustment Act and industry through the National Industrial Recovery Act. Centralized planning and direction seemed necessary to guide the functioning of markets, stabilize prices, and increase employment. Confronted by the failure of the economy to recover, economists of the 1930s sought explanations for the causes of the Great Depression in John Maynard Keynes's The General Theory of Employment, Interest and Money (1936). These economists emphasized the collapse of investment produced by the end of the frontier, lower population growth from immigration restriction, the drop in residential construction, reduced wealth from the stock market crash, and the collapse of foreign trade following the Smoot-Hawley Tariff Act of 1930.6 Most economists today agree that the depression was primarily a consequence of both domestic and international monetary forces, although some emphasize the real forces of demography, structural change in the economy, and technology The monetary explanation for the Great Depression in the United States focuses on the money supply (M2), which declined by more than 33 percent between August 1929 and March 1933. The key source of the decline was a series of banking panics that led to the closing of more 6. On the Keynesian approach, see Temin (1976). 7. The literature on the causes of the Great Depression is vast. On the subject of domestic and international monetary forces, see, e.g., Bordo (1986, 1989), Eichengreen (1992). Romer (1993). and Temin (1993). On real-side factors, see, e.g., Temin (1989).

9

The Defining Moment Hypothesis: The Editors’ Introduction

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

Year Fig. 5 Unemployment rate, 1900-1995 Sources: 1900-1970, U.S. Bureau of the Census (1975, series D-86); 1971-95, U.S. Council of Economic Advisers (1996, table B-39). Notes: Hatched area delineatesthe period of the Great Depression. Unemployment refers to people 14 years old or older prior to 1947 and 16 years old or older afterward.

than one-third of the nation’s banks in less than four years. The monetary collapse produced deflation and falling real output in the face of nominal rigidities.8 Friedman and Schwartz (1963) attribute monetary collapse and the banking panics to ineptitude by Federal Reserve officials who, according to their view, were unwilling and seemingly unable to use well-known tools of monetary policy to prevent the banking panics and to reverse the decline in money supply.9 In addition to domestic monetary forces, a recent revisionist view sees the gold standard as the cause of the Great Depression and its international propagator (see, e.g., Bernanke 1995; Eichengreen 1992; Temin 1989). According to this explanation, adherence to gold standard orthodoxy encouraged the monetary authorities of major countries, including the United States, to follow deflationary policies in the face of external shocks. Fixed exchange rates trans8. Alternative mechanisms by which monetary collapse may have led to depression include debt deflation (Fisher 1933) and credit disintermediation(Bernanke 1993). 9. Friedman and Schwartz (1963) explain the policy failure by institutional paralysis arising from a division between the Federal Reserve Board and the Reserve Banks, as well as the death of Benjamin Strong, the influential president of the Federal Reserve Bank of New York. For an alternative view-that the Federal Reserve followed a flawed policy strategy based on the real bills doctrine-see Meltzer (1995), Wheelock (1991). and Wicker (1965).

10

Michael D. Bordo, Claudia Goldin, and Eugene N. White

mitted the shock internationally, while maintenance of convertibility prevented the use of reflationary monetary policy domestically. Countries could extricate themselves from the depression only by cutting the link with gold (as Britain did in 1931 and the United States in 1933), devaluing their currencies, and following expansionary monetary and fiscal policies. l o The recovery that began in spring 1933, it is generally believed, reflected a series of Treasury-sponsored reflationary policies including the devaluation of the dollar in 1934 and expansionary fiscal policy and gold purchases (Friedman and Schwartz 1963; Romer 1992; De Long, chap. 2 in this volume). Much concerning the origins, propagation, and persistence of the Great Depression is still debated. One feature, and possible consequence, of the slow and painful recovery that no one appears to contest was the growth of government. Were the 1930s a Defining Moment for the U.S. Economy? The Role of Government The growth of large-scale government with a predilection for intervention is usually traced to the New Deal. Rockoff‘s chapter examines how the Great Depression produced a truly “great” expansion in the economic role of the federal government and created a propensity for further growth in subsequent decades. The transformation was a product of the depth and persistence of the unparalleled economic collapse. The federal government’s share of civilian employment and its spending as a fraction of GNP both appear to have been permanently doubled by the depression, with the depression also initiating a steady growth in government’s share of GNP.I1The change was not merely one of size but also one of scope, with the government participating in many new areas. New Deal agencies continued to grow after World War 11, and new agencies were formed to regulate more sectors of the economy. One critical ingredient of this revolution was the decisive shift in public opinion about the appropriate role of government. The apparent failure of capitalism conditioned the public to accept proposed government intervention in markets. Farmers who lost their property, depositors turned away from their banks, the elderly with no pensions, and workers without jobs made the new government programs seem sensible. The question was no longer whether the government should intervene but why it should not. 10. There is debate on these points as well. Although the gold standard was the likely transmitter around the world of the deflationary shock from the United States, some argue that the Federal Reserve had sufficient gold reserves to follow expansionary policies without being hampered by “golden fetters” (Meltzer 1995). 11. There is an obvious difficulty in making such statements because GNP was enormously depressed in the 1930s. In addition, the federal government grew at the expense of other levels of government. On both these issues, see Rockoff (chap. 4 in this volume) and Wallis and Oates (chap. 5 in this volume).

11

The Defining Moment Hypothesis: The Editors’ Introduction

And the question changed quickly. Given the limited role of government before 1929, the ability of the federal government to rapidly develop and deploy a wide variety of economic programs in response to the depression is striking. Rockoff shows that this rapid response was made possible by an ideological shift among economists from laissez-faire to interventionism that preceded the Great Depression by at least a decade. Only when its attention was focused by the calamity was the public ready to listen. “The New Deal,” writes Rockoff, “was just what the doctors (of economics) ordered and. . . they believed their advice was soundly based on clinical evidence” (Rockoff, chap. 4 in this volume). The overwhelming majority of articles analyzing New Dealtype programs in major economic journals between World War I and 1929 were favorable. While the economic collapse galvanized even University of Chicago economists behind public works programs, most economists, especially microeconomists, had long studied and found merit in government intervention to correct what they judged to be market failures and deficiencies. Proposals for a minimum wage, social security, unemployment compensation, public ownership, public works, securities regulation, and deposit insurance were already on the table. Having studied reforms instituted in Europe, Canada, Australia, and at the state level in the United States, economists generally found favorable, supporting empirical evidence. The experts were convinced of the need for more government intervention, but it took the depression to damage the public’s strong ideological bias against it. Once established, the public’s predisposition toward intervention endured for several decades. Beginning with the stagflation of the 1970s, skepticism about government intervention began to reappear. A shift in public opinion, like that in the 1930s, was preceded by a shift in opinion among economists. The 1930s was a defining moment in the conception of government’s role. Even though public and professional opinion has continued to evolve, the depression created an environment of opinion that allowed the establishment of many lasting institutions. A New Federalism Although the size and scope of government intervention grew throughout the industrialized world in response to the depression, the American experience was strongly conditioned by its federal political structure. Wallis and Oates show that the depression and New Deal accelerated the move from “coordinate” or “dual” federalism to “cooperative” federalism. Prior to the economic collapse, states and localities had operated with relative independence from the national government. The depression made the national government dominant. It grew mainly at the expense of local, not state, governments (see fig. 2 ) . The national government’s influence was made even greater during the New Deal because much of its spending took the form of intergovernmental grants to states and localities. Today, federal grants no longer have economic expansion as their objective. Rather they provide fiscal support for specific projects

12

Michael D. Bordo, Claudia Goldin, and Eugene N. White

or for income maintenance. These grants, by the estimates of various economists, have a highly stimulative-what is known as a “flypaper”-effect on recipient governments, thus serving to increase the overall size of government. Whereas some New Deal programs were uniform nationally, many had a truly federal character, allowing states authority in decision making when they accepted matching grants from Washington. This characteristic, which had few historical precedents, made the programs more palatable politically and helped ensure that they would become permanent fixtures by creating support not only among the programs’ ultimate beneficiaries but also among the administering state and local governments. During the 1960s and 1970s there was an explosion of social welfare programs; some were expansions of those begun during the New Deal. Others (like the War on Poverty) extended government in new directions, but not ones that would have been foreign to many New Dealers’ agendas. Beginning in the 1970s, and especially in the 1980s, Congress and the Republican administration gutted the War on Poverty. Yet many New Deal and New Deal-inspired programs remain part of the national landscape. Thus, for fiscal federalism, the 1930s were clearly a defining moment. Protection and Insurance for Industry, Agriculture, and Banking The Great Depression and the New Deal had profound effects on particular sectors. New Deal legislation altered the balance of power between special interest groups, and it locked in change in certain sectors. Innovations occurred in agriculture and banking. In those sectors, the New Deal increased the level and altered the character of intervention by providing new safety nets for farmers and for bank depositors. But this intervention stands in contrast to the decisive movement, at the same time, away from protectionism in U.S. tariff policy. In all three cases, the 1930s were a defining moment, but for different reasons. Federal agricultural policy during the New Deal, according to Libecap, shifted from the provision of public goods and transfers on a limited basis to programs designed to raise prices and support farm incomes by controlling supply and enhancing demand. Although some of these programs find precedent in World War I policies, the intent of government purchases then was to aid the war effort, not raise prices. But farmers could not help noticing the influence government had on agricultural prices during the Great War, and in the 1920s, they lobbied for relief. Yet little came of their efforts. The major legislative emphasis was through the McNary-Haugen bills of the 1920s, which offered modest support for agricultural prices. These bills, however, were seen as special interest legislation and were not enacted into law. The more extensive and far-reaching farm crisis of the 1930s, with its plummeting incomes and soaring foreclosures, altered the perception of protective farm programs. Agricultural legislation eventually covered virtually all domestically produced commodities. The broad safety net and array of services offered to farmers helped to create a powerful and politically active farm lobby. In banking, the New Deal created a corset of regulations that reduced corn-

13

The Defining Moment Hypothesis: The Editors’ Introduction

petition and insured deposits, protecting both depositors and bankers. Much of the legislation-notably deposit insurance-had been proposed before the depression, but it took the banking collapse of 1930-33 to alter the perception that deposit insurance was simply special interest legislation. Federalism, coupled with nineteenth-century banking legislation, had long before created a system of distinctly separate financial intermediaries numbering in the thousands. These, in turn, spawned interest groups concerned with preserving the structure of this system. When the New Dealers considered remedies for the collapse of the banking system, they did not attempt to alter the basic structure of the banking and financial system. Instead, they introduced anticompetitive legislation, such as controlled entry, prices, and products, and they instituted deposit insurance, all to preserve the status quo ante. Other countries also enacted various types of anticompetitive legislation to assist their weakened financial systems during the depression, but deposit insurance was a uniquely American invention designed to protect the small banks that had proliferated because of our federal structure and a host of pre-1930s regulations. The U.S. public’s general approval of insurance and its apparent success here made the idea of deposit insurance an exportable policy remedy in the 1960s, despite its specifically U.S. origin. White shows in his chapter that banks, beginning in the 1950s, pressed Congress to increase the level of deposit insurance. Other financial intermediaries, concerned about the competitive advantages bestowed by deposit insurance, followed their lead and convinced Congress to broaden the insurance of financial liabilities, far beyond the initial boundaries conceived by the New Dealers. The introduction of new legislation-more deposit insurance in 1950, 1966, 1969, 1974, and 1980 plus insurance of thrifts, pension funds, and brokerage accounts in the 1970s-rarely occasioned any outcry because the public was conditioned by experience to regard the insurance of deposits and other liabilities as an appropriate task for government. The financial structure generated by the New Deal and subsequent, complementary regulation allowed the macroeconomic shocks of the 1970s and 1980s to wreak havoc on savings and loan associations and banks. In their aftermath, most anticompetitive regulation was swept away. What remain are the policies that provide protection to concentrated groups of beneficiaries-banks and other financial intermediaries-while spreading the cost to the wider public. As Irwin shows, the depression was also a defining moment for American commercial policy. But, instead of expanding protectionism, a formal system for tariff reduction was inaugurated. Throughout the nineteenth century, the nation cycled through high- and low-tariff regimes, as elections alternated power between the high-tariff Republicans and the low-tariff Democrats. The overwhelming Democratic victory in 1932 would traditionally have led Congress to revise tariff schedules downward. But worldwide depression produced a rise in foreign trade barriers that made unilateral tariff reduction an unattractive political option. The consequence of worldwide economic and politi-

14

Michael D. Bordo, Claudia Goldin, and Eugene N. White

cal chaos was a startling innovation-the Reciprocal Trade Agreements Act (RTAA) passed in 1934-whereby Congress granted the president the authority to negotiate reciprocal trade agreements without seeking congressional approval. Although there were no real pre- 1930s antecedents, the way for reciprocal trade agreements negotiated by the president was eased by the introduction of the income tax in 1913 and the adoption, in 1923, of an unconditional mostfavored-nation trade policy. The RTAA locked in a process that has served to lower trade barriers. Rent seeking may have shifted to other forms of trade barriers, such as quotas, but under the RTAA industries are far less able to secure higher protective tariffs than under the previous system. This New Deal change is in direct contrast to the protection garnered in the 1930s by agriculture and banking. The RTAA survived the return of a Republican president and Congress in the 1950s. It became a fixture, as business and labor groups supported freer trade, which appeared not to threaten the dominant position of the United States in the postWorld War I1 economy when open markets were linked to national security in the cold war. A New Deal for Labor

At first glance, the misery visited upon American workers duringthe depression years was at least partly compensated for by the New Deal’s granting of various types of social insurance and a “bill of rights” to organized labor. The unemployed gained some protection; an old-age pension scheme was created; the elderly poor, the disabled, and others in need were to receive federally encouraged state aid; and legitimation in the eyes of the law was awarded to unions. But the chapters on these issues reveal that the Great Depression was a defining moment for only some aspects of these enduring programs and legislation. Most of the social programs of the New Deal were not invented in the 1930s. Various states had debated unemployment insurance (UI) schemes, for example, long before the Great Depression. Massachusetts did in 1916, as did Wisconsin in 1921. But, although six state legislatures introduced bills prior to the depression, UI generated little interest at the national level. Similarly, various state-level old-age assistance programs were in force before the depression. But these noncontributory programs provided scant income to few individuals and had strict means tests. According to Baicker, Goldin, and Katz, unemployment compensation was destined to be adopted even in the absence of the depression, although the actual form taken by UI was crucially shaped by the environment of the 1930s. Miron and Weil echo the notion of eventual passage with regard to old-age pensions but do not see the 1930s as having substantially shaped the structure of the social security program. The design of UI reflected, in part, Roosevelt’s long struggle with the Supreme Court, and in that sense the depression was a defining moment in this

15

The Defining Moment Hypothesis: The Editors’ Introduction

case study. The American UI system is unique. It is operated separately by each of the states and is experience rated within each state, which means that employers are generally penalized for causing unemployment. Although the essential characteristics of the program have not changed since 1935, its coverage has expanded from just over one-half, at its inception, to virtually the entire workforce today. UI was given a federal-state, rather than a national, structure because even though the Roosevelt administration might have generated sufficient congressional support to pass a national UI system, it feared the Supreme Court would invalidate it. Thus, UI was designed as a federal-state system to be upheld by the Court, and this structure has become a permanent feature. The experience rating of the system is linked by Baicker et al. to the federalstate structure, although the depression need not have been a defining moment here. Similarly, other portions of the Social Security Act, such as old-age assistance, aid to disabled persons, and aid to dependent children were instituted at the state level with partial support by federal contributions. Court precedent regarding conditional grants assured that such programs would survive the scrutiny of even a staunchly conservative Supreme Court. The old-age insurance part of social security, however, was forced to deviate from the structure of these other programs. It was created as a national system because there was no other way to achieve actuarial soundness. At the time of its passage in 1935, 20 countries had compulsory, contributory, non-meanstested old-age insurance programs. But old-age pensions would not have passed Congress in 1935 had it not been for Roosevelt’s steadfast desire to be the president who gave the American people their old-age security. UI, not oldage pensions, had broad-based support with the public and Congress. Roosevelt creatively tied the two together. His administration further ensured that social security would be locked in place forever by making it self-financing out of payroll taxes, thereby creating the perception that benefits were paid as a matter of employee right. Roosevelt stated his intentions best: “We put those payroll contributions there so as to give the contributors a legal, moral, and political right to collect their pensions. . . . With those taxes in there, no damn politician can ever scrap my social security program” (Schlesinger 1958, 308-9). Furthermore, the program expanded modestly to cover spouses and widows, in the 1950s to cover all private employees and the self-employed, and finally in 1983 to cover all federal civilian employees-meeting the architects’ goal of universal coverage. Although the founders of social security did not anticipate the better health, higher life expectation, and lower labor force participation rates of future generations of American workers, Miron and Weil show in their chapter that the program has grown according to plan, with little change in the projected replacement rates and benefits. Whereas social security and unemployment compensation have become permanent features of the economic landscape, unionization, which boomed during the depression, has not. The singular rise and long duration of unemploy-

16

Michael D. Bordo, Claudia Goldin, and Eugene N. White

ment helped to drive the surge in union membership. Previous spurts of union growth were modest by comparison, but then the economic contractions were also relatively modest. Most labor historians have attributed the burst of unionization during the 1930s to the passage of the National Industrial Relations (Wagner) Act in 1935 coupled with the dynamic leadership of the Congress of Industrial Organizations unions. The depression was certainly the catalyst for union formation and the rise in union membership, but Freeman shows that economic conditions, not new legislation, drove much unionizing activity. Recognition strikes and the growth of existing unions were the key factors, not the use of the new National Labor Relations Board election rules. The Wagner Act did not guarantee a unionized economy, but it did prevent a sharp decline in union membership after the depression. A return to stable economic growth, the changing nature of work, and greater international trade all led to a long, slow decline in private sector unionization. Except for the public sector, labor market changes have reduced unionization to its predepression level. The act’s primary effect, according to Freeman, has not been as much to aid the union cause as to reinforce labor-management conflict. Unionization became a legalistic business fought in the courts and before the National Labor Relations Board, with no provision for unions of supervisors, professionals, and managers or intermediate organizations-staff associations, work councils, or company unions-to give a voice to workers. Macroeconomic Policy At the macroeconomic level, the Great Depression inaugurated a new period of instability and policy activism. The birth of fiscal and monetary policy activism greatly complicated the choices of policymakers. As Obstfeld and Taylor point out, the “trilemma” of twentieth-century policymakers has been the incompatibility of fixed exchange rates, capital mobility, and discretionary monetary and fiscal policy. Before World War I, fixed exchange rates and capital mobility provided a stable international regime for the major economic powers, conducive to a century of steady growth and prosperity. The regime precluded the systematic use of discretionary fiscal policy. World War I shut down the gold standard, halted capital mobility, and was witness to extraordinary macroeconomic measures. Afterward, an attempt was made to reestablish the gold standard. The restoration was short lived as the dislocations of the Great Depression and World War I1 washed away the ideological attachment to gold. Exchange rates were fixed under Bretton Woods, but free convertibility was postponed, capital controls continued, and realignments permitted. This seismic shift of opinion reflected the deep-seated belief that domestic concerns must, at times, take precedence over adherence to a fixed exchange rate. In the 1930s, the public in the United States and Western Europe learned to judge a government not solely by its promise to maintain a fixed parity and price stability but also by its ability to maintain full employment and growth. Discretion-

17

The Defining Moment Hypothesis: The Editors’ Introduction

ary domestic policy became a form of macroeconomic insurance against a new economic disaster. The implications of this important change were largely cloaked by the disintegration of capital markets, which began during World War I, underwent a brief hiatus in the 1920s, and was completed during World War 11. Fixed exchange rates were set under the Bretton Woods agreement, but capital controls remained in place. Yet these controls-like domestic price controls-could not endure market pressures indefinitely. Domestic political coalitions for capital controls in the United States, and elsewhere, were too feeble to frustrate the public’s interest in the pursuit of full-employment policies by the government. The gradual reintegration of capital markets, begun in the 1960s, brought the trilemma to the fore. And when governments could not abstain from discretionary macroeconomic policy, the system of fixed exchange rates collapsed in the 1970s. The Great Depression’s legacy of domestic policy activism ultimately led to an international policy regime of floating exchange rates and international capital mobility. Bordo and Eichengreen address the counterfactual question of whether a fixed-rate regime could have survived in the absence of the Great Depression. This regime would have avoided the Bretton Woods innovations and would have been a gold exchange standard of pegged rates and unlimited capital mobility. It would have been suspended during World War I1 then reinstated at the original parities after the war, following the post-World War I example. It is unlikely, however, that this resumption would have been achieved by a big deflation and recession, as had occurred in 1919-21, because of the memory of that negative experience. The liquidity generated during the war obviated the need to repeat the post-World War I experience of deflation. Rapid economic growth would then have produced a gold scarcity, leading to a pure dollar-gold exchange standard. The unwillingness of many countries to accept this evolution would have precipitated a move to floating exchange rates. Thus, the Great Depression was not a defining moment for the international monetary system because it did not alter the system’s ultimate evolution. Like other developed nations, the United States had traditionally adhered to an orthodox fiscal policy of balanced budgets in which macroeconomic management of the business cycle was nonexistent. Deficits happened only in times of war, and in the peace that followed, surpluses reduced the debt incurred during the war. Hence, in the 1920s, the U.S. government debt was halved. The depression created a large peacetime deficit, which seemed impossible to cure by tax increases or cuts in spending. As De Long’s chapter shows, the government was forced, during the depression, to change its ideology and trumpet the fiscal advantages of unbalanced budgets. The depression also made fiscal policy potentially effective by increasing the size of government to the point at which the new automatic stabilizers could contribute significantly. The failure of the economy to rapidly recover

18

Michael D. Bordo, Claudia Goldin, and Eugene N. White

convinced the public that cyclical deficits were an acceptable fact of fiscal life. New Deal relief programs, unemployment compensation (which began paying out in 1938), and other automatic stabilizers were regarded by the public as important guarantees or insurance. Congress attempted to codify this momentous shift in the nation’s opinion with the Employment Act of 1946. Neither the economic crises nor the response of replacing fiscal conservatism with macroeconomic activism were unique to the United States. Rather, they were common to most industrialized nations. The Keynesian fiscal revolution, initiated by the Great Depression, pushed beyond these changes in its second generation to an acceptance of not just cyclical but also structural deficits in the 1980s. Recent demands for a balanced budget amendment suggest that peacetime budget deficits may not be permanent features of the macroeconomic landscape. Whereas the precepts of fiscal policy had remained unchanged for a century before the depression, the establishment of the Federal Reserve System in 1913 was a recent institutional innovation. In the decade before @e Great Depression, the Federal Reserve had pursued an active countercyclical monetary policy. Calomiris and Wheelock argue that the depression did not change the Federal Reserve’s goals and tactics as much as it altered its tools and external environment. The Fed’s pre- 1929 procyclically biased operating procedure continued, first during the depression and then during World War 11. The gravity of the depression and the obtuseness of Federal Reserve officials, however, weakened the public’s long-standing unwillingness to accept political pressure on monetary policy. By increasing the power of the presidentially appointed Federal Reserve Board over the Federal Reserve Banks, New Deal banking legislation reduced the Fed‘s independence vis-a-vis the Treasury. The Fed was also granted authority to use government securities to back note issues, removing a key limit on the monetization of debt. Most important, abandonment of the gold standard in 1933 helped permanently eliminate a crucial restraint on policy. The continuance of the Fed‘s pre- 1930s operating methods, pressures to monetize government debt, and the lifting of the nominal gold anchor allowed the Fed to acquiesce to low-level inflation in the 1960s. Once the last vestiges of the gold standard were erased, when Nixon closed the gold window in 1971, inflation spiraled upward.

Conclusion The chapters in this volume offer testimony to the legacy of the Great Depression. Without the depression, there would not have been a flood of New Deal-style legislation. Some innovations would have occurred following the dictates of economic growth, the two world wars, and the nation’s political economy. But, lacking the catalyst that jarred public attitudes and demanded action, the new economic institutions would have been more modest and dif-

19

The Defining Moment Hypothesis: The Editors’ Introduction

ferent in character. The large role of today’s government and its methods of intervention-from the pursuit of more activist monetary policy to the maintenance and extension of a wide range of insurance for labor and businessderive from the crisis years of the 1930s. Not all programs inaugurated by the New Deal have survived. But 60 years later, the basic imprint of the defining moment is still visible.

References Balke, Nathan S., and Robert J. Gordon. 1986. Appendix B: Historical data. In The American business cycle: Continuity and change, ed. Robert J. Gordon. Chicago: University of Chicago Press. Bemanke, Ben S. 1993. Non-monetary effects o f the financial crisis in the propagation of the Great Depression. American Economic Review 73 (June): 257-76. . 1995. The macroeconomics of the Great Depression: A comparative approach. Journal of Money, Credit, and Banking 27 ( 1 ) : 1-28. Bordo, Michael D. 1986. Explorations in monetary history: A survey of the literature. Explorations in Economic History 23 (October): 339-41 5. . 1989. The contribution o f AMonetary History of the United States, 1867-1960, to monetary history In Money, history and internationalfinance: Essays in honor of Anna J. Schwartz, ed. Michael D. Bordo, 15-70. Chicago: University of Chicago Press. Brownlee, W. Elliot. 1996. Federal taxation in America: A short history. New York: Woodrow Wilson Center Press and Cambridge University Press. Eichengreen, Barry. 1992. Goldenfetters: The gold standard and the Great Depression, 19/9-1939. New York: Oxford University Press. Fisher, Irving. 1933. The debt deflation theory of Great Depressions. Econometrica 1 (October): 339-57. Friedman, Milton, and Anna J. Schwartz. 1963. A monetary history of the United States, 1867-1960. Princeton, N.J.: Princeton University Press. Higgs, Robert. 1987. Crisis and Leviathan: Critical episodes in the growth ofAmerican government. New York: Oxford University Press. Hofstadter, Richard. 1955. The age of reform: From Bryan to ED.R. New York: Knopf. Hughes, Jonathan R. T. 1977. The governmental habit: Economic controlsfiom colonial times to the present. New York Basic Books. Keynes, John Maynard. 1936. The general theory of employment, interest and money. New York: Harcourt, Brace. Kolko, Gabriel. 1967. The triumph of conservatism: A reinterpretation of American history, 1900-1916. Chicago: Quadrangle. Libecap, Gary. 1992. The rise o f the Chicago packers and the origins o f meat inspection and antitrust. Economic Inquiry 32 (April): 242-62. Meltzer, Allan H. 1995. Why Did Monetary Policy Fail in the Thirties? In A history of the Federal Reserve. Pittsburgh, Pa.: Carnegie-Mellon University. Unpublished manuscript. Patterson, James T. 1969. The New Deal and the states: Federalism in transition. Princeton, N.J.: Princeton University Press. Romer, Christina D. 1992. What ended the Great Depression. Journal of Economic History 52 (December): 757-84.

20

Michael D. Bordo, Claudia Goldin, and Eugene N. White

. 1993. The nation in depression. Journal of Economic Perspectives 7 (Spring): 19-39. Schlesinger, Arthur M., Jr. 1958. The coming of the New Deal. New York: Houghton Mifflin. Temin, Peter. 1976. Did monetary forces cause the Great Depression? New York: Norton. . 1989. Lessonsfrom the Great Depression. Cambridge, Mass.: MIT Press. . 1993. Transmission of the Great Depression. Journal of Economic Perspectives 7 (Spring): 87-102. U.S. Bureau o f the Census. 1975. Historical statistics of the United States: Colonial rimes to 1970. Washington, D.C.: Government Printing Office. U.S. Council of Economic Advisers. Various years. The economic report of the president. Washington, D.C.: Government Printing Office. Wheelock, David C. 1991. The strategy and consistency of Federal Reserve monetary policy 1924-1933. Cambridge: Cambridge University Press. Wicker, Elmus. 1965. Federal Reserve monetary policy, 1922-33: A reinterpretation. Journal of Political Economy 73 (August): 325-43.

1

Was the Great Depression a Watershed for American Monetary Policy? Charles W. Calomiris and David C. Wheelock

The Great Depression witnessed many substantial changes in the monetary policy environment of the United States. These changes included a relaxation of the gold standard, an opening of a new avenue for monetizing government debt, changes in the structure of the Federal Reserve System, and the birth of a new policy ideology that doubted the stability of private markets and prescribed government management of aggregate demand. Did these changes have lasting effects? To address this question, we begin with a brief review of Federal Reserve policy before and during the Great Depression, which we compare with monetary policy since the depression. The legacy of the Great Depression is difficult to identify from immediate post-World War I1 policy outcomes because of the massive expansion of federal debt during the war and the Fed's commitment to funding that debt at low cost during and well after the war. The period between 1934 and 1941 is thus especially important for gauging the effects on monetary policy of the cyclical downturn of 1929-33 and the institutional reforms it produced between 1932 and 1935. Moreover, 1934-41 was a period of transition for the international monetary regime and, thus, provides important evidence on the impact of changed international linkages on domestic monetary policy. Accordingly, we explore in some depth the monetary policy of 1934-41. We discuss how the tools of Fed policy, but not its goals or tactics, changed in the mid-1930s. At the same time, we show how structural reforms weakened the Federal Reserve relative to the Treasury and removed a key limit on the monetization of government debt. While Fed policy goals and tactics did not Charles W. Calomiris is the Paul M. Montrone Professor of Finance and Economics at Columbia University's Graduate School of Business and a research associate of the National Bureau of Economic Research. David C. Wheelock is a research officer at the Federal Reserve Bank of St. Louis. The authors thank Michael Bordo, Claudia Goldin, Allan Meltzer, Anna Schwartz, Eugene White, and conference participants for their helpful comments and suggestions. The views expressed in this paper do not necessarily reflect those of the Federal Reserve Bank of St. Louis or the Federal Reserve System.

23

24

Charles W. Calomiris and David C. Wheelock

change in the mid-1930s, the increased power of the Treasury to determine the direction of policy, along with the departure from gold and the new “technology” for monetizing government debt, produced a new (albeit small) inflationary bias in monetary policy that lasted until the Treasury-Federal Reserve Accord of 195 1. The second half of the paper focuses on the postwar era. The Fed regained some independence with the accord of 195 1 and became increasingly aware of the inflationary potential of excessive monetary growth. The Fed failed to break with its traditional (predepression)operating methods, however, and the procyclical bias in these procedures-along with pressures to monetize government debt-explains how the Fed stumbled into an inflationary policy in the 1960s. Such a policy could not have been sustained under the gold standard as it existed before the depression. Under that regime, an incipient excess supply of dollars would have produced gold outflows, signaling the need for monetary contraction. Under the new dollar standard of Bretton Woods, the inflationary bias in monetary policy remained latent until it had gathered substantial momentum. That momentum, in turn, caused the external imbalance that underlay the collapse of the Bretton Woods system in the 1970s and the abandonment of external constraints on domestic monetary policy. The most important legacy of the Great Depression for monetary policy was thus its initiation of America’s departure from the gold standard.’ Thus the institutional changes occurring between 1932 and 1935 mattered in two ways. First, they had immediate consequences for monetary policy, largely via the temporary increase in the power of the Treasury. Second, depression-era changes removed preexistent checks on inflationary tendencies in monetary policy. These “latent” consequences became important in the 1960s.

Our analysis of the Fed’s powers, policy intentions, and effects relies as much on qualitative historical evidence-Federal Open Market Committee (FOMC) minutes and the like-as on statistical evidence of breaks or turning points in relevant time series. The two approaches are complementary. For example, statistical evidence is essential for documenting the timing of changes in free reserves, money, and inflation in the 1960s and the role of policy in producing inflation; but without qualitative evidence of policymakers’ intentions, it would be difficult to distinguish intended from unintended outcomes or to explain the timing of important statistical changes. 1. Our discussion is somewhat at odds with the counterfactual modeling of the gold standard by Bordo and Eichengreen (chap. 12 in this volume). As we will argue below, we believe that going off the gold standard in 1933 produced a fundamentally different, and more persistent, departure than going off the gold standard during wartime. Thus, while we believe that the United States probably would have left gold in 1941 even if it had remained on gold in 1933, we will argue that a hard-money standard would have been restored after the war and would have persisted long afterward.

25

The Great Depression as a Watershed for American Monetary Policy

A qualitative historical approach to identifying important turning points in policy is particularly useful for understanding the influence of latent institutional changes-those exerting their influence on policy outcomes only after long delays or under particular circumstances.For example, the long-run inflationary potential of the decoupling of the dollar from the gold standard that began in 1932 was not immediately visible in any data from the 1930s and became relevant later only under a particular combination of circumstances. Our historical approach eschews purely theoretical analyses of Fed policy shifts based either on welfare economics or on political economy. While such considerations surely play a part in our history, we believe that the role of theory for understanding policy changes is limited. In many cases, the outcomes of policy are different from the intent of policymakers. While economic or political economic influences may shape the intent of policy, institutional constraints and mistaken beliefs of policymakers place a wedge between intents and outcomes. The origins of the inflation of the 1960s, for example, should not be traced to shifts in economic or political fundamentals affecting the taste for inflation. Nor was inflation an inevitable consequence of the departure from gold. Rather, it reflected the combination of economic circumstances, traditional (flawed) policy rules, new theories about the role of government in the economy, and the absence of gold standard discipline, which together produced an unintended inflation.

1.1 Monetary Policy during the Great Depression and Before To gauge how the Great Depression might have altered monetary policy making, we first review monetary policy during the depression. By almost any measure, the monetary policy of 1930-33 was a disaster: the money supply and price level both fell by one-third, ex post real interest rates rose well into double digits, and banks failed by the thousands. How could the Fed have let this happen? Explanations for the Fed's disastrous monetary policy during the Great Depression largely fall into two categories. One attributes policy failures to innocent mistakes or neglect, while the other contends that the Fed willfully engineered contractionary monetary policy to foster bureaucratic objectives, or in response to interest group pressure. Although some political scientists and public choice economists favor the latter explanation (e.g., Epstein and Ferguson 1984;Anderson, Shughart, and Tollison 1988),most economists and economic historians blame the Fed's policy on misguided policy rules, as well as on petty jealousies that limited the Fed's ability to respond decisively to rapidly changing conditions. The most prominent explanation of Federal Reserve behavior during the Great Depression is that of Friedman and Schwartz (1963), who argue that a distinct shift in policy occurred with the death in 1928 of Benjamin Strong, governor of the Federal Reserve Bank of New York. Like Fisher (1935) before

26

Charles W. Calomiris and David C. Wheelock

them, Friedman and Schwartz (1963) contend that Strong understood how to employ the tools of monetary policy to minimize cyclical fluctuations in output and prices and to prevent or limit financial panics. His death created a void of both leadership and understanding that left the Fed unresponsive to financial crises, bank runs, and their contractionary effects. Under Strong’s leadership, the Fed had used the tools at its disposal to conduct an activist monetary policy aimed at both domestic and international objectives (Wheelock 1991). Large open market purchases and discount rate reductions in 1924 and 1927 were apparent attempts both to encourage domestic economic growth and to enable Great Britain to attract gold reserves. Open market sales and discount rate hikes in 1928 and 1929, on the other hand, were intended to discourage stock market speculation, which at least some Fed officials viewed as a manifestation of inflation. On the surface, the Fed seems to have been less responsive to the depression than it had been to earlier, smaller, cyclical downturns. Although the Fed made some open market purchases and discount rate adjustments between 1929 and 1933, relative to the declines in output and prices these operations were small in comparison with those of the 1920s. Despite the Fed‘s apparent lack of vigor between 1929 and 1933, however, some researchers argue that policy changed little, if at all, with Benjamin Strong’s death (e.g., Wicker 1966; Brunner and Meltzer 1968; Wheelock 1991). Monetary policy between 1929 and 1933 appears largely consistent with the policy strategy Strong outlined in the 1920s (Wheelock 1991). That strategy used borrowed reserves and market interest rates as short-run control variables, or policy “instruments.”2When member banks borrowed relatively little from Federal Reserve Banks or market interest rates were unusually low, Fed officials interpreted monetary conditions as “easy.” Conversely, high levels of borrowed reserves or high interest rates signaled that money was “tight.” Once the depression began, both borrowed reserves and interest rates fell sharply and generally remained low, giving Fed officials the impression that money was “cheap” and p l e n t i f ~ l . ~ Many economists have noted that rigid use of borrowed reserves or interest rates as policy instruments will cause the money supply to rise and fall procy2. The use of open market operations for objectives other than to secure earning assets evolved in the early 1920s, but their use to manipulate instruments or operating targets, such as borrowed reserves, evolved only gradually as the Fed gained experience. Well into the depression, directions to the Fed‘s trading desk from the FOMC specified the dollar amounts of securities the desk was authorized to buy or sell. By 1932, however, discussion at FOMC meetings turned more toward the desired level of excess reserves and focused less on the specific dollar volume of securities to buy or sell. Later in the 1930s, the committee targeted yields on Treasury securities, as well as excess reserves. 3. Borrowed reserves increased markedly in the fourth quarter of 1931, when the United States suffered a large outflow of gold following Britain’s abandonment of the gold standard. Fed officials understood that money was tight but did not make significant open market purchases. The Fed publicly justified its policy by a lack of gold reserves. Wicker (1966) argues that officials feared exacerbating gold outflows by pursuing “inflationist” policies.

27

The Great Depression as a Watershed for American Monetary Policy

clically because borrowed reserves and interest rates are positively correlated with economic activity. Moreover, Wheelock (1991) finds that the banking crises of 1929-33 made borrowed reserves an especially poor indicator of monetary conditions during the depression because the crises made banks reluctant to borrow. Although a few system officials questioned the reliability of borrowed reserves as a policy guide during the depression, the prevailing view was that monetary conditions were exceptionally easy and that the economy’s failure to expand was not the fault of monetary policy. We cannot say for certain whether monetary policy would have been different between 1929 and 1933 had Benjamin Strong lived, but it does seem to have been consistent with Strong’s response to business cycle downturns in 1924 and 1927 and with the guidelines for assessing the stance of monetary policy he had ~ u t l i n e dBut . ~ no matter what one’s view of the importance of Strong’s death, this much is clear: if the Great Depression was a watershed for monetary policy, the reason was not that it made policy active or responsive to the business cycle, for the Fed had long had such a policy.

1.2 Institutional Changes of the 1930s The year 1932 marked the beginning of a series of institutional reforms with potentially large consequences for monetary policy (table 1.1). Among the most significant were the Glass-Steagall Act of 1932, which permitted the Federal Reserve to use government securities to back its note issues; suspension of the international gold standard by executive order on 6 March 1933 (ratified by Congress on 9 March); the Thomas amendment to the Agricultural Adjustment Act of 1933, which, among other things, permitted the Federal Reserve to adjust commercial bank reserve requirements; the Gold Reserve Act of 1934, which authorized the president to fix the dollar price of gold and established the Treasury’s Exchange Stabilization Fund; and the Banking Act of 1935, which markedly altered the structure of the Federal Reserve System and expanded the Fed’s authority to adjust reserve requirements. By permitting the backing of Federal Reserve note issues with U.S. government securities, the Glass-Steagall Act of 1932 removed an important constraint on discretionary monetary policy. The Federal Reserve Act (as amended in 1917) had required the Reserve Banks to maintain gold reserves equal to 40 4.We make no attempt here to consider all aspects of monetary policy between 1929 and 1933 or the extent to which it was consistent with Strong’s policies. Meltzer, however, reaches many of the same conclusions about the failure of monetary policy during the Great Depression that we do, writing that “the main reason for the failure of monetary policy in the depression was the reliance on an inappropriate set of beliefs about speculative excesses and real bills. This set of beliefs, embodied in the Riefler-Burgess framework, directed attention to short-term market interest rates and member bank borrowing and encouraged their use as indicators of the magnitude and direction of monetary policy” (1994, 15). See Wheelock (1991) for analysis of the consistency issue, and the references therein for general histories of Federal Reserve policy during the 1920s and early 1930s.

28

Charles W. Calomiris and David C. Wheelock Key Institutional Changes in Monentary Policy during the Early 1930s

Table 1.1 1932

1933

1934

1935

Glass-Steagall Act (27 February): temporarily made U.S. government securities eligible collateral for Federal Reserve note issues, thereby expanding the Fed‘s ability to make open market purchases (made permanent in 1933); also temporarily relaxed rules on discount-window lending (extended in 1933, made permanent in 1935) Emergency Banking Act (9 March): ratified suspension of the gold standard Thomas amendment to Agricultural Adjustment Act (12 May): authorized the Fed to set reserve requirements; gave the president authority to require open market purchases by the Fed, and to fix the weights of gold and silver dollars Banking Act of 1933 (16 June): enhanced Federal Reserve Board control of discount-window lending; made technical adjustments to Federal Reserve System organization GoldReserve Act (30 January): authorized transfer of monetary gold stock to the U.S. Treasury; amended the president’s authority to fix dollar prices of gold and silver; established the Exchange Stabilization Fund Silver Purchase Act (19 June): authorized the president to purchase and nationalize monetary silver: authorized limited Fed lending to industrial and commercial firms Banking Act of I935 (23 August): reorganized Federal Reserve’s Open Market Committee and otherwise enhanced the authority of the Board of Governors of the Federal Reserve System relative to the Federal Reserve Banks; extended Fed authority to adjust member bank reserve requirements

percent of their note issues (and 35 percent of their deposit liabilities), with the remaining 60 percent in the form of either gold or “eligible paper,” consisting mainly of commercial loans that banks could rediscount or sell outright to the Reserve Banks. The Federal Reserve Act authorized the Reserve Banks to purchase government securities but did not permit their use as collateral for Fed liabilities. The Glass-Steagall Act of 1932 expanded the definition of “eligible paper” to include U.S. government securities in the Fed‘s portfolio, thereby enhancing the Federal Reserve’s ability to initiate transactions that monetized government debt.5Although he lent his name to the enabling legislation, Carter Glass, who had sponsored the original Federal Reserve Act, apparently voiced considerable worry about the inflationary potential of permitting government obligations to serve as collateral for Federal Reserve notes 5. Previously, member commercial banks could use government securities as collateral for discount-window borrowing, and during World War I the Fed offered a preferential discount rate on such borrowing, with the effect of substantially monetizing the fiscal deficit. These transactions, however, were initiated by commercial banks, rather than by the Federal Reserve.

29

The Great Depression as a Watershed for American Monetary Policy

(Chandler 1971, 189). We argue below that Glass was prescient in his concerns.6 Suspension of the gold standard removed another substantial barrier to discretionary monetary policy. Although the Gold Reserve Act of January 1934 restored gold convertibility for international payments to other gold standard countries, the gold standard imposed no constraint on money supply growth through the remainder of the 1930s. Countries that remained on the gold standard at their predepression par levels, such as France, tended to lose gold to the United States and other countries until finally devaluing or going off gold themselves. The coming war in Europe also precipitated gold flows to the United States. Thus, after 1933 the United States was never forced to restrict money growth or raise interest rates to protect its international reserves. Before 1933, the Federal Reserve, like many central banks, failed to play consistently by the rules of the gold standard. The Fed sterilized gold flows, allowing offsetting changes in other components of bank reserves in the face of gold inflows or outflows. Twice, however, the Federal Reserve reacted to international gold flows in a manner consistent with restoring “external balance.” On those occasions, between November 1919 and June 1920, and during September and October 1931, Fed officials believed that gold outflows threatened their own reserve position. On each occasion, the Fed increased its discount rate and maintained it at a high level until the crisis had passed, and each time the policy was sharply contractionary. The Fed clearly believed that the perceived long-term benefits of defending the gold standard were worth the short-term contractionary impact of a defense of their reserve position (Wheelock 1991). The Roosevelt administration,by contrast, was willing to relax the gold standard to permit reflation. Roosevelt suspended gold payments by executive order on 6 March 1933, which the Emergency Banking Act ratified on 9 March. On 10 March Roosevelt prohibited private transactions in gold and foreign exchange unless licensed by the secretary of the treasury. On 5 April, Roosevelt ordered private individuals and banks to deliver their holdings of gold to Federal Reserve Banks for transfer to the U.S. Treasury and prohibited future use of gold for domestic transactions and private ownership of gold. The Gold Reserve Act of January 1934 authorized the president to fix the dollar price of gold for internationalpayments, which Roosevelt did at $35 per ounce (in contrast to the previous parity of $20.67 per ounce). The United States thus returned to the gold standard for the settlement of international payments with other countries also on the gold standard. The restored gold standard, however, differed fundamentally from the previous standard in the degree to which its operation was removed from private markets and placed 6. The Glass-SteagallAct of 1932 was intended as an expedient to ensure that sufficient Federal Reserve notes could enter circulation in an emergency. It was set to expire after one year but was made permanent in March 1933. The Glass-Steagall Act of 1932 should not be confused with the Banking Act of 1933, which is commonly referred to as the “Glass-Steagall Act.”

30

Charles W. Calomiris and David C. Wheelock

under control of government authorities. Gold was no longer regarded as an absolute exogenous check on government manipulation of the supply of money. Under the weight of the Great Depression, the ideology of the gold standard, which viewed gold as fundamental to a country’s economic prosperity, had cracked. Although the dollar remained linked to gold, the link was weakened and, perhaps more important, government authorities had demonstrated a willingness to manipulate the gold standard to limit the extent to which it would interfere with discretionary monetary policy. Thereafter, when the Fed’s gold reserve requirement threatened to limit money supply growth, the reserve requirements were reduced and then eliminated with apparently little debate or fanfare.’ The gold standard as it existed after 1933 was thus fundamentally different from its precursor and foreshadowed the Bretton Woods gold standard that was to replace it after World War II.8 In addition to marking a fundamental shift in the degree to which gold served as a constraint on domestic monetary policy, the revaluation of gold in 1934 left the U.S. Treasury with a capital gain of some $2.8 billion on its gold holdings. Under authority conveyed by the Gold Reserve Act of 1934, the Treasury used $2 billion of its windfall to establish the Exchange Stabilization Fund: “For the purpose of stabilizing the exchange value of the dollar, the Secretary of the Treasury . . . is authorized . . . to deal in gold and foreign exchange and such other instruments of credit and securities as he may deem necessary.” 9 Although the operations of the Exchange Stabilization Fund during the 1930s had little effect on the quantity or growth of bank reserves, the size and open-ended authority of the fund was widely viewed as a threat to the Federal Reserve System and its ability to effect monetary policy. For example, Roy Young, then governor of the Federal Reserve Bank of Boston, argued that the Gold Reserve Act “gives the Secretary of the Treasury such powers, of a permanent nature, that he could nullify anything we [the Federal Reserve] could do” (quoted in Johnson 1939, 36). The Commercial and Financial Chronicle had a similar reaction: “The Reserve authorities have been reduced to shadowy nonentities, the Federal Reserve System having become simply an adjunct of the United States Treasury and the Federal Government, to do what they are told to do” (20 January 1934, 367). In addition to the Exchange Stabilization Fund, authorities granted the president and treasury secretary included the right to “request” the Federal Reserve to use open market purchases to increase bank reserves by up to $3 billion and, 7. E.g., the Fed’s gold reserve requirement was reduced from 40 percent (against its note issues) and 35 percent (against deposits) to 25 percent in 1945. Foreshadowing the ultimate collapse of the Bretton Woods system, the reserve requirement against deposits was eliminated in 1965 and that against Federal Reserve notes was eliminated in 1968. 8. Friedman and Schwartz write that “perhaps the best description of the role of gold in the United States since 1934 is that, rather than being the basis of the monetary system, it is a commodity whose price is officially supported in the same way as the price of wheat, for example, has been under various agricultural programs” (1963,472). 9. See Schwartz (1997) for a history of the Exchange Stabilization Fund.

31

The Great Depression as a Watershed for American Monetary Policy

if the Fed refused, to issue a commensurate amount of fiat currency. This power was granted by the Thomas amendment to the Agricultural Adjustment Act of 1933, which, along with the Silver Purchase Act of 1934, also authorized the purchase of silver and permitted the president to devalue the silver dollar. Between 1933 and 1938, the Treasury purchased 1.8 billion ounces of silver, thereby increasing bank reserves by $1 billion (some 20 percent of the total increase in reserves during the period). Had the president chosen to devalue the dollar in terms of silver, the Treasury would have reaped a $2.2 billion windfall on its silver holdings (Johnson 1939, 195-98). In summarizing the various new authorities given the administration, Johnson concludes, “The President could double or triple bank reserves, had complete discretion over the gold value-and consequently the foreign exchange value-of the dollar, and could establish bimetallism by proclamation, in other words, he could completely refashion the monetary system of the country, and the sole criteria required were his own subjective evaluations of the situation” (1939, 202). In the event, the president never invoked his authority to “request” Federal Reserve purchases or to devalue the silver dollar-he did not have to since Treasury pressure generally kept the Fed in line. Organizationalchanges to the Federal Reserve System may have contributed to the Fed‘s willingness to accept the Treasury’s desired monetary policy. The authors of the Federal Reserve Act agreed that the Federal Reserve System should not be a “central bank” on the European model, but a federal system of semiautonomous Reserve Banks with an overseeing board. Dissatisfaction with the subsequent performance of the Federal Reserve, both during the 1920s and during 1929-33, led to reforms that enhanced the authority of the Federal Reserve Board at the expense of the Reserve Banks. Marriner Eccles accepted the chairmanship of the Federal Reserve Board in 1933 with the understanding that he would have freedom to redesign the Federal Reserve System. His reforms included limits on the power of the Federal Reserve Bank of New York, which he viewed as an instrument of the private interests of New York bankers, and measures to ensure oversight and coordination of the activities of the regional Reserve Banks in pursuit of the national interest (Eccles 1966, 170-72). Under Eccles’s plan, which the Banking Act of 1935 substantially brought into effect, the Board of Governors was given substantial control over open market operations and Federal Reserve Bank discount rates. The FOMC was reconstituted to include all seven members of the Board of Governors and just five of the twelve Reserve Bank presidents.’O The legislation thereby increased the authority and stature of the Federal Reserve officials located in Washington 10. The Banking Act of 1935 also changed the titles of the chief executive officers of the Federal Reserve Banks from the more prestigious “governor” to “president,” while discontinuing the Federal Reserve Board in favor of the Board of Governors, whose members all held the title “governor.” The Board of Governors was further authorized to approve the appointments of Federal Reserve Bank presidents and first vice presidents, and to generally supervise Reserve Bank operations.

32

Charles W. Calomiris and David C. Wheelock

and appointed by the president. On the other hand, it also sought to limit the influence of the president by removing the secretary of the treasury and comptroller of the currency as ex officio FOMC members. With his reforms, Eccles intended that monetary policy would be made by professionals whose allegiance was solely to the national interest. As we will discuss below, however, these changes increased political pressures on the Fed at the same time that establishment of the Exchange Stabilization Fund and other measures increased the administration’spower to conduct monetary policy. consequently, these reforms shifted power away from the Fed toward the Treasury and promoted an inflationary bias in monetary policy.

1.3 Monetary Policy after the Contraction The Federal Reserve lay nearly dormant between April 1933 and February 1936, when the structural changes to the Federal Reserve System imposed by the Banking Act of 1935 took effect. Between April and December 1933, however, the Fed did buy nearly $600 million of government securities, which, along with currency inflows, increased member bank reserves by some $700 million. The Fed made no further changes to the size of its open market portfolio until 1937. The Fed also permitted discount-window loans outstanding to fall from some $100 million at the end of 1933 to less than $10 million at the end of 1934, where they remained for the rest of the decade. Over time, Fed officials became increasingly concerned about substantial increases in bank reserves, especially excess reserves. During 1934 and 1935, gold inflows of some $3 billion contributed to a doubling of member bank total reserves (from $2.76 billion in January 1934 to $5.72 billion in December 1935) and more than a tripling of excess reserves (from $866 million to $2.98 billion; Board of Governors of the Federal Reserve System 1943, 371). The buildup of excess reserves alarmed Fed officials, who feared that these “idle” balances might permit a wave of speculation and inflation. Using its traditional tools the Fed would have reduced reserves (or slowed their rate of growth) by selling securities and raising the discount rate. But this was not feasible in the mid-1930s. A discount rate increase would have had no effect on reserves since discount-window borrowing already was trivial, even at a discount rate of just 1.5 percent. Similarly, by mid-1935, member bank excess reserves alone equaled the Fed‘s total security holdings, leaving the Fed unable to slow significantly the growth of total reserves through open market sales. The Fed was given an important new policy tool, however, by the Thomas amendment to the Agricultural Adjustment Act of 1933, which permitted the Fed to alter member bank reserve requirements with the approval of the president. The Banking Act of 1935 broadened the Fed’s authority by permitting the Board of Governors to alter reserve requirements within certain limits independently “in order to prevent injurious credit expansion or contraction.” With

33

The Great Depression as a Watershed for American Monetary Policy

its traditional means of monetary control becoming ineffective, the Fed came to rely on reserve requirements for regulating the supply of money and credit. The system’s first major policy initiative after 1933 was to increase reserve requirements in three steps in 1936 and 1937. 1.3.1 The Reserve Requirement Increases of 1936 and 1937 Alarmed at the sharp increase in excess reserves that had taken place since 1933, and viewing it as potentially inflationary, the Board of Governors increased required reserve ratios in August 1936, and again in March and May 1937. In total, the reserve requirements on time deposits were increased from 3 percent to 6 percent. Requirements on demand deposits were increased from 7, 10, and 13 percent to 14, 20, and 26 percent for country, reserve city, and central reserve city banks, respectively. The increases, according to the Annual Report of the Board of Governors for 1936, were intended to eliminate those excess reserves the board deemed ‘‘superfluous for prospective needs of commerce, industry, and agriculture, and, if permitted to become the basis of a multiple expansion of bank credit, might have resulted in an injurious credit expansion” (14). We find certain aspects of monetary policy between 1936 and 1938 striking, both in how the practice of monetary policy had changed and in how it had not changed since 1933. Gold inflows had made the old tools of monetary policythe discount rate and open market operations-ineffective, and so the Fed used its new tool-reserve requirements. In raising reserve requirements, however, the Fed sought to restore the effectiveness of the discount rate and open market operations: “With excess reserves reduced to a manageable figure, the Reserve System would be in a position to take prompt action to bring about current adjustments of the reserve position of member banks to credit needs by employing the more flexible instrument of open-market operations to ease or tighten conditions in the money market” (Board of Governors, Annual Report 1937, 5 ) . In other words, by reducing the amount of excess reserves (and thereby increasing the Fed’s security holdings relative to excess reserves), the Fed sought to make usable its traditional tools-open market operations and the discount rate-in the future if contractionary moves became desirable. A second objective of the reserve requirement increases of 1936 and 1937 was to limit the banking system’s ability to increase loans and deposits without Fed acquiescence. The Fed interpreted the excess reserves of commercial banks as largely superfluous balances with inflationary potential. The Fed wrote in its annual report for 1936 that ‘(inraising reserve requirements it was not the intention of the Board to reverse the policy of monetary ease which has been pursued by the System since the beginning of the Depression” (Board of Governors 1936, 15). By reducing excess reserves, the Fed sought to move from a policy stance that it viewed as dangerously easy to one that was sufficiently easy to promote continued, but noninflationary, economic growth. From this perspective, the reserve requirement hikes of the mid-1930s are best

34

Charles W. Calomiris and David C. Wheelock

seen as an attempt by the Fed to prevent banks from being the engine of inflation and to return to a status quo in which its traditional rules of thumb and traditional policy tools could be used as before. The Fed‘s policy was similar to previous policies in two respects. First, Fed officials judged the stance of monetary policy with the same indicator-free reserves (excess less borrowed reserves)-that they had used to gauge monetary conditions during the early 1930s and before. Second, in seeking to prevent inflation (which they defined broadly to include “speculation”), policymakers were influenced by a view that excessively easy monetary policy in the 1920s had fueled financial speculation and produced the stock market crash and depression. Extreme proponents of this view held that the Fed should withdraw credit during recessions so as to discourage renewed speculation or inflation and hasten a process of liquidation that they deemed a prerequisite of recovery. Although a majority of Fed policymakers, both during 1929-33 and later, decided to encourage economic recovery by promoting monetary ease, the “liquidationist”view undoubtedly influenced the ultimate outcome of Fed deliberations both early in the depression and after 1935.

1.3.2 The Free Reserves Policy Guide Throughout the 1920s and 1930s the Fed used free reserves as its main guide to monetary conditions. Free reserves is the difference between excess and borrowed reserves, but because banks held little or no excess reserves before the depression, officials spoke of monetary conditions in terms of borrowed reserves. For example, in 1926, Benjamin Strong suggested that as a guide to the timing and extent of any [open market] purchases which might appear desirable, one of our best guides would be the amount of borrowing by member banks in principal centers. . . . Our experiencehas shown that when New York City banks are borrowing in the neighborhood of 100 million dollars or more, there is then some real pressure for reducing loans, and money rates tend to be markedly higher than the discount rate. . . . When member banks are owing us about 50 million dollars or less the situation appears to be comfortable, with no marked pressure for liquidation. (Quoted in Chandler 1958,239-40) The Fed’s lack of vigor in responding to the depression between 1929 and 1933 was largely consistent with the policy guidelines Strong outlined. Member bank borrowed reserves plummeted during 1930 and 1931, and again after gold outflows and the banking panic in the fourth quarter of 1931 had subsided. In addition, for the first time, banks accumulated significant excess reserves. Some system officials argued that monetary conditions had been made too easy-that the Fed was pushing reserves on a saturated market and thereby delaying needed and inevitable economic adjustments. Free reserves continued to serve as the Fed’s principal indicator of monetary conditions in the mid-to-late 1930s, though because banks borrowed almost

35

The Great Depression as a Watershed for American Monetary Policy

no reserves after 1933, monetary conditions were measured in terms of excess reserves. At an FOMC meeting in September 1937, E. A. Goldenweiser, the Fed’s chief economist, recalled Strong’s rule of thumb that monetary conditions were neutral when New York City banks were borrowing $50 million of reserves from the Fed. Now, Goldenweiser claimed, a neutral policy stance was achieved when New York City banks held $250 million of excess reserves and all banks held $700-$800 million of excess reserves (FOMC, Minutes 11 September 1937).” Goldenweiser’s redefinition of monetary neutrality to be a free reserve position of New York banks some $300 million higher than in the past undoubtedly reflected the experience of intervening years, including, perhaps, an understanding that banks held excess reserves as a long-run portfolio choice (compensatingfor a decline in bank capital). Free reserves, however, remained the metric for judging monetary conditions.Total reserves, the monetary base, the money supply, and real interest rates were scarcely considered. 1.3.3 The “Liquidationist”Policy View The Fed‘s use of free reserves as its principal policy guide during 1936 and 1937 was consistent with previous policy. We also find precedent in the claim of many policymakers during 1936 and 1937 that excess reserves were superfluous, even dangerous. Between 1930 and 1933, some Fed officials argued that monetary conditions were excessively easy-that money was even “sloppy.” Chandler (1971) describes two schools of thought prevailing among system officials in the early 1930s. One school, which Chandler terms the “accommodationists,” believed that during a recession the Fed should accommodate recovery by promoting monetary ease, defined as low interest rates and little discount-window borrowing. The other school, termed the “liquidationists,” believed that monetary ease interfered with recovery by delaying inevitable adjustments of prices, wages, and resource utilization that are necessary before a sustainable recovery can begin. Between 1930 and 1933, Fed officials rarely debated whether monetary conditions could be described as “easy” but frequently considered whether monetary ease was the appropriate policy to follow. Proponents of the liquidationist view tended to blame the depression on financial speculation that, they argued, the Fed had supported with easy money policies during recessions in 1924 and 1927. As Adolph Miller, a member of the Federal Reserve Board, put it, open market purchases had been made in 1927 during “a time of business recession. Business could not use and was not asking for increased money at that time” (Miller 1931, 134). It was commonly believed, both within and outside the Fed, that the volume of Federal Reserve 11. The Minutes ofthe Federal Open Market Committee are not verbatim transcripts, but rather summaries of the discussion, sometimes with attribution and sometimes not. For exposition, when we quote directly from the Minutes, we put such text in quotation marks and attributethe text to the individual referred to in the Minutes. It may be, however, that the quote is not the exact statement of the individual.

36

Charles W. Calomiris and David C. Wheelock

credit outstanding would be appropriate if extended only at the initiative of commercial banks in the rediscount of short-term commercial paper (the “real bills doctrine”). Accordingly, Fed credit should contract during recessions, when the demand for commercial loans fell and thus when banks required fewer reserves. This explains why officials of the Federal Reserve Bank of Richmond responded as follows to a question from the Senate Banking Committee in 1931 about the system’s open market purchases in 1924 and 1927: “We think United States securities should not have been purchased in these periods, and the aim should have been to decrease rather than augment the total supply of Federal Reserve credit” (U.S. Senate 1931, 817). Like Miller, these officials believed that open market purchases supplied reserves that were unneeded to support productive loans and, hence, fueled the speculation that ultimately and inevitably led to a crash and depression.‘2 Opponents of open market purchases during the depression often warned that not only would monetary ease prolong the downturn but it might reignite speculation. For example, Frederic Curtis, chairman of the Federal Reserve Bank of Boston, opposed open market purchases “on the ground that they were likely to feed the stock market rather than the bond market” (quoted in Friedman and Schwartz 1963,373). And James McDougal, governor of the Federal Reserve Bank of Chicago, warned that if open market purchases did not reignite stock market speculation, they might fuel speculation “in some other direction” (371). Still another view was put forward by the governor of the Federal Reserve Bank of Philadelphia at a meeting of the FOMC in September 1930: We believe that the correction must come about through reduced production, reduced inventories, the gradual reduction of consumer credit, the liquidation of security loans, and the accumulation of savings through the exercise of thrift. These are slow and simple remedies, but just as there is no “royal road to knowledge,” we believe there is no short cut or panacea for the rectification of existing conditions. . . . We have been putting out credit in a period of depression, when it is not wanted and could not be used, and will have to withdraw credit when it is wanted and can be used. (Quoted in Chandler 1971, 137) The views of Miller, McDougal, and other liquidationists did not dominate the Fed during the 1930s, but they were heard and probably kept the system from responding more vigorously to the depression. Many of the most vocal critics of easy money policies in the 1920s, such as Miller, were no longer in the system in 1936 and 1937, when reserve requirements were being debated. Still, the view remained among policymakers that 12. Although today many of the liquidationist arguments may seem strange, under the gold standard the price level tended to be stationary, and thus, there is logic in the view that an inflation would be followed at some point by deflation, and the greater the inflation, the deeper the deflation.

37

The Great Depression as a Watershed for American Monetary Policy

the depression had been in large part caused by financial speculation and the stock market crash. Many officials believed that the Fed was too slow to check speculation in 1928 and 1929, and the consequences of not reacting quickly and decisively to nip speculation in the bud was one lesson these individuals took from the depression. This view might explain why in January 1937 Goldenweiseradvocated further increases in reserve requirements,despite noting that unemployment remained high. He argued that it was better to check “unsound and speculative situations” in their early stages while control is still possible, with now being that time (FOMC, Minutes 26 January 1937). The board gave a similar explanation for the timing of the reserve requirement increases in its annual report for 1937: “Notwithstanding the fact that recovery was far from complete and that there was still a large amount of unemployment, boom conditions were developing in particular industries and boom psychology began to be manifested” (Board of Governors 1937,2). To summarize, we find that the Fed‘s methods of interpreting the stance of its policy, as well as the fears policymakers had about potential future financial speculation and inflation, changed little between the two halves of the 1930s. In each period, Fed officials sought to achieve a degree of monetary ease that would promote economic recovery but not reignite the sort of inflation or speculation that they thought might hasten another collapse. In 1926, Strong advocated a borrowed reserves total of $50 million as an appropriate free reserves target during a recession. His replacement at the New York Fed, George Harrison, advocated a target for excess reserves of $250 million at a meeting of the FOMC in July 1932 (cited in Chandler 1971,200).And, finally, at a FOMC meeting in March 1937, Marriner Eccles argued that $500 million of excess reserves would make “ample funds available for legitimate business use” (FOMC, Minutes 15 March 1937). The target level was thus changed from time to time, but never the fundamental policy framework. We conclude therefore that the first major policy initiative following the revamping of the gold standard regime and the restructuring of the Federal Reserve System between 1933 and 1935 reflected remarkably little change in the Fed’s goals or tactics. 1.3.4 The Recession of 1937 and 1938 The doubling of reserve requirements between 1936 and 1937 suggests the absence of an inflation bias in monetary policy and even raises the question of whether the Great Depression created a deJutionury bias-an excessive concern with containing speculation that inspired contractionary policy in 1936 and 1937. We think that such a view is not warranted by the intent or consequences of Fed actions in those years. The Fed was willing to double reserve requirements when the unemployment rate was well over 10 percent and the price level and real output stood far below their 1929 levels, but the Fed did not see itself as initiating a contraction of credit. Rather it sought to prevent a future expansion of credit that threatened to bring inflation and also derail eco-

38

Charles W. Calomiris and David C. Wheelock

nomic recovery. Even after reducing excess reserves by some $2 billion, the Fed believed that monetary conditions would remain easy, with ample reserves available to support continued economic growth.l 3 A recession did follow the reserve requirement increases, however, and the Fed received considerable criticism from both contemporaries and more recent analyses of monetary policy. For our purposes, the Fed‘s intent and its perceptions of policy are more important than the impact of the reserve requirement increases, although elsewhere we argue that the reserve requirement increases of 1936 and 1937 did not precipitate the recession of 1937 and 1938 (Calomiris and Wheelock 1996; see also Frost 1971; Cume 1980; Calomiris and Wilson 1996). 1.3.5 Treasury Pressure and an Easing of Policy in 1937 and 1938 We have argued that Federal Reserve policy in 1936 and 1937 did not reflect a major change in system goals or targets. Nevertheless, New Deal reforms enhancing the administration’s ability to influence monetary policy affected Federal Reserve behavior in the 1930s and, in concert with fundamental changes in the gold standard, undermined stationarity of the price level over the long term. The Banking Act of 1935 removed the secretary of the treasury from the Federal Reserve Board but gave the newly formed Board of Governors majority control of the FOMC. Previously, the 12 Reserve Bank presidents had formed the FOMC and the board‘s role had been limited to an ambiguous supervisory role. The Banking Act of 1935 put control over open market policy in the hands of officials appointed by the president. Furthermore, only those members of the Federal Reserve Board who had been Roosevelt appointees were placed on the new Board of Governors. The new FOMC structure was a compromise between Eccles’s original plan, under which open market policy would have been vested solely in the Board of Governors, and that of Carter Glass, who argued that open market operations should remain the province of the Reserve Banks. Treasury Secretary Morgenthau favored the ultimate outcome, intending thus to weaken Eccles and the Fed. Referring to the Fed’s structure before 1935, Morgenthau concluded in a diary entry that “the way the Federal Reserve Board is set up now they can suggest but have very little power to enforce their will. . . . [The Treasury’s] power has been the Stabilization Fund plus the many other funds that I have at my disposal and this power has kept the open market committee in line and afraid of me.” Favoring the new makeup of the FOMC committee, Morgenthau wrote in his diary, “I prophesy that . . . with the seven members of the Federal Reserve Board and the five governors of the Federal Reserve 13. Friedman and Schwartz reach a similar conclusion: “Even so drastic a use of that new power [over reserve requirements] does not contradict the view that the Federal Reserve System was following a largely passive policy. The rise in reserve requirements was not imposed primarily to affect current conditions but to enable the System to control future developments it feared might be set in train by the large excess reserves” (1963,5 17).

39

The Great Depression as a Watershed for American Monetary Policy

Banks forming an open market committee, that one group will be fighting the other and that consequently they will not be able to do anything constructive, and that therefore if the financial situation should go sour the chances are that the public will blame them rather than the Treasury” (quoted in Blum 1959, 352). Regardless of how the Federal Reserve System was organized, the Treasury had the power to intimidate the Federal Reserve into carrying out the administration’s policies. By 1936, Morgenthau believed strongly in sterilizing gold inflows to have a weapon to offset outflows should they occur. When Eccles and Morgenthau clashed over whether sterilizing operations should be done at the initiative of the Treasury or the Federal Reserve, Eccles was forced to back down when it became clear that the Treasury had the means and the political capital to conduct policy as it saw fit (see Blum 1959, 360-67).14 The Banking Act of 1935 eliminated a provision imposed by the Thomas Amendment of 1933 requiring presidential approval of changes in commercial bank reserve requirements. Still, prior to increasing reserve requirements in 1936 and 1937, the Fed consulted President Roosevelt and Treasury Secretary Morgenthau about possible reserve requirement changes.15 The administration was not pleased, however, with the apparent impact of those increases on the market for government securities and pressured the Federal Reserve into undertaking offsetting open market operations and ultimately reversing a portion of the reserve requirement increase. The Treasury’s new monetary clout set the stage for a decade of explicit Treasury dominance of monetary policy and an even longer period in which its objective of “maintaining stability” of government securities prices remained an important goal of monetary policy. Following the Fed’s doubling of reserve requirements, Secretary Morgenthau increasingly sought to influence the conduct of monetary policy. At a meeting of the FOMC Executive Committee on 13 March 1937, Chairman Eccles reported that Morgenthau had expressed displeasure about a decline in government bond prices following the 1 March increase in reserve requirements and wanted to know how the Fed intended to rectify the situation. Eccles also reported Morgenthau as saying that the Treasury stood ready to use the Exchange Stabilization Fund, or possibly end its sterilization of gold inflows, to raise bond prices (FOMC, Minutes 13 March 1937). Although members of 14. Ordinarily, the Treasury paid for its gold purchases by transferring funds from its Federal Reserve accounts to those of member commercial banks, the effect being to increase aggregate commercial bank reserves. The Treasury could prevent an inflow from affecting reserves, i.e., “sterilize” it, by simultaneously transferring funds from its accounts with commercial banks to those it held with the Federal Reserve. Alternatively, the Treasury could reduce reserves by issuing new debt to the public. 15. Whether Roosevelt and Morgenthau had approved reserve requirement changes is unclear. Some accounts state that they had, while Blum writes that Morgenthau “was taken by surprise on July 15 when he picked up the morning paper and read that the Board of Governors had increased requirements 50 percent” and subsequently had a “blistering” conversation with Eccles about it (Blum 1959,356-57).

40

Charles W. Calomiris and David C. Wheelock

the FOMC agreed that “the drop in prices in the bond market was in the nature of a natural adjustment which was due primarily to non-monetary causes” (FOMC, Minutes 13 March 1937, 2), the FOMC Executive Committee met with Morgenthau and assured him of their desire to maintain an “orderly” market for government securities while not pegging prices at specific levels. On 15 March, the full FOMC met to review the economic outlook and discuss Morgenthau’s criticism of Fed policy. The committee agreed to authorize open market purchases to preempt more invasive action by the Treasury, though Eccles felt that excess reserves remained “ample for legitimate business use” (FOMC, Minutes 15 March 1937, 7).16 Midway through the meeting, Morgenthau telephoned Eccles to find out what the FOMC had decided to do, and Eccles reported that the committee had voted to maintain an orderly market for government securities and make additional open market purchases in the event of an “emergency.” Over the subsequent week, and under continuing pressure from Morgenthau, Eccles changed his mind. He now regarded previous policy as having been inappropriate and argued that reserve requirement increases had been responsible for declines in government bond prices. Although absent from a meeting of the FOMC Executive Committee on 23 March, Eccles communicated his desire to buy government securities because their yields were “quite out of line” with reasonable levels. Other members of the Executive Committee disagreed, however, deciding that the situation was not an “emergency” and therefore did not meet the criterion established at the 15 March FOMC meeting for increasing the system’s portfolio. A majority of the Executive Committee also argued that open market purchases simply were not needed. At the FOMC meeting of 3 April, Eccles relayed Morgenthau’s continued displeasure with weakness in the government bond market. He indicated that Morgenthau had requested the cancellation of the third increase in reserve requirements, set for 1 May, and had threatened to desterilize $500 million of gold, thereby increasing coinmercial bank reserves by that amount. Morgenthau also indicated, however, that the Treasury would not act if the Fed bought securities in the open market. Eccles proposed that the Fed make the open market purchases Morgenthau wanted, claiming that the increases in required reserve ratios had “drastically” reduced excess reserves and that “it would take the banks some time to accus16. Suggesting the tone of Morgenthau’s meetings with Eccles, Morgenthau claims to have told Eccles at one point that the President suggested that I should say to the Federal Reserve: “Now Congress gave you the job of managing the money market and that is your responsibility. You muffed it. You haven’t done it. You have not maintained an orderly market, and this thing is getting steadily worse. . . . Now I, Henry Morgenthau, Jr., talking for the U.S. Government, serve notice on the Federal Reserve Board that I ask you to do what Congress has given you the power to do, namely, to increase your portfolio. If you don’t do it, the Treasury will step in. . . . We are putting you on notice.” (Quoted in Blum 1959, 373-74)

41

The Great Depression as a Watershed for American Monetary Policy

tom themselves to operating with a smaller amount of excess reserves, as evidenced by the fact that they had sold earning assets rather than reduce their balances with correspondents” (FOMC, Minutes 3 April 1937,7). Eccles went on to argue that open market purchases were justified because the government securities market had been “disturbed” by the increases in reserve requirements. As the FOMC continued to discuss open market policy and the advisability of proceeding with the third increase of required reserve ratios, their meeting was twice interrupted by calls from Secretary Morgenthau inquiring about the committee’s decisions. Following the second call from Morgenthau, a meeting of the FOMC Executive Committee and Morgenthau was arranged for that evening. The FOMC reconvened on Sunday 4 April, at which time Eccles reported that Treasury Secretary Morgenthau had requested that the Fed support bond prices and stated that the Treasury would take action if the Fed‘s moves failed to stabilize the market for government securities. Eccles proposed to the FOMC that open market purchases be used to increase the excess reserves of member banks by as much as $250 million but argued that the board should go forward with plans to increase required reserve ratios on 1 May. Majorities of the FOMC and Board of Governors agreed with these proposals. For Eccles and Morgenthau, the prime motivation for this policy was not a fear of economic contraction, but rather a perceived need to support the price of government securities at a time when the Treasury had a continuing need to float debt. Monetary policy had thus become an instrument of Treasury debt management. By its 11 September meeting, the FOMC recognized that the economy was slowing, which policymakers attributed to an unusually large seasonal downturn rather than to a recession. Goldenweiser forecast that the excess reserves of New York City banks would soon decline to zero and suggested that a target of $250 million of excess reserves might be appropriate to maintain a neutral monetary policy. After reviewing their options, the committee agreed that Eccles would propose a policy to Morgenthau in which the Treasury would desterilize gold and the Fed would purchase securities. At a continuation of the FOMC meeting on 12 September, Eccles reported Morgenthau’s enthusiasm for the plan, whereupon it was agreed that the Fed would purchase up to $300 million of securities in conjunction with a desterilization operation of $300 million by the Treasury. The Fed‘s close cooperation with (or capitulation to) the Treasury continued over the autumn and into early 1938. At a meeting of the FOMC on 21 April 1938, Eccles indicated that the Board of Governors had agreed to lower reserve requirements as part of an administration plan for economic stimulus (which also involved desterilization of gold). Goldenweiser explained that the result of the joint action would be to restore excess reserves to a level above where they had been when the first increase in reserve requirements was made in

42

Charles W. Calomiris and David C. Wheelock

1936, but at that time the economy had been “booming,” while now there was a “depression” with no end in sight (FOMC, Minutes 21 April 1938, 10). As the details of monetary policy discussions during the mid-to-late 1930s make clear, the Fed faced considerable pressure from the Treasury. Removal of the treasury secretary from the Federal Reserve Board did not eliminate administration influence on monetary policy, while the restructuring of the FOMC probably increased such influence. Perhaps the clearest indication of how this affected policy is revealed by the views of George Harrison, president of the Federal Reserve Bank of New York, who headed the FOMC before 1935 and became its vice chairman in 1935. Between 1930 and 1933, Harrison frequently favored more expansionary open market operations than were accepted by the full committee. During 1936 and 1937, he often disagreed with Eccles and other FOMC members who favored a more conciliatory response to the treasury secretary’s call for using open market operations to stabilize the government securities market. Then, during 1938 and 1939, Harrison again disagreed when a majority of the FOMC so feared the appearance of tightening that they opposed letting the Fed‘s portfolio decline as securities matured. We believe that Harrison’s shift from being an FOMC “dove” to being a committee “hawk” reflected not a change in his views, but rather changes in committee membership caused in large part by the restructuring of the Federal Reserve System. Administration pressure on the Fed also helps explain a shift in the focus of monetary policy away from the markets for private securities, such as commercial paper and bankers acceptances, toward the market for government securities. The depression brought a decline of private economic activity and increased government spending that made the market for government securities larger, both in absolute terms and relative to private markets. More important, the Glass-Steagall Act of 1932 and the absence of a gold standard constraint left monetary policy free to monetize government debt, while diminished Federal Reserve independence made monetization more likely to occur. In effect, deficit monetization was the outcome of the Fed’s new policy of stabilizing government securities prices during the late 1930s. The maintenance of high bond prices then became the only policy objective during World War 11. 1.4 Federal Reserve Monetary Policy and the Accord With American entry into World War 11, the Federal Reserve announced that it would peg the yield on Treasury bills at 3/8 percent and maintain ceiling yields on government securities of longer maturities of up to 2.5 percent on bonds. The Fed stood ready to buy securities without limit to maintain this pattern of rates, and by June 1946 the system held 85 percent of all outstanding Treasury bills (Federal Reserve Bulletin, December 1948, 1400). The Fed ended its T-bill yield peg in mid-1947 but continued to maintain

43

The Great Depression as a Watershed for American Monetary Policy

ceiling yields on other government securities. Between 1947 and 1949, the Fed was able to maintain this policy without excessive growth of bank reserves or the money supply, but by 1950 inflationary pressures caused the Fed to question the wisdom of holding government security yields artificially low. The Treasury, however, remained firmly committed to funding government debt issues at low cost, a commitment that only hardened with growing tensions in Korea. Increased government spending, full employment, and the necessity of furnishing bank reserves at an increasingly rapid rate to maintain ceiling yields on government securities prompted concern at FOMC meetings in 1950. On several occasions, the Fed communicated to Treasury officials its view that the Treasury should seek “nonbank” sources for funding their debt issues. In reply, the secretary of the treasury always made clear his view that the Fed‘s support of government securities prices was crucial to the financing of federal government spending. Fed officials became increasingly frustrated with Treasury intransigence, and some argued that it was time for the Fed to act independently. For example, at a FOMC meeting on 18 August 1950, Marriner Eccles, who had been replaced as chairman of the Board of Governors in 1948 by Thomas McCabe, argued that he “felt it was time the System, if it expected to survive as an agency with any independence whatsoever, should exercise some independence” (FOMC, Minutes 18 August 1950, 12). On 21 August, the FOMC took Eccles’s advice and decided to allow a rise in short-term interest rates along with a one-quarter percentage point increase in the discount rate to 1.75 percent. This infuriated the Treasury as the Fed‘s action came during a major debt-refunding operation. Treasury Secretary John Snyder described the resulting decline in demand for government debt as “a significant financial failure for the Federal Government” (Snyder 1969, 305 1). In subsequent FOMC meetings, Fed officials debated whether to continue to buck Treasury policy, with McCabe telling the committee on 11 October that they faced “one of the most important decisions [the FOMC] had been called upon to make” (FOMC, Minutes 11 October 1950,5-6). The Fed continued to argue with Treasury officials that rapid growth of bank reserves and money would cause inflation and do more harm to the market for government securities (via the Fisher effect) than modest increases in market yields attendant on the withdrawal of Fed price support. These arguments fell on deaf ears. As the treasury secretary’s account of the struggle makes clear, the Treasury and the Fed disagreed about the potential problems of continuing securities price support and the likely consequences of removing Fed support. The Treasury interpreted the Fed’s statements of concern about inflation as a desire to pursue contractionary monetary policy and, in effect, raise real interest rates. The Fed’s main concern, however, was less the current money supply than the money supply process. The Fed felt that automatic conversion of Treasury securities into money at a fixed price would lead to an unsustainable process

44

Charles W. Calomiris and David C. Wheelock

of monetization and inflation, which ultimately would lead to higher nominal interest rates.” The impasse finally came to a head early in 1951. Following a meeting with the FOMC on 3 1 January, President Truman issued a statement that the FOMC had agreed to cooperate fully with the Treasury to maintain the current level of interest rates. In fact, the FOMC had made no such commitment.According to the internal Fed account of the meeting, except for asking for the Fed‘s assurance that it shared his goal of maintaining confidence in the government’s credit, Truman seemed unconcerned about minor fluctuations in bond prices (FOMC, Minutes 3 1 January 1951).18The White House press release, however, stated that the Fed had agreed to hold government security prices at their current levels as long as international tensions remained high. Fed officials were much distressed by the White House account of their meeting with the president and again communicated to Treasury officials that the best way to protect the creditworthiness of the government was to prevent inflation. Fed protests led to the formation of a joint Fed-Treasury committee to work out their differences on policy matters. In early March 1951, the Treasury and the Fed reached an agreement, the “accord,” which explicitly recognized the Fed‘s independence and allowed some freeing of government security yields to find their market levels. The Fed‘s vague commitment to “support” the government securities market suggests that the Fed gave up little to win its independence. Underlying the Fed’s new independence was the growth in its balance sheet and a consequent reduction in the potential importance of Treasury threats. Open market purchases during World War I1 and in the postwar period had given the Fed considerable scope to alter the level of bank reserves, with Federal Reserve credit outstanding rising from some 16 percent of total member bank reserves in January 1941 to well in excess of total reserves by war’s end. Thus, the accord may have merely reflected the Fed‘s power to forge a more independent policy whether or not the administration approved. The agreement, however, was not entirely one sided. Coincident with the accord, the Truman administrationinformed Thomas McCabe, chairman of the Fed’s Board of Governors, that he was no longer “effective,” in effect calling for his resignation. McCabe’s replacement was William McChesney Martin, who had been the Treasury’s chief negotiator on the accord. Moreover, as we will show, throughout the 1950s and 1960s the Fed accepted the Treasury’s call for low and stable yields on government securities as one objective of its discretionary monetary policy. 17. This discussion draws on Krooss (1969, 3026-71). For a theoretical discussion of the Fed’s concerns, see Eichengreen and Garber (1991). 18. According to this account of the meeting, Truman had explained how, as a young veteran of World War I, he had been forced to cash in his Liberty Bonds at 80 cents on the dollar to buy necessities, and “he did not want the people who hold our bonds now to have done to them what was done to him” (FOMC, Minutes 31 January 1951,24-25).

45

The Great Depression as a Watershed for American Monetary Policy

In the postaccord period, however, FOMC discussion and policy statements came to place increased emphasis on the growth of reserves and money, which Friedman and Schwartz (1963,628) refer to as a “near-revolutionary” change in system policy (see also Ahearn 1963). Perhaps this development came in recognition of how rapid increases in bank reserves and the money supply can cause inflation, or with the Fed‘s desire to make clear its intention to forge an independent policy. But, immediately after the accord, the Fed also sought to restore the workability of the policy strategies that had guided system operations in the 1920s. As we will show, during the 1950s the focus of monetary policy targeting moved increasingly away from the behavior of total reserves and the money supply and toward the cost and availability of credit. Thus, despite new appreciation of the relationshipbetween growth rates of the money supply and price level, monetary policy during the 1950s and 1960s was largely consistent in terms of goals and strategies with the Fed’s long-standing behavior. That policy, pursued in a post-1933 environment of weakened political independence, a large outstanding government debt, a relaxed gold constraint, and the new Keynesian ideology, accounts for the inflationary burst that began in the mid-1960s.

1.5 Did the Great Depression Cause the Great Infiation? We now explore the fundamental causes of the accelerating inflation of the 1960s and 1970s. We place special emphasis on two aspects of the monetary regime: (1) the Fed‘s operating strategy, which tended to exacerbate, rather than counteract, movements of the price level, and (2) the absence of effective long-run constraints on monetary policy, which allowed errors in operating strategy to persist. The Fed‘s operating strategy, which used free reserves and nominal shortterm interest rates as policy instruments, was not significantly altered by the Great Depression. But the tolerance for long-run policy errors as a result of relying on that operating strategy had increased. Most important, the international monetary regime changed substantially both during the 1930s and after World War 11. The gold standard provided a source of discipline on monetary policy, which helped ensure long-run stability of the price level. Monetary policy that was inconsistent with the long-run maintenance of price stability would cause an outflow of gold. If gold outflows were not halted by contractionary policy, ultimately the monetary authority would have to abandon gold convertibility. Under the Bretton Woods dollar standard, inflationary monetary policy that was not consistent with a long-run fixed exchange rate could (and did) continue for decades before finally producing the inevitable collapse. Expanded power to monetize government debt also contributed toward inflation in the long run. The Glass-SteagallAct of 1932 relaxed a key constraint on money supply growth by permitting the partial backing of Federal Reserve

46

Charles W. Calomiris and David C. Wheelock

liabilities with government securities. Although a gold reserve requirement remained, the new authority in effect promoted monetization of government debt. This encouraged increases in government deficits and their monetization, as government officials pressured the Fed to maintain low yields on government securities. Even after the accord, the Fed continued to place a great deal of emphasis on maintaining “stability” in the market for government securities, which translated in practice into monetizing the deficits of the 1950s and 1960s. We argue that the history of deficits and Fed reactions to them during the 1950s and 1960s is at least consistent with the view that money supply growth, and perhaps the deficits themselves, would have been lower if the Fed had lacked the power to monetize government debt. Two other byproducts of the Great Depression also helped to remove longrun constraints on inflationary monetary policy. The diminution of the Federal Reserve’s political independence and a concomitant increase in the power of the Treasury contributed to the policy of deficit monetization by giving greater influence to Treasury objectives in Fed policy making. Another outcome of the depression was a new macroeconomic ideology that called for the active use of fiscal and monetary policy to achieve full employment and maximum growth of real output. Our argument that the institutional and ideological legacy of the Great Depression mattered for the history of postwar inflation presumes an implicit counterfactual argument-that similar changes would not have come about in the absence of the depression. Counterfactual institutional history is treacherous ground, and reasonable people can differ about what the world would have looked like by the 1960s if the depression had not occurred. Nevertheless, we believe that the depression produced unique changes that would not have taken place otherwise. While one could plausibly argue that the gold standard would have been suspended during World War 11 (as in previous wars) even in the absence of the depression, that is largely beside the point of our argument. Wars had always produced temporary departures from fixed exchange rates, but the end of wars have brought a return to fixed exchange regimes (Calomiris 1991; Bordo and Kydland 1995). The international departure from gold in 1931 and 1933 was unique because it was associated with a growing belief that the gold standard was a pernicious institution that had contributed to the severity of the depression. The depression produced not only a departure from gold but a departure from the classical ideology of gold. The ideology of gold-including the view that long-run price stability warranted the sacrifice of control over short-run macroeconomic fluctuations-had withstood over a century of peacetime recessions and wartime inflations but now crumbled under the unprecedented economic collapse of the 1930s. In its place, the new Keynesian ideology of government intervention to manage business cycles would now reign. Thus, the depression can be credited with producing a permanent departure from a hard-money standard that would not have occurred otherwise.

47

The Great Depression as a Watershed for American Monetary Policy

Similarly, it is likely that permanent monetization of deficits would not have been produced by World War I1 alone. No doubt special temporary laws or other measures would have produced wartime monetization during World War 11, as in World War I. But the expansion in the Fed‘s power to monetize deficits in 1932 was important because it did not originate in war and thus would not disappear with the end of war. It reflected a new view that monetary powers should be broadened permanently-a view that reflected the perceived need to strengthen the Fed’s hand against deflation. Finally, it seems unlikely that the diminution of Fed power that was produced by the 1935 restructuring of the Fed would have occurred without the Great Depression. Major restructurings of institutions are typically byproducts of a crisis mentality-the perception that something is drastically wrong. Absent the depression, we doubt it would have been so easy for the Treasuly and the president to obtain their new powers over monetary policy. In the remainder of this section we first document the return to the Fed‘s traditional operating strategy after the accord. Then we show how the absence of long-run constraints on policy allowed the Fed to stumble into the unintended inflationary surge of the 1960s. 1.5.1 A Return to Traditional Instruments: Interest Rate and Free Reserves Targeting

Following the accord, the Fed placed new emphasis on the growth of bank reserves and money supply as sources of inflation, in both public statements and internal discussion. In formulating an operating strategy in the postaccord era, however, the Fed returned once again to a strategy of using open market operations to affect the level of discount-window borrowing and excess reserves. In the Board of Governors’ annual report for 1953, the Fed stated that “during the preceding two years, the Federal Reserve has moved toward greater reliance on influencing the cost, availability, and supply of credit through the discount mechanism, that is by making it necessary for member banks to borrow from the Federal Reserve Banks a portion of the additional reserves required to meet credit growth. This mechanism limits credit expansion, puts pressure on banks, and makes them more responsive to changes in the discount rate” (quoted in Ahearn 1963, 122). And, the April 1953 Federal Reserve Bulletin indicated that “open market operations and the discount rate are again being used . . . as twin reserve banking measures. . . . Open market operations can be employed when needed to condition the current tone in credit markets and the general availability of credit. By these operations the Federal Reserve can tighten or ease the pressure on member bank reserve positions and thus cause banks to borrow or enable them to reduce borrowings at the Reserve Banks. Subsequently, this tightness or ease is transmitted and magnified in money and credit markets” (quoted in Ahearn 1963, 123). These statements are virtually identical to operating guidelines explained by Benjamin Strong and other Fed officials in the 1920s. For example, in testimony in 1926, Strong indicated that “it is a more effective program . . . to

48

Charles W. Calomiris and David C. Wheelock

begin to sell our government securities. It lays the foundation for an advance in our discount rate. . . . The purchase of securities eases the money market and permits the reduction of our discount rate” (Strong 1926, 332-33). Moreover, “the influence of the reserve banks upon the volume of credit is . . . felt not directly, but indirectly through the member banks. The reserve banks do not ‘push’ credit into use” (468). Winfield Riefler, who was to be the FOMC secretary during the 1950s, explained how the Fed could affect monetary conditions by influencing the volume of discount-window borrowing: “Various monetary factors-such as gold movements, changes in currency demand, and open-market operations by the Reserve Banks . . . determine the volume of indebtedness [of member banks to the Reserve Banks]. . . . And changes in this indebtedness appear to be the initiating force in corresponding changes in money rates” (1930, 27). According to this view, banks are reluctant to borrow from the Reserve Banks, and when forced to borrow, banks will sell securities, call loans, and otherwise reduce credit. Strong argued that “experience in the past has indicated that member banks when indebted to the Federal Reserve Bank of New York . . . constantly endeavor to free themselves from the indebtedness, as a consequence such pressure as arises is in the direction of curtailing loans” (quoted in Chandler 1958, 239). The similarity of these views with those outlined by Fed officials in the 1950s illustrates that at an operational level, monetary policy was little changed between the 1920s and 1950s. As in the 1930s, Federal Reserve officials in the 1950s sought to make operational the tools and strategies the Fed had used in the 1920s. Their goal was to reestablish a regime in which free reserves served as the operating target, with market interest rates and the availability of credit being the intermediate objectives of policy. At FOMC meetings in early 1952, one topic of discussion was how to define monetary policy “neutrality.” The definition of a neutral policy put forward at the meeting of 21 April was a policy “under which System operations endeavor to maintain a situation where there is a moderate amount of borrowing at the Federal Reserve Bank [sic]” (FOMC, Minutes 21 April 1952, 15-16). New York Fed President Allen Sproul went on to argue that “a borrowing level of $400 million and a high level of excess reserves was not the same as an equal amount of borrowing with no excess reserves” (16). In other words, the key variable to focus on was not discount-window borrowing, but rather the difference between excess reserves and borrowed reserves, that is, free reserves. As Fed Chairman Martin put it, free reserves “had to be used as an indication, for that was the framework within which the [FOMC] Account Manager had to work” (FOMC, Minutes 6 May 1958, 52). By early 1953, the Fed had become sufficiently concerned about inflation to increase the discount rate for the first time since 1950 and to carry out open market transactions to reduce free reserves. Despite the accord, Fed officials remained sensitive to the administration’s desire to keep the rates on government securities relatively low and stable, particularly near issuing dates. Gen-

The Great Depression as a Watershed for American Monetary Policy

49

q I I

,I.!.

CPI

. .I

._

..I

erally, the Fed sought to avoid substantial fluctuations in T-bill yields as issue dates approached, a policy officials referred to as maintaining an “even keel.” At the 6 January 1953 FOMC Executive Committee meeting, policymakers decided to tighten, however, even though a Treasury issue was forthcoming. Chairman Martin argued for this policy because he saw “danger signs of speculation . . . not least of which was activity in the stock market itself” (FOMC, Minutes 6 January 1953,5-6). How reminiscent this was of policy during 1928 and 1929, and again during 1936 and 1937, when the Fed also manipulated free reserves to quell perceived speculation. The behavior of free reserves throughout the 1950s and 1960s is consistent with its use by the Fed in response to large movements in prices and real output. For illustration, we plot in figure 1.1 the growth rate of the consumer price index (CPI) and the level of free reserves (both smoothed using a 13-month centered moving average filter) from the accord (March 1951) through 1969. The Fed acted to increase free reserves in 1954 when industrial production declined and the price level held constant. Free reserves were brought down as the economy recovered and inflation rose. Free reserves were increased in 1958 with another cyclical downturn, lowered in 1958 and 1959, and increased during the recession of 1960. During the 1960s, free reserves generally fell, with two notable exceptions (see fig. 1.1). Those exceptions were the rises between 1966 and 1967 and between 1969 and 1970. The latter reflected the Fed’s response to a distinct slowing of economic activity. The 1966 increase reflected the Fed’s response to a “credit crunch,” which, according to some Fed critics (including President

Charles W. Calomiris and David C. Wheelock

50

a,

M I (Annualized Percent Change)

Free Reserves (millionsof $) ,1500

0

-750

-8

1951

1953

1955

1957

1959

1961

1963

1965

1967

1969

Fig. 1.2 Free reserves and money supply growth Source: Board of Governors of the Federal Reserve System.

Johnson), had occurred because of an excessively tight monetary policy. More important, however, than these fluctuations were the declining trend of net free reserves and the rising trend of interest rates. These trends created the impression within the Fed that monetary policy was neutral or contractionary-that the Fed was “leaning against the wind”--when in fact policy was quite inflationary. Academic critics of the Fed, especially monetarists, argued that monetary policy was inflationary because the money supply growth rate generally accelerated over the decade. Figure 1.2 shows that the growth rate of demand deposits and currency held by the public (Ml), accelerated during the 1960s even as the level of free reserves de~1ined.I~ By 1965, inflation had risen to 4 percent, and Fed officials considered anew the stance of policy. At the FOMC meeting of 12 October, Chairman Martin expressed considerable worry about inflation but argued against a tighter policy: “With a divided [open market] committee and in the face of strong Administration opposition he did not believe it would be appropriate for him to lend his support to those who favored a change in policy now” (FOMC, Minutes 12 October 1965,69). By the November FOMC meeting, Martin was even more convinced that “the country was in a period of creeping inflation . . . and the balance of payments situation would be benefited by more restraint in the overall economy. In short, he thought the economy was growing too fast at the moment” 19. Meigs (1962) and Brunner and Meltzer (1964) were among the earliest critics of the free reserves interpretation.

51

The Great Depression as a Watershed for American Monetary Policy

(FOMC, Minutes 23 November 1965, 85). On 5 December, the Federal Reserve discount rate was increased from 4 percent to 4.5 percent. Had the Fed finally moved to stop inflation? Not according to one member of the Board of Governors, Sherman Maisel, who argued the following in May 1966: When members looked at total reserves or nonborrowed reserves, either of which he took to be the principal measure of the [open market] committee’s actions, they must be appalled at the committee’s results. . . . In the five months since December 1 [1965], the committee had poured more reserves into the banking system than were furnished in the entire previous year. In fact . . . since December 1, the total increase in reserves had been larger than in any full year since 1951. . . . These results did not accord with either the committee’s intent, its statements or sound policy. . . . The committee apparently had followed sub-goals such as feel of the market, net [free] reserves, or the need to offset shocks, and as a result it had moved in a direction opposite to its real policy aim. (FOMC, Minutes 10 May 1966, 62-63) Martin responded to this criticism with the view that “he did not think the committee should be too dissatisfied with monetary policy because of the recent trends in banking aggregates. . . . The Committee should not press too hard in the belief that monetary policy alone could achieve price stability. . . . Perhaps all members of the Committee-and he included himself-tended to think at times that monetary policy could do more than it in fact could.” Interestingly, Martin also expressed the “hope that the System would not find itself in the position of having raised the discount rate after the crest of the cycle had been passed. If it did, it was likely to bear all the blame for subsequent developments.” Martin went on to recall the Fed‘s moves to tighten policy in 1957, and how the subsequent recession had been blamed on the Fed. Martin then cautioned that “he thought the committee should bear that in mind” (FOMC, Minutes 10 May 1966,94-95). As Maisel succinctly explained to the FOMC in May 1966, the inflationary surge had been the result of poor targeting rules, not an inflationary intent by the Fed. Fed officials believed they were pursuing appropriate policies to contain inflation and assist in maintaining equilibrium in the balance of payments but had stumbled into inflation because of their attachment to free reserves targeting and misinterpretation of rising interest rates. Despite the forceful arguments of Maisel and a few other members of the FOMC, the Fed did not reject the use of free reserves as its operating target (or nominal interest rates as a policy guide).20In our view, the acceleration of 20. The Fed never used a free reserves operating target mechanistically-discretion always was important, and by the late 195Os, doubts about the reliability of free reserves led the Fed to adopt a more eclectic approach to policy. Much as the Fed appears to be groping for policy guides today, in the 1960s Fed policymakers discussed an array of potential candidates, settling on none as paramount. This lack of specificity is reflected in a discussion between a Reserve Bank president and the FOMC account manager in 1959. The president remarked that he “did not know precisely

52

Charles W. Calomiris and David C. Wheelock

inflation in the mid-1960s cannot be attributed plausibly to perceived or actual shifts in the Fed’s taste for inflation. Clearly, the FOMC did not intend to increase inflation when it permitted the money supply to rise before 1965, or when it stepped once again on the monetary gas peddle in 1967. Declining levels of free reserves during the 1960s reflected Fed efforts to prevent inflation. Similarly, it is important to note that nominal interest rates rose sharply, roughly doubling from 1962 to 1966. The inflationary surge of the mid-1960s thus was not the consequence of a rigid adherence to a nominal interest rate target like that of the World War I1 era.*’ How much of the onset of the Great Inflation of the 1960s and 1970s can be attributed to the impact of the Great Depression? As we have shown, the depression was not a watershed in terms of the Fed‘s fundamental objectives and targeting strategy. Under Benjamin Strong’s leadership, by 1924 the Fed was using open market operations and discount rate adjustments to pursue both domestic and international goals. At an operational level, the Fed targeted the borrowed reserve positions of member banks and judged the stance of its policy by the levels of member bank borrowing and market interest rates. Over time, the Fed became more cognizant of the behavior of the supply of money, but throughout the 1950s and 1960s it never rejected free reserve or interest rate targeting in favor of a monetary aggregate operating strategy. Consequently, as in the 1920s and early 1930s, in the 1950s and 1960s monetary policy tended to exacerbate swings in economic activity and the price level, which explains to a great extent how the Fed “stumbled” into an inflationary monetary policy during the early 1960s. 1.5.2 The New International Monetary Regime The Bretton Woods system represented a substantially new international monetary regime in which excessive U.S. money growth (and concomitant external imbalance) could continue for several years. Had the Fed been subject what was meant by an even keel policy. Should it be measured by net free reserves, net borrowed reserves, the feel of the market, or the intuition of the Account Manager?” The manager replied that he “thought it was a mixture of the things [the president] had mentioned” (FOMC, Minutes 5 May 1959, 45). On another occasion, Chairman Martin argued that when carrying out policy, “reference should be made primarily to the color, tone, and feel of the [government securities] market . . . rather than to a specific level of free reserves” (FOMC, Minutes 11 September 1962,48). 21. We differ with Goodfriend (1991). who argues that the Fed used free reserves to disguise its true policy of controlling interest rates during both the 1920s and in the postaccord era. Free reserves was regarded by Fed officials as their appropriate operating target and also as a useful independent signal of conditions in the money market. Goodfriend is correct, however, that the Fed‘s policy regime was fundamentally nonmonetarist and permitted money supply growth to exacerbate fluctuations in the growth rate of nominal GDP. Another view, offered by Chari, Christiano, and Eichenbaum (1996), is that the inflationary surge was the product of self-fulfilling expectations of higher inflation by the public in an interest rate targeting regime. This seems implausible, however, because of Caskey’s (1985) finding that public perceptions of inflation persistently lagged actual inflation during the 1960s and 1970s. In other words, in the mid-l960s, expected inflation remained consistently lower than actual inflation.

53

The Great Depression as a Watershed for American Monetary Policy

to the discipline of the gold standard as it existed before 1933, inflationary money growth could not have continued for so long, and it would not have been possible for the Fed to stumble into the inflation of the 1960s. The Bretton Woods system operated as a set of gold-dollar fixed exchange rate standards. The system began in 1946 and came fully into place as a worldwide monetary system in 1959. Under the Bretton Woods system, the United States was the only country that maintained a commitment to convertibility into gold, via the “gold window.” Other countries pegged their currencies to the dollar. Bretton Woods was thus like the interwar gold exchange standard in that most countries held their international reserves in the form of the currency of the country or countries at the core.22The mechanism for maintaining gold convertibility under Bretton Woods, however, was fundamentally different from that of the interwar gold exchange standard. It was this mechanism that explains how inflation could persist in the United States during a period in which a fixed exchange rate between the dollar and gold was maintained. Neither the interwar gold exchange standard nor the Bretton Woods system was a pure gold standard-both permitted some discretion. Under Bretton Woods, however, the mechanism by which the balance of payments exerted discipline on central banks was largely removed from private markets and given to central banks themselves. This mechanism reflected a fundamental shift in ideology, from one that saw maintaining gold convertibility as paramount for long-run prosperity to an ideology that viewed fixed exchange rates and gold convertibility as desirable but not so important as to warrant the sacrifice of short-term economic stability in defense of the international system. Discretionary monetary policy-“managed money”-was permitted under Bretton Woods to a degree never before achieved under a gold standard. Under Bretton Woods, a balance-of-payments deficit (surplus) for the United States would be reflected in a buildup (contraction) of foreign central bank holdings of U.S. dollars, unless the foreign central banks and the United States exchanged dollars for gold. The Bretton Woods system imposed only limited constraints on member countries (including the United States) to pursue monetary and fiscal policies that were consistent with a long-run commitment to a fixed exchange rate. While foreign central banks could have enforced monetary discipline on the United States by converting dollars into gold, in practice they refrained from doing so until 1965, when the French began to convert dollars into gold in the face of large and persisting American balanceof-payments deficits. By that time, a large dollar imbalance had been created by expansionary U.S. monetary policy. The relative world supplies of gold and dollars trended in opposite directions over this period, and clearly a fixed dollar-gold exchange rate was not sustainable under these circumstances. In 22. Redish (1993) argues that Bretton Woods was thus an “extension” of the interwar gold standard, which, in turn, was one step on a long path delinking the world’s currencies from gold (with Bretton Woods being the last). See Bordo and Eichengreen (1993) and Eichengreen (1992) for comprehensive treatments of the evolution of the Bretton Woods system.

54

Charles W. Calomiris and David C. Wheelock

the event, the Bretton Woods system collapsed from 1971 to 1973 following closure of the gold window on 15 August 1971. As we have seen, the Fed‘s reaction to domestic inflation, persistent U.S. balance-of-payments deficits, and French demands for redemption of dollars caused only a brief interruption in the upward trend in money supply growth in the mid-1960s. Bordo (1993,57) argues that part of the reason for the Fed’s lack of sustained vigor was the widely held belief that the Bretton Woods system could be sustained without monetary discipline. Influential economists argued that monetary expansion could continue, while the balance-of-payments problem could be addressed by other policies (Bordo 1993, 58-59; Yeager 1976, 567-76). Federal Reserve officials became increasingly concerned about the balance of payments in the late 1950s. But Fed officials were also wary of combating a balance-of-payments deficit with policies that might interfere with other goals. For example, Alfred Hayes, president of the Federal Reserve Bank of New York, argued that “I would think it unwise to let the gold outflow itself affect our monetary policy directly, i.e., in the way of using a tightening move directed specifically toward stemming the flow and unrelated to domestic economic developments” (FOMC, Minutes 10 November 1958, 14-15). On another occasion, a Reserve Bank president expressed concern about the balanceof-payments deficit but was reluctant to advocate a tighter policy for fear of disrupting the market for government securities: “Generally, he felt that the course of monetary policy should be moving toward a more restrictive posture. At the same time, he was quite concerned about the rate picture in the government securities market and the problems facing the Treasury in the future” (FOMC, Minutes 5 May 1959, 34). This reluctance to face squarely gold outflows and a balance-of-payments deficit stands in marked contrast to the Fed’s swift reaction to gold outflows in 1931. At that time, Fed officials agreed that maintaining convertibility of the dollar into gold at a constant price was fundamental to long-run economic stability, and they were willing to tighten monetary policy in the middle of a depression in order to preserve the international monetary regime. By contrast, in the 1950s and 1960s, Fed officials viewed the balance of payments with concern but hesitated to make it the sole, or even the primary, focus of policy. This change in philosophy, attaching less importance to the gold standard rule and more to discretionary policy was an important legacy of the Great Depression. Although Fed officials were unwilling to tighten sufficiently to arrest the balance-of-payments deficit, they did see the deficit as influencing their ability to promote domestic economic activity. Chairman Martin, for example, argued that “if the Federal Reserve got the reputation of following a cheap money policy just for the sake of doing so, people abroad would be encouraged to think the System was not concerned with the balance of payments or the soundness of the dollar” (FOMC, Minutes 13 December 1960,40). Martin also ar-

55

The Great Depression as a Watershed for American Monetary Policy

gued that “the balance of payments problem . . . was a vital factor in the unemployment situation. Foreign capital was finding the United States less and less attractive, there were pressures for movement of capital abroad, and this was having a deleterious effect on employment in this country” (FOMC, Minutes 6 March 1962,56). Martin often spoke of the international payments deficit as the most important problem facing the Fed. As often, however, he seems to have found excuses for not taking stronger action. For example, in November 1964 Martin expressed a preference for a tighter policy because of the worsening balanceof-payments deficit and the potential for increased inflation. He was unwilling, however, to advocate a tightening without broad committee support. Martin also pointed out that a new Treasury issue was forthcoming, which generally called for an even keel policy. Reflecting the frustration of pursuing multiple objectives with a single tool, New York Fed President Hayes then complained that “there seemed never to be a right time to make a difficult decision, and that the need for maintaining an even keel during Treasury financings inhibited action by the Committee much of the time” (FOMC, Minutes 10 November 1964, 82). Fed officials also understood that the balance-of-payments deficit stemmed from differences in the macroeconomic policies of different countries. At an FOMC meeting in 1959, a Fed staff member reported that “the net result of attempts in this country to validate our wage and price policies through monetary expansion could succeed only if we could inflate the whole world.” The staff member went on to argue that expansionary monetary and fiscal policy could “price United States’ goods out of world markets” because officials of other countries, notably Germany and the Netherlands, surely would not permit inflation in their domestic prices (FOMC, Minutes 5 May 1959, 14). The same official, however, was unwilling to blame monetary policy alone for the balance-of-payments deficit. In arguing that gold outflows “call for a generally restrictive credit policy . . . more effective corrections . . . would be moves to reduce the budgetary deficit and the checking of price rises due to wage and other cost increases” (FOMC, Minutes 21 October 1958, 17-18). The Fed‘s unwillingness to tighten sufficiently to stem the balance-ofpayments deficit led it to consider other actions it might take. One of the earliest of the policies intended to restore external balance was Operation Twistan attempt to twist the term structure of interest rates, to keep short-term interest rates high enough to attract foreign capital while keeping long-term interest rates low enough to favor domestic expansion. This unsuccessful policy is now a popular textbook example of government failure to understand the concept of arbitrage in financial markets. In fact, Fed officials were skeptical about Operation Twist from the beginning. Martin, for example, believed that “when an attempt was made to determine the short rate against the long rate except for a very short period of time,

56

Charles W. Calomiris and David C. Wheelock

the System would be in trouble” (FOMC, Minutes 22 November 1960, 42). Nevertheless, Martin believed that the tactic was worth trying (see FOMC, Minutes 20 August 1963,58). Another Fed initiative to combat the balance-of-payments deficit and the weak dollar was direct intervention in foreign exchange markets. The Gold Reserve Act of 1934, which established the Treasury’s Exchange Stabilization Fund, had been interpreted by counsel to the Board of Governors as precluding Federal Reserve intervention in foreign exchange markets, except as agent for the Treasury. In 1961, the Fed reconsidered the efficacy of intervention and whether it had the legal standing to intervene on its own. This time, the board’s counsel argued that the Fed did have the right to intervene. Concern on the part of individual FOMC members, however, led the Fed to seek the advice of the Treasury and the banking committees of Congress. The treasury secretary wrote to the FOMC that “it is surely a proper central banking function to engage in temporary operations that will help to buffer and moderate tendencies toward volatile flows of funds” (FOMC, Minutes 19 December 1961, 85-86). With that endorsement, the FOMC began its independent intervention in foreign exchange markets. Other policies intended to correct international payments imbalances without slowing domestic activity included new agreements with foreign central banks to forebear from demanding gold, the issuance of foreign-currency-denominated U.S. bonds (“Roosa bonds”), requests of early repayment by foreign governments of debts to the U.S. government, the removal of interest rate ceilings on U.S. bank time deposits, capital outflow constraints imposed in the United States, and changes in U.S. tax treatment of foreign earnings. Balance-of-payments deficits continued, however, and the long-term feasibility of the existing dollar-gold standard increasingly became a mathematical improbability. Such delay could not have happened under the classical gold standard. That system imposed strict limits on the ability of central banks to ignore external imbalance in the interest of internal growth. It is true that limited and temporary sterilization of gold outflows was possible under the classical gold standard, and there is some evidence that central banks resisted the so-called rules of the game (by which is meant a policy of reinforcing rather than sterilizing the effects of short-run gold flows on money). Nevertheless, a sustained outflow of reserves over the years was not feasible under the discipline of the gold standard. Under Bretton Woods, by contrast, the absence of market-determined gold flows deprived the Fed of an early warning sign of unsustainable money policy and enabled greater monetary policy discretion.

1S.3 Stabilization of Government Security Prices Although important, the Fed’s flawed policy framework and the absence of gold standard discipline cannot account entirely for the monetary policy outcomes of the 1960s and 1970s, especially for the Fed‘s unwillingness to quell

57

The Great Depression as a Watershed for American Monetary Policy

inflation once it had clearly begun. Another contributor to these outcomes, which did represent a break with predepression policy, was the Fed‘s policy of stabilizing the prices of government securities. This policy was begun in the mid-I930s, became the sole objective of policy during World War 11, and continued even after the 1951 accord. The effect of this policy in the 1960s was the partial monetization of that decade’s growing fiscal deficits. The Federal Reserve System was established with two guiding principlesthe gold standard and a theory of banking that later became known as the real bills doctrine. Although Federal Reserve Banks were permitted to buy U.S. government securities, the extension of Federal Reserve credit was limited by their holdings of gold and short-term commercial securities acquired mainly by making discount-window Hence, the Federal Reserve Act implicitly limited the extent to which federal debt could be monetized. The Fed first departed from real bills principles during World War I, when it encouraged banks to buy government securities by offering to lend them reserves against their security holdings at a rate that guaranteed the banks a profit. During the war, however, the Fed always held sufficient gold to back its liabilities. Another apparent departure from real bills principles came when the Fed made open market purchases of government securities in the 1920s, but again, the Fed’s gold and commercial paper reserves were never jeopardized. The Glass-Steagall Act of 1932, however, constituted a fundamental departure from the principles underlying the Federal Reserve Act by expanding the permissible assets backing Fed liabilities to include government securities. That legislation, and its permanent replacement in 1933, substantially weakened the checks on Federal Reserve monetization of government debt. As we have seen, the Fed came under considerable pressure to maintain low yields on government securities in the mid-l930s, and this policy became paramount during World War 11. Whether or not the Fed’s policy of maintaining ceiling yields on federal government debt during World War I1 was justifiable, the Fed‘s continuing commitment to stabilization of government securities prices after the 1951 accord is puzzling. While the Fed had agreed as part of its accord with the Treasury to “maintain an orderly market” for government securities (Krooss 1969, 3056), such a vague commitment had no real force. Furthermore, in contrast to the 1930s, the Fed was sufficiently large by the 1950s to easily offset any interventions by the Treasury, thus removing any credible threat by the Treasury to force the Fed’s hand. 23. The Federal Reserve was founded in part to provide a mechanism for increasing and decreasing the supply of bank reserves and currency in response to fluctuations in the needs of banks for accommodating the demands of their customers for loans and currency. The real bills doctrine holds that when banks restrict their lending to short-term commercial notes, the supply of credit will be sufficient but not inflationary. Thus, the extension of Federal Reserve credit on the basis of the rediscount of commercial paper will ensure the noninflationary accommodation of credit demand associated with real economic activity. See Wheelock (1991) and the references therein for additional detail about the founding principles of the Fed.

58

Charles W. Calomiris and David C. Wheelock

What then accounts for the Fed’s continuing commitment to support government securities prices in the postaccord period? Part of the answer lies in the magnitude of Treasury borrowings in the early 1950s, which coincided with the Korean War. The burden of the war grew in the years immediately following the accord. While national security expenditures accounted for only 33 percent of the federal budget in 1950, they reached 67 percent in 1952 (Blum et al. 1973,726). It would have been tantamount to “unpatriotic” for the FOMC to have completely neglected the government’s financing needs at this time of fiscal strain. President Truman made clear to FOMC members his view of their responsibilities in a meeting with them on 31 January 1951: “The President emphasized that we must combat Communist influence on many fronts. He said one way to do this is to maintain confidence in the government’s credit and in government securities” (FOMC, Minutes 3 1 January 1951,24). As during World War 11, Fed officials undoubtedly believed that the national interest would be served by easing the government’s debt management problems during the Korean conflict. Perhaps similar thinking about the role of monetary policy during armed conflicts partly explains the Fed’s bond price stabilization policy during the Vietnam War. Some evidence of this is revealed by meetings between Fed and Treasury officials about securities markets in conjunction with the war’s escalation. For example, in describing a 1965 meeting between Fed officials and Paul Volcker, then deputy undersecretary of the treasury for monetary affairs, the Fed’s account manager reported that the Fed had not been asked to peg the prices of government securities but nevertheless to lend general market “support” (FOMC, Minutes 10 October 1965,26). The view that Korea and Vietnam contributed to deficit monetization, however, does not imply that monetization was an unavoidable outcome of those wars. The scale of the wars was probably not great enough to have produced special wartime monetary powers (deficit monetization or the suspension of the gold standard). If those powers had not been in place prior to the wars (as a consequence of depression-era changes), monetization of government debt might have been largely avoided. Furthermore, in the absence of an automatic mechanism for monetizing deficits, the Johnson administration likely would have felt more pressure to choose between war expenditures and Great Society programs. It is possible that both the Korean and Vietnam eras would have seen shrinkage in non-warrelated federal expenditures if the post-World War I1 regime had been characterized by a return to the gold standard and the absence of Fed authority to monetize deficits. The desire to placate the Treasury by maintaining low and stable yields on government securities, especially during wartime, was only one reason the Fed continued to attempt to manipulate government securities prices. A second reason, revealed in the minutes of FOMC meetings during the 1950s and 1960s, was that Fed officials viewed the market for government securities as inher-

59

The Great Depression as a Watershed for American Monetary Policy

ently inefficient or unstable. Not only was the policy of even keel near dates of new Treasury issues accepted virtually without question, but Fed intervention seems to have been widely viewed as necessary to preserve stability in the market. Later, a similar view was taken toward the foreign exchange market. While no direct reference is made in the minutes to the experience of the depression, we wonder whether the view that financial markets were inherently unstable and market prices were in need of “management” was itself in part a reaction to the d e p r e s s i ~ n . ~ ~

1S.4 A Diminution of Federal Reserve Independence and a New Economic Ideology Although some New Deal legislation expanded the scope of Federal Reserve powers, other legislation served to lessen the Fed‘s independence. We have argued that diminished political independence contributed to the refocus of monetary policy toward the stability of the government securities market in the 1930s. Although the accord restored a measure of independence to the Fed, the Fed remained under pressure to keep market yields on government securities low and to attempt to engineer higher output and employment in the near term, even at the expense of higher inflation down the road (e.g., see De Long 1995). A monetary policy framework that interpreted rising nominal interest rates as evidence of tighter monetary policy, coupled with pressure to fund federal government borrowing at the lowest possible cost, caused monetization of government debt. Over time, this produced higher inflation. 24. The use of monetary policy to stabilize the market for government debt represented one sharp departure from the Fed’s founding principles and the policies it pursued before 1933. A second departure, also first appearing with the Glass-Steagall Act of 1932, involved collateral requirements for discount-window loans and, later, the uses of Federal Reserve credit. In line with real bills principles, the Federal Reserve Act restricted the collateral for discount-window loans to high-quality, short-term commercial loans. In relaxing the collateral provisions, first in 1932, and then by subsequent legislation during 1933-35, Congress accepted the argument that many banks with sound assets had been denied access to the discount window because they had lacked eligible collateral and that this had worsened the banking collapse and depression. In addition, it was widely believed that inadequate access to credit had forced the bankruptcy of many firms, and so the Fed was authorized to lend directly to individuals, partnerships, and corporations and to make industrial advances of up to five years. Although these lending provisions have never resulted in large extensions of Federal Reserve credit, they represented a fundamental break with the principles underlying the Federal Reserve Act as it was originally conceived. The expanded scope for Federal Reserve lending reflects a view that legal restrictions had limited the Fed’s ability to protect the banking system from collapse during 1929-33. In essence, the Fed had failed to act as lender of last resort, and relaxed discount-window borrowing rules were one attempt to ensure that the Fed would not fail again in its responsibilities for lack of legal authority. In recent years, the Fed has shown considerable willingness to act in the name of preserving financial stability. Examples include its guarantees of liquidity to banks to preserve stability of the commercial paper market in the wake of Penn Central‘s default in 1970, and of financial markets in general following the stock market collapse of 1987. Whether or not such interventions constitute appropriate lender-of-last-resort behavior or are otherwise socially justifiable (for opposing views, see Schwartz 1992; Calomiris 1994), the Fed seems unlikely to allow a collapse of the financial system on the scale of the Great Depression without exerting considerable effort to prevent it.

60

Charles W. Calomiris and David C. Wheelock

By reducing the role of Federal Reserve Bank presidents in favor of the Board of Governors, the Banking Act of 1935 may have contributed to the system’s willingness to stabilize the yields of government securities and certainly subjected the Fed to greater political pressures. Did this make the Fed more willing to attempt fine-tuning operations or to accept higher inflation over the long term in exchange for faster real growth and higher employment in the near term? Or was that policy choice more the product of Keynesian ideology, which prescribed interest rate manipulation to promote growth and employment objectives? Others have written extensively on the new economic ideology born of the Great Depression and associated with Keynes, which increasingly dominated policy discussions, both within and outside the Federal Reserve during the 1960s (e.g., De Long 1995). This macroeconomic paradigm held that monetary policy could increase output and employment without producing inflation so long as resources were not fully employed. Some proponents acknowledged that expansionary monetary policy might increase the rate of inflation but argued that modest inflation was an acceptable cost of faster growth and higher employment. For example, Fed governor Maisel argued that “there is a tradeoff between idle men and a more stable value for the dollar. A conscious decision must be made as to how much unemployment and loss of output must be made in order to get smaller price rises” (Maisel 1973, 14). Another cornerstone of the Keynesian ideology was that inflation was not always the product of monetary policy. As Maisel puts it, “Some price increases originate on the cost side or in particular industries. These cannot be halted by monetary policy” (1973, 14). Accelerating inflation in the late 1960s and early 1970s, he argues, was caused by “government deficits; . . . speculative investment in plant, equipment, and labor by business corporations; . . . use of economic power to raise wages and profits.. . . But most significant were the government deficits” (Maisel 1973, 12). Although not all economists or policymakers held these views, enough did to influence monetary policy. Even today, high-ranking officials of the Federal Reserve System contend that monetary policy can and should be used to maintain full employment ( e g , Blinder 1996). The Great Depression clearly influenced mainstream economic thought in the 1960s, especially among key economic policy advisors. Lucas writes that one of the “the main features of the Keynesian Revolution and the neoclassical synthesis into which it evolved in the United States . . . [was] the onset of the Great Depression and the consequent shift of attention from explaining a recurrent pattern of ups and downs to explaining an economy apparently stuck in an interminable down” (1980, 704). Even though much of today’s mainstream theory rejects many Keynesian ideas, according to Lucas, the Great Depression continues “to defy explanation by existing [new] economic analysis” (1980,706). For macroeconomics, and to some extent for macroeconomic policymakers, the Great Depression was a watershed.

61

The Great Depression as a Watershed for American Monetary Policy

1.5.5

The Final Depression Legacy: Disinflation in the 1980s and Price Targeting in the 1990s

We have argued that various legacies of the Great Depression were important in allowing the Fed’s traditional (and flawed) operating strategy to produce the inflationary surge of the 1960s and 1970s. These include the permanent collapse of the gold standard, debt monetization powers, the decline in Fed independence, and a new macroeconomic ideology that regarded financial markets as unstable and government intervention as desirable. We would emphasize that these various legacies were not equally important in explaining the origins of inflation or its persistence. With respect to the origins of inflation, the collapse of the gold standard was central. By itself, a change in macroeconomic thought could not have changed monetary policy fundamentally, nor could reduced Federal Reserve independence or even the opening of an explicit avenue of debt monetization. The necessary condition that made these other influences important was the replacement of the interwar gold standard with the Bretton Woods system, which allowed price-level drift that otherwise would have been i m p o ~ s i b l e . ~ ~ While the collapse of the gold standard was essential in the origins of inflation, other legacies of the depression were probably more important for explaining the persistence of inflationary policy in the late 1960s and 1970s. That postponement of disinflation until the 1980s imposed large costs on the economy. As Goodfriend (1993) emphasizes in his discussion of the disinflation of the early 1980s, the costs of avoiding inflation are much lower than the costs of disinflating when the central bank’s commitment to end inflation lacks credibility. As Caskey (1985) has shown, inflationary expectations during the late 1960s and early 1970s were consistently below actual inflation. Caskey explains that fact in the context of a Bayesian model of inflation expectations. He argues that the public was slow to grasp the change toward inflation bias in monetary policy. He finds that the public’s view of the Fed‘s reaction function significantly lagged the actual changes because learning about central bank intentions takes time. From that perspective, the Fed probably could have reduced inflation in the 1960s at little cost, relative to the costs borne from 1980 to 1982. The inflation of the 1960s and 1970s, and the economic costs of disinflation in the 1980s, have undermined some of the legacies of the depression. In this sense, the consequences of the institutional legacies of the depression have weakened the long-run (future) importance of the institutional and ideological 25. Unlike Bordo and Eichengreen (chap. 12 in this volume) we do not believe that, in the absence of the depression and the Bretton Woods system, the gold standard would necessarily have collapsed permanently during the postwar period as the result of gold scarcity. Many other counterfactual outcomes seem as plausible to us-including the adoption of bimetallic or symmetallic standards, or other methods for reestablishing a balance between the real reserves of nations adhering to a hard-currency standard and the real supply of reserves.

62

Charles W. Calomiris and David C. Wheelock

changes wrought in the 1930s. Clearly, the new emphasis on price stability as a central short-run concern for monetary policy in the 1990s is something of a throwback to the predepression gold standard. One also sees a rebirth of “longrun” thinking in macroeconomics, including a new emphasis on the need for credibility in long-run control of deficits and in deficit monetization. Even the “cleansing benefits of recessions” is enjoying a rebirth in macroeconomics. Thus, looking forward to the twenty-first century, it may be that the influence of the depression-especially the potential inflation bias produced by depression-era policies-may lie more in the past than in the future. On the other hand, despite the current vigor of the anti-inflation hawks and the decline in the ideological legacy of the Great Depression, it is possible to argue that the most important legacy of the depression remains intact-the destruction of international monetary linkages has not been undone or replaced with new constraints on discretionary policy.26The lack of an institutional check on discretionary money supply growth suggests that inflation bias may again be latent, waiting to reemerge to finance the widely predicted explosion of government deficits in the twenty-first century.

1.6 Conclusion The Great Depression brought little change in Fed thinking about the appropriate goals, targets, and instruments of domestic monetary policy. At least as late as the 1960s, Fed behavior was much like that of the 1920s. The free reserves operating framework, which included the use of nominal interest rates as a policy guide, contributed both to the economic collapse of the depression and to the inflation of the 1960s and 1970s.In this sense, the Great Depression did not change monetary policy. The Great Depression did result in important institutional and ideological changes, however, and these changes made possible the sustained inflation of the 1960s.The fundamental change was the abandonment of the interwar gold standard and its subsequent replacement by the Bretton Woods dollar system. The Fed had carried out a discretionary monetary policy during the 1920s and early 1930s under the interwar gold exchange standard. Gold, however, remained an important barrier to discretion-a barrier that Fed officials were unwilling to breach, even as the economy collapsed in 1931. A new political regime was required, first to devalue, and ultimately to replace the interwar gold standard with the Bretton Woods system. Under Bretton Woods, monetization of government debts and Keynesian policy experiments could take place without the swift disciplinary response of 26. In contrast to the United States, some countries have adopted new institutional restraints on discretionary policy, such as legislated mandates on central hanks to limit price-level fluctuations. The European Monetary Union, with its common currency, can also be viewed as a fundamental institutional change designed to prevent discretionary monetary policy at the level of the individual country.

63

The Great Depression as a Watershed for American Monetary Policy

gold outflows. Once inflation appeared, policymakers lacked the willingness to disinflate. This unwillingness stemmed in part from the Fed‘s reduced independence, the continuing role of money supply growth as an instrument of government debt management, and new economic views about the causes of inflation and the appropriate role of monetary policy in the economy. Given the persisting fiscal deficits of the postwar era and the new institutional environment, born of the Great Depression, Bretton Woods was doomed and a great inflation was almost inevitable. One legacy of the Great Depression-the dominance of Keynesian macroeconomic theory among professional economists-has now faded. Although some policymakers remain enamored with the possibility of using monetary policy to increase economic growth, such views no longer have the professional respectability that they had during the 1960s. Moreover, the high inflation of the 1970s made both policymakers and their economic advisors wary of policies that might ignite another inflationary spiral. Still, the permanent absence of a “nominal anchor” for the dollar provides an enduring legacy of uncertainty about the long-run path of monetary policy in the twenty-first century.

References Aheam, Daniel S . 1963. Federal Reserve policy reappraised, 1951-1959. New York: Columbia University Press. Anderson, Gary M., William F. Shughart, and Robert D. Tollison. 1988.A public choice theory of the Great Contraction. Public Choice 59 (October): 3-23. Blinder, Alan S . 1996. On the conceptual basis of monetary policy. Remarks at the Senior Executives Conference of the Mortgage Bankers Association, New York, 10 January. Blum, John M. 1959. From the Morgenthau diaries: Years of crisis, 1928-1938. Boston: Houghton Mifflin. Blum, John M., Edmund S. Morgan, Willie Lee Rose, Arthur M. Schlesinger,Jr., Kenneth M. Stampp, and C. Vann Woodward. 1973. The national experience: A history of rhe United States, 3d ed. New York: Harcourt Brace Jovanovich. Board of Governors of the Federal Reserve System. Various years. Annual report of the Board of Governors. Washington, D.C.: Board of Governors of the Federal Reserve System. . 1943. Banking and monetary statistics, 1914-41. Washington, D.C.: Board of Governors of the Federal Reserve System. Bordo, Michael D. 1993. The Bretton Woods international monetary system: A historical overview. In A retrospective on the Bretton Woods system: Lessons for international monetaiy reform, ed. Michael D. Bordo and Barry Eichengreen, 3-98. Chicago: University of Chicago Press. Bordo, Michael D., and Barry Eichengreen, eds. 1993. A retrospective on the Bretton Woods system: Lessons for international monetary reform. Chicago: University of Chicago Press.

64

Charles W. Calomiris and David C. Wheelock

Bordo, Michael D., and Finn E. Kydland. 1995. The gold standard as a rule: An essay in exploration. Explorations in Economic History 32 (October): 423-64. Brunner, Karl, and Allan H. Meltzer. 1964. The Federal Reserve Z attachment to the fiee reserves concept. Staff Analysis, House Committee on Banking and Currency. 88th Cong., 2d sess. . 1968. What did we learn from the monetary experience of the United States in the Great Depression? Canadian Journal of Economics 1 (May): 334-48. Calomiris, Charles W. 1991. The motives of debt-management policy, 1790-1880. Research in Economic History 13: 67-105. . 1994. Is the discount window necessary? A Penn Central perspective. Federal Reserve Bank of St. Louis Review 76 (May/June): 31-55. Calomiris, Charles W., and David C. Wheelock. 1996. The neutrality of reserve requirement changes in the 1930s. New York Columbia University. Manuscript. Calomiris, Charles W., and Berry Wilson. 1996. Bank capital and risk management: The 1930s scramble to shed risk. New York: Columbia University. Manuscript. Caskey, John. 1985. Modeling the formation of price expectations: A Bayesian approach. American Economic Review 75 (September): 768-76. Chandler, Lester V. 1958. Benjamin Strong: Central banker: Washington, D.C.: Brookings Institution. . 1971. American monetarypolicy, 1928-1941. New York: Harper and Row. Chari, V. V., Lawrence J. Christiano, and Martin Eichenbaum. 1996. Expectation traps and discretion. NBER Working Paper no. 5541. Cambridge, Mass.: National Bureau of Economic Research. Cume, Lauchlin B. 1980. Causes of the recession. History of Political Economy 12 (Fall): 316-35. De Long, J. Bradford. 1995. America’s only peacetime inflation: The 1970s. Berkeley: University of California. Manuscript. Eccles, Marriner S. 1966. Beckoningfiontiers. New York: Knopf. Eichengreen, Barry. 1992. Golden fetters: The gold standard and the Great Depression, 1919-1939. New York: Oxford University Press. Eichengreen, Bany, and Peter M. Garber. 1991. Before the accord: U.S. monetaryfinancial policy, 1945-51. In Financial markets and financial crises, ed. R. Glenn Hubbard. Chicago: University of Chicago Press. Epstein, Gerald, and Thomas Ferguson. 1984. Monetary policy, loan liquidation, and industrial conflict: The Federal Reserve and the open market operations of 1932. Journal of Economic History 44 (December): 957-83. Federal Open Market Committee (FOMC). Various issues. Minutes of the Federal Open Market Committee. Washington, D.C.: Federal Reserve System, Federal Open Market Committee. Internal document. Fisher, Irving. 1935. Testimony before the House Banking and Currency Committee. Banking Act of 1935. 74th Cong., 1st sess. Friedman, Milton, and Anna J. Schwartz. 1963. A monetary history of the United States, 1867-1960. Princeton, N.J.: Princeton University Press. Frost, Peter A. 1971. Banks’ demand for excess reserves. Journal of Political Economy 79 (July/August): 805-25. Goodfriend, Marvin. 1991. Interest rates and the conduct of monetary policy. CarnegieRochester Conference Series on Public Policy 34 (spring): 7-30. . 1993. Interest rate policy and the inflation scare problem: 1974-1992. Federal Reserve Bank of Richmond Economic Quarterly 79 (Winter): 1-24. Johnson, G. Griffith. 1939. The Treasury and monetary policy, 1933-1938. Cambridge, Mass.: Harvard University Press. Krooss, Herman E., ed. 1969. A documentary history of banking and currency in the United States, vol. 4. New York: McGraw-Hill.

65

The Great Depression as a Watershed for American Monetary Policy

Lucas, Robert E. 1980. Methods and problems in business cycle theory. Journal of Money, Credit, and Banking 12 (November, pt. 2): 696-715. Maisel, Sherman J. 1973. Managing the dollal: New York: Norton. Meigs, A. James. 1962. Free reserves and the money supply. Chicago: University of Chicago Press. Meltzer, Allan H. 1994. Why did monetary policy fail in the thirties? Pittsburgh: Carnegie-Mellon University. Manuscript. Miller, Adolph. 1931. Testimony before the Senate Banking and Currency Committee. Operation of the national and Federal Reserve banking systems. 7 1st Cong., 3d sess. Redish, Angela. 1993. Anchors away: The transition from commodity to fiat money in Western economies. Canadian Journal of Economics 26 (November): 777-95. Riefler, Winfield W. 1930. Money rates and money markets in the United States. New York: McGraw-Hill. Schwartz, Anna J. 1992. The misuse of the Fed's discount window. Federal Reserve Bank of St. Louis Review 74 (September/October): 58-69. . 1997. From obscurity to notoriety: A biography of the Exchange Stabilization Fund. Journal of Money, Credit, and Banking 27 (May): 135-53. Snyder, John W. 1969. Annual Treasury report on the events leading to the accord of 1951 and on the Treasury's monetary powers, 1951. In Documentary history of banking and currency in the United States, vol. 4, ed. Herman E. Krooss, 3026-7 1. New York: McGraw-Hill. Strong, Benjamin. 1926. Testimony before the House Banhng and Currency Committee. Stabilization. 69th Cong., 1st sess. U.S. Senate. 1931. Operation of the national and Federal Reserve banking systems. 71st Cong., 3d sess. Wheelock, David C. 1991. The strategy and consistency of Federal Reserve monetary policy, 1924-1 933. Cambridge: Cambridge University Press. Wicker, Elmus. 1966. Federal Reserve monetary policy, 1917-1933. New York: Random House. Yeager, Leland B. 1976. International monetary relations: Theory, history and policy, 2d ed. New York: Harper and Row.

This Page Intentionally Left Blank

2

Fiscal Policy in the Shadow of the Great Depression J. Bradford De Long

Before the Great Depression the U.S. government did not have a fiscal policy, at least not in the sense that economists have meant for the past two generations. The government did not attempt to tune its deficit or surplus to achieve the goal of full employment or low inflation. This is not to say that the federal budget was typically in balance. The federal government did borrow, and borrow on a very large scale in wartime: a typical pre-World War I1 war would end with total federal debt equal to some three-tenths of a year’s national product. But after a typical war was over the debt would rapidly be redeemed: the War of 1812 debt had been paid off by the 1830s. The Mexican War debt had been extinguished by the early 1850s. The enormous Civil War debt and the less enormous Spanish-American War debt had been extinguished by the eve of World War I. Thus U.S. “fiscal policy” before the Great Depression was simple: the federal government borrowed what it could during wartime. It strove thereafter to run peacetime surpluses to reduce the ratio of debt to national product. All this changed with the Great Depression. The first depression-era federal deficits were involuntary: both Herbert Hoover and (in his first term at least) Franklin Roosevelt tried to achieve balanced budgets. But they each failed to find politically feasible policies that could pass the Congress and would balance the budget. Later depression-era federal deficits were more voluntary: the government came to make a virtue out of necessity and to trumpet the potential macroeconomic benefits of a depression J. Bradford De Long is associate professor of economics at the University of California, Berkeley, a research associate of the National Bureau of Economic Research, and a visiting scholar at the Federal Reserve Bank of San Francisco. The author thanks the National Science Foundation and the Alfred P. Sloan Foundation for financial support. He also thanks John Auten, Robert Barsky, Bany Eichengreen, Claudia Goldin, Richard Grossman, Robert Hall, Chris Hanes, Chad Jones, Lars Jonung, Greg Mankiw, Alicia Munnell, Charles Schultze, Andrei Shleifer, Larry Summers, Richard Sutch, Robert Waldmann, David Wilcox, Eugene White, and especially Michael Bordo for helpful discussions. 67

68

J. Bradford De Long

deficit. Thus the U S . government abandoned the principle that the only good peacetime budget was a balanced budget. And the depression-era deficits were continued and vastly expanded during World War 11. The shadow cast by the Great Depression prevented a return to the predepression principle of budget balance after World War 11: the political economic folk wisdom, not completely wrong, was that addiction to budget balance had deepened the depression. But what was to replace the predepression commitment to peacetime balance as an operational goal? For the first generation or so after World War 11, economists’ and politicians’ views appeared to be converging on a possible consensus, a position that had been set out by the Committee on Economic Development (CED) under the intellectual leadership of future Nixon-era Council of Economic Advisers (CEA) chair Herbert Stein.’ The CED position was to set tax rates and expenditure programs so that the budget would be in surplus or in balance at high employment (whichever was desired) but to leave unhindered the operation of the “automatic stabilizers” set in motion as recession causes revenues to fall, spending to rise, and the budget to swing into deficit. These automatic stabilizers take effect within a single calendar quarter, as claims for unemployment insurance and food stamps are filed and as tax withholdings are received (or not received) by the Federal Reserve system. They reduce the sensitivity of aggregate demand to shifts in autonomous spending or monetary velocity, thus diminishing the magnitude of cyclical fluctuations. And they operate much more quickly than discretionary fiscal policies or shifts in monetary policy can-thus they are an irreplaceable tool of macroeconomic management. But neither Herbert Stein’s nor any other principle managed to become the basis of an enduring consensus. And no logic can be seen in the pattern of U.S. fiscal policy across decades: whatever logic might have been seen vanished with the emergence of long-term “structural” deficits during the 1980s under Ronald Reagan’s presidency, and with the appearance of very long term projections of large deficits as a social insurance system designed before 1973 collided with slower real revenue after the 1973 beginning of the productivity slowdown (see Auerbach 1994). This paper traces, first, the breakdown of the predepression consensus on fiscal rectitude under the pressure of the depression. It then considers the shadow cast by the memory of the depression on post-World War I1 fiscal policy, as economists, bureaucrats, and politicians struggled to learn the right lessons from the depression. It concludes on a note of pessimism: pressures on the U S . fiscal balance are strong, political understanding of the benefits of alternative policies are weak, and the attention span of the political system is short. Even though the consen1. Stein’s The Fiscal Revolution in America: Policy in Pursuit of Reality was and remains a classic.

Fiscal Policy in the Shadow of the Great Depression

69

sus of economists has provided and will continue to provide good advice, fiscal policy in the future is likely to display a similar lack of logic and have similar damaging effects on the economy as in the recent past.

2.1 Predepression Fiscal Policy 2.1.1

Peacetime Surpluses

Before the Great Depression, the idea that the government should tune its fiscal policy to control and moderate the business cycle was far from the center of political and economic discourse. The government did borrow, but its borrowings were confined to wartime. Wartime borrowings were large relative to the size of the economy: wars are very expensive. The Revolutionary War debt assumed by Treasury Secretary Alexander Hamilton amounted to perhaps onefifth of the then-United States’ annual national product. The Civil War debt accumulated under Treasury Secretary Salmon Chase and the World War I debt borrowed under Treasury Secretary William G . McAdoo each amounted to roughly three-tenths of annual national product (see fig. 2.1). But after the wars were over, the debt invariably shrank as a share of national product. Some of the debt was retired by budget surpluses. The remaining debt, constant in nominal terms, shrank relative to GDP because of real per capita income growth, population growth, and inflation. The Revolutionary War debt (both that issued under the Articles of Confederation and that assumed from the states on Hamilton’s initiative) had been reduced from roughly 20 to perhaps 5 percent of national product by the eve of the War of 1812. The War of 1812 debt was steadily reduced throughout the 120%

,

100%

80%

40%

20%

0%

1790

1820

1850

1880

1910

Fig. 2.1 Federal debt as a share of national product Sources: U.S. Bureau of the Census (1975) and CEA (1996).

1940

1970

J. Bradford De Long

70 5%

0% -5% -1 0%

-1 5%

-20% -25%

1870

1890

1910

1930

1950

1970

1990

Fig. 2.2 Budget surplus (or deficit) as a share of national product Sources: U.S. Bureau of the Census (1975) and CEA (1996).

1820s, and President Andrew Jackson paid off virtually the entire national debt in the 1830s. The small debt run up during the Mexican War was similarly erased by the end of the 1850s. The large Civil War debt and the small (relative to national product) Spanish-American War debt were similarly paid off over the decades: the ratio of federal debt to national product was less than 3 percent on the eve of World War I. The pattern was the same after World War I. The decade of the 1920s saw a near halving of the federal debt as a share of national product. Before the Great Depression there were peacetime decades-the 192Os, the 1880s, and the 1820s-in which the nominal federal budget was in surplus to the extent of 2 percent of national product or so. There were peacetime decades in which the nominal federal budget was in rough balance, and in which the growth of per capita income, the growth of population, and (usually) the slow progress of inflation reduced the relative size of the debt measured as a share of national product. There were no peacetime decades in which the federal budget was in more than trivial deficit (see fig. 2.2). 2.1.2

Ideologies and Doctrines

The rationale for the predepression consensus against countercyclical fiscal policy had at least five facets. First came a fear of the impact of large federal debts on the economic health of the country. A large debt meant large interest payments to service the debt. Relatively large interest payments to service the debt required high taxes-and, perhaps, a significant excess burden, especially given the limited record-keeping capacities of pre-twentieth-century govemments and thus the limited range of taxes that they could effectively administer.

71

Fiscal Policy in the Shadow of the Great Depression

Adam Smith, for example, surveyed the rising trend in war-induced debts, writing that rising debt as a share of national product “has gradually enfeebled every state which has adopted it. . . . Spain seems to have learned the practice from the Italian republicks, and (its taxes being probably less judicious than theirs) it has, in proportion to its natural strength, been still more enfeebled. . . . [The same is true of] France . . . [and] the United Provinces [of the Netherlands]” ([ 17761 1937,880-81). Smith believed that the conclusion of this “progress of the enormous debts” was “probably ruin.” He showed little patience with those who argued that the eighteenth-century British Empire need not fear the economic consequences of the accumulation of debt:*“Another war . . . may . . . render the British system of taxation as oppressive as that of Holland, or even as that of Spain. . . . Let us not. . . rashly conclude that [the British economy] is capable of supporting any burden; nor even be too confident that she could support without great distress a burden a little greater” ([1776] 1937, 881). Second came a fear of the political consequences of a high national debt. The government becomes more concerned with keeping its debt holders happy when the ratio of debt to national-product is high. It fears the consequences of capital flight: the government becomes, in a sense, the property of its debt holders when this ratio rises. Perhaps followers of Alexander Hamilton were not unhappy with the idea of a government whose every step was taken with a worried backward glance to see whether it pleased the bondholders. But followers of Thomas Jefferson or Andrew Jackson were very unhappy. Third-and perhaps most important-was that few believed that deficits in recession would be successful stabilization policy. Running a peacetime deficit to try to alleviate a recession would increase the tax burden, depress production and wealth in the long run, and reinforce dangerous tendencies in politics, and it would fail to boost employment and production in the short run. One important reason to doubt that deficits would be effective at fighting recessions was the operation of the pre-World War I gold standard. Under the prevailing international monetary system-the gold standard-capital flowed freely across national borders. To a first approximation, the domestic interest rate was set in world markets: equal to the world interest rate plus whatever risk premium international investors demanded for placing their wealth in the home country,Thus, there seemed to be little reason to think that deficits would be a powerful stimulative policy. First, deficits reduce the net supply of loanable funds to the private market: money borrowed by the government is not available to be loaned to business. Second, deficits might well increase-perhaps sharply increase-the risk premium investors would demand for lending their capital on the domestic market. The net effect? Ifprivate domestic demand for loanable funds is inelastic, 2. On the British political military state, see Brewer (1990). Also relevant is De Long and Shleifer (1993).

72

J. Bradford De Long

Interest Rate

newtotal supply

demand for loanable funds

old total supply

I

old private investment

Loanable Funds

new private investment newtotal (pnvate plus government) investment

Fig. 2.3 Effect of deficit spending on demand and supply of investment funds under the gold standard

and ifthe shift in the risk premium charged by overseas investors is small, then it might be that the sum of private investment and public spending if the deficitspending program were undertaken would be larger than otherwise. But it might not. And there was no theoretical presumption that it would (see fig. 2.3). Indeed, these fears made coherent sense in an impeccably Keynesian framework: that of Keynes’s (1936) General Theory. Stein quotes Keynes that deficit spending “may have the effect of increasing the rate of interest, and so retarding investment . . . whilst . . . the increased cost of capital goods will reduce their marginal efficiency. . . . With the confused psychology that often prevails, the Government programme may through its effects on ‘confidence’ increase liquidity-preference or diminish the marginal efficency of capital” (quoted in Stein 1969, 36-37). And economists working in the Keynesian tradition have at times argued that deficit spending is not expansionary. Consider Alan Blinder’s claim that the 1993 fiscal policy contraction in the United States boosted production and employment: “The remarkable decline in long-term interest rates from the fall of 1992 to the fall of 1993 is what kick-started a previously lackluster economic recovery into sustained growth. Remember. . .

73

Fiscal Policy in the Shadow of the Great Depression

that the bond market rallied with no change in monetary policy. Does anyone seriously doubt that the Clinton administration’s large, credible, deficitreduction plan of 1993 was the driving force behind the lower long-term interest rate^?"^ The predepression fears that fiscal policy multipliers could be zero or negative were real fears. Overinvestment Theories

A fourth argument against countercyclical fiscal policy was not that it would not work but that it would-and that depressions were good for you. Many economists explained the business cycle in general and the Great Depression in particular as consequences of “overinvestment.” For example, Joseph Schumpeter, writing from Harvard in the middle of the Great Depression, claimed that there was a “presumption against remedial [stimulative policy] measures [because] policies of this class produce additional trouble for the future. . . . [Depressions are] not simply evils, which we might attempt to suppress, but . . . forms of something which has to be done, namely, adjustment to change . . . [and] most of what would be effective in remedying a depression would be equally effective in preventing this adjustment” (quoted in Brown et al. 1934, 138). In what Haberler (1937) classified as “monetary overinvestment” theories of the business cycle, depressions were born either because of excessively easy monetary policy or because of ex post overoptimistic expectations of economic growth. When monetary policy ceased to be easy, or when investors and businesses recognized that their forecasts of future growth had been overoptimistic, the economy was left with a large inventory of investment projects that were unprofitable. True and sustainable economic recovery was not possible until the economy’s overinvestment overhang had been ‘‘liquidated’’-and the painful depression was this process of liquidation. Monetary and fiscal policies to moderate the depression would, in this conceptual framework, keep workers and firms producing in unsustainable lines of business and levels of capital intensity. Such attempts to alleviate the depression would make the depression less deep only at the price of making it longer and would add to the total sum of human misery (Hayek 1935). Indeed, economists like Lionel Robbins (1934) went as far as to blame the tiny steps toward moderating the decline in the money stock and boosting fiscal demand that governments undertook over 1929-33 for the persistence of the Great Depression into the mid- 1930s. Limits to Predepression Action

Fifth came yet another reason to be skeptical about predepression countercyclical fiscal policy: the problem of implementation. How might the government 3. See http://www.slate.codCoC/96-09-09Monday.asp.

J. Bradford De Long

74

40% 3 0%

20%

10%

0% 1 a70

1890

1910

1930

1950

1970

1990

Fig. 2.4 Federal revenues and expenditures as shares of national product Sources: U.S. Bureau of the Census (1975) and CEA (1996).

manage to spend money in a timely fashion on a large enough scale to do any significant good? It is not clear how the predepression federal government could have used the budget as a tool of business cycle management, even if it had very strongly wanted to (see fig. 2.4). Before World War I, federal receipts and expenditures were roughly 3 percent of national product. To obtain the same magnitude of fiscal “automatic stabilizers” that we have today-when a $70 billion fall in annual national product is associated with a $26 billion dollar increase in the federal deficitwould in pre-World War I circumstances require that the government respond to a recession carrying with it a 4 percentage point increase in the rate of unemployment by either doubling federal spending or eliminating taxation entirely. Fiscal policy can be stabilizing only if government spending is large enough to act as a plausible sea anchor for aggregate demand.

2.2 Fiscal Policy during the Depression 2.2.1 Striving for Budget Balance

Before the Great Depression, the government’s first instinct when the economy turned down was to do nothing. President Hoover sought to break this pattern. He stressed that he did assume a governmental responsibility to fight the depression-in sharp contrast to many others in his own party. He reserved special scorn for his own economic policy team, whom he termed the “ ‘leave it alone liquidationists’ headed by Secretary of the Treasury Andrew Mellon, who felt that government must keep its hands off and let the slump liquidate

75

Fiscal Policy in the Shadow of the Great Depression

itself. Mr. Mellon had only one formula: ‘Liquidate labor, liquidate stocks, liquidate the farmers, liquidate real estate.’ He insisted that, when the people get an inflation brainstorm, the only way to get it out of their blood is to let it collapse . . . even a panic was not altogether a bad thing” (1952, 3: 30).4 But Hoover’s idea of how activist government policy should fight the Great Depression was for the government to make sure that its budget remained in surplus. Hoover sought tax increases because “our major sources of revenues, income taxes and corporation profits, were going out from under us with appalling speed. . . . National stability required that we balance the budget. To do this, we had to increase taxes on the one hand and, on the other, to reduce drastically government expenditures” (1952, 3: 132).5 Thus Hoover’s December 1931 State of the Union message called for stringent action to balance the budget: Our first step toward recovery is to reestablish confidence and thus restore the flow of credit, which is the very basis of our economic life. The first requirement of confidence and of economic recovery is financial stability of the United States government. Even with increased taxation, the Government will reach the utmost safe limit of its borrowing capacity by the expenditures for which we are already obligated.. . . To go further than these limits . . . will destroy confidence, denude commerce and industry of their resources, jeopardize the financial system, and actually extend unemployment. (Hoover 1952, 3: 132-33) From our perspective, Hoover’s fears appear wrongheaded. The federal debt at the end of World War I had been nearly twice as large as a share of GDP as at the end of the 1920s. There was no sign in higher interest rates that the government had reached the “ultimate safe limit” of its borrowing capacity.6 Moreover, reasons that led monetary economists to fear deficit spending quickly vanished during the depression. What could do more to discourage 4. Four paragraphs later, Hoover insists that “Secretary Mellon was not hard-hearted.” 5. Hoover seems to have assumed the worst about the motives and aims of his political adversaries-both politicians in Washington and grassroots demonstrators and advocates: whether Democrats “set in their determination to delay recovery”; “old-guard Republican leaders in the Senate and the House . . . defeated in their Presidential ambition in 1928 . . . [who] certainly did not exert themselves energetically in their traditional duty to counterattack and expose [Democratic political] misrepresentations”; or bonus marchers seeking early payment of the World War I veterans’ bonus, “organized and promoted by the Communists [who] included a large number of hoodlums and ex-convicts . . . frequently addressed by Democratic Congressmen seeking to inflame them against me . . . given financial support by some of the publishers of the sensationalist press.. . . 5,000 mixed hoodlums, ex-convicts, Communists, and a minority of veterans” (1952, 3: 225-26). 6. Herbert Stein sees two causes of Hoover’s campaign to raise taxes and balance the budget in 1932. The first is that “Hoover and his close advisers had prejudices in favor of balancing the budget.” The second is the combination of the “gold outflow of 1931 [following Britain’s abandonment of the gold standard,] . . . rising interest rates, falling bond prices, increasing bank suspensions,” all of which led some solidity to Hoover’s fears.

76

J. Bradford De Long

private investment than the 8 percent per year deflation seemover 1929-33? To where could capital flee when other countries were likely to devalue further than the United States? Republicans were not alone in seeking budget balance: recall that Franklin Roosevelt made Hoover’s inability to balance the federal budget during the depression a campaign issue in 1932. And the overinvestment doctrine-that in the long run the Great Depression would turn out to have been “good medicine” for the economy and that proponents of stimulative monetary and fiscal policies in the 1930s were shortsighted enemies of the public welfare-drew relatively broad support. There were proponents of government action to expand demand in the Great Depression. Keynes tried to ridicule the “overinvestment” view, which Salant (1989) terms the “crime and punishment” view of business cycle^.^ Ralph Hawtrey, an advisor to the British Treasury and the Bank of England, called it the equivalent of “crying, ‘Fire! Fire!’ in Noah’s flood’ (1938, 145; see also Temin 1989).Much later, Milton Friedman would recall that at Chicago, where he went to graduate school, such dangerous nonsense was not taught-but he would speculate that perhaps the presence of such doctrines at other universities like Harvard was what induced otherwise bright economists to rebel and become Keynesians (see Gordon 1972). But President Hoover, at least, swallowed the claim that the Great Depression was due to overinvestment brought about by the Federal Reserve’s supposed “credit inflation” of the late 1920s hook, line, and sinker. Never mind that Friedman and Schwartz (1963) find no sign of too-loose monetary policy in the 1920s, as both monetary aggregates and prices followed their normal long-run growth path. Never mind that economic historians like Jeffrey Miron have argued that the Federal Reserve in the 1920s was too contractionary and set the Great Depression in motion by its attempts to cool off the economy out of the fear that the stock market boom of the 1920sreflected “overspeculation.” Hoover had no doubt that a Federal Reserve governed by “political appointee[~]. . . utterly devoid of global economic or banking sense . . . mediocrities . . . mental annex[es] to Europe” had set the Great Depression in motion by not causing absolute deflation in the 1920s. He came close to calling for the 7. From Keynes: Some austere and puritanical souls regard [the Great Depression] both as an inevitable and a desirable nemesis on so much [late 1920~1overexpansion, as they call it; a nemesis on man’s speculative spirit. It would, they feel, be a victory for the mammon of unrighteousness if so much prosperity was not subsequently balanced by universal bankruptcy. We need, they say, what they politely call a “prolonged liquidation” to put us right. The liquidation, they tell us, is not yet complete. But in time it will be. And when sufficient time has elapsed for the completion of the liquidation, all will be well with us again. . . . I do not take this view. I find the explanation of the current business losses, of the reduction in output, and of the unemployment which necessarily ensues not in the high level of investment which was proceeding up to the spring of 1929, but in the subsequent cessation of this investment. I see no hope of a recovery except in a revival of the high level of investment. And I do not understand how universal bankruptcy can do any good or bring us nearer to prosperity. ([1931] 1973, 349)

77

Fiscal Policy in the Shadow of the Great Depression

Budget Surplus or Deficit, as a Percentage of National Product

2%

1--

1 4

-2%

+

-4

f -4%

--

4

4 -6%

--

4

4 ‘ 4

7

Fig. 2.5 Unemployment and the budget, 1920-40 Sources: U.S. Bureau of the Census (1975) and CEA (1996).

execution of his entire Federal Reserve Board, ending the section in his Memoirs dealing with 1920s Federal Reserve policy with the accusatory: “There are crimes far worse than murder, for which men should be reviled and punished” (1952, 3: 9, 14).

2.2.2 Depression-Era Deficits If Herbert Hoover regarded his presidency as a success or failure depending on whether he managed to keep the budget balanced, his presidency was a failure. The federal budget swung from substantial surplus at the start of Hoover’s presidency into a deficit of 3 percent of GDP or more-partly as a result of congressional override of Hoover’s veto of the veterans’ bonus; partly as a result of “extraordinary” relief expenditures; and mostly as a result of the collapse in the nominal collections of a relatively progressive tax system, as real national income and the price level fell in the slide into the Great Depression. Figure 2.5 shows the pattern of deficits and official unemployment rates during the interwar period. The 1920s see relatively low official unemployment, in the 3-8 percent range, with government surpluses established to reduce the World War I debt in the range of approximately 1 percent of national product. By contrast, the 1930s see unemployment in the 15-25 percent range. And the federal deficit varies from near 1 up to 6 percent of national product. Swings in the deficit associated with swings in the unemployment rate in the interwar period are substantial. On average, a 5 percentage point increase in the unemployment rate is associated with a 1.6 percentage point increase in the deficit, measured as a share of actual national product. During the Hoover and the first Roosevelt administrations, these deficits

J. Bradford De Long

78 2%

1 \

Full Emdovment Balance

1%

0% -1 %

-2% -3%

-4% -5% 1922

1926

1930

1934

1938

Fig. 2.6 Full employment and actual budget balance as a share of national product, 192241 Sources: U S . Bureau of the Census (1975), CEA (1996). and author’s estimates.

were by and large involuntary The collapse of nominal revenues outran the ability of the executive branch to propose, and the Congress to enact, cuts in expenditures. The deficit rose in spite of discretionary fiscal policy action, due to the operation of the automatic stabilizers built into the government’s budget. By automatic stabilizers economists mean the natural tendency of revenues to fall-and for entitlement social welfare spending to rise-in a recession. As businesses lay off workers and lose profits, tax collections fall. As more people apply for relief programs of various kinds, expenditures rise. This “automatic” swing in the federal budget generates a shift toward a deficit that “stabilizes” the economy. It cushions the fall in disposable income that accompanies a recession, helps keep consumption from falling, and reduces the Keynesian multiplier. It has become conventional to conceptually divide fiscal policy into two parts: first, the operation of the automatic stabilizers that swing into action without any legislated changes in spending programs or tax schedules; second, changes in legislated policy that affect what the government surplus or deficit would be if economic activity were unchanged relative to potential output. This second component is usually measured by the full-employment deficit-what the federal deficit would be if the economy was near “full employment” (see fig. 2.6). Such estimates of the full-employment deficit are hazardous for the period of the depression. They require extrapolating tax collections and spending programs down to levels far outside previous experience. Brown’s (1956) esti-

79

Fiscal Policy in the Shadow of the Great Depression

mates were the first to show that fiscal stimulus on a full-employment basis, as opposed to automatic stabilizers, was rare during the Great Depression. Compared to a Great Depression that at its nadir involved a 40 percent reduction in output relative to potential, the 1 or 2 percent of GDP full-employment deficits of the mid-1930s delivered very little discretionary fiscal stimulus. By the second Roosevelt administration the government had begun to make a virtue of necessity, and to trumpet the potential fiscal advantages of unbalanced budgets. As Stein (1969) points out, the widespread belief in this politically palatable doctrine came well before the spread of the Keynesian tools of analysis that were to provide it with a respectable theory. Note that the title of Stein’s book is not the Keynesian Revolution but the Fiscal Revolution in America; for the fiscal revolution, the acceptance of deficit spending as an appropriate depression-fighting tool, was largely accomplished in America before the arrival of Keynesian doctrines.

2.3 The Keynesian Age 2.3.1 The Employment Act of 1946 The Employment Act of 1946 established Congress’s Joint Economic Committee, established the Council of Economic Advisers, called on the president to estimate and forecast the current and future level of economic activity in the United States and announced that it was the “continuing policy and responsibility” of the federal government to “coordinate and utilize all its plans, functions, and resources . . . to foster and promote free competitive enterprise and the general welfare; conditions under which there will be afforded useful employment for those able, willing, and seeking to work; and to promote maximum employment, production, and purchasing power” (see Heller 1966; Bailey 1950). Did it commit the government to the business of managing the macroeconomy? Perhaps. But recall that the 1978 Humphrey-Hawkins Act committed the federal government to reducing the unemployment rate to 4 percent by 1983 and to maintaining it thereafter; committed the federal government to reducing the inflation rate to zero by 1988, and required the chairman of the Federal Reserve to testify before Congress twice a year on the state of the macroeconomy.The Humphrey-HawkinsAct has had no effects-save to trigger the semiannual Humphrey-Hawkins testimony of the Federal Reserve chair. On the other hand, sometimes we refer to laws as boundary stones: shorthand markers of deeper changes in attitudes and predispositions. In this sense the Employment Act of 1946 certainly marked the commitment of the federal

80

J. Bradford De Long

government to the macroeconomic management business. As introduced, the Full Employment Act required the president to submit a National Production and Employment Budget (NPEB) that “assure[ed] a full employment volume of production” in the followingfiscal year-a fiscal year that would begin approximately six months after the submission of the NPEB. Congress does not move that fast; as written, the Full Employment Act could not have been implemented. As enacted, the Employment Act called for an annual economic report “setting forth . . . current and foreseeable trends in the levels of employment, production, and purchasing power . . . and a program for carrying out the policy” of the federal government to promote ”conditions under which there will be afforded useful employment for those able, willing, and seeking to work.” The enacted bill is not a cause but a signal of the federal government’s commitment to macroeconomic management. However, it is a powerful signal.*

2.3.2 Post-World War I1 Automatic Stabilizers Belief that automatic stabilizers have a significant effect on business cycle variability depends on the arguable proposition that liquidity constraints are pervasive in the economy (see De Long and Summers 1986). The procyclicality of taxes and transfers can work to reduce the size of recessions only if a significant fraction of consumers depend at the margin on their disposable income to finance their spending. Granting that movements in disposable income have significant effects on consumption, the shift over the past 70 years in the cyclical behavior of the federal budget considered as a sea anchor for the economy’s level of total spending is impressive. A good deal of this increase in the magnitude of automatic stabilizers comes from the increase in the size of the government as a share of national product. The post-World War I1 federal government taxes and spends one-fifth or more of national product in peacetime. The depression-era federal government taxed 5-7 percent and spent 8-10 percent of national product. The predepression federal government taxed and spent 5 percent of national product in peak peacetime periods. In other periods-the first peacetime presidency of Woodrow Wilson, for example-federal revenues and federal expenditures were little more than one-fiftieth of national product (see fig. 2.7). 8. The Employment Act of 1946 also created the Council of Economic Advisers. But its creation of the CEA as we know it today-one chair, two deputies, and a senior staff of 15, almost invariably drawn from and planning to return to the professoriate-is best described as an accident. The original proposal was for an expansion of the Budget Bureau, or perhaps for an “Office of the Director of the National Budget.” My reading of the legislative history is that the Truman administration dropped the ball on issues of Executive Office of the President organization on which it ought to have had strong views, and that Congress-I believe incorrectly-saw use of the Senate’s advise-and-consent power (along with the existence of an economic policy appropriation line item) as tools it could use to influence macroeconomic policy. The institution building of primarily Arthur Bums and secondarily Walter Heller gave the CEA the recruitment and staffing patterns that it has today.

81

Fiscal Policy in the Shadow of the Great Depression

Budget Surplus or Deficit, Percent of National Product (diamonds)*

0%

-2%

-4%

-6%

t 0%

1983-1995 (squares). 2%

4%

6%

8% 10% Unemployment Rate

Fig. 2.7 Unemployment and the deficit, 1950-95 Sources; U.S. Bureau of the Census (1975), CEA (1996), and author’s estimates.

With a large federal government, automatic stabilizers can easily be large: a 2.5 percentage point increase in the unemployment rate associated with a 5 percent fall in output relative to previous forecasts would “automatically” produce a deficit of 1 percent of national product if revenues and spending are one-fifth of national product, even if revenues have a unit elasticity with respect to shocks to production, and even if expenditures do not rise in response to unexpected increases in unemployment. When revenues are only 6 percent of national product, the smaller size of the government alone makes automatic stabilizers only one-third as large. And when, as before the Great Depression, both spending and revenues are 5 percent of national product or less, it is hard to see how automatic stabilizers could have any macroeconomic significance. Table 2.1 presents simple regressions of the level of the government’s fiscal balance as a share of national product on the unemployment rate (for 18901916, the nonfarm unemployment rate). Think of these regressions as summary statistics for how fiscal policy stabilized the economy, either through automatic stabilizers that were allowed to operate (and not offset by moves toward budget balance in recession) or by legislated changes in fiscal policy. The pre-World War I period shows little sign of automatic stabilizers: the small size of federal spending prohibits the deficit from acting as a sea anchor to stabilize the economy. Thus, post-World War I1 fiscal policy has been very different from what it would have been in the absence of the Great Depression. The growth in domestic federal spending can be traced to the expectations of the proper role of government and the social insurance state set in motion by the Great Depression. And it is hard to imagine that automatic stabilizers of

82

J. Bradford De Long

Table 2.1

Period

Regressions of Federal Budget Balance as a Share of GDP on the Unemployment Rate Unemployment Rate

Post-1982 Period ~~

1890-1916” 1920-40 1950-95 1950-95

-0.083 (0.043) -0.317 (0.045) -0.894 (0.188) -0.680 (0,127)

R2

SEE

0.257

0.002

0.847

0.010

0.526

0.013

0.746

0.010

~~

-0.021 (0.004)

”Using nonfarm unemployment only

the magnitude engendered by this growth in government would have been allowed to function, were it not for the shadow of the Great Depression.

2.3.3 Discretionary Fiscal Policy It is difficult to argue that “discretionary” fiscal policy has played any stabilizing role at all in the post-World War I1 period (see Gordon 1980; De Long and Summers 1986).Even Walter Heller could find only one case of successful discretionary stabilization policy-the Kennedy-Johnson tax cut (Heller 1966). Recessions of the size seen in the post-World War I1 era are not anticipated, are not forecasted, and develop quickly. Economic policymakers are working with shaky data from one quarter or so in the past. The legislative process takes at least two or three quarters. Appropriated funds require at least two more quarters before they can be spent on any substantial scale. And by that time the “need” for economic stimulus has passed. The U.S. government simply lacks the knowledge to design and the institutional capacity to exercise discretionary fiscal policy in response to any macroeconomic cycle of shorter duration than the Great Depression itself. By contrast, automatic stabilizers swing into action within the current quarter. A fall in incomes leads to a shortfall in revenues-and an increase in the deficit-as soon as witholdings reach the Federal Reserve. It is difficult to imagine how alternative policy instruments could deliver such a within-thequarter response to shifts in spending and employment. Thus, a sensible rule of thumb to adopt for fiscal policy would be the recommendation of the CED: let the economy’s automatic stabilizers work unhindered and set the cyclically adjusted full-employment or high-employment deficit at whatever level is felt to be consistent with views on the desirable long-run level of national savings. Indeed, for the first generation or so after World War I1 economists’ and politicians’ views did appear to be converging on this position. But the second post-World War I1 generation saw the disappearance of all possible consensus. The 1980s saw the unbalancing of fiscal policy, as slowed bracket creep from the fall in inflation, rising defense spend-

83

Fiscal Policy in the Shadow of the Great Depression

ing, and broad-based tax cuts together produced large deficits not just in recessions but in expansions as well. The 1980s saw the reversal of the peacetime rule that the debt-to-GDP ratio fell: the debt-to-GDP ratio rose from near 25 percent at the end of the 1970s to 50 percent by the early 1990s.

2.4 Long-Run Effects of Short-Run Policies There is an argument that use of fiscal policy as a stabilization tool has had harmful side effects. Buchanan and Wagner (1977) argue that deficits are dangerous because voters are highly myopic: when spending is raised and taxes are raised to finance the extra government spending, voters feel both the pain of reduced after-tax incomes and the benefits of spending programs and can judge whether the one is worth the other; but when spending is raised and financed by borrowing, voters feel the benefits from spending but do not sense the true resource cost of added indebtedness. The consequence, Buchanan and Wagner argue, is a government that engages in spending programs that at the margin do not provide social benefits equal to their resource costs. Democratic politics applied to government spending functions well only as long as deficits are effectively prohibited. Arguments that deficit spending could be used for stabilization managed, in Buchanan and Wagner’s view, to undermine the polity’s immune system, which had prevented the emergence of borrow-and-spend as a standard operating procedure of political parties. And the adoption of borrow-and-spend as a policy threatens to have evil consequences for economic growth: in the United States, at least, there is no sign that any increase in the government’s deficit sets in motion an endogenous rise in private savings to offset the fall in national wealth accumulation. It is hard to look back at America’s federal deficit since 1980 (see Auerbach 1994) without concluding that there is a good deal of truth in Buchanan and Wagner’s argument. “Cyclical deficit: good, structural deficit: bad” appears to be a message that is just a little bit too hard for the political nation to grasp. 2.5

Conclusion

Before the Great Depression the U.S. government borrowed in time of war and ran peacetime surpluses to pay off war debt. The use of the government deficit as a tool of macroeconomic management was rarely considered, andif considered-was rejected as inconsistent with international exchange rate arrangements. The depression broke this pattern: both Hoover and Roosevelt wished to maintain surpluses, but both recoiled at the austerity required in the midst of the depression. In the end, the political nation made a virtue of necessity: it concluded that deficits in time of recession helped alleviate the downturn. Later on, a theoretical rationale-John Maynard Keynes’s General Theorywas advanced to provide underpinnings for this shift in policy. But in Herbert

84

J. Bradford De Long

Stein’s formulation, the fiscal revolution in America was broader based than and was completed before the Keynesian revolution. In the generation after World War 11, economists and politicians moved toward a consensus view of fiscal policy: set tax rates and expenditure plans so that the high-employment budget would be in surplus, but do not take any steps to neutralize automatic stabilizers set in motion by recession. Set the highemployment budget surplus or deficit to conform to views about the desired impact of the federal budget on national saving over the business cycle. Do not attempt to aggressively use discretionary fiscal policy because the lags make it impossible to do so in an effective manner. But the American political system could not hold more than one idea in its mind at a time. The idea that cyclical deficits in recession could be good has weakened the belief that structural deficits that permanently reduce the national savings rate are bad. Some share of the deficits that began under Ronald Reagan’s presidency is attributable to the shadow of the Great Depression. And it may turn out that fear of continued structural deficits has undermined support for allowing fiscal automatic stabilizers to work smoothly: few who advocate balanced budget amendment proposals think about the implications of such proposals for stabilization policy. Thus, the shadow cast over fiscal policy by the Great Depression may be fading. A world in which the Great Depression had not cast its shadow over postWorld War I1 fiscal policy would be a different world. It might be a world in which many post-World War I1 macroeconomists called themselves Hayekians rather than Keynesians, discoursed on “monetary overinvestment,” and argued that deep recessions were a necessary price for the dynamic growth efficiencies of market-led economic development. It might be a world in which many governments responded to depressions by cutting spending and raising tax rates to keep the budget in balance and so prevent investors from losing confidence and making the depression worse. In such a world an outbreak of inflation-like that seen in the 1970swould have been highly unlikely. In such a world a repeat of the Great Depression would have been somewhat more likely. To the extent that the social and economic costs of the outbreak of inflation in the 1970s were low relative to the probability and cost of another episode like the Great Depression, we are indeed fortunate that post-World War I1 fiscal policy has been made in the shadow of the Great Depression.

References Auerbach, Alan. 1994. The U.S. fiscal problem: Where we are, how we got here, and where we’re going. In NBER Macroeconomics Annual 1994, ed. Olivier Blanchard and Julio Rotemberg, 141-74. Cambridge, Mass.: MIT Press.

85

Fiscal Policy in the Shadow of the Great Depression

Bailey, Stephen. 1950. Congress makes a law: The story behind the EmploymentAct of 1946. New York: Columbia University Press. Brewer, John. 1990. Sinews ofpower: War; money, and the English state 1688-1783. Cambridge, Mass.: Harvard University Press. Brown, Douglass V., et al. 1934. The economics of the recoveryprograrn. New York: McGraw-Hill. Brown, E. Cay. 1956. Fiscal policy in the ‘thirties: A reappraisal. American Economic Review 46, no. 5 (December): 857-79. Buchanan, James, and Richard Wagner. 1977. Democracy in deficit: The political legacy of Lord Keynes. New York: Academic Press. De Long, J. Bradford, and Andrei M. Shleifer. 1993. Princes and merchants: City growth before the Industrial Revolution. Journal of Law and Economics 36 (October): 671-702. De Long, J. Bradford, and Lawrence H. Summers. 1986. The changing cyclical variability of economic activity in the United States. In TheAmerican business cycle, ed. Robert J. Gordon. Chicago: University of Chicago Press. EmploymentAct of 1946. Public Law 304.79th Cong., 2d sess., 20 February 1946. Gordon, Robert J., ed. 1972. Milton Friedmank monetaryframework: A debate with his critics. Chicago: University of Chicago Press. . 1980. Postwar macroeconomics: The evolution of events and ideas. In The American economy in transition, ed. Martin Feldstein. Chicago: University of Chicago Press. Haberler, Gottfried. 1937. ProsperiQ and depression: A theoretical analysis of cyclical movements. Geneva: League of Nations. Hawtrey, Ralph. 1938. A century of bank rate. London: Longmans, Green. Hayek, Friedrich. 1935. Prices andproduction. London: Routledge and Kegan Paul. Heller, Walter. 1966. New dimensions of political economy. New York: Norton. Hoover, Herbert. 1952. The memoirs of Herbert Hoover: The Great Depression, 19291941. New York Macmillan. Keynes, John Maynard. (1931) 1973. An economic analysis of unemployment. In John Maynard Keynes: Collected works, vol. 13, The General Theory and after; pt. 1, Preparation, 343-67. London: Macmillan. . 1936. The general theory of employment, interest, and money. London: Macmillan. Robbins, Lionel. 1934. The Great Depression. New York: Macmillan. Salant, Walter. 1989. The spread of Keynesian doctrines and practices in the United States. In The political power of economic ideas: Keynesianism across nations, ed. Peter Hall. Princeton, N.J.: Princeton University Press. Smith, Adam. (1776) 1937. An inquiry into the nature and causes of the wealth of nations, ed. Edwin Cannan. New York: Modern Library. Stein, Herbert. 1969. The Jiscal revolution in America. Chicago: University of Chicago Press. Temin, Peter. 1989. Lessonsfrom the Great Depression. Cambridge, Mass.: MIT Press. U.S. Bureau of the Census. 1975. Historical statistics of the United States: Colonial times to 1970, Washington, D.C.: Government Printing Office. U.S. Council of Economic Advisers (CEA). 1996. The economic report of thepresident. Washington, D.C.: Government Printing Office.

This Page Intentionally Left Blank

3

The Legacy of Deposit Insurance: The Growth, Spread, and Cost of Insuring Financial Intermediaries Eugene N. White

One enduring legacy of the Great Depression was the creation of deposit insurance for financial intermediaries. Deposit insurance was a real innovation in federal regulation of the financial system. While the New Deal's anticompetitive barriers have largely collapsed, deposit insurance has become deeply rooted.' Coverage of the banking system has expanded steadily, and it has spread to other financial sectors. Economists have inveighed against government insurance of financial intermediaries' liabilities; and yet even in the wake of costly insurance disasters, there is little political interest in altering this pillar of the New Deal. Without the Great Depression, the United States would not have adopted the New Deal package of financial regulations that prominently featured deposit insurance. The New Deal regulations limiting competition had profound effects on the financial system; however, these regulations, with some exceptions, have disappeared while insurance of financial intermediaries appears to be permanent. Insurance began with the New Deal's limited explicit guarantee of bank deposits. This protection has grown considerably and is now granted implicitly to protect the deposits of all large banks. Furthermore, as table 3.1 shows, insurance has spread to other financial sectors. Although some features of deposit insurance have changed recently, there is no evidence of a rollback. With the important exceptions of mutual funds and money market mutual funds, the insurance of financial institutions' liabilities is pervasive. There is no ready model to explain the growth and spread of federal insurEugene N. White is professor of economics at Rutgers University and a research associate of the National Bureau of Economic Research. The author is especially grateful for helpful suggestions from Michael Bordo, Hugh Rockoff, Anna Schwartz, Lawrence J. White, and conference participants. I . For complete descriptions of New Deal banking regulations and their evolution over time, see Golembe and Holland (1986). Macey and Miller (1992), and White (1991).

87

Table 3.1

Spread of Financial Intermediary Insurance (nominal value of insurance per customer)

Liability Insured and Intermediary Deposits of commercial banks and mutual savings banks Deposits of savings and loan associations Shares in credit unions Customer accounts held by broker-dealers Cash Cash and securities Life insurance policies from life insurance companies Defined-benefit pensions

Coverage Begun or Increased Insurer

Jan 1934

Sept 1934

1950

I966

1969

FDIC

$2,500

$5,000

$1O,ooO

$15,000

$5,000

$10,000

$15,000

FSLIC

1970

1974

1980

$20,000

$40,000

$100,000

$20,000

$40,000

$100,000

NCUSIF SIPC

State funds

PBGC

$20,000 $40,000 $100,000 $50,000 $100,000 $500,000 Established in the 1970s Established in 1974

Sources: Federal Deposit Insurance Corporation, Annual Report (various years), Federal Home Loan Bank Board (various years), National Credit Union Administration (various years), Securities Investor Protection Corporation (various years), Brewer and Mondschean (1993), Pension Benefit Guaranty Corporation (1996).

89

The Legacy of Deposit Insurance

ance of intermediaries. Political economy offers models of logrolling (Schattschneider 1935) and cascading regulation (Hoekman and Leidy 1992; Feinberg and Kaplan 1993) that are not applicable here. In “logrolling,”sectors of an industry or related industries bargain in Congress for favorable legislation combined in one bill. Regulation “cascades” when one industry upstream secures protection, inducing downstream firms to follow them later and push for their own protection. Cascading regulation in international trade moves vertically from industry to industry. In contrast, the spread of insurance in the financial sector from banks and thrifts to credit unions, broker-dealers, life insurance companies, and pension funds represents horizontal movement. Although competition between types of intermediaries had increased in the 1920s, New Deal regulations tried to ensure very imperfect competition between the various sectors of the financial industry. Over time, competition within each segment and between each type of intermediary increased. The advantages conferred on banks by deposit insurance were then eagerly sought by uninsured intermediaries, and weaker institutions pushed up the level of insurance. In this paper, I examine how insurance spread from one group of institutions to the next and how the level of insurance was gradually raised. Although deposit insurance has often been discussed as an important guarantor of the stability of the banking system and hence the economy (Friedman 1959), the expansion of deposit insurance cannot be justified on macroeconomicgrounds. The general view today is that while the failure of individual banks might begin a panic, a systematic collapse may be prevented by proper intervention by the Federal Reserve as the lender of last resort (Friedman and Schwartz 1986). Instead, it is its redistributive features that have made insurance a permanent feature of the financial system while other New Deal regulations disappeared. Redistribution of the costs of failures, hidden in the insurance premiums, has gained public acceptance and allowed financial intermediaries to successfully lobby for expanded coverage. If insurance was not necessary for securing macroeconomic stability, substantial costs may have been incurred. I explore the cost of insurance with a counterfactual analysis of an insurance-free postGreat Depression financial system in order to assess the burden imposed by this legacy of the New Deal.

3.1 The Origins and Establishment of Deposit Insurance While deposit insurance today enjoys broad public support, proposals for federal insurance before the Great Depression were viewed as special interest legislation. States had experimented with insurance of bank liabilities before the Civil War and after the panic of 1907. These state systems had, at best, mixed results, establishing a strong policy prejudice against federal insurance (Golembe 1960; White 1983; Calomiris 1990;Wheelock 1992). Nevertheless, a well-motivated lobby of predominantly rural, unit bankers was keen on se-

90

Eugene N. White

curing a federal guarantee system. Hoping to increase depositor confidence while preserving the existing banking structure, these bankers opposed the liberalization of branching laws and other regulations, which could have produced a more stable banking system of larger, diversified institutions (Calomiris 1993). Studies of the origins of deposit insurance from Golembe (1960) to Calomiris and White (1994) emphasize that deposit insurance would have had little chance of adoption if the 1930-33 banking collapse had not frightened the public into supporting the proinsurance bankers’ cause in Congress. Even so, the hurdles faced by backers of deposit insurance were high. From earlier state experiments, the problems of moral hazard and adverse selection were well known and debated in Congress (Flood 1991). Aware of the potential problems, the Roosevelt administration, the bank regulatory agencies, and the larger banks resisted any proposal. In the face of such opposition, credit for the adoption of deposit insurance belongs largely to Rep. Henry Steagall (DAla.), chairman of the House Banking and Currency Committee, who refused to permit the passage of any banking legislation unless it included an insurance system. Far from being a high-minded policy aimed at protecting the depositor, the design of the Federal Deposit Insurance Corporation (FDIC) was the product of a lengthy legislative struggle, pitting smaller state-chartered, often unit banks against larger banks, often members of the Federal Reserve System. Under the Banking Act of 1933 (often called the Glass-Steagall Act), the Temporary Deposit Insurance Fund was organized and scheduled to begin operations on 1 January 1934. The coverage per depositor was set at a maximum of $2,500.2All Federal Reserve member banks were required to join. Nonmember banks could receive insurance only if they joined the Fed within two years. The last provision was resented by nonmember banks because they would be forced to meet the higher requirements and stricter regulations imposed on members. Banks joining the system were to pay a 0.5 percent assessment of insurable deposits, half upon joining and half subject to call (FDIC 1984, 56-57). When the temporary fund was extended for a year in 1934, Steagall attempted to increase coverage to $10,000 against Roosevelt’s objection that 97 percent of depositors were already covered. Congress raised the limit to $5,000 and postponed compulsory Federal Reserve membership until 1 July 1937-a victory for the small banks (Bums 1974). The temporary system became permanent under Title I of the Banking Act of 1935, which created the FDIC. All Federal Reserve members were still required to join; but in a major concession, nonmembers, while subject to approval of the FDIC, were no longer required 2. A depositor in a bank is provided with insurance up to the legal maximum. Insurance coverage is determined by how accounts are legally titled. A husband and wife might each have individual accounts and also share a joint account in one bank. All three accounts would be entitled to insurance up to the maximum because the legal title for each is unique.

91

The Legacy of Deposit Insurance

to become members of the Fed. The permanent plan required an annual assessment on total, not just insured, deposits. This shift was opposed by the larger banks, whose shares of uninsured deposits were much greater.3 The Banking Act of 1935, based largely on the draft legislation of the FDIC staff, set a flat annual assessment rate of one-twelfth, or 0.0833 percent, of a bank’s total deposits, eliminating the original capital contribution by banks. To ensure that the insurance fund was not depleted, the FDIC was given authority to borrow up to $975 million from the Treasury. Banks contributed premiums as a fraction of all their deposits but only received protection on deposits up to a maximum of $5,000 per account. Small banks and lower-income individuals with small deposit accounts benefited, while bigger banks with larger depositors provided a subsidy. The smaller banks’ competitive position was improved, and there was less pressure to build stronger, larger banks. The requirement that all Federal Reserve members join the new FDIC guaranteed that the bigger banks, many of whom had opposed federal deposit insurance, joined the system rather than lose the benefits of Fed membership. Nonmember banks, almost all smaller, state-chartered banks, had pushed for deposit insurance. Happy with the design, they signed up immediately. In 1935, 91 percent of the 15,488 commercial banks, with 86 percent of assets, joined the system. Only mutual savings bank membership was low. Of the 566 mutual savings banks, 11 percent, with 11 percent of total assets, took out membership. Most mutual savings banks preferred to remain in existing state insurance systems that offered higher levels of coverage. Nevertheless, the nearly universal coverage of commercial banks and the subsequent disappearance of bank failures was seen as triumph for the New Deal.

3.2 The Growth of Deposit Insurance For the next 15 years, the FDIC’s insurance of commercial banks and mutual savings banks appeared to be an unqualified success. By 1949, commercial bank membership crept up to 95 percent, accounting for 49 percent of deposits; mutual savings bank membership increased to 36 percent, holding 70 percent of all assets. Bank failures declined, no panics occurred, banks were more profitable, and the insurance fund grew. At the same time, inflation had reduced the real value of insurance. World War I1 inflation shrank the real value of coverage per depositor from $5,000 in 1934 to $2,807 by 1949. Figure 3.1 depicts the real value of the maximum coverage offered per depositor from 1934 to 1995, with the changes in the nominal levels of coverage indicated by vertical lines. However, this decline in protection elicited no outcry by depositors for more protection. As seen in figure 3.2, the percentage of total deposits 3. In 1936, the 10,014banks with deposits of under $1 million had 85 percent of their deposits insured, while the 209 largest banks with deposits over $25 million had only 28 percent of their deposits covered. See table 3.2 below.

92

Eugene N. White $zo,ooo

\

$15,000

0 0

# 0 3

m 0 bl

$5,000

L

0 1934

1940

1950

1960

1970

1980

1990

Fig. 3.1 Real value of deposit insurance per depositor, 1934-95 Sources: Federal Deposit Insurance Corporation, Annual Report (various years) and the consumer price index. Note: Changes in nominal coverage are indicated by the vertical lines.

covered by FDIC insurance had climbed from 45 percent in 1934 to 50 percent in 1950.4The absence of big failures and the growth of deposits kept the total insurance fund at about 1.5 percent of all insured deposits, as shown in figure 3.3, in spite of repayment of the initial contributions by Treasury and Federal Reserve Banks in 1949 (FDIC 1984,5-7). By any measure, the vast majority of “small depositors” were well protected by this level of insurance, and there was no public demand for a big increase in coverage. In 1949, only 4.4 million of the 104 million bank accounts were not fully protected (FDIC,Annual Report 1949). Some of these accounts were government (293,000) and interbank deposits (127,000), which had high average balances of $40,000 and $90,000, in contrast to the average demand deposit balance of $1,911 and savings and time deposit balance of $824. The FDIC (Annual Report 1949) calculated that any increase in coverage would offer little additional protection. A rise in coverage to $10,000-which would have returned coverage to its real 1934 value-would have fully covered another 3 million accounts, reaching 97 percent of the total. The percentage of insured banks’ deposits covered would have risen from 50 percent to 57 per4. The large decline of coverage in the 1940s, from 45 percent to 35 percent in 1942, was a consequence of the rapid growth of total deposits. Coverage bounced back to 50 percent thanks to the account-creating activity of depositors. Between 1941 and 1949, total deposits increased 117 percent and insured deposits by 174 percent, with the number of fully protected accounts rising by 47 percent (FDIC, Annual Report, various years).

"' 70

65

-

60

-5

55

-

50

-

45

-

40

-

35

- , 1 1 1 1 1 1 1 1 1 1 1 1 [ ,

0 0

0 0 0" In

3

#

x

0

3

#

u3

#

1934

1940

1 1 1 1 1 1 1 1

1950

.A,

ll[l[rr,

1960

1970

1980

1990

Fig. 3.2 Percentage of bank deposits insured by the FDIC, 1934-94

Sources: See fig. 3.1. Nore: Changes in nominal coverage are indicated by the vertical lines.

Fig. 3.3 FDIC insurance fund as a percentage of insured deposits, 1934-94 Sources: See fig. 3.1.

94

Eugene N. White

99'5

I +

5

4

W L

a

+

95.5

+

1934

1944

+

1954

1964

1

1984

1994

Accounts Fully Insured

Fig. 3.4 Insurance of accounts in FDIC banks, 1934-94 Sources: See fig. 3.1. Note: Changes in nominal coverage are indicated by the vertical lines.

cent. An increase to $25,000 would have covered 99.5 percent of all accounts and 65 percent of all deposits. Mutual savings banks were a shrinking component of the banking industry and played no significant role in the politics of deposit insurance. By 1949, the FDIC insured only 192 of the 531 mutual savings banks. Most of the remainder (190 of 339) were in Massachusetts and were insured by a state fund. FDICinsured mutual savings banks had 12.6 million accounts in 1949, with $13 billion in deposit^.^ Ninety-four percent of these accounts were fully insured, and 61 percent of all deposits were insured. While this profile looks similar to that of commercial banks, mutual savings banks were not at the same risk. In commercial banks, 68 percent of all deposits were held in the 3 percent of accounts with over $5,000; in mutual savings banks, only 39 percent of all deposits were in the 6 percent of accounts with over $5,000. Very few accounts, representing 3.6 percent of deposits, exceeded $10,000, whereas 57.7 percent of commercial banks' deposits were in accounts in excess of $10,000. Mutual savings banks were not as vulnerable as commercial banks and did not join the demand for a rise in insurance. Demand for an increase in coverage was driven by the small banks' fear of losing deposits. Figure 3.4 shows the drop in fully insured accounts, from 98.5 5. Non-FDIC-insured mutual savings banks had $5.7 billion in deposits.

95

The Legacy of Deposit Insurance Percentage of Insured Commercial Banks by Their Percentage of Insured Deposits

Table 3.2

Percentage of Insured Deposits Year

Less than 20

20-59

60-89

90-100

Number of Insured Banks

1936 1938 1941 1945 1949 1951 1955 1964

0.8 0.7 1.3 1.3 1.2 0.8 0.2 0.1

6.1 5.2 6.4 9.3 13.5 6.7 8.2 12.2

57.7 57.2 59.4 76.5 79.6 69.3 72.5 77.8

35.4 36.9 32.9 12.9 5.7 23.2 19.1 9.9

14,085 13,705 13,434 13,289 13,440 13,451 13,278 13,468

Sources; FDIC, Annual Report (1951,76; 1955,68; 1964, 102-3).

percent at the inception of insurance to just under 96 percent by 1949. The smallest banks felt this change acutely. Table 3.2 shows that in 1936 35.4 percent of banks had 90-100 percent of their deposits insured. The number of banks enjoying this high level of coverage collapsed to 5.7 percent in 1949. The search for protection by large depositors threatened smaller banks. In the 1950 Senate hearings on deposit insurance, Sydney J. Hughes of the Industrial Bank of Commerce of New York City and member of the Consumer Bankers Association explained that “when a depositor’s balance exceeds the $5,000 insured maximum, he shifts the surplus to another bank and becomes one of what must be millions of multiple deposits” (U.S. Senate 1950, 90). There were good reasons for deposits in excess of the insured maximum to worry bankers, as one recent study suggests. Using a special sample of wealthy households from the 1992 Survey of Consumer Finances, Kennickell, Kwast, and Starr-McCluer (1996) found that while large depositors keep substantial shares of their assets in insured depositories, they often fail to keep them within insurance limits. According to the survey, a sizable 17.3 percent of household deposits were uninsured. Kennickell et al. found that any reduction or restriction in insurance coverage would substantially increase the uninsured deposits of households and increase the likelihood of withdrawak6 In 1950, bills to raise coverage were introduced by Senators John W. Bricker (R-Ohio), Claude D. Pepper (D-Fla.), Charles W. Tobey (R-N.H.), Hugh A. Butler (R-Nebr.), Willian Larger (R-N.D.), and Burnet R. Maybank (D-S.C.), who was chairman of the banking committee. All of these bills contained increases up to $15,000, and Pepper’s would have removed the limit altogether. 6. Employing a probit model, Kennickell et al. (1996) found that lowering the deposit insurance ceiling from $lOO,OOO to $75,000would increase total household uninsured deposits by 29 percent. Smaller effects were found for eliminating separate coverage of existing IRA and Keogh accounts.

96

Eugene N. White

In his plea for a rise to $10,000, Senator Butler noted that “from my correspondence, I judge that it is primarily the smaller country banks that are anxious for this change. It seems that under the present system a good many depositors maintain part or all of their funds in the city banks at some distances, perhaps from their homes” (US. Senate 1950, 101). Ben Dubois, the secretary of the Independent Bankers Association, made an explicit appeal to protect the small banks, stating that “the Federal Deposit Insurance Corporation has been a powerful instrument in the perpetuation of independent banking. It has put the small bank on a par with the large bank in the eyes of the average depositor” (U.S. Senate 1950,87-88). Federal regulators supported the increase but tended to cloak their support in terms of the ideology of guaranteeing continued protection of the small depositor. In the 1950 hearings, there was general support from federal regulators to raise the ceiling to $10,000. The secretary of the treasury, John W. Snyder, and the comptroller of the currency, Preston Delano, favored an increase to $10,000. Delano argued that $10,000was justified on the grounds that prices had risen, lowering effective coverage, even though he admitted that $5,000 still covered 96 percent of accounts. The chairman of the FDIC, Maple T. Harl, also supported the increase on the grounds that protection of the small depositor required it; but he was also clear that “the preservation of the American banking system. . . . As you very well know, the survival of the dual banking system in large measure depends on Federal deposit insurance” (U.S. Senate 1950, 22-23). Former FDIC chairman Leo Crowley testified that he favored the increase from $5,000 to $10,000 because it would help small savers and the small banks in their home communities. Larger banks were generally willing to support a rise, but they were less enthusiastic and were more concerned about the fact they subsidized the system. American Bankers Association officials testified in favor of $10,000 coverage but warned that any further increase would endanger the system (U.S. Senate 1950, 66). Frederick A. Potts, president of Philadelphia National Bank and a representative of the Reserve Bankers Association, testified that limited deposit insurance was a sound idea. He warned, however, that a rise in protection to $10,000 would undermine good bank management and stimulate demands for more coverage (U.S. Senate 1950, 80-81). The most striking testimony against the proposal came from one of the founding fathers of the FDIC, Senator Arthur Vandenberg. In a letter, he denounced the proposed rise to $10,000 coverage, arguing that it was imprudent: “There is no general public demand for this increased coverage. It is chiefly requested by banker demand in some quarters for increased competitive advantage in bidding for deposits.” He predicted that “if we extend the coverage to $10,000, how long will it be before we confront demands for total coverage? Total coverage would virtually socialize our private banking system. It could involve many of the vices which so often wrecked previous well-meaning adventures in this field” (U.S. Senate 1950, 50-5 1).

97

The Legacy of Deposit Insurance

The willingness of larger banks to support an increase in the level of coverage did not arise from any hope to improve their competitive position by insuring more deposits. Their position changed very little in terms of insurance coverage after the 1950 act went into effect. At the very beginning, in 1936, large banks received very little protection, as seen in table 3.3. While the more than 10,000 banks with under $1 million in deposits had 86 percent of their deposits insured and the banks with $1-$5 million in deposits had 74 percent of their deposits protected by the FDIC, the 200 largest banks had only 28 percent of their deposits insured. Coverage for them grew; yet by 1949, coverage was still only 36 percent. What concerned the larger banks was not the fact that they still had large uninsured deposits but that they were assessed on their total, not just their insured, deposits. To cover a much larger fraction of their deposits would have required a huge increase in coverage that would have interested few smaller banks. Furthermore, a big increase in coverage would have decreased the ratio of the insurance fund to insured deposits, depicted in figure 3.3, perhaps requiring an increase in assessments. The insurance fund had grown thanks to the virtual disappearance of bank failures. The fund easily repaid the initial contributions ($289 million) of the Treasury and the Federal Reserve Banks (FDIC 1984, 58-60). There was concern that the assessment rate was too high, not too low, cutting into bank profits. Although banks’ net earnings rose steadily during the 1940s, net profits had declined from $906 million in 1945 to $831 million in 1949. At the same time, the FDIC assessment climbed from $86 million to $109 million, following the rise in total deposits (FDIC, Annual Report 1949, 40). Cutting the assessment could easily buoy profits. Not surprisingly, the larger banks lobbied Congress for a reduction in assessments while they grumbled about the increase in coverage. The smaller banks returned the favor. The Independent Bankers Association was set against any reduction in the premium and protested that big banks had no right to complain as they had obtained the interest prohibition on demand deposits under the New Deal. The end result was a compromise: an increase in coverage and a change in assessment that satisfied both parties and ensured swift passage of the 1950 act. Figure 3.2 shows that the new level of $10,000 coverage protected an additional 5 percent of deposits. More important for banks concerned about protection, the shares of fully protected accounts, shown in table 3.2, returned to their earlier level. The more exposed banks that had fallen from their high level, 90-100 percent, of protected deposits, regained ground lost in the previous decade. The larger banks also benefited. The basic assessment rate was not reduced because the FDIC feared this might set the stage for a depletion of the fund. Instead, it was lowered by a rebate system. The FDIC deducted operating expenses and insurance losses from gross assessment income then shared the remainder, returning 60 percent to the banks and keeping 40 percent. As seen in figure 3.5, this rule produced some fluctuation in the assessment rate around

Table 3.3

Insurance Coverage by Size of Bank Under $1 million

$1-$5 million

$5-$25 million

Over $25 million ~

1936 Number of banks Deposits insured (%) Accounts insured (%) 1949 Number of banks Deposits insured (%) Accounts insured (%) 1951 Number of banks Deposits insured (%) Accounts insured (%) 1964 Number of banks Deposits insured (%) Accounts insured (%) 1966 Number of banks Deposits insured (%) Accounts insured (%) 1968 Number of banks Deposits insured (%) Accounts insured (70)

Over $100 million ~

Over $1 billion

~~

10,014 86.1 99.4

3,231 74.2 98.8

694 58.0 98.4

209 28.4 97.4

2,554 81.3 97.6

735 1 73.6 97.1

2,812 63.7 96.4

50 1 52.7 95.5

213 36. I 94.4

2,349 89.3 99.1

7,463 81.9 99.0

3,035 70.9 98.6

564

60.0 98.4

24 1 41.3 98.3

102 85.0 98.6

7,082 79.6 98.3

5,124 71.3 98.1

1,030 63.2 96.9

760 55.4 96.9

39 36.9 96.4

672 73.2 99.0

6,101 64.9 97.9

5,499 56.9 98.2

1,075 44.9 97.3

372 33.0 97 .o

38 21.4 97.3

470 76.7 99.7

5,268 80.1 99.2

6,143 68.6 99.2

1,359 59.0 98.9

429 42.6 98.4

45 28.0 98.3

1970 Number of banks Deposits insured (%) Accounts insured (%) 1972 Number of banks Deposits insured (%) Accounts insured (%) 1975 Number of banks Deposits insured (%) Accounts insured (%) 1980 Number of banks Deposits insured (%) At $40,000 At $100,000 Accounts insured (%) At $4O,ooO At $100,000

358 80.0 99.5

4,562 79.2 99.4

6.65 1 73.0 99.4

1,610 61.3 99.0

468 48.1 98.9

48 30.8 98.8

213 78.6 99.5

3,637 76.8 99.7

7,151 70.8 99.0

2,154 59.9 99.0

585 44.0 98.3

63 29.4 98.4

137 83.1 98.5

2,568 82.0 99.7

7,878 79.8 99.6

3,075 69.4 99.4

779 53.7 99.3

79 28.6 99.0

87

1,064

7,177

4,881

1,232

140

60.0 80.0

77.8 88.9

77.6 86.7

70.2 79.8

54.0 63.1

34.1 41.5

99.2 99.7

99.4 99.4

99.5 99.7

99.4 99.7

99.2 99.4

98.9 99.4

Sources: FDIC, Annual Reports (various years) and Report of Deposits (various years).

100 -

Eugene N. White

0.26

Fig. 3.5 FDIC effective assessment rate, 1934-94 Sources: See fig. 3.1.

0.035 and 0.037 percent of total deposits, far below the original 0.0833 percent. Total assessments in 1951 reached $124 million, but $70 million was rebated to the banks (FDIC, Annual Report 1951). Net profits for 1951 were $908 million, but they would have stood at only $838 million without this change. The 1950 act was a well-crafted compromise. Insurance coverage of all deposits was on the rise. Larger banks, which had initially opposed deposit insurance, now “signed on” to support insurance thanks to the reduction in the effective assessment rate. The 1950 increase in insurance coverage was the last time that commercial banks appear to have been the primary movers behind insurance legislation. While commercial and mutual savings banks covered by the FDIC saw nominal coverage rise and the percentage of funds insured increase, greater competition and inflation put more pressure on other financial intermediaries who clamored more loudly for higher coverage. 3.3 Evaluating the Rise in Coverage for Commercial Banks

Legislation raising the level of coverage is only one factor leading to a higher level of protection. To explain the growing percentage of deposits in FDIC-insured institutions that were covered by FDIC insurance, shown in figure 3.2, four factors were considered: (1) If the real maximum deposit insurance coverage per depositor is increased, the percentage of covered deposits should rise. (2) Failures, measured either as the number of failing banks or the

101

The Legacy of Deposit Insurance

percentage of deposits in failing banks, might induce depositors to shift their uninsured deposits to new accounts or banks for complete coverage. (3) A rapid growth in deposits might decrease coverage if individuals’ balances quickly rise above insured levels. (4) If individuals open new accounts to ensure coverage of their deposits, the increase in the number of accounts should raise the percentage of deposits covered.’ Data on the number of accounts were difficult to obtain, as this information was only collected by the FDIC in occasional special reports until 1981. This is displayed in table 3.4. Beginning in 1990, some data on commercial banks’ accounts were collected by the FDIC.*Accounts of all banks appear to have grown at a very rapid rate between 1934 and the mid-1970s. The average rate of growth exceeded the real rate of growth of the economy. Starting in the late 1970s and certainly in the early 1980s, this growth slows down, with some years of decline. The stagnation between 1981 and 1990 may be attributable to the high level of coverage provided by the jump from $40,000 to $100,000 insurance and the increase in alternatives to bank deposits, such as money market mutual funds. A continuous time series of accounts for the period 1934-81 was constructed by regressing the number of accounts on time and timesquared to fill in the missing observations, but no attempt was made to fill the gap between 1981 and 1990 when the trend growth abruptly changed. Unit root tests and an examination of the partial autocorrelations indicated that the percentage of insured deposits, the real insurance per depositor, and the measures of bank failures needed first-differencing for stationarity. It was difficult to judge whether the growth of deposits also required firstdifferencing, but the results were similar so only the first-differenced results were reported in table 3.5. Regressions ( 1 ) and (2) are for the whole period, 1934-94, and exclude the variable for accounts. As hypothesized, an increase in the real value of maximum deposit insurance coverage per depositor raises the percentage of covered deposits. A rise in real coverage of $10,000 would drive the percentage of insured deposits up by about 6 percent, suggesting that this factor alone can only account for a modest portion of the increase. Also, as conjectured, an increase in deposits tends to lower the percentage of covered deposits. An acceleration in deposit growth of 1 percent pushed down coverage approximately 1.7 percent. The most notable example of this effect was during World War 11, when the rapid growth of deposits outweighed other influences and temporarily halted the upward trend in coverage. Neither variable for bank failures helps to explain the rising coverage of deposits, probably because there is little variation in failures. For depositors, it may have been a minor consideration given the FDIC’s practice of frequently providing full insurance to depositors whose accounts were over the limit (FDIC 1984). 7 . The data were obtained from the FDIC’s annual reports. The consumer price index was used to obtain the real value of deposit insurance. 8. The 1990-96 data were obtained through the FDIC’s website (R. Drozdowski, correspondence with author, [email protected]).

Table 3.4

Number, Growth, and Insurance of Bank Accounts, 1934-96 All Insured Banks All Accounts Number (millions)

Year

(1)

1934 1936 1938 1941 1945 1949 1951 1955 1964 1966 1968 1969

51.2 58.8 62.7 69.5 92.3 104.0 111.6 129.7 174.8 193.0 212.0

Annual Growth Rate (%) (2)

7.1 3.3 3.5 7.4 3.0 2.4 3.8 3.4 5.1

4.8

All Commercial Banks

Fully Insured Accounts

Fully Insured Accounts

All Accounts

Number (millions) (3)

Annual Growth Rate (%) (4)

Percentage of Accounts Fully Insured (3)/(1) (5)

50.4 57.8 61.7 68.2 89.0 99.6 109.9 127.4 169.8 191.1 208.8

7.1 3.3 3.4 6.9 2.8 3.4 3.8 3.2 6.1 4.5

98.4 98.4 98.3 98.1 96.4 95.7 98.5 98.2 97.2 99.0 98.5 99.1

Number (millions)

(6)

98.2 114.6 155.0 173.0 191.0

Annual Growth Rate (%) (7)

3.9 3.4 5.7 5.1

Number (millions) (8)

Annual Growth Rate (%) (9)

Percentage of Accounts Fully Insured (8)/(6) (10)

1970 1971 1972 1975 1980 1981 1990 1991 1992 1993 1994 1995 1996

231.7

4.5

229.4

4.8

244.5 302.5 322.9 323.4

2.7 7.4 1.3 0.1

241.3

2.6

99.0 99.0 98.7

99.2 99.2 99.2 99.2 99.1

209.3

4.7

220.2 276.5 289.9 290.0 277.4 287.9 289.7 285.8 287.1 303.8 306.2

2.6 7.9 1.o 0.0 -0.5 1.9 0.3 -0.7 0.2 2.9 0.4

287.5 283.6 284.7 301.3 303.5

-0.7 0.2 2.9 0.4

99.2 99.2 99.2 99.2 99.1

Sources: 1934-8 1, FDIC, Annual Reports (various years) and Reporrs ofDeposits (various years); 1990-96, R. Drozdowski, correspondence with author, rodrozdowski @fdic.gov.

104 Table 3.5

Eugene N. White Determinants of FDIC Insurance Coverage of Commercial Bank Deposits 1934-94

1934-81

Variable

(1)

(2)

lntercept

0.48 1 (2.06) 0.566 (3.51) 0.188

0.486 (2.10) 0.571 (3.59)

Real value of insurance per depositor Percentage of deposits in failing banks Number of failing banks Growth in bank deposits Growth in number of bank accounts Adjusted R’ F-statistic Durbin-Watson statistic

(0.00) -16.518 (-4.01)

0.302 9.381 2.119

0.014 (1.12) - 16.507 (-4.06)

0.3 18 10.013 2.138

(3) 0.338 (0.44) 0.598 (3.371) -346.486 (-0.87)

- 16.046

(-3.53) 0.016 (0.14) 0.380 6.300 2.134

(4)

0.44 I (0.57) 0.602 (3.38)

0.020 (0.67) - 16.087 (-3.53) -0.029 (-0.02) 0.315 6.175 2.178

Note: Numbers in parentheses are t-statistics.

The constructed time series on number of accounts was used in regressions (3) and (4) for the years 1934-81. The variable does not help explain the behavior of the dependent variable. However, this should not be taken as evidence that account-creating activity of depositors had no effect on coverage. The correlation between the number of accounts and the percentage of covered deposits from 1934 to 1981 is high, 0.93, reflecting a common trend. In the regression, the percentage of insured deposits is first-differenced, but the application of this procedure to accounts is first-differencing a variable, many of whose observations are fitted to the trend, thus rendering it relatively weak in the regression. While the quality of the data on accounts does not permit very robust tests of the effects of individual account-creating activity, qualitative evidence implicates account creation as an important factor from the beginning of the FDIC until at least the mid-1970s. Although account creation may have been more important, the regressions only identify the FDIC’s increased coverage per account in 1950, 1966, 1969, 1974, and 1980 as a key factor. The first increase in real coverage in 1950 was the product of lobbying by unhappy sectors of commercial banking. Afterward, it was not commercial banks but rather their rivals that pushed for expanded coverage. 3.4 Raising Deposit Insurance in 1966 and 1969: The Role of the S&Ls Savings and loan associations (S&Ls) originally had little interest in deposit insurance. They were very cautious about advocating any guarantee system

105

The Legacy of Deposit Insurance

and probably would never have supported one if commercial banks had not obtained the FDIC (Ewalt 1962). S&Ls were given the opportunity to obtain federal deposit insurance at the same time as Congress established the FDIC. The National Housing Act of 1934 established the Federal Savings and Loan Insurance Corporation (FSLIC), almost as an afterthought, to provide a full set of institutions to S&Ls to parallel those for banks9 Many thrifts had found it advantageous to join the Federal Home Loan Bank (FHLB) System. The purchase of shares in one of the 12 regional Federal Home Loan Banks gave them access to FHLB credit facilities but did not impose any additional regulations on them (Grossman 1992). Many fewer took out charters to become federal mutual savings and loan associations. Supervised by the Federal Home Loan Bank Board (FHLBB) and narrowly constrained in lending, a federal charter appeared relatively unattractive to most S&Ls. Although federally chartered S&Ls were required to join the FSLIC, members of the FHLB system were not so obliged. This regulation contrasted with insurance provisions for banks, under which all Federal Reserve members-national banks and state bankswere required to obtain FDIC insurance. Thus, by 1940, half of all S&Ls had joined a Federal Home Loan Bank, but only 20 percent took out federal charters. Unlike the banks, for which deposit insurance was almost universal from the outset, only 30 percent of the S&Ls, with 50 percent of assets (see fig. 3.6), had obtained FSLIC insurance by 1940. The initial responses of banks and S&Ls to deposit insurance reflected their different experiences during the Great Depression and the costs and benefits of insurance they faced. Both industries suffered severe withdrawals of deposits between 1929 and 1933. Commercial banks lost 17 percent of their deposits and S&Ls 28 percent. S&Ls were forced to endure a larger contraction, but it was more orderly. Between 1929 and the end of 1933, the number of banks fell from 24,504 to 14,440; yet the number of S&Ls only declined from 12,342 to 10,596. Unlike the banks, which had to wait for state and then federal bank holidays to refuse customers payment, the S&Ls had the right to put depositors “on notice” and refuse to meet demands for withdrawals until loan repayments came in. Thus, S&Ls had a device to ward off the runs that devastated banks and saw less advantage to insurance that required acceptance of more federal regulation. In addition, FSLIC insurance came at a higher price. The FSLIC premium was 0.125 percent of deposits, whereas FDIC insurance was 0.0833 percent. The FSLIC rate was only reduced to the FDIC level in 1951 (Grossman 1992). After World War 11, the thrifts were one of the fastest growing groups of financial intermediaries. The New Deal conferred a variety of advantages on thrifts, whose share of all financial intermediaries assets rose from 6 percent 9. In 1932, Congress passed the Federal Home Loan Bank Act, which created the Federal Home Loan Banks and Federal Home Loan Bank Board, which paralleled the Federal Reserve System. The Home Owners Loan Act of 1933 gave the board authority to charter a new class of intermediary, federal mutual S&Ls, thus creating for the thrift industry a dual federal-state regulatory system that paralleled the dual banking system.

106

Eugene N. White

-

A l l S&Ls

----

A s s e t s of all S&Ls

Fig. 3.6 Membership in the FSLIC, 1935-73

Sources: Federal Home Loan Bank Board (various years).

in 1950 to 13 percent in 1970. Although imperfect substitutes for commercial banks’ demand deposits, which paid no interest, S&L interest-bearing passbooks were attractive to small savers and competed with banks’ time deposits. By 1950, 50 percent of S&Ls, with 80 percent of all assets, had joined the system (see fig. 3.6). However, unlike banks, FSLIC-insured institutions had almost all their accounts insured. In 1941, 86 percent of all savings capital (deposits) in S&Ls were insured, rising to 94 percent by 1947. This high level of insurance is attributable to the predominance of small savers with balances averaging well below $1,000 (FHLBB 1947). This nearly complete coverage helps to explain why the insured S&Ls did not participate in the 1950 deposit insurance debate. Ten years later, conditions had changed dramatically. When the flow of savings deposits surging into S&Ls came to an abrupt halt in the credit crunch of 1966, S&Ls became interested in deposit insurance. The similarity of coverage among thrifts assured a fairly uniform view of the desirability of increased insurance in contrast to the wide divergence of opinion among banks. Neither banks nor S&Ls saw the erosion in the real value of deposit insurance per depositor as a threat. The decline in real deposit insurance shown in figure 3.1 was slight compared to what happened before the 1950 increase. Furthermore, total coverage of deposits, shown in figure 3.2, was fairly stable. However, there was a significant drop in the number of fully insured accounts that especially affected small banks. Although coverage dropped for most

107

The Legacy of Deposit Insurance

classes of banks, the three smallest categories of banks in table 3.3 show very large declines in coverage of deposits between 1951 and 1966. Inflation and the shift between groups make comparisons between years difficult, yet the danger posed by this decline in coverage is clear in table 3.2. Here, the percentage of commercial banks with 90-100 percent of their deposits insured by the FDIC dropped from 23.2 percent in 195 1 to 9.9 percent in 1964. Thus, a small but significant fraction of the banking industry was feeling increasingly exposed. By the mid-1960s, banks and thrifts were also worried that interest rate restrictions reduced their ability to attract deposits. While banks had been subject to Regulation Q interest rate ceilings since 1935, FHLB member thrifts were constrained by FHLBB rules, which imposed a variety of restrictions on the “dividends” paid on savings account (FHLBB 1965). In 1965, the limit on the interest charged on banks’ time deposits stood at 5.5 percent, yet very few S&Ls could offer rates in excess of 5 percent. In this year, market rates moved above the ceilings, and both banks and thrifts began to lose funds.I0 Thrifts experienced a 4 percent fall in funds available for new investment, followed by a 28 percent fall in 1966, when savings inflows and loan repayments fell off. The big demand for advances from the FHLBB, led the board to ration lending to S&Ls, which then slashed mortgage lending by one-third. In response, the board adopted a more flexible dividend policy; and by the end of 1966, over 20 percent of S&L deposits were paying 5.25 percent (FHLBB 1966). Interest rate regulations needed some unification to preserve the system. The Treasury and the FDIC proposed that the FHLBB be given more supervisory authority and power to set maximum interest and dividend rates. Many S&Ls were not enthusiastic about the prospect of new FHLBB regulation, but they were willing to countenance more control if it would ensure that deposit inflows resumed. Of considerable concern to the S&Ls was that savers were showing great reluctance to hold deposits in excess of the $10,000 level of coverage. One board study showed that there was an “artificial bulge” in the number of S&L accounts at the $10,000 level, indicating that people were limiting their deposits (Congressional Record 1966, 112, pt. 15: 2035 1-52). Efforts to raise insurance predated the 1966 credit crunch, but demands had become more urgent. Hearings in Congress were held in 1963 to consider a rise in insurance coverage to $25,000 for banks and S&Ls. Over the next three years, congressmen wrangled over the level of coverage and whether the FHLBB should be granted additional regulatory powers. In congressional hearings, there was no protest by the FHLBB, the Board of Governors of the Federal Reserve, or the FDIC about the increase in insurance. They were much more concerned about the effects of changing supervisory practices. Rep. 10. Some aggressive thrifts employed brokers to advertise and collect funds for them (Marvel1 1969, 133-36).

108

Eugene N. White

Wright Patman (D-Tex.), chairman of the House Banking and Currency Committee, vigorously argued for a simple increase in coverage. He brushed aside arguments that individuals could easily secure coverage by creating multiple accounts and claimed that the current $10,000 maximum coverage encouraged everyone from businessmen to widows to firemens' funds to put their money in out-of-town banks once the ceiling was reached in local banks. Patman slammed the big banks for pressuring their correspondent banks to block an increase in insurance, portraying them as predators anxious to drive the S&Ls out of business (Congressional Record 1966, 112, pt. 13:20354). Congress navigated through these complex, competing interests in writing the Interest Rate Control Act of 1966. The act extended Regulation Q to thrifts but gave them a favorable differential. Thrifts were allowed to pay 3/4 of 1 percent in interest more than banks, in the hope of channeling funds back to the mortgage market. Congress also gave more supervisory authority to the FHLBB. The legislators settled on increasing deposit insurance for persons holding accounts in banks and thrifts to $15,000, a relatively low number as far as many thrifts were concerned. Following the 1950 deal, the 1966 package provided a sweetener for the larger banks in the form of an increase in the assessment rebate to 66 2/3 percent. Although interest rate flexibility was clearly of greatest concern to intermediaries, the rise in insurance did help. No data exist on insurance coverage among thrifts, but the level of insured deposits rose for all sizes of banks in table 3.3. These adjustments to the New Deal system did little to alleviate the underlying problems. Once again in 1969-70, tighter monetary policy pushed market rates above the Regulation Q ceilings. S&Ls saw virtually no net inflow of new funds, while commercial banks lost funds, in contrast to the 1966 experience when S&Ls were in greater distress. S&Ls were in better shape thanks to the favorable differential in interest rate ceilings. Still, there were gaps in the interest rate controls. A substantial number of mutual savings banks in the Northeast that were not members of the Federal Reserve or the FDIC avoided the controls, as did non-FSLIC-member thrifts. These institutions' higher rates were drawing funds away. Congress responded to the complaints of controlled banks and thrifts by extending Federal authorities' control of all institutions in states where over 20 percent of savings were held by non-federally regulated institutions (FHLBB 1969). With no debate, Congress also raised deposit insurance coverage for banks and thrifts on 23 December 1969, from $15,000 to $20,000.'' This hike halted the new decline in fully insured accounts depicted in figure 3.4. The real value of coverage was now higher than it had ever been (fig. 3.1), reaching approxi11. Initially, the FDIC and FSLIC relied on state laws to define what constituteddifferent deposit ownership, allowing people in some states to set up multiple accounts within banks and attain coverage many times the limit intended for individuals. In 1968, the FDIC and FSLIC joined together to produce a consistent set of rules on how to treat multiple accounts, placing some limits on protection (Marvel1 1969, 106-11).

109

The Legacy of Deposit Insurance

mately $7,000 in 1934 dollars. The percentage of insured accounts and deposits of FDIC institutions were at all-time highs of 99.1 and 63.1 percent (fig. 3.2), with institutions of all sizes (table 3.3) benefiting from the increase. By 1969 there were only 208 noninsured commercial banks and nondeposit trust companies and 166 noninsured mutual savings banks, virtually all of the latter being located in Massachusetts and covered by its deposit insurance system. (FDIC, Annual Report 1969). In the thrift industry, over 70 percent of the S&Ls, with over 90 percent of assets, were covered by 1969. Deposit insurance coverage in the 1960s had grown considerably for banks and thrifts, well beyond the initial intentions of the New Deal. 3.5

The Spread of Insurance

Greater interest rate volatility and increased competition in the 1960s created difficulties for all financial intermediaries. Facing these new challenges, credit unions, broker-dealers, pension funds, and insurance companies sought the benefits of government-provided insurance for their liabilities. As they held relatively modest or no funds on deposit and no claim could be made that insurance would serve to prevent a panic, the history of these intermediaries demonstrates how, even in the absence of concern about macroeconomic instability, new classes of intermediaries were successful in lobbying Congress to expand insurance far beyond its New Deal boundaries. Designed to assist the small saver, credit unions grew rapidly in the postwar period. The Federal Credit Union Act of 1934 made federal credit union charters available, as an alternative to state charters, and they soon dominated the industry. The number of all credit unions-federal and state-more than doubled from 10,571 in 1950 to 23,656 in 1970, with deposits climbing from $880 million to $15.5 billion.I2Like the S&Ls, credit unions were initially reluctant to press Congress to create institutions for them. But competition from federally assisted and protected banks and thrifts coupled with increased financial difficulties led the credit unions to aspire to parity with banks and thrifts. Between 1934 and 1969, over 5,600 federal credit unions were 1iq~idated.l~ Failures were increasing, and in 1969, 274 federal credit unions were closed, 35 of them at a total loss of $95,000 to their members. Some assistance for failing firms came from credit union leagues, which bailed out 280 other credit unions; but these private reserve funds were very small. Failures induced Massachusetts in 1961, and later Wisconsin and Rhode Island, to create state funds; but they were restricted to state-chartered credit unions, a small fraction of the industry (Congressional Record 1970, 116, pt. 23:30734-47). Prompted by the credit crunches of 1966 and 1969, credit unions pressed 12. However, they were small by comparison with the 5,669 S&Ls, which held $146 billion in deposits, and the 14,187 banks, which held $505 billion in deposits in 1970. 13. Adequate data exist only for federally regulated credit unions.

110

Eugene N. White

for lending and insurance institutions to parallel the Federal Reserve, the FHLB system, the FDIC, and the FSLIC. In 1970, Congress obliged them with the Federal Credit Union Act, creating the National Credit Union Administration (NCUA), an analogue to the Federal Reserve and FHLBB. While this bill was making its way through Congress, an amendment was made to create a system of insurance for credit unions.I4 The amendment was initially sponsored by several senators, and there was no apparent opposition from either banks or thrifts. A simple rationale was given by one sponsor, Sen. Wallace F. Bennett (R-Utah), who pointed out that federally chartered credit unions were the only depository institutions not covered by a federal insurance program. Bennett admitted that the absence of insurance posed no threat to the stability of the financial system and that the losses of credit unions had been small. Insurance coverage for credit unions was almost a matter of pure competitive equity. Established in 1970, the National Credit Union Share Insurance Fund (NCUSIF) gave the credit unions an insurance system. At the same coverage per depositor of $20,000 as the FDIC and the FSLIC, 22 million credit union members, who had an average of $650 on deposit, gained ample protection. Like the FDIC and the FSLIC, the NCUSIF was mandatory for federally chartered credit unions and optional for state institutions. Administered by the NCUA, the fund charged an annual premium of 0.0833 on the aggregate of members’ accounts and creditor obligations. Adoption of federal insurance was not initially universal. Many state-chartered credit unions did not want to accept the federal regulations necessary to obtain NCUSIF insurance. At the behest of these institutions, more states created their own insurance funds. In 1981, when California established the California Credit Union Share Guaranty Corporation, there were 16 state funds, covering 3,150 credit unions with $12 billion in deposits (NCUA 1982). However, in the wake of widespread failures of banks, S&Ls, and credit unions in the 1980s, there was a flight to the NCUSIF, which afforded greater protection. In 1981, NCUSIF-insured credit unions held 82 percent of all credit union shares. By 1985 this figure jumped to 92 percent, rising to a nearly universal 99 percent in 1995 with the demise of the state insurance plans (NCUA 1989, 1995). In the same year that credit unions secured federal protection for their depositors, customers of broker-dealers received guarantees for their funds on deposit-protection that the original New Deal had never countenanced. The Securities Exchange Act of 1934 tried to protect customers from brokers’ dishonesty but not their incompetence. Protection from incompetence was the responsibility of the relevant self-regulatory organization-the New York Stock Exchange (NYSE) or the National Association of Securities Dealers (NASD). These organizations could intervene and transfer customer accounts 14. The legislation enjoyed wide support among credit unions: when the Credit Union National Association surveyed its membership, 92 percent supported the bill.

111

The Legacy of Deposit Insurance

from a weak to a strong member firm, liquidate failing members, or merge weak firms with stronger ones (Teweles and Bradley 1987). The rising volume of activity on the exchanges during the 1960s bull market put an enormous strain on brokerages’ ability to handle the complex paperwork that accompanied every transaction. The number of “fails” or failures to deliver security certificates or complete transactions produced a “back office” crisis. Many firms were swamped by business and could not manage their operations well. Firms used customers’ free credit balances for any business purpose, including trading or underwriting, putting these funds on deposit at risk. When Ira Haupt and Company failed in 1963, as a result of a huge default on commodity contracts, the NYSE stepped in and assisted with the firm’s liquidation (Teweles and Bradley 1987). Anticipating more problems, the NYSE created a special trust fund of $10 million and a $15 million line of credit in 1964 to assist troubled members and protect customers (Sowards and Mofslcy 1971). The American Stock Exchange (AMEX) followed the NYSE’s lead, and by 1968 all the exchanges had established special funds (Sobel 1972; U.S. Senate 1970). When the bull market broke in 1969 and prices and volume fell, many brokerages held large inventories. Falling revenues and costly inventory losses led 129 NYSE member firms to be liquidated, merged, or acquired, and another 70 required some assistance from the exchange. The special fund ran out of funds in the summer of 1970 and was unable to pay out customers’ accounts in failed brokerages. In this emergency, the NYSE transferred $30 million from its building fund to its special fund. However, it was clear that if a large brokerage went uhder, the resources of the special fund would be inadequate (Sobel 1975). The free credit balances-in effect, the funds customers held on deposit with broker-dealer firms-stood at $3 billion in 1970 for NYSE member firms. In addition, broker-dealers had custody of $50 billion in customer securities ( US . Senate 1970). Although there were no runs on brokerages, the exchanges appeared unable to provide sufficient protection on their own. Insurance, equivalent to FDIC and the FSLIC, was viewed as a reasonable solution by the securities industry and the public (Seligman 1982). The House report on insurance legislation was explicit: “Failures may lead to loss of customers’ funds and securities with an inevitable weakening of confidence in the U.S. securities markets. Such lessened confidence has an effect on the entire economy. . . . The need is similar in many respects to that which prompted the establishment of the Federal Deposit Insurance Corporation and the Federal Savings and Loan Insurance Corporation” ( U S House 1970, 2). This misreading of history identified macroeconomic stability as the prime reason for insurance, when special interests in the financial industry always had the keenest interest in the establishment of insurance funds. A proposal was put before Congress to establish a Securities Investor Protection Corporation (SIPC) to act as an FDIC or FSLIC for the securities industry. The bill had the support of the Securities and Exchange Commission (SEC),

112

Eugene N. White

the Department of the Treasury, and Congress’ Joint Securities Industry Task Force. An old New Dealer, Emanuel Celler (D-N.Y.) questioned the intention of insuring all firms registered with the SEC without any inspection or further regulation. These qualms were repeated by other congressmen; but like the bill for credit unions, the idea of insuring customer accounts had wide support in Congress (Congressional Record 1970, 116, pt. 29:39345-70). Congress passed the Securities Investor Protection Act (SIPA) in December 1970. This act created the SIPC, which was charged to administer a fund providing protection up to a maximum of $50,000 for both cash and securities, with a limit of $20,000 for cash. This insurance was mandatory for brokerdealers registered with the SEC, making coverage nearly universal from the outset. All SIPC members were assessed 3/16 of 1 percent per year of gross revenues from the securities business for the SIPC fund (Matthews 1994). If needed, the corporation could borrow up to $1 billion from the U.S. Treasury with the approval of the SEC (Seligman 1982).15 Under SEC oversight, the SIPC has no authority to examine or inspect its members. Instead, the securities exchanges and the NASD are the examining authorities for their members, and the SIPA gave the SEC additional authority to adopt rules relating to the acceptance, custody, and use of customers’ securities, deposits, and credit balances.l6 The examples of insurance for credit unions and broker-dealers reflect the low tolerance for even small losses to the customers of financial intermediaries and the drive for equal competitive advantage. Although concern about the effects of failures on the stability of the financial system was often discussed, it motivated few participants in the legislative process. The spread of insurance to nondepository intermediaries, where a financial panic or run is not a concern, highlights this fact. Both pension funds and insurance companies responded to the favorable political circumstances to demand insurance. Underfunding of private defined-benefit pension plans left workers without pensions when their employers went bankrupt. The Pension Benefit Guaranty Corporation (PBGC) was established by Title IV of the Employee Retirement Income Security Act in 1974 to protect retirement incomes from definedbenefit pension plans. Financed by premiums collected from companies, the PBGC’s coverage of pensions reached over one-third of the workforce by 1995 (PBGC 1995). While insurance of pensions became a federal responsibility, the guarantee of life insurance became a state responsibility as the federal government had never ventured to regulate life insurance. Before 1970, only New York had a guarantee system to protect policyholders. A rise in failures of life 15. Some brokerage firms carry additional commercial insurance on accounts exceeding SIPC coverage. 16. Before this act, there were no SEC or exchange rules regarding the use of customers’ credit balances or other balances in possession of broker-dealers.After 1973, the SEC limited the use of customers’ funds to finance margin loans to other customers and other customer-related activities (Matthews 1994,55-56).

113

The Legacy of Deposit Insurance

insurance companies prompted the National Association of Insurance Commissioners to recommend a model guarantee system to state legislatures in 1970.Although the plans vary from state to state, funds guarantee insurance in all 50 states (Brewer and Mondschean 1993). By the early 1970s, financial pressures had pushed the insurance of liabilities beyond the banking system to the securities, pension, and insurance industries. There was no anticipation that the FDIC, FSLIC, NCUSIF, SIPC, PBGC, or state insurance systems could fall into trouble. In fact, the spread of insurance helped to prompt new demands from depositories for increased protection.

3.6 The 1974 Increase in Insurance In 1973, Fernand St. Germain (D-R.I.) offered a bill to increase deposit insurance from $20,000 to $50,000 and provide 100 percent insurance for all government deposits, amending the FDIC act, the National Housing Act, and the Federal Credit Union Act. Where did this demand for more protection come from? Once again, there was no cry by the public for increased protection. As seen in figure 3.1, inflation had reduced the real value of insurance after the 1969 increase, but a $50,000 increase would have been a huge increase in real coverage. Total FDIC coverage of deposits in figure 3.2 had sagged a bit, but it was slight for all sizes of banks in table 3.3. The interest group at work behind this new proposal was the thrift industry, although some banks also were eager for higher levels of coverage. An appeal was made to raise coverage to $50,000 to achieve parity with the securities industry-even though brokerage accounts only had insurance of $20,000 for cash. Frank Willie, chairman of the FDIC, took the view of the small banks in testifying that more insurance was required because “depositors seem to believe that their money is safest in the largest institutions. . . . A depositor is more likely to put funds exceeding the insured limit in a large commercial bank than a small one” (U.S. House 1973, 14). In addition, he pointed out that more insurance would reduce the flight of funds from depository institutions to nondeposit institutions and markets. The thrifts appeared to be especially eager to attract state and local deposits and were relentless in their congressional testimony about the need for 100 percent insurance of government deposits (U.S. House 1973). The representative of the U.S. Savings and Loan League described the cumbersome process of depositing county or state funds into multiple accounts, none exceeding the $20,000 limit, to ensure full protection. In addition, many states required that bonds be used to collateralize deposits, with requirements varying from one locality to the next. Donald P. Lindsay of the National League of Insured Savings Associations described the task of the King County treasurer in Washington State, who kept 552 S&L passbook accounts to ensure that county funds were fully protected. He also gave the example of a city treasurer in Washing-

114

Eugene N. White

ton State who mistakenly calculated the FSLIC coverage within one S&L and lost funds. The National Association of Mutual Savings Banks supported 100 percent insurance of government deposits, hoping for more business (U.S. House 1973). The vice president of the Credit Union National Association, William D. Heier, supported the St. Germain bill. Since the Federal Credit Union Act prohibited federal credit unions from receiving funds from state and local governments, he proposed an amendment to allow credit unions to receive such funds. While Willie favored higher individual coverage, he resisted full coverage for government deposits. The chairman of the FDIC pointed out that public depositors’ losses had been very small and they had recovered 99 percent of funds from failed banks. He was concerned that this innovation would imperil the insurance fund. An increase in coverage for all depositors to $50,000 would have caused the ratio of the insurance fund to insured deposits to fall from 1.28 to 1.13 percent. Willie did not find this alarming, except when coupled with 100 percent insurance for public units. The full coverage for public units would have driven the coverage of the insurance fund to 1.04 percent. At such a level, the fund might easily be exhausted if large banks continued to fail. In contrast, Thomas R. Bomar, head of the FWLBB, was more sanguine and fully supported the position of the thrift industry, testifying that the FSLIC fund would not be put at risk by 100 percent insurance of government deposits (U.S. House 1973). In spite of the growing demands from many parts of the financial industry for more and more insurance, some sectors resisted. One official of the American Bankers Association, H. Phelps Brooks, Jr., president of Peoples National Bank of Chester, South Carolina, made their case: “Full insurance coverage of public accounts will open the door to pressure for 100 percent insurance of all accounts.Account holders with quasi-public responsibility could well ask why their savings or checking accounts above $20,000 are any less important than Government funds. . . . When the county sewer district promptly receives 100 percent of its deposits upon closing of the institution, the officials at the local private hospital will certainly feel entitled to special consideration. Then other depositories with large accounts would not understand why their accounts are not fully covered.” Brooks concluded that 100 percent coverage would have detrimental effects on the sound management of depository institutions (U.S. House 1973, 114). Faced with these strongly held conflicting positions, Congress passed compromise legislation in 1974. Insurance of deposits for individuals and businesses was lifted to $40,000, while the government deposit guarantee was hiked to $100,000. This legislation applied to commercial banks, mutual savings banks, and thrifts. SIPC protection was raised to $100,000 in cash and securities, with a $40,000 maximum for cash. The result was a dramatic rise in real protection as seen in the data for the FDIC. The real value of insurance rose in figure 3.1, as did the total coverage of deposits in figure 3.2. All sizes

115

The Legacy of Deposit Insurance

of banks, except the very largest, as seen in table 3.2, achieved much higher rates of protection for their deposits and accounts. Five years of legislation, beginning in 1970, had spread insurance to institutions beyond the banking system and dramatically raised the level of insurance. Until the S&L crisis broke, a further increase in insurance appeared unlikely.

3.7 $100,000 Insurance and “Too Big To Fail” The collapse of the S&L industry has been extensively chronicled (Barth 1991; Kane 1989; White 1991). By the end of the 1970s, the income and net worth of the thrift industry was plummeting. Measured by book value, the net worth of the thrift industry fell from 5.5 to 0.5 percent of assets between 1977 and 1982, but any market value method showed the industry as whole to be insolvent by about $100 billion. The FSLIC possessed only $6.5 billion of reserves and could have paid off only a fraction of the deposits of insolvent thrifts. The housing industry did not want massive S&L closures, and the Reagan administration had no desire to see a doubling of the federal deficit. A militant S&L lobby pressured the FSLIC into a policy of forbearance-putting off any serious attempt to discipline or close thrifts. With generous PAC money, the thrifts also helped to persuade Congress to give them another chance to recover. The results of intense lobbying by the thrifts and other financial institutions were the Depository Institutions Deregulation and Monetary Control Act of 1980 (DIDMCA) and the Garn-St. Germain Act of 1982. All financial institutions, banks and thrifts, began a phased eliminated of interest rate ceilings over the next six years. The 1982 act authorized banks and thrifts to offer money market deposit accounts to compete with money market mutual funds. Furthermore, S&Ls were given a whole new range of powers. They were released from their traditional portfolio constraints and permitted to increase consumer loans, commercial real estate mortgages, and business loans. In addition to this legislation, the FHLBB diluted capital requirements. Congress did not openly discuss the issue of deposit insurance. There was considerable opposition to any further protection. Federal bank regulators strongly opposed an increase in coverage, emphasizing that it would cause some institutions to take more risks. Instead, the increase in coverage was added quietly and quickly to DIDMCA in a House-Senate conference session to placate the thrifts, which feared the impact of interest rate deregulation (Litan 1994). The 1980 act raised federal deposit insurance coverage for individual deposits from $40,000 to $100,000 for banks, thrifts, and credit unions. Customer accounts with broker-dealers were now insured up to a maximum of $100,000 in cash and $500,000 in both cash and securities. The result of this legislation was a big increase in the real value of insurance (fig. 3.1) to approximately three times the 1935 level. The percentage of insured deposits was racheted up

116

Eugene N. White

(fig. 3.2); and, as table 3.2 shows, the leap from $40,000 to $100,000 brought a much higher rate of protection for all classes of banks. Deposit insurance was locked firmly in place, yet since 1980, there has been no further increase in deposit insurance. As of 1996, it has been 16 years since there was any nominal increase in coverage. Unlike the end of two periods of similar length, 1934-50 and 1950-66, there is no swelling demand for a new rise. The real value of insurance has declined with inflation, but it is still more than 50 percent higher than the 1934 level. Some constraints have been placed on insurance. There have been some additional limits placed on the coverage of accounts to limit the creation of joint and multiple accounts to expand coverage.” Following the 1986 increase in the minimum capital ratio to 6 percent (White 1991), the Federal Deposit Insurance Corporation Improvement Act of 1991 mandated the creation of risk-based insurance premiums in an attempt to control the problem of moral hazard. While the high real level of coverage may have reduced the demand for insurance, there are other important factors at work, most notably, the “too big to fail” policy providing de facto 100percent insurance. Deposit insurance was a useful instrument for guaranteeing relatively small deposits. The advent of very large denomination, uninsured certificates of deposit (CDs) allowed banks greater ability to manage their liabilities. But it left them subject to the judgment of the money market. Rumors of insolvency panicked large CD holders into a run on Continental Illinois in 1984. The Federal Reserve and the FDIC intervened to protect all depositors, large and small, because they feared that losses would precipitate runs on other banks, generating a systemwide crisis. The bailout of Continental Illinois certified the too-big-to-fail policy that had been evoked in the early 1970s in the case of selected banks, like Franklin National (Sprague 1986). Although initially aimed at only the money center banks, the doctrine was extended in varying degrees to other big banks (Boyd and Gertler 1993). This subsidization of risk taking by large banks produced an incentive to grow. When combined with the reduction in geographic barriers to branching and holding companies, a wave of mergers and acquisitions began in the 1980s. The winnowing of weak institutions in the bank and thrift crises of the decade and this consolidation of the banking industry has reduced the lobbies that previously pushed for higher coverage while leaving deposit insurance firmly in place.

3.8

Conjecture and Conclusion

In the public’s eye, deposit insurance is still considered one of the great successes of the New Deal. While many economists no longer hold it in such high regard, any serious rollback is politically inconceivable. Public acceptance of deposit insurance for banks and thrifts, even with numerous costly 17. See http://www.fdic.gov for the details of these restrictions.

117

The Legacy of Deposit Insurance

failures, has enabled these intermediaries to obtain higher levels of real coverage and made it easier for other institutions to press their claims for insuring their liabilities. A reasonable policy question is whether the cost of deposit insurance exceeded the cost of bank failures in the absence of deposit insurance, following the Great Depression. This counterfactual question is potentially complex, and I will only consider here the case with the available complete data for the FDIC. The New Deal greatly altered the structure of the financial system. The constraints that were placed on banks allowed other intermediaries to capture what otherwise might have been banking business. Thus, the size of the banking sector is smaller than it would have been in the absence of the New Deal. Similarly, the regulations on bank portfolios altered the liquidity and risk of banks, affecting the probability of bank failure. Any attempt at constructing what the banking system would have looked like and how many failures would have occurred in the absence of the New Deal requires grand simplifying assumptions. Aware of these difficulties, I offer here a simple, suggestive counterfactual where macroeconomic policy continued to be generally stabilizing after World War 11, preventing any new great depression. First, I estimated the real cost of bank losses under the FDIC from 1945 to 1994. The cost here is taken to be the administrative and operating expenses of the FDIC plus the losses from bank failures. To estimate the latter, I considered the losses from the three types of FDIC interventions: deposit payoffs, deposit assumptions, and assistance transactions. For payoffs, I took the estimated losses (disbursements less recoveries) plus the deposits not reimbursed by the FDIC (estimated by the total deposits times the fraction of uninsured deposits).18For assumptions and assistance transactions, the estimated losses to the FDIC were used. The total losses for each year were converted into real dollars, employing the consumer price index with 1982-84 as base year. As presented in table 3.6, the total cost of resolving bank failures with the FDIC was $39 billion for 1945-94, or an annual cost of $770 million. The present discounted value of the cost of bank failures from the beginning of the postwar era, 1945, was $7.8 billion. This starting date was selected to omit the chaos and cleanup of the 1930s. What the bank failures would have looked like in the absence of the New Deal is difficult to estimate. Banking and Monetary Statistics (Board of Governors 1943) reported the estimated losses to depositors for all bank failures from 1921 to 1941. The average annual loss rate on total bank deposits for 1921-28, the nearest period of stability without insurance, was 0.1032 percent. If we assume that the structure of the banking system after 1945 remained essentially the same as it was in the 1920s and the shocks to the economy were the same, then we could use the loss rate to estimate the losses to depositors in the ab18. Data on losses to customers whose accounts were over the maximum level of coverage were not apparently obtainable.

118

Eugene N. White

Table 3.6

Cost of FDIC Insurance and Counterfactual Deposit Losses, 1945-94 (billion 1982-84 dollars)

Assumption Under the FDIC No FDIC Loss rates of 1921-28 Bank failure rates 0.05 percent 0.1 percent 0.2 percent 0.283 percent

Estimated cost AAClPDV

Recovery Rates (%)

80

65

50

48

,1412 ,2615 .52/10 .74114

,2414 ,4619 ,92117 1.08124

,6616 .66/12 1.32124 1.86134

,68116 ,68113 1.38125 1.94136

.7717.8 .96/12.8

Note: Table reports estimated cost of bank failures. Entries list average annual cost (AAC)/present discounted value (PDV).

sence of the FDIC. Multiplying the loss rate times the real deposits of insured banks for each year of 1945-94 yields a potential annual loss of $960 million, or a present discounted value of $12.8 billion. There is also reason to think that estimate is high because the banking system was undergoing a shakeout in the 1920s, as many small banks were disappearing. The Great Depression accelerated this process and eliminated virtually any bank showing signs of weakness. The recession of 1936-37 would have further winnowed the banking system. Furthermore, the New Deal halted the process of merger and consolidation that had started in the 1920s. This development would certainly have continued more vigorously in the post-World War I1 period in the absence of New Deal banking regulation. Both the destruction of weak banks and the formation of larger banks would have produced a stronger banking system with fewer losses. An alternative approach to estimating the losses to depositors in the absence of the FDIC is to use varying bank failure rates and recovery rates. The ratio of deposits in suspending banks to total deposits for the period 1921-28 was 0.29 1 percent. For a slightly longer period (1907-29) and for national banks only, it was 0.283. Table 3.6 offers four possible failure rates. Beginning in 1907, the comptroller of the currency (U.S. Comptroller of the Currency, vanous years; and see Calomiris and White 1994) produced detailed records of recoveries and losses for national banks. No single detailed source exists for state-chartered banks. The recovery rates used are percentages paid out on proved claims three years after suspension. After three years, recoveries are very low. The average recovery rate for suspended national banks from 1907 to 1927, weighted by bank deposits, was 48 percent.19 The recovery rate for the FDIC on its disbursements for failed banks from 1934 to 1994 was 65 19. If the rates are weighted by the number of banks, the average rate is 42 percent. I stop in 1927 because any later year collections were being made during the depression.

119

The Legacy of Deposit Insurance

percent (FDIC, Annual Report 1994). Whether the FDIC was more efficient than the receivers under the national banking system or whether the nature of the failures or economic conditions were different is difficult to determine. Rather than hazard a guess, table 3.6 offers several recovery rates, ranging from 48 to 80 percent and including the FDIC and national bank suspension rates. Table 3.6 provides a range of counterfactual estimates. If banks in the post1945 period continued to fail at the same rate as national banks had in 1921-27 and had the same low recovery rate, depositors might have been hit with losses of $1.86 billion per year, much more than under the FDIC. However, this estimate is certainly an upper bound. If failure rates were lower and recovery rates were higher-both plausible assumptions with a stronger banking systemthen costs to depositors would have been similar or even lower than under the FDIC. For a broad range of estimates, it appears that the FDIC did not reduce costs and may have raised them. Unfortunately, given the absence of comparable data, it was not possible to conduct this exercise for the FSLIC. However, the sheer magnitude of the S&L disaster of the 1980s relative to the calm of the 1920s strongly suggests that the FSLIC imposed very high costs compared to an uninsured system. Even given the tenuous nature of these estimates, it is hard to escape the conclusion that deposit insurance did not substantially reduce aggregate losses from bank failures and may have raised them. What it did do was to alter the distribution of losses. Instead of a small number of depositors bearing the losses of a relatively small number of banks, costs were distributed to all depositors and hidden in the premia levied on banks. While these costs remained large in aggregate, they appeared to have vanished to the individual depositor. The new distribution of the costs of failure made the FDIC a widely accepted program and has ensured the continuance of deposit insurance into the next century.

References Barth, James R. 1991. The great savings and loan debacle. Washington, D.C.: American Enterprise Institute Press. Board of Governors of the Federal Reserve System. 1943. Banking and monetary statistics. Washington, D.C.: Board of Governors of the Federal Reserve System. Boyd, John H., and Mark Gertler. 1993. In NBER macroeconomics annual 1993, ed. Olivier J. Blanchard and Stanley Fischer. Cambridge, Mass.: MIT Press. Brewer, Elijah, 111, and Thomas H. Mondschean. 1993. Life insurance company risk exposure: Market evidence and policy implications. Contemporary Policy Issues 1 1 :56-69.

Bums, Helen M. 1974. The American banking community and New Deal banking reforms, 1933-1935. Westport, Conn.: Greenwood.

120

Eugene N. White

Calomiris, Charles W. 1990. Is deposit insurance necessary? Journal of Economic History 50 (June): 283-95. . 1993. Regulation, industrial structure, and instability in U.S. banking: An historical perspective. In Structural change in banking, ed. Michael Klausner and Lawrence J. White. Homewood, Ill.: Irwin. Calomiris, Charles W., and Eugene N. White. 1994. The origins of Federal Deposit Insurance. In The regulated economy: A historical approach to political economy,ed. Claudia Goldin and Gary D. Libecap, 145-88. Chicago: University of Chicago Press. Congressional Record. 1966. 89th Cong., 2d sess. Washington, D.C. . 1970. 91st Cong., 2d sess. Washington, D.C. Ewalt, Josephine H. 1962. A business reborn: The savings and loan story, 1930-1960. Chicago: American Savings and Loan Institute Press. Federal Deposit Insurance Corporation (FDIC). Various years. Annual report. Washington, D.C.: Federal Deposit Insurance Corporation. . 1984. The first fifry years: A history of the FDIC, 1933-1983. Washington, D.C.: Federal Deposit Insurance Corporation. . Website. http:llwww.fdic.gov. Federal Home Loan Bank Board (FHLBB). Various years. Annual report. Washington, D.C.: Federal Home Loan Bank Board. Feinberg, Robert M., and Seth Kaplan. 1993. Fishing downstream: The political economy of effective administered protection. Canadian Journal of Economics 26 (February): 150-58. Flood, Mark D. 1991. The great deposit insurance debate. Federal Reserve Bank of St. Louis Review 74 (July-August): 51-77. Friedman, Milton. 1959.A program for monetary stabilify. New York: Fordham University Press. Friedman, Milton, and Anna J. Schwartz. 1986. Has government any role in money? Journal of Monetary Economics 17:37-62. Golembe, Carter H. 1960. The deposit insurance legislation of 1933. Political Science Quarterly 76 (June): 181-95. Golembe, Carter H., and David S. Holland. 1986. Federal regulation of banking, 198687. Washington, D.C.: Golembe Associates, Inc. Grossman, Richard S. 1992. Deposit insurance, regulation, and moral hazard in the thrift industry: Evidence from the 1930s. American Economic Review 82 (September): 800-821. Hoekman, Bernard M., and Michael P. Leidy. 1992. Cascading contingent protection. European Economic Review 36:883-92. Kane, Edward J. 1989. The S&L insurance mess: How did it happen? Washington, D.C.: Urban Institute Press. Kennickell, Arthur, Myron L. Kwast, and Martha Starr-McCluer. 1996. Households’ deposit insurance coverage: Evidence and analysis of potential reforms. Journal of Money, Credit, and Banking 28 (August, pt. 1): 3 11-22. Litan, Robert E. 1994. Financial regulation. In American economic policy in the 1980s, ed. Martin Feldstein. Chicago: University of Chicago Press. Macey, Jonathan R., and Geoffrey P. Miller. 1992. Banking law and regulation. Boston: Little, Brown. Marvell, Thomas B. 1969. The Federal Home Loan Bank Board. New York:Praeger. Matthews, John 0. 1994. Struggle andsurvival on WallStreet: The economics of competition among securities firms. New York: Oxford University Press. National Credit Union Administration (NCUA). Various years. Annual report. Washington, D.C.: National Credit Union Administration. Pension Benefit Guaranty Corporation (PBGC). Various years. Annual report. Washington, D.C.: Pension Benefit Guaranty Corporation.

121

The Legacy of Deposit Insurance

Schattschneider, Elmer E. 1935. Politics, pressure, and the tariff New York: PrenticeHall. Seligman, Joel. 1982. The transformation of Wall Street. Boston: Houghton Mifflin. Sobel, Robert. 1972. Amex: A history of the American Stock Exchange, 1921-1971. New York: Weybright and Talley. . 1975. N.W.E.: A history of the New York Stock Exchange, 1935-1975. New York: Weybright and Talley. Sowards, Hugh L., and James S. Mofsky. 1971. The Securities Investor Protection Act of 1970. Business Lawyer 25 (April): 1271-88. Sprague, Irvine H. 1986. Bailout. New York: Basic Books. Teweles, Richard, and Edward S. Bradley. 1987. The stock market, 5th ed. New York: Wiley. U.S. Comptroller of the Currency. Various years. Annual report. Waslungton, D.C.: Government Printing Office. U.S. House. 1970. Report no. 91-1613: Securities Investor Protection Act of 1970. 91st Cong., 2d sess. . Committee on Banking and Currency. 1966. Financial Institutions Supervisory and Insurance Act of 1966: Hearings. 89th Cong., 2d sess., 15,20,22 September. . Subcommittee on Bank Supervision and Insurance. 1973. To providefill deposit insurance for public units and to increase deposit insurance from $20,000 to $50,000: Hearings. 93d Cong., 1st sess., $ 6 , 26 November. U S . Senate. 1950. Amendments to Federal Deposit Insurance Act. Hearings before a subcommittee of the Committee on Banking and Currency. 81st Cong., 2d sess., 11, 23,30 January. . 1970. Report No. 91-1218: Securities Investor Protection Corporation. 91st Cong., 2d sess. Welfling, Weldon. 1968. Mutual savings banks. Cleveland, Ohio: Press of Case Western Reserve University. Wheelock, David C. 1992. Deposit insurance and bank failures: New evidence from the 1920s. Economic Inquiry 30 (July): 530-43. White, Eugene N. 1983. The regulation and reform of the American banking system, 1900-1929. Princeton, N.J.: Princeton University Press. White, Lawrence J. 1989. The reform of federal deposit insurance. Journal of Economic Perspectives 3 (fall): 11-29. . 1991. The S&L debacle. New York: Oxford University Press.

This Page Intentionally Left Blank

4

By Way of Analogy: The Expansion of the Federal Government in the 1930s Hugh Rockoff

4.1 Ideological Change and the Growth of the Federal Bureaucracy The major turning point in the growth of the federal government was, of course, the New Deal. A host of programs were added that in themselves account for a substantial share of the growth of government in the twentieth century, and the propensity to add new programs increased. The New Deal was the result of a unique concatenation of forces: the unprecedented magnitude of the contraction, the political accident that the party favoring bigger government was out of power when the contraction began, and the unique personalities of Hoover and Roosevelt were among the most important. Moreover, as many historians of the Great Depression have recognized, there was an important ideological factor in the equation: intellectuals had already been converted to the cause of an expanded federal sector. It is hard, of course, to be certain about the role of this factor. But there were earlier crises, in the 1830s and especially the 1890s, that were also severe in terms of depressed incomes and unemployment. Ideology, and in particular the ideology dominant among intellectuals and opinion makers, appears to be a good candidate explanation for the failure of earlier crises to produce a change in the role of government of the same order that occurred in the 1930s. While it is easy to see that there was an ideological shift-from widespread skepticism about the ability of the central government to improve the functioning of the economy to widespread faith in the competence of government-it is harder to see what produced it. To shed additional light on the ideological preconditions for the New Deal, I will explore the attitude of economists toHugh Rockoff is professor of economics at Rutgers University and a research associate of the National Bureau of Economic Research. The author thanks his discussant, Stanley L. Engerman, and the editors for their detailed and thoughtful criticisms. The remaining errors are the author’s.

125

126

Hugh Rockoff

ward government intervention in the decade preceding the depression. As it turns out, virtually all of the reforms adopted in the 1930s-minimum wages, social security, unemployment compensation, the Civilian Conservation Corps, and so on-had been championed by economists. When the depression came, moreover, economists played a major role in bringing their reforms into being by writing the legislation, testifying in support of it before Congress, and implementing the new programs (Barber 1996). What had persuaded the intellectual descendants of Adam Smith to favor Big Government? One answer was given by George Stigler in his famous essay “The Economist and the State.” Stigler argued that economists had switched their support first from mercantilism to laissez-faire and then from laissez-faire to Big Government without any discernible reasons! Economists should have based their support for or opposition to government intervention on empirical studies of the effects of intervention, but they had failed to do so. As usual, its better to let George say it: “The economic role of the state has managed to hold the attention of scholars for over two centuries without arousing their curiosity” (Stigler [1965] 1986, 111). Here, however, I will argue that on the eve of the New Deal American economists believed that there was an abundance of empirical evidence from experiments in Germany, Scandinavia, and (especially) English-speaking countries proving that an expanded role for government would work well. The studies based on those experiments would not satisfy modern standards of rigor, certainly not Stigler’s, and their authors, with a few exceptions, have been forgotten. But the quantity and consistent thrust of those studies, I believe, moved the profession toward support of the modern mixed economy. In other words, I will argue that the New Deal was just what the doctors (of economics) ordered and that they believed their advice was soundly based on the clinical evidence. Liberal economists were only one small part of the intellectual support for an expanded federal sector that was in place before the depression. It is conceivable that events would have unfolded along lines similar to what actually transpired even if the economists had remained loyal to Adam Smith, although there might have been many differences in the details. My hope is that by looking at what persuaded economists to support the New Deal we may find a clue to the arguments that were broadly persuasive. Before turning to the economists, however, I want to take a closer look at just how important the programs created by the New Deal were to the growth of the federal government in the 1930s and beyond.

4.2 The New Deal and the Expansion of the Federal Government It is a commonplace that the growth of government power cannot be measured merely by looking at the number of federal employees or the amount of money they spend. To take what is perhaps the most extreme example, a change in the membership of the Supreme Court can permanently alter the

127

The Expansion of the Federal Government in the 1930s

effective role of government, even though quantitative measures of the size of government are unchanged. Typically, economists note this problem and then discuss employment and spending on the feeble grounds that these are the only statistics available, a procedure I shall follow. However, I can plead in my defense that the portrait of the growth of the federal government that results from looking at the quantitative data is consistent with portraits drawn on the basis of qualitative data such as changes in the legal system, for example, Higgs (1987) and Hughes (1991). 4.2.1 The Federal Labor Force The ratio of federal employees (excluding the military and the postal service) to the total civilian labor force is plotted in figure 4.1. This series shows a dramatic upward ratchet in the 1930s and a second upward ratchet in the 1940s. A good part of the World War I1 ratchet, however, is to be found in the Veterans Administration. When employees of the Veterans Administration are excluded (the lower line) the wartime ratchet drops by one-half. The remaining World War I1 ratchet, moreover, is lost in the postwar period. By the end of the 1980s the federal labor force share is back to where it was at the end of the 1930s. Was the expansion of the federal labor force in the 1930s mostly the result of adding new agencies or expanding old agencies? Skepticism about the motives of bureaucrats might lead us to expect that it was older agencies that expanded the most. In the expansive atmosphere that prevailed in the early days of the New Deal, we would expect bureaucrats in established agencies to try to increase their workforces. The established bureaus would try to formulate plans to fight the depression calling for more employees, plans that would be funded during the emergency because the Congress was controlled by legislators favorable to an expansion. Later, the inherent inertia in government would prevent conservative legislators from removing all of the workers and canceling all of the new programs added during the emergency. Porter (1980) asserted that this happened during World War 11; and it would not be unreasonable to think that it also happened during the depression. Figure 4.2 provides a test. I divided federal agencies into three categories: (1) New Deal agencies (the Works Progress Administration, the Federal Deposit Insurance Corporation, and so on), ( 2 ) old agencies that produced a final product (the Post Office, Veterans Administration, and so on), and (3) old agencies that produced intermediate services (the Treasury Department, Government Printing Office, and so on). The assumption is that since the demand for the services of category 3 was derived partly from the new agencies, we would expect the agencies in category 3 to expand during the depression, even in the absence of an exploitation of the emergency by the bureaucracy. Figure 4.2 displays the annual increase in employment in each category from 1934 to 1938. Civilians working for the military and emergency workers hired by the Civilian Conservation Corps are excluded. Two features stand out. First, the increase in the federal labor force was concentrated in 1934 and 1935. By

128

Hugh Rockoff

1.4

-

1.3

-

1.2 -

1.1

1 c

C

0

e

0.9

-

L

0.8

-

0.7

-

1

0.6

0.2

'

I

1910

1920

1930

1-

1940

- ; 9 - 7

1960

Fig. 4.1 Federal civilian employment, 1908-90 (percentage of civilian labor force) Sources: U S . Bureau of the Census, Historical Statistics of the United States, Colonial Times to 1970 (Washington, D.C., 1975). series D4, D14, Y308, and Y313; and the continuations of these series reported in U.S. Bureau of the Census, Statistical Abstract ofthe United States (Washington, D.C., various issues). Note: Dashed line shows employment excluding the Veterans Administration (VA).

1936 the expansion was over. In that year, there was little net change in the size of the government labor force; instead, what we see is an absorption by traditional agencies of some of the agencies created in the first years of the New Deal. Indeed, one can say that the increase in the share of the labor force working for the federal government in the twentieth century is basically the product of 1934 and 1935. Second, the increase was dominated by employees hired by new agencies (category l), or by existing agencies that could claim an increase in the derived demand for their services (category 3).

The Expansion of the Federal Government in the 1930s

129

/

.. - ..

I

im

5

2I-

_ -_ -

I

-lo -20

Old (intermediate)

t

40 ’

’

1934

I

1935

I

-

1936

-

1937

L I

1936

Fig. 4.2 Increases in the federal labor force by type of agency, 1934-38 Source: U S . Bureau of the Census, Statistical Abstract of the United States (Washington, D.C., 1939), table 166, p. 157. Notes: The figure plots changes from one year to the next. “Old (final product)” agencies include the Departments of Agriculture, Commerce, Interior, Labor, State, and War, the Navy, the Panama Canal, the Post Office, the Veterans Administration, and smaller agencies. “Old (intermediate)” agencies include the Civil Service Commission, the Departments of Justice and Treasury, the General Accounting Office, and smaller agencies. “New Deal” agencies include the Agricultural Adjustment Administration, Civil Aeronautics Authority, Farm Credit Administration, Federal Deposit Insurance Corporation, Federal Emergency Administration of Public Works, Home Owners Loan Corporation, Reconstruction Finance Corporation, Social Security Board, Tennessee Valley Authority, Works hogress Administration, Department of Agriculture Emergency Conservation Work, and smaller agencies. Emergency workers are excluded.

Existing bureaus producing final products, however, were able to benefit when they could formulate acceptable plans or happened to be in the right place at the right time. Employment at the Post Office, for example, increased only 6.2 percent between 1933 and 1938; but employment at the Veterans Administration increased 15 percent.

130

Hugh Rockoff

CI

C 0

D

n

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

Fig. 4.3 Federal civilian spending, 1900-90 (percentage of GNP) Sources: U S . Bureau of the Census, Historical Statistics of the United States, Colonial Times to 1970 (Washington, D.C., 1975), series Y463, Y466-Y469, Y472-Y474, Y476, Y485, Y494; and the continuations of these series reported in U.S. Bureau of the Census, Statistical Abstract of the United Srates (Washington, D.C., various issues). Note: To go from total federal spending to federal civilian spending, deductions were made for spending by the military services and the Veterans Administration, for spending on international affairs, and for interest on federal debt incurred during wars.

4.2.2 Federal Spending Figure 4.3 is a version of the familiar plot of total civilian spending as a percentage of GNP. In this case, in addition to deducting military expenditures from total outlays, I have also deducted expenditures on veterans’ benefits, international affairs, and interest on the national debt incurred during the world wars and the Korean War. The goal was to get a measure of civilian spending as free as possible from war-related spending. Obviously, these adjustments could be refined: not all expenditures on international affairs are defense related, and my adjustments for interest incurred in wartime were relatively sim-

131

The Expansion of the Federal Government in the 1930s

ple. But the broad picture resulting from alternative adjustments would be similar. The spending data, like the employment data, show a sharp upward ratchet during the New Deal. Then there is a decline in the ratio during the 1940s, produced by the elimination of emergency relief programs and the rapid growth of the economy. Despite this decline, the ratio is substantially higher in the postwar years than it was before. The difference between the employment and spending plots lies in the period after 1950. Spending as a percentage of GNP continues to rise at a fairly rapid rate until the 1980s, driven mainly by the expansion of New Deal transfer programs (the most important being social security) and by the addition of transfer programs (the most important being Medicare and Medicaid) during the Little New Deal of the 1960s. The spending data, to sum up, suggest that the New Deal created an upward ratchet in both the level and rate of growth of the share of federal spending in GNP. Again we can ask whether the increase in spending in the 1930s was concentrated in the new programs or whether it was spread throughout the government. In this case, as in the case of employment, New Deal agencies accounted for the bulk of the increase. I then divided civilian spending into five categories: (1) unemployment relief, which was new and was carried out mainly by new agencies such as the Civilian Conservation Corps; (2) pensions, which includes both new programs such as social security and older programs such as veterans’ benefits that were expanded during the emergency; (3) spending on public works, which was camed out by new agencies and by a number of older agencies; (4) spending on agricultural relief, which was carried on by new agencies that were quickly scooped up by the Department of Agriculture; and (5) all other spending, which includes mostly programs that existed prior to the depression. The lines in figure 4.4 show the difference between civilian spending in each year and spending in 1931. Evidently, relief expenditures (unemployment and agriculture) were the major source of increased spending in the 1930s. Even as late as 1942, relief expenditures are the largest source of the increase over 1931. Public worksto some extent another form of relief-were important in the middle years of the depression but were cut back sharply as the economy improved at the end of the decade and military construction increased. Pensions, although they would be the major spending legacy, are important mainly in the years when veterans’ bonuses were paid. Traditional programs (all other) show relatively little growth in the 1930s, and some of this growth could be viewed as derived from the growth of the relief agencies. All in all, the message here is similar to the message in the employment graphs: the federal government increased in size because it increased in scope. The hard question is, Why did the scope of the federal government increase so much in the 1930s?

Hugh Rockoff

132

2.8

1

2.6

2.4

2.2 2 1.0 1.6 1.4

1.2 1 0.8

0.6

0.4 0.2

0 -0.2 -0.4

Fig. 4.4 Increase in federal civilian expenditures from 1931 levels

Source: U.S. Bureau of the Budget, Budgetfor 1942 (Washington, D.C., 1942), table 8, p. 1046.

4.3 The Economists and the State, from World War I to 1929 The New Deal, as noted in the introduction, resulted from a unique concatenation of forces (see Higgs 1987, 170-71; Hughes 1991, 140-42): (1) the major factor, of course, was the sheer fury of the contraction-twenty-five percent unemployment created a not-to-be-denied demand that somebody do something; (2) the accident that the party generally favoring bigger government was out of power when the contraction began-The depression of the 1890s, by way of contrast, weakened the Democrats and strengthened the Republicans; (3) the personality of Hoover, who was unable to communicate his concern for the suffering produced by the depression, and the personality of Roosevelt, who proved to be a masterful tactician and strategist on behalf of an expanded role for government; and (4) the previous conversion of intellectuals to a favorable opinion of government intervention in the economy. As noted in the introduction, it is with respect to item 4 that I hope to shed

133

The Expansion of the Federal Government in the 1930s

some additional light by looking at the attitude of economists in the years leading up to 1932 toward the reforms that became the New Deal. Why is it important to look at what economists believed and how they came to those beliefs? In ordinary times, it is true, the opinions of economists got short shrift. Over 1,000 economists, including almost all the leading economists, signed a petition against the Smoot-Hawley tariff (Dorfman 1959, 674)-to no avail. But in the unique circumstances created by the depression the ideas of economists became important. Many of the new programs, including the Civilian Conservation Corps, unemployment insurance, and minimum wages, were advocated and shaped by economists. It is important, moreover, to learn what persuaded economists that these programs would work because the evidence that persuaded them was probably broadly persuasive. Other groups of intellectuals were, if anything, even more enthusiastic than economists. Social workers, for example, were enthusiastic and influential advocates of social insurance (Berkowitz and McQuaid 1992, 38-39). But precisely because economists were the heirs of an intellectual tradition that took laissez-faire seriously, the arguments that convinced them that government intervention would be beneficial were likely to have been especially potent. 4.3.1 Just What the Doctor Ordered There is no justification at all for viewing the economists in the 1920s as doctrinaire defenders of laissez-faire. That view is easily rejected by even a cursory study of the history of economic thought prior to the New Deal. Professional opinion, it is true, varied from subdiscipline to subdiscipline. Macroeconomists, typically, were more conservative. The gold standard was generally accepted as the best long-run monetary framework. Even in the case of macroeconomics, however, there was considerable support for spending on public works to cure cyclical unemployment (Blaug 1978, 684-86). Chicago economists-Jacob Viner, Henry Simons, Paul Douglas, and Aaron Director, among others-were particularly vociferous about the need to create employment in the early 1930s through public works programs (Davis 1968; Tavlas 1976; Rockoff 1996).' Nevertheless, it is fair to say that the Keynesian revolution was to a great extent a product of the 1930s.' Keynesian economics, in turn, had a major impact on the structure of the debate over new government programs. Before the Keynesian revolution, advocates of a new program had to overcome the argument that taxes would have 1 . Although a number of leading economists favored deficit-financed public works, they did not agree on why public works would reduce unemployment. Douglas, e.g., argued simply that the direct effect of employing additional workers was unlikely to be offset by any plausible employment decreases in the private sector. Simons argued from the quantity theory of money: deficits meant more cash (or more bonds, which were practically money) and hence more total spending. 2. Stein (1969) and Barber (1996) describe the evolution of macroeconomic thought in the depression.

134

Hugh Rockoff

to be raised to pay for the new program and someone would have to consume less private goods. After the Keynesian revolution, the cost of a new program became (except in the special case of full employment) zero or negative. Among microeconomists, the major focus in this paper, however, there was considerably more support for an expanded role for government before the depre~sion.~ At the far end of the spectrum from the macroeconomists were the labor economists, who “characteristically were among the staunch advocates of an expanded program of social insurance” (Dorfman 1959, 592). In general, as Dorfman shows, microeconomists championed a long list of reforms such as minimum wages, employment exchanges, old-age pensions, publicly owned regional power utilities, and so on: the programs that became the enduring achievements of the New Deal. One piece of evidence for this generalization is explored below: the overwhelming majority of the articles on New Deal-type reforms published in major economic journals between World War I and 1929 were favorable. We can conclude, confidently, that there was a large and active group of economists who favored New Deal reforms. The extent to which their arguments were persuasive with the remainder of the profession, the silent majority, is less certain. But as McCloskey (1994, 71-84) has argued, persuasion is itself an economic activity. Economists invested considerable time and effort in these studies over several decades. It is reasonable to believe that they persisted because they earned an adequate return on their investment. In any case, I will show that any economist who claimed to be guided by the evidence reported in the major journals would have been forced to conclude that the New Deal was made up of a valuable group of workable programs. For the most part economists did not try to make the case for reform by relying on theoretical arguments. Even A. C. Pigou, who developed the concept of externalities into a formidable weapon for criticizing unfettered competition, and whose book The Economics of Welfare (1932) advocated most of the programs that were adopted in the 1930s, relied to a great extent on empirical evidence and “common sense” to make his case for a mixed economy. Where did this empirical evidence come from? Before World War I, social insurance plans had been advocated on the basis of German models (Skocpol 1992, 160-76); but after the war the German experiments became the basis for labeling social insurance plans as foreign and subversive (Dorfman 1959, 11l), so other examples had to be used. Some radical economists were strongly influenced by the U.S. experience in World War I. Rexford Tugwell, for example, thought that the War Industries Board had made the industrial economy per3. In an important book, Socializing Security, which came to my attention too late for me to incorporate its findings into this paper, David A. Moss (1996) discusses the American Association for Labor Legislation. Many of the economists mentioned below played a key role in the association, which vigorously promoted a range of measures intended to improve the security of workers. This paper complements Moss’s work by examining the arguments for the feasibility of reform that the enthusiasts of reform found persuasive and with which they hoped to persuade their professional colleagues.

135

The Expansion of the Federal Government in the 1930s

form like a “well oiled machine” and had “compressed ten years of economic progress into one.” Or to take a more mainstream example, Frank W. Taussig in a famous paper in the Quarterly Journal of Economics (1921) asked “Is Market Price Determinate?’ His experience on the price-fixing committee of the War Industries Board convinced him that the answer was no: there was a range of prices acceptable to the market, and hence scope for positive government intervention. Tugwell, like other radicals, was also strongly influenced by the Soviet Union. During the 1920s he returned from a visit appalled by the political persecution but at the same time convinced that in central planning the Soviets were onto a good idea (Dorfman 1959,509-10). And some economists must have been influenced in their thinking about the welfare state by the experiments and writings of progressive employers such as Edward Filene and Gerald Swope. Certainly, progressive employers had an influence on politicians and policymakers, as documented by Berkowitz and McQuaid (1992). But to judge from the major journals, American economists, typically, were not easily persuaded by experiments that held so little constant. For them, the most telling experiments were run in peacetime in economies similar to that of the United States. If need be, an experiment undertaken on the European continent, particularly in Germany, would do. But best, to judge from the journals, was an experiment in an English-speaking country: at the national, state, or local levels in the United States, Canada, Britain, Australia, or New Zealand.

4.3.2 The Evidence Supporting Intervention To illustrate these points I will review a number of New Deal reforms and the evidence that economists believed they had that these reforms would be beneficial. I chose reforms, such as minimum wages, that could be challenged using fairly simple economic analysis, such as supply and demand curves. These would be the reforms, presumably, that would meet the most professional resistance. I have steered clear (for the most part) of monetary and fiscal policy because those issues are being covered in other chapters; and I have steered clear of antitrust because the attitude of the New Deal toward antitrust was ambiguous and its legacy limited. The articles (mainly in the American Economic Review,the Quarterly Journal of Economics, and the Journal of Political Economy) and books discussed below had a very different rhetorical structure from their modem counterpart^.^ Typically, these articles focused on institutional details: who supported and who opposed reform, when laws were passed, what the content of the law was, what the penalties were for violating it, and so forth. Occasionally, some attempt at quantitative assessment was made, say by making a before-and-after 4. I will focus mainly on articles published between the end of World War I and 1929, but in a number of cases I discuss articles from before the war that appeared to have been especially influential.

136

Hugh Rockoff

comparison, but often a conclusion about the success of the law would be drawn without any explicit supporting evidence. The logic was that the author, having shown his extensive familiarity with the experiment, should be trusted. There were only a handful of attempts, among the articles I examined, to use statistical techniques to hold other things constant. The Tennessee Valley Authority (TVA), 1933. What evidence was there that the government could successfully manage a huge regional power project? One could always cite the Prussian railroads, an example that had apparently become, for socialists, the classic example of successful government management of a large-scale enterprise (Marshall 1907,23). Taussig (1924,431) cites the German post office, telegraph and telephone, railways (perhaps), and military organization (unmistakably) as examples of public enterprises that were well managed.5Only in Switzerland, however, does Taussig see (possibly)public industry with “a spirit of progress comparable to that of private industry.” Taussig also makes his point by reference to Australia? “He who compares, for example, the railways of the United States and Australia will undoubtedly find some serious defects in Australia. But he will find crying evils in the United States. He will find greater efficiency in our country, but also tortuous management, and ominous consequences in the greater inequality of wealth; and he will not render an unhesitating verdict against the state railways of Australia” (1924, 434). The evidence most directly relevant to the TVA concerned the Ontario Power Commission, a publicly owned company that sold electricity generated at Niagara Falls. Emerson Biggar (1921) presented a comparison of the Ontario Power Commission with privately owned companies in New York in the Journal of Political Economy. Biggar was adamant that public ownership in Canada had proved superior to private ownership in New York State. Biggar even offered a shred of quantitative evidence: rates for electricity were lower in Toronto, which was 90 miles from the falls, than in Buffalo, which was closer, or even in the city of Niagara itself (1921, 5 1-52). Biggar also noted that the Ontario Power Commission had promoted rural electrification,while a private power company that supplied an area near Niagara known as “the garden of Canada” refused to supply farmers with electricity (1921, 52-53). The precedent here for the TVA and for the Rural Electrification Administration (1935) is obvious. Eight years later H. W. Peck (1929), armed with far more quantitative data, returned to the comparison in a paper presented at the annual meetings of the American Economic Association. Peck showed that production costs per kilowatt hour were slightly lower in Ontario, although similar to production costs 5. Taussig wrote frequently on Germany and railroads, and in one case (1894) on the Prussian railroads, a piece in which he evaluated a German pamphlet critical of certain administrative policies. 6. Schumpeter says that Taussig’s Principles “will help us appraise ‘what students got’ at that time,” although Schumpeter probably had economic theory in mind (1974,871).

137

The Expansion of the Federal Government in the 1930s

for a comparable group of utilities in New York, while prices charged to consumers of electricity were substantially lower in Canada. Peck then considered several possible explanations for the difference. Taxes, for example, were higher for the New York utilities, and interest costs were lower for the Ontario Power Commission, possibly because the Canadian government guaranteed the bonds. But these differences could account for only a fraction of the difference in costs to consumers. In the end Peck concluded that public ownership had proved better all around for consumers in Canada mainly because the philosophy that guided the Ontario Power Commission was service to the public, rather than maximizing profits. His discussant, J. Bauer (1929), offered no objection. Bauer’s main point was that the regulatory system in the United States could probably be reformed and made sufficiently effective to eliminate most of the advantage from public ownership. R. L. Dewey (1931) addressed the issue two years later in another paper presented at the annual meetings. Dewey did not undertake the detailed comparison of rates that Peck had performed. He cited, however, the Ontario-New York comparison, and a number of other cases in which publicly owned municipal power companies in the United States competed with privately owned companies, to show that public ownership produced lower rates. One swallow doesn’t make a summer. John Maurice Clark (1926,422), who cited the Ontario system among the successes of public ownership (along with the Glasgow street railway system and, of course, the Prussian railways),’ claimed that the Ontario Power Commission owed much of its success to one engineer who was in charge during its formative years, Sir Adam Beck. Clark concluded that while examples such as the Ontario Power Commission showed that public ownership was not necessarily disastrous, general considerations, such as the absence of a profit motive to hold down costs, suggested that one should make haste slowly with public ownership. But the point I want to emphasize here is that there was no rejoinder to Biggar and Peck, and while other examples of successful government-owned and -operated utilities were reported in the major journals, there were few examples of failures.8An economist who claimed to base his or her opinion on the evidence reported in the leading journals would have been forced to conclude that the empirical evidence supported public o ~ n e r s h i p . ~

7. Biggar (1917) compared private and public railroads in Canada, arguing that public railroads had proved superior because they had provided more honest administration, lower rates, and less legislative corruption. 8. Paul Douglas (1921) investigated the purchase of the Seattle street railway system by the city. This case had been cited in the press as an example of the folly of public ownership because the city had been forced to raise fares. But Douglas concluded that “the city has probably done as well or better than any private company would have done” (1921,47). 9. During the 1920s there was a major controversy in New York over the development of power sites along the St. Lawrence, the controversy that had motivated Peck‘s study. Governor Roosevelt had strongly supported public development.

138

Hugh Rockoff

The minimum wage, 1938. The minimum wage is a particularly good test of the attitude of economists toward intervention. What evidence was there that minimum wages would work? Lots! Harvard professor Arthur Holcombe’s (1910) article in the Quarterly Journal of Economics discussed the British Act of 1909 and noted that it was modeled on the successful law of the province of Victoria in Australia, and he expressed the hope that a minimum wage law would soon be adopted in the United States. His 1912 article in the American Economic Review cited experiments in Victoria, New South Wales, Great Britain, Minnesota, Wisconsin, and Massachusetts. There was no mistaking the conclusion: “Some immediate protection, however, for the American standard of living is necessary and an appropriate means is the establishment of a minimum wage” (Holcombe 1912,37). Sidney Webb’s (1912) article in the Journal of Political Economy, “The Economic Theory of a Legal Minimum Wage,” was also, as you might expect, extremely enthusiastic. The title is misleading. The article is essentially a presentation of the case for minimum wages based on the Australian experience:’O “In this matter of the Legal Minimum Wage the sixteen years’ actual trial in Victoria is full of instruction. . . . In the five sweated trades to which the law was first applied sixteen years ago, wages have gone up from 12 to 35 per cent, the hours of labor have invariably been reduced, and the actual number of persons employed, far from falling, has in all cases, relatively to the total population, greatly increased’ (973). How did minimum wages produce such positive results? One of Webb’s central arguments had a modem ring to it: minimum wages forced workers to augment their human capital. “The young workman, knowing that he cannot secure a preference for employment by offering to put up with worse conditions than the standard, seeks to commend himself by a good character, technical skill, and general intelligence” (1912,979). Three years later Edith Abbott (1915) provided a largely factual, but clearly positive description of British minimum wage laws in the Journal of Political Economy. After the war, Kathleen Derry and Paul Douglas (1922) reported on the minimum wage laws (for women) in Canada, again in the Journal of Political Economy. Much of the article was descriptive. One section, on British Columbia, however, provided what modem economists would recognize as an elementary “events study.” Deny and Douglas showed that wages for women increased, that the minimum wage did not become a maximum, that the distribution of wages was compressed, and that the employment of minors did not increase in the wake of the law. They did not comment explicitly on the decrease in hours worked, equally 10. Webb’s article was followed immediately in the journal by an article by Florence Kelly (1912) of the National Consumers’s League that described the current state of minimum wage legislation in the United States.

139

The Expansion of the Federal Government in the 1930s

apparent in their data. Implicitly, they seemed to be suggesting that the decrease in hours worked could have been prevented by a proposed regulation (which was disallowed by the courts) that would have forced payment of the minimum wage on a weekly basis, even if the employee only worked part time (Derry and Douglas 1922, 168). The possibility that this regulation would lead to employment at a lower wage in an uncovered sector, or unemployment, was not addressed. Other reports, such as Douglas (1919), Lucas (1924), and Feis (1926), tended to be mostly descriptive, but they shared a favorable view of the effects of minimum wages. Lucas described Massachusetts’s experiment and concluded that despite weaknesses that could be corrected by making the law mandatory, “it is apparent that the Massachusetts law has done much to ameliorate the condition of working women” (1924, 50). There is no suggestion of any negative effects. Aryness Joy described the state of Washington’s experiment and concluded that “it has, of course, with all its limitations, been a great gain” (1926,714). The last article about minimum wages to appear in a major journal before the depression was by G. V. Portus (1929) in the American Economic Review. Writing with the authority of a professor at the University of Sydney, Portus provided a critical history of labor market regulation in Australia. Although noting a number of problems-for example, that regulation had tended to “stereotype” wages-he found more to praise. Wage regulation had eliminated “sweating,” had maintained real wages at close to their prewar level during the war and its aftermath, and had increased real wages since that time. His policy conclusion was that wage regulation should be carried out at the national level rather than at the provincial level. Pigou’s Economics of Welfare (1932) devoted considerable space to minimum wages and the lessons to be learned from various experiments. The first question he addressed was whether evasion will be so persuasive that regulation is rendered ineffective. Pigou concluded that minimum wage regulation can be successful. The argument is that violations can be detected and that governments have a variety of powerful sanctions that can be brought to bear against those who violate the act. To illustrate Pigou cited a range of sanctions employed in wartime British legislation, and from legislation in Australia and the United States ([1932] 1962, 537-40). Pigou argued, however, that minimum wages would reduce real GDP by driving some workers to lower wage occupations. Interestingly, he drew attention to a weakness of the Australian analogue ([ 19321 1962,603). In Australia agriculture was still the largest sector. Workers displaced by minimum wages in the industrial sector could find work in the agricultural sector without significantly driving down wages. This was unlikely in Britain, where the agricultural sector could not easily absorb displaced industrial workers. What about the argument that minimum wages will force workers to augment their human capital? Again, Pigou was skeptical. Little can be learned,

140

Hugh Rockoff

he said, from the fact, for example, that people who move from low-wage to high-wage districts are soon earning the high wage: the more able workers choose to migrate. What is needed is a natural experiment. He cited the improvements in the quality of the men drafted in the war and, more to the point, some wartime British experiments with minimum wages in tailoring and box making as hopeful signs, but he did not consider these experiments adequate to clinch the case ([I9321 1962,608-9). In the end, Pigou provided a relatively modest endorsement for a national minimum wage. It will lower national income, and it will lower total labor income from wages. But ifthere is an adequate system of government support for the unemployed-his own favorite was a national minimum standard of income-then total labor income, including both the wages of those employed and the transfers received by those left unemployed, might be higher than in the absence of a national minimum wage, and that outcome would justify the policy ([1932] 1962, 692-93). The Securities and Exchange Commission, 1934. What evidence was there that the government could regulate successfully the issue of securities? In this case, there were studies that pointed out difficulties with government intervention. In the end, however, most, although not all, of the studies argued for regulation by a government bureau in Washington. Emil Friend (1908), writing in the Journal of Political Economy, discussed German stock exchange regulation. It was one of the most negative assessments of regulation to appear in the journals before the depression. Friend conceded that the German regulations were a response to genuine abuses: attempts to comer the grain market. But he concluded that speculation is inevitable in security markets and that attempts to curb it are likely in the long run to do more harm than good. Other discussions of security regulation focused on American experiments, and although some of these articles noted problems, they usually endorsed more regulation. E. T. Miller (1907) discussed the Texas Stock and Bond Law passed in 1893, which regulated the sale of railroad securities. Miller thought that the system had discouraged investment in improvements. His major recommendation, however, was not an abandonment of the system, but an increase in the fares charged by railroads, which were also regulated in Texas. James W. Angell (1920) in a paper in the Journal of Political Economy reported, somewhat hastily one would think, on the results of the Illinois Blue Sky Law passed in June 1919. It was not an empirical study in the typical modem sense-there were no statistics; instead, it was based mainly on interviews and correspondence with regulators, security dealers, and so on. Angell was not happy with the results of the Blue Sky Law. He worried that economic development might be reduced and that fraud would not be reduced. He viewed these problems, however, as the result of the structure of the law with its em-

141

The Expansion of the Federal Government in the 1930s

phasis on punishing securities dealers and its limitation to Illinois. He recommended a national law that would control the issue of securities at their “source” and so prevent, through government supervision, the possibility that a fraudulent security would ever be issued. William Z . Ripley (1914) covered an impressive range of evidence in an article in the American Economic Review. His procedure was to cite case after case of railway securities regulation, in each case grading the state regulators on their performances. Some regulators earned high marks and others failed. His policy conclusion was that the state regulatory agencies should be eliminated in favor of regulation by (a more powerful) Interstate Commerce Commission. Moreover, to eliminate state regulation, the commission would have to be given powers commensurate with those held by the most powerful state regulators. The last relevant study published in one of the major journals before the depression (Frederick 1929) looked at the reform that Ripley had advocated; at how well the regulation of railway securities by the Interstate Commerce Commission (a power granted in 1920) had worked in practice. Frederick concluded that “on the whole, the regulation of railway securities by the Interstate Commerce Commission has been beneficial to all [investors,railways, and consumers of railway services] concerned” (1929,201).” Employment exchanges, 1933. What evidence was there that national employment exchanges would improve the allocation of labor? There was little discussion in the journals of this question, in part I suspect because it was accepted as almost self-evident that the adoption of a nationwide system of publicly owned employment exchanges would improve on the existing system that relied mostly on private employment bureaus.IZThe argument was that a public system would not be tempted to defraud job seekers. Frankel discussed the German example in the Journal of Political Economy: “In passing the Public Employment Exchange Law of July 13, 1922, Germany . . . has come to have an instrument which, while it cannot of itself create work opportunities, can help minimize greatly the ever present unemployment problem” (1924, 207). Pigou agreed that employment exchanges should be run by the government. “Experience seems to show that, if they are to win an extensive clientele, they should be public-not run as a private speculation by possibly fraudulent private persons” ([1932] 1962, 516). Most of his analysis was concerned with 1 1 . Walker (1 927) criticized securities regulation. It was a “think piece” and did not discuss actual experiments with regulation. 12. Ohio established the first state employment agency in 1890. The federal government established an employment service in 1907 within the Bureau of Immigration and Naturalization. After 1918 it was an independent unit within the Department of Labor. But it was generally regarded as underfunded and ineffective.

142

Hugh Rockoff

how to create incentives for maximizing their use by potential employers and employees ([ 19321 1962, 5 12-18). He considered, in this context, the experience of labor exchanges in Britain, France, Germany, and South Africa. Bank deposit insurance, 1934. What evidence was there that bank deposit insurance would prevent bank failures? Here at last we come to a reform that was not endorsed by economists in the 1920s on the basis of empirical evidence. There were a number of experiments with state-run deposit guarantee schemes in the early years of the century. Cooke traced the progress of these experiments for the readers of the Quarterly Journal ofEconomics in a half-dozen articles (Cooke 1909-10, 1913, 1915, 1921, 1924). Surveying the wreckage in his final article, Cooke (1924) considered the possibility that the national banking system might be successful where the state-run systems had failed because of the comptroller's tradition of close and effective supervision. But he shrank from recommending it. The vision of the national banking system going down in flames, even if a remote possibility, was too much. There were, however, no foreign experiments with nationwide deposit insurance to decide the viability of a nationwide system. (Deposit insurance, in fact, reversed the normal process of diffusion: it was the apparent success of the American example that led subsequently to adoptions in other countries.) From the vantage point of the 1920s, the possibility remained open that a nationwide system-particularly if it avoided some of the problems of the state systems, such as allowing banks to withdraw when large losses and special assessment loomed on the horizon-could be made to work. This possibility was exploited by the proinsurance forces in Congress during the depre~sion.'~ The absence of a strong professional consensus prior to the New Deal may explain one anomaly. Deposit insurance was the only major New Deal reform that Roosevelt opposed (Calomiris and White 1994, 146).

Unemployment insurance, 1935. How did economists know that unemployment insurance financed by an employer tax would protect workers without seriously reducing the incentive to work and bankrupting employers? This question is especially pertinent because economists played a major role in the enactment of unemployment compensation. The first state unemployment insurance act was passed in a special session of the Wisconsin legislature in 1932. It was introduced by Harold Groves, a colleague and former student of John R. Commons at the University of Wisconsin. A similar bill, the work of Isaac M. Rubinow, an economist at Ohio State, Paul Douglas, and other economists, was introduced at about the same time in the Ohio legislature. The Ohio bill, however, was defeated partly because of opposition from employers 13. See Calomiris and White (1994) for an explanation of how deposit insurance came to be adopted.

143

The Expansion of the Federal Government in the 1930s

on the grounds that a state plan would leave them at a competitive disadvantage (Dorfman 1959,628-29). Those economists were activating reforms that a large number of American labor economists believed had been shown to be workable by experiments in Europe. But this confidence was not centuries old. One of the first discussions of unemployment compensation at the annual meetings of the American Economics Association occurred not long after the turn of the century when Charles Tuttle (1902) presented his case for unemployment compensation in a paper entitled “The Workman’s Position in the Light of Economic Progress.” Tuttle’s paper was devoted almost entirely to establishing the worker’s ethical claim to compensation from his employer, and from the state, when the worker was unemployed as a result of technological progress. Tuttle suggested, tentatively, that compensation might take one or more of three forms: (1) “free public employment bureaus” and “perhaps free railroad transportation,” (2) “making the business owner liable to a money damage for the summary dismissal of a faithful and efficient workman,” and (3) an “indemnity fund” financed in part from “general tax revenue” and in part from “some sort of special tax” upon the entrepreneur (Tuttle 1902,210-11). In the tradition of the day, the journal recorded the opinions of eight discussants. Most sympathized with the ethical case for compensation but worried about whether compensation was practical. What would be the effect, for example, on the incentive to work? David Kinley, one of the leading economists of the day, simply declared that Tuttle’s “proposal would be difficult if not impossible to put into effect” (1902, 218). And William W. Folwell declared the idea “wholly impractical” (1902,230). Several discussants tried to provide examples to show that unemployment compensation was practical, but they could only come up with rough approximations. Samuel L. Lindsay, for example, pointed to the British “Workingman’s Compensation Act” (for workrelated injuries) as a possible precedent (1902, 222). Perhaps the most interesting try was by John R. Commons, because Commons would become an influential advocate of unemployment compensation in the decades to come. Commons offered an example from “some five or six years ago,” when Massachusetts had indemnified workers who lost their jobs as the result of a large water project by paying them six months’ wages (1902, 231). Commons assured his colleagues that the initiative of the workers was not undermined in this experiment. He acknowledged, however, that this example was not entirely persuasive, and at the end of his remarks he expressed the hope that “if we set about as economists trying to discover practical means for indemnifying the laborer I think we can find lessons from other countries and different states, and draw upon our own economic foundation for plans which would be practical and would combine theory and practice” (Commons 1902, 232). Two decades later this challenge had been largely met, at least to the satisfaction of many liberal economists. Frank W. Taussig summarized some of the evidence for a system of unemployment insurance in his Principles

144

Hugh Rockoff

Economics: “A number of cities in Belgium and elsewhere have adopted the ‘Ghent system’ (first developed in that place with apparent success) of offering a supplement to the trade-union unemployed benefit; . . . the same system has been adopted on a national basis in Denmark and Norway” (1924,366-67). The ablest discussion of the Ghent system, in the major journals was by Rubinow (1913) in the Journal of Political Economy. Rubinow first discussed a number of older European plans in which workmen could elect to insure with a fund run by a city. His conclusion was that these plans suffered from two common insurance problems: moral hazard (workers were tempted to choose unemployment) and selective risks (workers who were more likely to lose their jobs were more likely to sign up for insurance). The Ghent system, which began in 1900 and spread rapidly across Europe, Rubinow explained, avoided these problems by using the trade unions to set rules and supervise adherence. “It is fair to say that the Ghent system was the first successful method of organizing social unemployment insurance” (19 13,422). Rubinow, however, was not satisfied with the Ghent system because it was restricted to trade unions and therefore excluded the most destitute workers. Some form of compulsory insurance, Rubinow argued, was required. The British system established in 1911 provided, of course, a much closer analogue to a nationwide state-run system of unemployment compensation. In 1925 Newman A. Tolles provided a critical summary of the British literature for readers of the Quarterly Journal of Economics. Tolles’s evaluation was highly favorable. True, the system had to be bailed out with general funds after the war, but that was only because the benefits had been recklessly expanded just before a depression hit. The cost of the bailout was large, but the suffering alleviated made the cost w0rthwhi1e.I~Maintenance of the unemployed had not undermined their will to work, and overall the prevention of unemployment had been neither helped nor hindered. In discussing the British act (which provided for an array of social insurance besides unemployment compensation) Taussig concluded that “an extraordinary forward step was taken in the field of social reform.” Social insurance would be expensive, but Taussig was optimistic because “when a country plunges into war, treasure is poured out on a scale that would cover, many times, the expenditure needed for the contested social reforms” (1924, 367). The only negative reference to the British system in the major journals was by Leo Wolman (1929). Wolman believed that certain aspects of the British system had retarded recovery from the last recession by discouraging labor mobility. These features included the use of labor exchanges (which were supposed to find people jobs) to administer unemployment benefits,15rigid definiof

14. Despite Tolles, the perception remained that the British system had been too generous. For that reason, the U.S. system limited the length of time a worker was eligible (Baicker, Goldin, and Katz, chap. 7 in this volume). 15. Wolman’s discussant, Squires (1929), pointed out that it would be simple enough to separate the two functions.

145

The Expansion of the Federal Government in the 1930s

tions of suitable alternative employment (which discouraged reemployment in expanding industries), and excessive benefits. As a model for American reform Wolman preferred a private, union-based system, inaugurated in Chicago in 1923 that covered the men’s clothing industry. In retrospect, perhaps the most interesting point about Wolman’s paper is that the discussion had moved far beyond whether there should be unemployment insurance, the subject of Tuttle’s 1902 paper, to how any negative side effects could be minimized.I6 Although the obvious precedents were the Ghent system and the British system of 1911, I should note that the Soviet Union was not ignored completely. In an article in the Journal of Political Economy (anxious as always in those days to publish an article celebrating government intervention in the economy) A. Abramson (1929) reviewed social insurance in Russia. He concluded that “it must be admitted that they have certainly been successful in so far as social insurance is concerned” (398).

The Civilian Conservation Corps, 1933. What evidence was there that an “industrial reserve army” would relieve unemployment? And what evidence was there that it would improve the character of the young men it employed, that it would not be just a dole? Nothing like the Civilian Conservation Corps, as far as I have been able to determine, was discussed in the journals. Nevertheless, it is a relevant example of the commitment of economists to an expanded role for government because the idea for the Civilian Conservation Corps, according to Dorfman (1959, 671), came from Richard T. Ely’s Hard Times: The Way In and the Way Out (1931). Ely was then one of the famous old men of economics; he had been president of the American Economic Association in 1900-1901, and his textbook Outline of Economics, first published in 1893, was still among the leaders in the 1930s. Ely’s plan was very similar to the actual Civilian Conservation Corps. His “industrial reserve army” would have a permanent “general staff” that in good times would concentrate on planning conservation projects, such as reforestation, and would formulate its plans so that in “hard times” the industrial reserve army could be rapidly and indefinitely mobilized to incorporate the unemployed. The permanent general staff would earn salaries commensurate with administrative salaries in the private sector. But the ordinary worker would only earn a salary similar to base pay in the regular army. When good times returned, employers would simply call on the peacetime army for men, who would immediately be mustered out and be fit and ready to report for active duty in the private sector. Ely was indebted to William James’s ([1910] 1912) essay “The Moral Equivalent of War” for the basic idea. James proposed a universal national 16. Similarly, Yoder (1931) was concerned mainly with getting the incentives right, and Witmer (1931) with getting the benefits structure correct.

146

Hugh Rockoff

service. All young men, including especially our “gilded youth,” would be conscripted and set to work, for a time, in mines, forests, and so on. In this way they would learn martial values, the one good thing about militarism, without becoming part of a war machine. Ely’s twist was to turn James’s peacetime army into a countercyclical policy. As empirical models of successful peacetime armies Ely cited the Salvation Army and the Jesuits, but the most influential model was the German army before World War 1.l’ German militarism before the World War was so repellant to most thoughtful Americans that they generally failed to see great benefits brought to Germany through military training. I was a student in Germany for three years, and during various visits to Germany afterwards I was strongly impressed by the bright side of military training. . . . The army was a school of many different kinds. Those who had not attained a certain minimum of school education had regular classes to bring up their deficiencies. The discipline of the army was on the whole an excellent thing. It taught order, it taught men to work together and to do the things that they were told to do without question, simply because it was their duty. Loyalty was cultivated and men were taught to work hard. In short it afforded a discipline of life. All of this was a preparation for the economic life, and was a help in making Germany prosperous before the War. (1 93 1, 100) How was the peacetime army to be financed, especially when hard times required a significant mobilization? Ely argued that it could be financed by issuing long-term government bonds. “The experience of the World War . . . shows the almost undreamed of and almost limitless capacity to raise funds by millions and by billions when needed to save our selves from disaster” (Ely 1931, 109). He added that it would be a good idea if a reserve fund-modeled on the Prussian War Chest-were kept in Washington and the states. The president (the Kaiser?) could then be given the authority to declare an emergency and use the reserve funds for immediate mobilization (Ely 1931, 111-12). Ely was more frank than most would have been in citing his German inspiration. After the war any plan based on a German model could easily be attacked as the product of a militaristic and undemocratic state and therefore inappropriate for the United States. But at least in Ely’s case we can see a process at work that is familiar today: Japanese institutions are recommended because the economy as a whole has been successful. Thus, to sum up, on the eve of the depression the economics profession (or at least an important segment of it) was ready with an arsenal of reform plans, ranging from minimum wage laws to industrial reserve armies, for the New Deal to use in its war against the depression. The general assumption that lay 17. Ely earned his Ph.D. at Heidelberg and was an apostle of the German historical school.

147

The Expansion of the Federal Government in the 1930s

behind these plans was that markets frequently fail to produce socially desirable results, and that the central government was normally competent to diagnose and correct the problems created by unfettered markets. This assumption, massively reinforced by the depression itself, inevitably structured the debate over subsequent reforms, and constituted one of the most enduring legacies of the depression.

4.4 The Postwar Era In the postwar era the evolution of opinion (both professional and public) about the appropriate economic role of the federal government followed the inverse of the path described in the preceding sections. In particular, just as the depression of the 1930s encouraged the public to adopt the enthusiasm for government intervention already prevalent among leading economists, the stagflation of 1970s encouraged the adoption of a new skepticism about government already evident among economists at an earlier date. 4.4.1

The Continuation of the New Deal Regime, 1946-76

The agencies and programs established in the 1930s survived and often as not expanded. In some cases the expansion in a particular area seemed almost inevitable given the initial federal commitment. This is well known in the cases of social security and bank deposit insurance. Another example is medical research. The National Institutes of Health can trace their origin to an amendment to the Social Security Act of 1937. A steady expansion followed in the postwar era: Mental Health (1946); Allergy and Infectious Disease (1948); Dental Health (1948); Heart, Lung, and Blood (1948); Arthritis, Metabolism, and Digestive Diseases (1950); Child Health and Human Development (1 962); General Medical Sciences, to cover diseases that did not have their own institutes (1963); Alcohol Prevention and Drug Abuse (1966); Environmental Health Science (1966); Eye Institute (1968); Aging (1974); and Neurological and Communicative Disorders (1976). Part of the explanation for the persistence and expansion of the institutional structure created in the 1930s was the interest groups created. Once in place, an agency or program created a group of constituents, whether arthritis researchers or peanut farmers, who lobbied continually for their program’s expansion, and who could be counted on to launch a vigorous counterattack against any attempt at curtailment. A congressman who voted to reduce or simply limit the spending of an agency immediately felt the wrath of the offended interest group; the benefits created by such a reduction, however, would be spread over a large group of taxpayers and go unnoticed and unrewarded at election time. This phenomenon has been described by Milton Friedman and Rose Friedman as an “iron triangle” of beneficiaries, bureaucrats, and legislators (1984, 42-5 1). But, of course, the terms one uses reflect one’s ideology. Arthur Schle-

148

Hugh Rockoff

singer, Jr., sees the persistence of liberal programs as evidence that “affirmative government” has on balance served the public well: agencies survive because they do good (1986,245-48). But the postwar years witnessed more than the survival and expansion of the programs begun in the 1930s,they also witnessed the continued dominance of an ideology that held that government was competent to enter new areas and could afford to do so,’*that the burden of proof was on those who opposed meeting problems with an expansion of government. The dominance of a liberal ideology reached its peak in the 1960s with the large number of new agencies and programs created in the Little New Deal of the Kennedy-Johnson era-the addition of Medicare to social security, creation of the Departments of Transportation and Housing and Urban Development, and so on. Opposition to the market, if not support for Big Government, reached its peak in the rebellion of college-age youths in the 1960s.The rebellion seemed to be a fulfillment of Schumpeter’s prediction that capitalism would undermine itself (1943, 146).An increasingly large fraction of the population, Schumpeter argued, would go on to higher education, where it would inevitably be indoctrinated in a withering criticism of capitalism developed by intellectuals. As the future was to show, however, the willingness of young people to act on their professors’ ideas would soon fade, and perhaps even more surprising, the characterization of the intellectual, or at least the economist-intellectual, as an unrelenting critic of the market was about to change.

4.4.2 Crisis and Ideological Change in the 1970s In the 1970s we had a period of economic crisis, followed by a shift to a new, and in this case more conservative, ideological regime. The economic crisis of the 1970s, of course, was not nearly as severe as the crisis of the 1930s, and partly for that reason, the shift in public opinion was not as great. While the 1930s could be described as a swing of the pendulum from right to left, the 1970s would be described as a swing from the left to the center. Nevertheless, there are a number of striking parallels between the 1930s and 1970s. As in the 1930s, a macroeconomic fluctuation, best addressed, one would think, with macroeconomic medicines, produced pressures for medicines to deal with long-term problems in specific industries. More to the point, for our purposes, the shift in public opinion in the 1970s, like the shift in the 193Os, was preceded by a shift among economists. And in the 1970s as in the 1930s the shift in professional opinion seems to have been based to on a reading of empirical evidence. In the 1970s, however, the role of evidence from abroad was much less important than it had been in 1900-1929. When programs were merely a gleam in an American reformer’s eye, the only available 18. Michael A. Bernstein (1994) has argued persuasively that the consensus among economists in favor of active government “planning” was strongly reinforced by the prominent role that economists played in the Office of Price Administration, the War Production Board, and other agencies in World War 11, and in cold war defense-related activities.

149

The Expansion of the Federal Government in the 1930s

evidence was from other countries that had already tried the experiment; after World War I1 one could examine how New Deal reforms had worked in the United States. Much of the leadership for the shift in professional opinion came from Chicago. In macroeconomics Milton Friedman revived monetarism, publishing (with Anna J. Schwartz) A Monetary History of the United States in 1963. I will not pursue the impact of the monetarist revolution here, except to note that the fall of Keynesian macroeconomics and the rise of neoclassical macroeconomics inevitably altered the terms of debate over any proposed expansion of the federal government. Before the 1970s it was possible to argue that the cost of adding a new’agency was essentially zero or negative, except when the economy was at full employment. In microeconomics, George Stigler, who produced influential empirical critiques of regulation, was the dominant figure. He published The Citizen and the State in 1975, collecting papers originally published in the 1960s and early 1970s. Perhaps, however, as Thomas McCraw (1984, chap. 7) has pointed out, Alfred Kahn better represents the “intellectual odyssey” of the profession in the postwar era. Kahn entered graduate school in the late 1930s, and his early work on antitrust, much of it published in the 1940s and 1950s, was heavily influenced by the institutionalist tradition of Veblen, Ely, and Commons. Kahn then held that neoclassical microeconomics did not provide a sound basis for evaluating antitrust policies, in part because consumer preferences, the satisfaction of which played such a large role in neoclassical economics, were easily manipulated. Gradually, however, Kahn’s opinions changed. His masterwork, The Economics of ReguZation (1970), although it did not abandon his institutionalist insights or his liberal politics, made a strong case for marginal cost pricing and for allowing market forces to operate. In 1974 he became chairman of the New York Public Service Commission, which regulated utility rates. Here he became a national figure by putting marginal cost pricing strategies into effect: for example, peak load charges for electricity and, more controversially, charges for directory assistance. In 1977 he became chairman of the Civil Aeronautics Board, and here he went even further in promoting market forces. Pressure for deregulation had been growing for some years and had broad political support. Much of the support for reform was based on the experience in California and Texas. There, intrastate carriers were permitted much greater latitude in setting fares, and as a result fares were much lower than on comparable interstate routes regulated by the Civil Aeronautics Board. Under Kahn the industry, which had become a tightly regulated cartel, was deregulated: special fares and cuts in rates were permitted, and new entrants were encouraged. It was the first major effort to deregulate an American industry. In the pre-1939 period we had no way of knowing the opinions of the “silent majority” of economists. The best that I could offer was the argument that the profession as a whole was likely to have been persuaded, at least to some ex-

150

Hugh Rockoff

tent, by the articles appearing in the leading professional journals and that in any case what mattered most was the opinion of the “weighty” economists who wrote for those journals. In the postwar era we can learn something more from opinion polls. Alston, Kearl, and Vaughan (1992) surveyed a stratified sample of 1,350 economists. Most relevant for our purpose is the breakdown by vintage-year in which they received their highest degrees. The change in macroeconomic ideas was the most dramatic: 9.4 percent of economists who received their highest degrees before 1961 agreed that “the macro economy is generally selfcorrecting”; but 27.5 percent of those who received their highest degrees between 1971 and 1980 agreed. On the microeconomic side the change was less dramatic, although still clearly present: 55.6 percent of the pre-1961 vintage agreed that a “minimum wage increases unemployment”; but 66.7 percent of the 1971-80 vintage agreed. Agreement with “marginal tax rates impact work effort” was 14.8 percent for the pre-1961 vintage, but 27.3 percent for the 1971-80 vintage. Agreement with “income redistribution is legitimate” was 64.2 percent for the pre-1961 vintage, but only 47.3 percent for the 1971-80 vintage. There appears to have been, it is interesting to note, a slight swing to the left for the 1981-90 vintage, presaging perhaps, a swing in public opinion. The Reagan administration hoped to attack the New Deal all along the line, by cutting taxes and government spending, as well as cutting government regulation. But its successes were limited (Higgs 1987, 255-56; Hughes 1991, 205-7). Tax rates were cut, and there was additional deregulation (most of which was already in the pipeline) in communications, transportation, and financial services. But the administration’s efforts to cut expenditure programs fizzled when proposed cuts were attacked by outraged constituencies and their congressional supporters. David Stockman, who drew up the list of proposed cuts, described the process in The Triumph of Politics (1986). Perhaps the main legacy of the Reagan Revolution was ideological: the shift in the burden of proof from those who oppose federal solutions to economic and social programs to those who favor them, something to be measured by the relative paucity of new programs since the 1980s, rather than by the elimination of older ones. 4.5

Ideology as Explanation

In the decades prior to the Great Depression microeconomists championed a wide range of social and economic reforms such as minimum wages, oldage pensions, unemployment compensation, large-scale public works to cure unemployment, industrial armies to be mobilized in depressions, and so on. When the Great Depression created a uniquely receptive political environment, those reforms were adopted, permanently altering the role of the central government. What had persuaded economists, the heirs of Adam Smith, to favor a major

151

The Expansion o f the Federal Government in the 1930s

expansion of government? Undoubtedly, the economists were responding to a range of social and political forces that lay outside their discipline. Empirical evidence developed by economists, however, was important in converting the profession to a liberal (in the American sense) view of the role of government in the economy. American economists were influenced by the apparent success of government economic intervention in a wide range of countries, including Germany, Belgium, Russia, and the Scandinavian countries. Perhaps most important, however, were experiments in English-speaking countries- Australia, Canada, Great Britain, New Zealand, and the United States. This was sensible. Experiments in wartime economies, in Russia, even in Germany, as persuasive as they were to many, did not hold enough things constant to persuade the skeptical. In retrospect there is nothing unique in the influence traced in this paper. Who can doubt that in recent decades the attitude of economists toward the role of the state has been influenced powerfully by natural experiments such as East Germany versus West Germany or mainland China versus the “tigers” of Southeast Asia, and by the economic success o f Japan? The growth of the federal bureaucracy in the 1930s illustrates the point made recently by a number of economic historians (Higgs 1987; Fogel 1989; North 1990) that major changes in economic institutions cannot be understood without taking ideology and ideological change into account. The history of economic thought, I believe, provides a lens for bringing the ideological dimension of economic change into focus.

References Abbott, Edith. 1915. Progress of the minimum wage in England. Journal of Political

Economy 23 (March): 268-77. Abramson, A. 1929. Social insurance in Soviet Russia. Journal of Political Economy

37 (August): 377-99. Alston, Richard M., J. R. Kearl, and Michael B. Vaughan. 1992. Is there a consensus among economists in the 1990s?American Economic Review: Papers and Proceedings 82 (May): 203-9. Angel], James W. 1920. The Illinois Blue Sky Law. Journal of Political Economy 28 (April): 307-21. Barber, William J. 1996. Designs within disorder: Franklin D. Roosevelt, the economists, and the shaping of American economic policy, 1933-1945. Cambridge: Cambridge University Press. Bauer, J. 1929. Electric power and light utilities: Discussion.American Economic Review: Supplement 19 (March): 219-25. Berkowitz, Edward D., and Kim McQuaid. 1992. Creating the welfare state: Thepolitical economy of twentieth century reform, rev. ed. Lawrence, Kans.: University of Kansas Press. Bernstein, Michael A. 1994. American economics and the American economy in the

152

Hugh Rockoff

American century: Doctrinal legacies and contemporary policy problems. In Understanding American economic decline, ed. Michael A. Bernstein and David E. Adler, 361-93. Cambridge: Cambridge University Press. Biggar, Emerson B. 1917. Government ownership and private ownership of railroads in Canada. Journal of Political Economy 25 (February): 148-82. . 1921.The Ontario Power Commission. Journal of Political Economy 29 (January): 29-56. Blaug, Mark. 1978. Economic theory in retrospect, 3d ed. Cambridge: Cambridge University Press. Calomiris, Charles W., and Eugene N. White. 1994. The origins of Federal Deposit Insurance. In The regulated economy: A historical approach to political economy, ed. Claudia Goldin and Gary D. Libecap, 145-88. Chicago: University of Chicago Press. Clark, John Maurice. 1926. Social control of business. Chicago: University of Chicago Press. Commons, John R. 1902. [The workman’s position in the light of economic progress]: Discussion. American Economic Review: Publications, 3d ser., 3:230-32. Cooke, T. 1909-10. The insurance of bank deposits in the West. Quarterly Journal of Economics 24 (November): 85-108; 25 (February): 327-91. . 1913. Four more years of deposit guaranty. Quarterly Journal of Economics 28 (November): 69-1 14. . 1915. Deposit guaranty in Mississippi. Quarterly Journal of Economics 29 (February): 4 19-25. . 1921. The Nebraska deposit guaranty fund. Quarterly Journal of Economics 36 (November): 162-66. . 1924. The collapse of bank-deposit guaranty in Oklahoma and its position in other states. Quarterly Journal of Economics 38 (November): 108-39. Davis, J. Ronnie. 1968. Chicago economists, deficit budgets, and the early 1930s. American Economic Review 58 (June): 476-82. Deny, Kathleen, and Paul H. Douglas. 1922. The minimum wage in Canada. Journal of Political Economy 30 (April): 155-88. Dewey, R. L. 1931. The failure of electric light and power regulation and some proposed remedies. American Economic Review: Supplement 21 (March): 242-58. Dorfman, Joseph. 1959. The economic mind in American civilization. Vol. 5, 19181929, Continued. New York: Viking. Douglas, Dorothy W. 1919. American minimum wage laws at work. American Economic Review 9 (December): 701-38. Douglas, Paul H. 1921. The Seattle municipal street-railway system. Journal of Political Economy 29 (June): 455-77. Ely, Richard T. 1931. Hard times: The way in and the way out. New York: Macmillan. Feis, Herbert. 1926. The report of the Economic Commission on the Queensland Basic Wage. Journal of Political Economy 34: 244-48. Fogel, Robert W. 1989. Without consent or contract: The rise and fall of American slavery. New York Norton. Folwell, William W. 1902. [The workman’s position in the light of economic progress]: Discussion. American Economic Review: Publications, 3d ser., 3: 219-22. Frankel, Emil. 1924. Germany’s regulation of the labor market. Journal of Political Economy 32 (April): 207-25. Frederick, John H. 1929. Federal regulation of railway securities under the Transportation Act of 1920. Journal of Political Economy 37, no. 2 (April): 175-202. Friedman, Milton, and Rose Friedman. 1984. The tyranny of the status quo. New York: Harcourt Brace Jovanovich. Friend, Emil. 1908. Stock-exchange regulation in Germany. Journal of Political Economy 16 (June): 369-74.

153

The Expansion of the Federal Government in the 1930s

Higgs, Robert. 1987. Crisis and Leviathan: Critical issues in the emergence of the mixed economy. New York Oxford University Press. Holcombe, Arthur N. 1910. The British Minimum Wages Act of 1909. Quarterly Journal of Economics 24 (May): 574-88. . 1912. The legal minimum wage in the United States. American Economic Review 2: 21-37. Hughes, Jonathan R. T. 1991. The governmental habit redux: Economic controls from colonial times to the present. Princeton, N.J.: Princeton University Press. James, William. (1910) 1912. The moral equivalent of war. In Memories and studies, 267-96. New York Longmans, Green. Joy, Aryness. 1926. Washington’s minimum-wage law and its operation. Journal of Political Economy 34: 691-716. Kahn, Alfred. 1970-71. The economics of regulation: Principles and institutions, 2 vols. New York: Wiley. Kelly, Florence. 1912. Minimum wage laws. Journal of Political Economy 20: 9991010. Kinley, David. 1902. [The workman’s position in the light of economic progress]: Discussion. American Economic Review: Publications, 3d ser., 3: 213-19. Lindsay, Samuel. 1902. [The workman’s position in the light of economic progress]: Discussion. American Economic Review: Publications, 3d ser., 3:219-22. Lucas, Arthur E 1924. A recommendatory minimum wage law: The first decade of the Massachusetts experiment. American Economic Review 14 (March): 39-5 1. Marshall, Alfred. 1907. The social possib es of economic chivalry. Economic Journal 17 (March): 7-29. McCloskey, Donald N . 1994. Knowledge and persuasion in economics. Cambridge: Cambridge University Press. McCraw, Thomas K. 1984. Prophets of regulation. Cambridge, Mass.: Harvard University Press. Miller, E. T. 1907. The Texas stock and bond law and its administration. Quarterly Journal of Economics 22 (November): 109-19. Moss, David A. 1996. Socializing security: Progressive-Era economists and the origins of American social policy. Cambridge, Mass.: Harvard University Press. North, Douglass C. 1990. Institutions, institutional change and economic performance. Cambridge: Cambridge University Press. Peck, H. W. 1929. An inductive study of publicly owned and operated versus privately owned but regulated electric utilities. American Economic Review: Supplement 19 (March): 197-218. Pigou, A. C. (1932) 1962. The economics of welfare, 4th ed. London: Macmillan. Porter, Bruce D. 1980. Parkinson’s law revisited: War and the growth of government. Public Interest 60 (summer): 50-68. Portus, G. V. 1929. The development of wage fixation in Australia. American Economic Review 19 (March): 59-75. Ripley, William Z. 1914. Public regulation of railroad issues. American Economic Review 4 (September): 541-64. Rockoff, Hugh. 1996. Henry Simons and the quantity theory of money. New Brunswick, N.J.: Rutgers University. Mimeograph. Rubinow, Isaac Max. 1913. Subsidized unemployment insurance. Journal of Political Economy 21 (May): 412-31. Schlesinger, Arthur M., Jr. 1986. The cycles of American history. Boston: Houghton Mifflin. Schumpeter, Joseph. 1943. Capitalism, socialism, and democracy. London: Allen and Unwin. . 1974. Hisrory of economic analysis. New York Oxford University Press.

154

Hugh Rockoff

Skocpol, Theda. 1992. Protecting soldiers and mothers: The political origins of social policy in the United States. Cambridge, Mass.: Haward University Press. Squires, B. M. 1929. Unemployment: Discussion. American Economic Review: Supplement 19 (March): 32-36. Stein, Herbert. 1969. The jiscal revolution in America. Chicago: University of Chicago Press. Stigler, George J. (1965) 1986. The economist and the state. In The essence of Stiglel; ed. Kurt R. Leube and Thomas Gale Moore, 99-116. Stanford, Calif.: Hoover Institution Press. . 1975. The citizen and the state: Essays on regulation. Chicago: University o f Chicago Press. Stockman, David A. 1986. The triumph of politics: How the Reagan revolutionfailed. New York: Harper and Row. Taussig, Frank W. 1894. Recent discussions on railway management in Prussia, review. Quarterly Journal of Economics 9 (October): 77-87. . 1921. Is market price determinate? Quarterly Journal of Economics 33: 394-4 11. . 1924. Principles of economics, 3d ed., rev. ed. New York: Macmillan. Tavlas, George S. 1976. Some further observations on the monetary economics of Chicagoans and non-Chicagoans. Southern Economic Journal 42 (April): 635-92. Tolles, Newman Arnold. 1925. Recent literature on British unemployment insurance. Quarterly Journal of Economics 39 (August): 65 1-62. Tuttle, Charles A. 1902. The workman’s position in the light of economic progress. American Economic Review: Publications, 3d ser., 3: 199-212. Walker, Forrest Q. 1927. Some practical objections to compulsory publicity of accounts of industrial enterprises. American Economic Review: Supplement 17 (March): 34-36. Webb, Sidney. 1912. The economic theory of a legal minimum wage. Journal of Political Economy 20: 973-98. Witmer, Helen Leland. 1931. Some effects of the English unemployment discussion. Quarterly Journal of Economics 45 (February): 262-88. Wolman, Leo. 1929. Some observations on unemployment insurance. American Economic Review: Supplement 19 (March): 23-29. Yoder, Dale. 1931. Some economic implications of unemployment insurance. Quarterly Journal of Economics 45 (August): 622-39.

5

The Impact of the New Deal on American Federalism John Joseph Wallis and Wallace E. Oates

5.1 The New Deal and Fiscal Centralization A cursory look at the course of federal fiscal structure in the United States might suggest that the Great Depression and the New Deal merely accelerated already existing tendencies toward centralization of the public sector. Indeed, at least two of the most insightful political observers of the nineteenth century had forecast just such a trend. Alexis de Tocqueville, writing in the first half of the century, was convinced that democratic “sentiments” exerted a powerful force encouraging the centralization of political power. From his analysis of these “natural tendencies,” Tocqueville concludes that “I am of the opinion that, in the democratic ages which are opening upon us . . . centralization will be the natural government” ([1835] 1945, 2:313). Later in the century, Lord Bryce, although more circumspect about such a broad generalization, reached a similar forecast for the United States. After considering both the “centrifugal” and “centripetal” forces at work in American government, Bryce found that while the centrifugal forces were “likely, as far as we can now see, to prove transitory . . . the centripetal forces are pemanent and secular forces, working from age to age” (1901, 2344). Bryce went on to predict that “the importance of the States will decline as the majesty and authority of the National government increase” ([1888] 1901,2:844). From such a perspective, the New Deal might be seen against the backdrop of history as a major event hastening the underlying tendencies toward centralization. But such a view, on closer inspection, is seriously incomplete if not John Joseph Wallis is associate professor of economics at the University of Maryland, College Park, and a research associate of the National Bureau of Economic Research. Wallace E. Oaks is professor of economics at the University of Maryland, College Park, and University Fellow with Resources for the Future. The authors would like to thank Robert Margo and Claudia Goldin for helpful comments.

156

John Joseph Wallis and Wallace E. Oates

quite misleading. Even a straightforward examination of the standard measures of fiscal centralization (such as the central government share of total public spending or revenues) does not lend unambiguous support to this view. But, as we shall contend, simply looking at such measures of fiscal centralization provides only a superficial view of the evolution of intergovernmental fiscal structure in the United States. The New Deal meant much more than just a movement toward centralization of the public sector. It brought with it some fundamental and dramatic changes in the very character of American federalism, changes that would leave a permanent imprint on the intergovernmental system. In this paper, we explore the impact of the New Deal on the structure and workings of U.S. fiscal federalism, drawing on an extensive database on intergovernmental fiscal flows. We try to place the New Deal programs in the context of earlier intergovernmental structure and to follow their legacy into the second half of the twentieth century. It is our sense that the New Deal irrevocably altered the evolution of American federalism and did this largely through the widespread introduction of a fiscal instrument that had seen only modest use earlier: intergovernmental grants. The introduction of major grant programs in a form that involved close cooperation between federal and state authorities marked the end of a period in which different levels of government functioned with a high degree of independence. To make use of the terminology of the political scientists, we moved from a system of “coordinate” (or “dual”) federalism, in which the various levels of government function in relatively independent spheres, to one of “cooperative” federalism, in which there is much more sharing of fiscal functions and greater interplay among levels of government in the management and funding of public programs.’ In this way, the New Deal set American federalism on a new course that has brought increasing interaction among levels of government; its legacy is unmistakable. Much of our work has involved the careful assembling of the relevant data on public expenditures, revenues, and intergovernmental financial flows. This, as we shall see, is a tricky business involving some critical issues of definition, timing, and assignment of spending and revenue data. But when properly assembled and interpreted, the data reveal a striking evolution of the structure of the public finances in the United States over the course of the twentieth century. While the rise to prominence of the central government is a major feature of the fiscal landscape, it does not come primarily at the expense of state and local governments. It is rather part of the growth of the public sector as a whole and the extension of public responsibilities into a number of new functions, including a major role in social insurance and the assistance of low-income households. Many of the programs established in the New Deal and certainly much of the new “spirit” of cooperative federalism that it introduced were not 1. See, e.g., Geoffrey Sawer (1969) and William Stewart (1984) for discussions of these concepts.

157

The Impact of the New Deal on American Federalism

temporary in character; they have become essential elements of fiscal federalism in the United States over the remainder of the century. We shall contend that without the New Deal, the course of American federalism would likely have been quite different from that which we have experienced.

5.2 U.S. Fiscal Federalism in the Twentieth Century: An Overview

To put the New Deal programs in perspective, it is helpful at the outset to take a broad look at some summary measures of tendencies in U.S. fiscal federalism over the present century. Table 5.1 reports the shares of public expenditure of federal, state, and local government for selected years. In addition, the table indicates for those years the share of the public sector in GNP and the percentage of revenues of state and local government coming from national intergovernmental grants. The numbers reveal some important tendencies in U.S. fiscal size and structure. Three such trends are of particular importance for us. First, we see that over this century the government sector as a whole has grown in size relative to the economy. At the turn of the century, public spending accounted for only 8 percent of GNP; this had grown to 38 percent by 1992 (where the considerable difference in revenue and expenditure shares is the result of deficit financTable 5.1

Government Fiscal Measures, 1902-92 Total Expenditures Share by Level (%)

Total Government as a Share of GNP (9%)

Year

Federal

State

Local

Revenues

Expenditures

Federal Grants as a Share of State and Local Revenues (%)

1902 1913 1922 1927

34.16 29.89 39.39 30.57

8.22 9.27 11.69 12.98

57.62 60.84 48.92 56.44

7.84 7.53 12.58 12.85

7.66 8.09 12.52 11.78

0.7 0.6 2.1 1.5

I934 1940 1946 1952

38.69 44.91 82.43 69.10

16.83 17.51 6.24 10.80

44.48 37.58 11.33 20.10

17.36 17.86 29.5 1 28.5 1

19.56 20.36 38.22 28.40

13.7 8.7 5.7 9.0

1957 1962 1967 1972

62.11 59.98 58.76 52.41

13.48 14.50 15.64 18.42

24.41 25.52 25.59 29.17

28.64 29.19 30.81 3 1.49

27.82 30.56 31.44 33.09

9.1 12.8 16.8 19.7

1977 1982 1987 1992

53.04 57.45 59.33 58.33

18.95 17.42 16.50 17.99

28.01 25.13 24.17 23.68

32.82 36.11 36.97 33.94

34.65 38.82 38.34 38.42

24.6 19.0 15.8 21.4

Sources; 1902-82, U.S. Census of Governments (1985); 1987 and 1992, Advisory Council on

Intergovernmental Relations (1995).

158

John Joseph Wallis and Wallace E. Oates

ing). Second, we find a striking trend in the public sector toward increased fiscal centralization. This trend is evident both in the growing role of the federal government and in the expansion of state governments relative to their local counterparts. In 1902, for instance, the federal share in public expenditure was only 34 percent; local government accounted for the lion’s share, 58 percent, of public expenditure in the United States. These shares have changed dramatically over the course of the century, with the federal and state sectors expanding at the expense of local government. By 1992, federal, state and local shares were 58, 18, and 24 percent, respectively. And third, we see an important change in the pattern of finance within the public sector. At the beginning of the century, state and local reliance on federal grants was minuscule; in 1902 less than 1 percent of state and local revenues came from transfers from the federal government. Now, of course, federal intergovernmental transfers are a major feature of the fiscal landscape; in 1992, for example, federal grants accounted for 21.4 percent of gross state-local revenues. In thinking about the role of the New Deal in this process, it is helpful to look more closely at table 5.1 to see what it suggests about the rate of change of these variables during the 1930s. Certain important fiscal changes stand out. First, although government’sshare of GNP does increase during the 1930s, this growth is not abnormally rapid. Government’s share of GNP slightly less than doubled between 1902 and 1922, and it slightly less than doubled again between 1922 and 1940. Although the comparison can be affected by the choice of years, if we take the period from 1913 to 1927 and compare it to the period from 1927 to 1940, we get roughly the same effect.*We see continuing growth in the public sector in the 1930s, but not an accelerating growth. Second, what does stand out about the 1930s is the sharp increase in the national share of government activity. Accepting that 1922 and 1946 (and to a lesser extent 1952) are exceptional postwar years, the national share of public expenditure prior to 1934 was roughly 30 percent, while after 1952 it had risen to about 60 percent. The local share, which was roughly 60 percent before the 1930s, fell to about 25 percent, while the state share rose from approximately 10 percent to 18 percent. Third, we see that in 1934 federal intergovernmental grants suddenly became an important source of revenue for state and local governments. And this new feature of the fiscal system became permanent (with certain wartime exceptions), with federal transfers accounting for a significant share of state-local revenues over the second half of the twentieth century. This overview of fiscal trends provides some perspective on the historical impact of the New Deal. But before turning to a more detailed examination of the changing fiscal structure of the U.S. federal system, it is important to treat at least briefly some fundamental issues regarding the data. As we shall see, 2. The data for 1922 include relatively large interest expenditures for the federal government.

This manifesis itself in table 5.1 in the large central share of expenditures in 1922.

159

The Impact of the New Deal on American Federalism

the various measures of government growth and structure depend in critical ways on just how the data are assembled and interpreted. Much of our effort in this study has involved a careful analysis of the data themselves and how they are to be understood. We find that it is impossible to get an accurate sense of the significance of the New Deal from these data without addressing some key issues of definition and timing.

5.3 Some Basic Issues Concerning the Fiscal Data Although some partial information on public finances had been collected since 1850, the first complete “census of governments” in the United States was undertaken in 1902. Subsequent censuses were taken in 1913,1922,1932, and 1942, but the coverage of these later censuses varied, leaving some serious gaps in the fiscal data. In 1950, the Congress enacted legislation providing for a quinquennial census in years ending in “2” and “7,” but funds were not appropriated for 1952. Since 1957, a census of governments has been conducted every five years. In spite of these problems with the infrequency and erratic coverage of the census of governments in the first half of the century, it is possible to construct consistent and reliable estimates of national aggregates for national, state, and local fiscal activity for most years. The census of governments regularly updates the twentieth-century figures in volumes entitled Historical Statistics on Government Finance and Employment. We begin with these data but examine several important modifications in the way the accounts can be presented.’ The issue before us is the structure of American government over the long term, and the selection of the years we chose is critical. Both world wars and the depression left deep footprints in the government data. Wars caused enormous increases in federal government military expenditures, financed in part by taxes and in part by borrowing. Debt repayment lingered in the record for decades. The Great Depression likewise had a major impact, in this instance leading to a sharp reduction in federal government revenues. Unfortunately, two of the benchmark census years, 1932 and 1942, are profoundly affected by depression and war. The use of these years can thus introduce some serious distortion when trying to develop a perspective both on the fiscal trends over the longer period and on the impact of the Great Depression. It is interesting in this respect that by the 1982 census of governments, the authors of the Historical Statistics volume chose not to present data for 1932 or 1942; instead they chose the years 1927, 1934, 1940, and 1946. This selection conveniently straddles the Great Contraction and the worst of the war years. We have followed their selection, but it is important to point out that, for some purposes, 3. We will gladly provide interested readers with an appendix, describing the census sources and alternative ways of measuring government activity, that contains detailed tables.

160

John Joseph Wallis and Wallace E. Oates

the selection of other years is more appropriate. The patterns and changes in the fiscal aggregates can look quite different depending on the years that are chosen. A second issue revolves around the categories of federal expenditure that we include when computing the fiscal shares of the different levels of government. The federal government has widely varying and large expenditure commitments to the military and increasingly for interest on its debt. But for the purpose of measuring the relative shares of the federal, state, and local sectors in carrying out domestic programs, it often makes sense to exclude federal spending on national defense and on interest payments from total expenditure and to focus attention on the shares for domestic expenditure programs. This adjustment can also make a major difference in our reading of both the pattern and trends in fiscal centralization in the United States. A third critical issue concerns a basic ambiguity in the way in which the relative fiscal sizes of the different levels of government are measured. This involves the treatment of intergovernmental grants. Although this issue is of minor importance early in the century, it looms much larger as these grants become a major feature of the intergovernmental fiscal system. The issue concerns the attribution of these grants to the revenues and expenditures of the different levels of governments. Gross revenue and expenditure totals involve substantial double-counting. And the way in which we choose to adjust for this double-counting affects our picture of what happened over the century. There are basically two choices: to attribute grants to the grantor or to the recipient. On the revenue side, grants are typically credited to the granting government. Under this approach, grants received by state or local governments are deleted from their gross revenues. Thus state and local revenues will include only the funds that they themselves raise through their own taxes, fees, and borrowing. These are often referred to as “own revenues.” Expenditure measures typically attribute grants to the recipient, the level at which they are actually spent. Using this definition, intergovernmental grants are excluded from the expenditures of the granting government and remain in the expenditures of the re~ipient.~ Thus, it will be important, as we shall see, whether we measure the fiscal shares of the different levels of government by the monies they raise (revenue shares) or by the monies that they spend (expenditure shares). The importance of these alternative measures can be seen by looking closely at the measure of expenditure used by the National Income and Product Accounts (NIPA). They report “expenditures from own funds,” an expenditure measure that attributes expenditures for grants to the granting government. Figure 5.1 depicts the share of total domestic expenditures at each level of government using the more traditional measure of expenditures, while figure 4. In fact, it would be virtually impossible to exclude certain lump-sum grants to governments from the recipient’s expenditures given the fungibility of such monies with own revenues.

The Impact of the New Deal on American Federalism

161

v,

?

3

c .-

U

a,

n X

b

0 a,

4 . -

E C

a

60

~

50 40

~

30-

2010 0

~

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

Year Fig. 5.1 Shares of domestic expenditures: intergovernmental grants attributed to receiving government, 1902-92 Sources; 1902-82, U.S. Census of Governments (1985); 1987 and 1992, Advisory Council on Intergovernmental Relations (1995). Nore; Federal domestic expenditures are total federal expenditures minus expenditures for defense, international relations, and interest on the government debt.

5.2 shows the shares based on the NIPA own-funds definition. Each figure incorporates data for the years listed in table 5.1.5 The reversal of the national and local shares in figure 5.1 occurs in the 1940s, while this reversal of national and local shares occurs in the 1930s in figure 5.2. The reason for the difference between the two graphs is straightforward. During the 1930s the central government incurred what for these times were relatively large peacetime deficits, while state and local governments actually ran surpluses. Revenues from new borrowing are not counted as revenues in the traditional measures, nor is retirement of principal included in expenditures. Since a large part of the growth in central fiscal activity in the 1930s took the form of intergovernmental grants, what the expenditures-fromown-funds measure (N1PA)picks up as an increase in the size of the central government during the 1930s is missed by the more traditional measure. And since most of the increased central expenditures were financed through borrowing, revenue measures produce results like those in figure 5.1. 5. As we noted, the selection of years excludes the massive military buildup in World War 11. Including the war years would introduce a bulge in the federal share in both figures but would not affect the arguments of the paper.

162

John Joseph Wallis and Wallace E. Oates

10

~

I

.

Figures 5.1 and 5.2 illustrate the importance of keeping definitions straight. Compare the national and local shares after 1940 in the two figures. In figure 5.1, where grants are credited to the receiving government, local expenditures are roughly equal to national expenditures until the late 1970s. In figure 5.2, where grants are credited to the granting government, the domestic expenditures of the national government not only exceed those of local government by 1940, but the difference continues to grow, with some interruptions, for the remainder of the century. The relative fiscal importance of national and local governments over the last half-century is critically dependent on how one attributes intergovernmental grants.

5.4 U.S. Federal Fiscal Structure in the Early Twentieth Century As we saw in table 5.1, the fiscal role of the federal government at the turn of the century was tiny. The federal government tended to operate in its own sphere of activity encompassing national defense, foreign relations, the postal service, and various judicial functions. States and their local governments pro-

163

The Impact of the New Deal on American Federalism

vided the bulk of public services including public education, police protection, roads and sanitation, public welfare, and health and hospitals. Moreover, the state and local sectors funded their expenditures almost exclusively from their own sources. Federal grants were less than 1 percent of state and local revenues as late as 1913. There had always been some central aid to state and local governments, beginning with the national assumption of state debts in the 1790s. But intergovernmental grants in existence by 1900 were a small part of the public sector. They included textbooks for the blind (1879), agricultural experiment stations (1887), state soldiers’ homes (1888), resident instruction in land grant colleges (1890), and irrigation (1894). These were followed by grants to state marine schools (1911), state and forestry operations (191l), the agricultural extension service (1914), vocational education (1917), and vocational rehabilitation (1920). But these programs were, by 1920, overshadowed in fiscal terms by the highway construction grants begun in 1916. By 1922, $92 million of the $11 8 million in federal grants, or 78 percent, was for highways. A maternity and infancy health plan was begun in 1921, which gave rise to the famous decision in Massachusetts v. MeZZon (262 U.S. 447 [1923]) that conditional grants did not impinge on state sovereignty, since states were free to forgo the grants. By 1930 there were 15 federal grant programs to state and local governments in operation, dominated by the highway construction grants. But they were still small in the aggregate; state grants to local governments were about five times as large as federal grants to state and local governments combined. The federal system in the early part of the century was one in which the central government operated largely independently of the state-local sector. It might be reasonably described, drawing on the political science literature, as “coordinate” or “dual” federalism. Lord Bryce provided a nice metaphor for this view when he described the federal system as “a great factory wherein two sets of machinery are at work, their revolving wheels apparently intermixed, their bands crossing one another, yet each set doing its own work without touching or hampering the other” ([1888] 1901,325). But the New Deal would change all this.

5.5

The Transformation of the Federal Fiscal System under the New Deal

Keeping track of national, state, and local government programs and activity during the Great Depression is complicated. During the contractionary phase from 1930 to 1932, national government revenues fell from $4,057 million to $1,923 million, while the national debt rose by $3,302 million (numbers are given in table 5.2, panel A). As noted earlier, this dependence on borrowing throughout the 1930s has to be accounted for when we calculate the shares of

164

John Joseph Wallis and Wallace E. Oates National Government Fiscal Activity, 1929-40 (million current dollars)

Table 5.2

A. Revenues, Outlays, and Debt

Year

National Revenues (1)

National Outlays (2)

Gross National Debt (3)

Change in Debt (4)

1929 1930 193I 1932 1933 1934 1935 1936 1937 1938 1939 1940

3,861 4,057 3,115 1,923 1,996 3,014 3,795 3,997 4,955 5,588 4,979 6,879

3,127 3,320 3,577 4,659 4,598 6,644 6,497 8,421 7,733 6,764 8,841 9,055

16.93 1 16,185 16,801 19,487 22,538 27,053 28,700 33,778 36,424 37,164 40,439 42,967

-673 -746 616 2,686 3,05 1 4,515 1,647 5,078 2,646 740 3,275 2,528

Outlays Outlays in 1933 (5)

0 2,046 1,899 3,823 3,135 2,166 4,243 4,457 21,769

Total

B. Distribution of Grants Census Grants

Cooperative Grants

Relief

Year

1932 1933 1934 1935 1936 1937 1938 1939 1940 Total

Works

Agriculture

214 190 1,803 2,197 1,015 818 790 1,031 967

250 432 2,857 3,649 3,969 4,273 3,518 4,794 3,922

154 2,126 2,221 2,343 2,405 2,047 2.67 1 2,188

27,414

16,155

108 196 356 459 618 624 504 69 1 52 1

Highway (11)

(9) 13 12 303 664 573 636 43 1 743 865

186 161 219 272 22 1 33 1 217 185 171

Sources: Panel A, US.Bureau of the Census (1975); panel B, Wallis (1984).

government activity undertaken at each level. Traditional revenue measures, which exclude borrowed funds, tend to understate the growth of the national government during the New Deal. Likewise, the procedures used in the census of governments for classifying New Deal programs are misleading. These procedures label as grant programs those that make “indirect expenditures” in contrast to programs involving “direct expenditures.” But even this classification was not applied entirely consis-

165

The Impact of the New Deal on American Federalism

tently over the decade. As we will discuss in more detail shortly, expenditures in the earliest New Deal programs were heavily weighted toward relief programs, particularly the Federal Emergency Relief Administration (FERA). After 1935 FERA was replaced by a combination of the Works Progress (later Projects) Administration (WPA), the categorical relief programs, and the unemployment insurance program created by the Social Security Act. Funds expended under the WPA were, technically, national government direct expenditures (not grants) because the WPA “employed” its relief recipients directly. Moreover, since WPA projects were typically for construction of various types (e.g., schools, highways, and parks), WPA expenditures were not classified by the census as relief or public assistance expenditures; instead, they were assigned to other functional categories, primarily natural resources.6 Panel B offers an alternative measure of grant activity and the functional distribution of these grants. Column (6) provides the census total for national government grants by fiscal year. Column (7) gives the total of national expenditures in programs “administered cooperatively with states.” The remaining columns give grants in cooperative programs by major function. Keeping in mind that fiscal 1933 ended on 30 June 1933, national government outlays in 1933 were primarily the result of the Hoover administration.’ Column (5) of table 5.2 gives the difference in national government outlays in each of the New Deal years compared with fiscal 1933. The total increase in national government outlays over 1933 levels was $21,769 million (in current dollars). Of that increase, $16,155 million went for grants in relief programs, 75 percent of the total. Total expenditures in cooperatively administered programs rose by $27,414 million. This total was larger than the rise in total expenditures. It indicates how important cooperative programs had become to the national government, which actually reduced the “noncooperative” part of its activities after 1933. Some programs were more cooperative than others. Of the four major groups shown in table 5.2, both relief and highways consisted primarily of programs that were jointly funded but that were administered by the state or local government, even the WPA with its odd accounting.8Agricultural price supports and other programs initiated under the Agricultural Adjustment Act (AAA) of 1933 were funded entirely by the national government and rarely involved the financial participation of state and local governments. But the AAA programs were often administered locally by county extension agents. For example, although crop allocations were determined nationally, within local areas the awarding of contracts was done by extension agents and commit6 . These issues are discussed in detail in Wallis (1984) and in U.S. Bureau of the Census (1955,5-7). 7. A small amount of New Deal grants were made in May and June 1933, particularly for relief under FERA. 8. There were exceptions. The Civilian Conservation Corps was a relief program run entirely by the national government.

166

John Joseph Wallis and Wallace E. Oates

tees of local farmers. Finally, public works programs involved a mix of national, state, and local arrangements. At one end were purely national projects under the Public Works Administration and at the other the direct grants made to local governments for public housing. The New Deal programs with the largest effect on state and local governments were the relief programs. This was not only the result of the large amount of funds expended for relief but also because work relief programs like the WPA made a significant contribution to a number of state and local functions through construction activity: education, highways, parks, and natural resources. Understanding the relief programs is central to understanding intergovernmental relations during the 1930s. These programs had both a major economic and a psychological impact. In fiscal 1934, the national government made over $2 billion in grants to state and local governments for relief, which in turn made more than $2 billion in relief grants to needy individuals and families. In 1933, $2 billion was 4 percent of GNP. Imagine what the effect would be today if the national government announced it would pass out $240 billion (4 percent of 1996 GNP) in relief to needy individuals, who need only apply to their local (typically county) relief office to see whether they qualified. And this in the depths of the nation’s greatest depression. Whereas granting $2 billion a year for relief was an unprecedented act of government largesse, it also created the possibility of unprecedented political patronage for the politicians in control of the money. The administration of public relief had always been intertwined with the issue of graft.9 Professional social workers were themselves deeply disturbed by the “politics” of relief. This partially explains the existence of two major professional organizations, the more prestigious of which had a membership of private sector social workers. When FDR appointed Harry Hopkins to head FERA in May 1933, it was not at all clear what the organization of relief would look like in coming years (see Brown 1940). The FERA legislation made it clear that FERA was to restrict itself to making grants to state governments. Half of the original $500 million appropriation was to be allocated among the states on the basis of matching grants. But the other half of the appropriation was to be allocated at the discretion of the relief administrator on the basis of need. The discretionary grants posed a problem for Hopkins: if he gave larger grants to states that spent less of their own (or local) funds, then states had an incentive to reduce their contributions. But the discretion allowed to him by Congress enabled him also to do the opposite, to reduce grants to states that made smaller contributions. The threat, and in a 9. As Howard put it, “So accustomed have the American people become to the infiltration of venal politics in many areas of public service, that doubt is frequently expressed whether public relief and politics can ever be divorced completely enough to assure decent care to those who are really in need and to prevent dissipation of resources among political favorites. As a result, there has been a great reluctance by many persons to expand relief programs lest they serve only to extend spheres of influence of corrupt political practices” ([1943] 1973.49).

167

The Impact of the New Deal on American Federalism

few cases the reality, that Hopkins would reduce national grants encouraged states to spend more of their own funds. State and local expenditures for relief expanded markedly after 1933. In extreme cases, the FERA legislation gave Hopkins the authority to “federalize” relief and take over the administration of relief in a state. This occurred seven times1” State governors, glad to get the national grants, were not happy to have Hopkins announce that grants would be reduced if the state government did not come up with a larger relief appropriation. State officials expressed their displeasure actively. Governor Davey of Ohio actually had an arrest warrant sworn out for Hopkins after he had charged Davey with using relief for political purposes. The original FERA appropriation had been $500 million for two years, and FERA was in charge of the nation’s entire relief effort. By the end of 1933, however, the funds were running low. Congress continued to appropriate funds for FERA on an emergency basis. FERA made roughly $2 billion a year in grants between May 1933 and August 1935.“ After the initial matching grants were exhausted, all further grants were discretionary The grants were conditional, and FERA promulgated an extensive set of regulations covering how relief programs were to be administered. Hopkins was able to enforce several simple and important regulations; for example, all relief funds had to be spent through public agencies. But FERA’s ability to affect personnel policies and recipient selection criteria was limited. As Williams (1939) noted, the ability to enforce these policies was much greater in states where the national contribution was larger.I2 Like many New Deal programs, FERA was an emergency program, intended to pass with time. Although FERA was given additional funds several times, Roosevelt and Hopkins were working on a more permanent solution to the relief problem. The Committee on Economic Security (CES), chaired by Frances Perkins, the secretary of labor, and of which Hopkins was member, began drafting a plan for an Economic Security Act (ESA) that would ultimately create the social security system. In his state of the union address in January 1935, Roosevelt announced that the national government “must and will quit this business of relief” and sent the committee report to Congress. A strong case can be made that the Social Security Act was the most important legislation not only of the decade but of the century. As it emerged from Congress, the act created three basic welfare programs. First, it introduced what we now call social security: old-age insurance (OAI, now OASDI) was a nationally administered program of old-age insurance. 10. The logic of how Hopkins could use discretionary grants to pry more funds out of state governments is developed in Wallis (1991). 1I. The $2 billion a year figure includes the amount spent by the Civil Works Administration in the winter of 1933. 12. For a discussion of the rules and regulations, and their enforceability, see Williams (1939) and Wallis (1981).

168

John Joseph Wallis and Wallace E. Oates

Second, the act established a program of unemployment insurance (UI). UI was funded through a 3 percent payroll tax, 2.9 percent of payrolls to be automatically credited to state UI trust funds in any state with an approved UI program. States were given wide latitude in setting up their programs. The third set of programs was categorical assistance: old-age assistance (OAA), aid to the blind, and aid to dependent children (ADC, the forerunner of aid to families with dependent children [AFDC]). Categorical programs were to be administered by the states and funded through automatic matching grants. National grants were open ended, but grants per relief case were capped. Although the independent Social Security Board created by the act had to approve each state’s categorical program, there were strict limits on the board’s ability to interfere with the actual administration of the programs. For example, the board was explicitly forbidden from withholding grants because of personnel decisions at the state level. The legislation submitted to Congress proposed each of these three programs but would have administered them in far different ways. The OAI program was national, but the categorical programs would have been administered by FERA, or a FERA-like agency, using discretionary rather than matching grants. The discretionary grants did not make it through the first committee hearings in the House. It was the strong state control over the categorical relief and UI programs, along with the fact that general relief (i.e., care of the needy who did not fit into one of the other categories) was left completely to state and local governments, that caused such a strong outcry among the social work professionals. “Returning relief to the states” was something they vigorously opposed.13Hopkins, who had gone to the professional social workers’ meetings in 1934 as a hero, came to the meetings in 1935 as a goat. Social work professionals opposed returning relief to the states because they saw that as returning control to the politicians. Hopkins, in fact, had been trained as a social worker and had done everything he could to improve relief administration in his tenure as the FERA administrator. Understanding why the social security system was set up as it was is critical to understanding the New Deal’s legacy and position in the political and economic history of the century. With the addition of Medicaid and Medicare in 1967, the social security system remains in place in an expanded form. In 1992, expenditures for social insurance, unemployment insurance, and public assistance were one-quarter of total government expenditures (at all levels). Medicaid was set up as a categorical assistance program, with matching grants and state administration. Medicare was set up like old-age insurance, with national administration and standards. 13. This story is told in a number of places. Brock (1988) and Bremer (1984) are very good on the details. The notion that returning relief to the states was an attempt to return control of the relief programs to local economic elites is elaborated in Piven and Cloward (1971) and Block et al. (1987). For a reformulation of this hypothesis based on the interest of Southern legislators, see Alston and Feme (1985).

169

The Impact of the New Deal on American Federalism

There are several places we can turn to find evidence on the motives of Congress and the president in 1935. One is to note the incredible opportunities that relief (and indeed all of the New Deal programs) offered for political patronage. Johnson and Libecap (1994) have pointed out that the New Deal is the one significant period since the Pendleton Act created the civil service system that the portion of national government employees who were not classified as civil service actually rose. Roosevelt and Hopkins asked Congress to put the relief program’s administrative employees under the civil service, but Congress refused until 1939. The attraction of patronage was simply too strong for the new Democratic majorities who had been out of power for decades. Johnson and Libecap (1994) argue that civil service reform was adopted in the first place because the cost of monitoring patronage employees had grown larger than the benefits from hiring them. Roosevelt and Hopkins learned this quickly during the New Deal. By 1935 Governor Langer of North Dakota was in jail for manipulating relief programs. While FDR stood to gain the most from thankful relief recipients, he also stood to lose the most when political abuses of relief were uncovered. Hopkins’s attempts to eliminate political manipulation through rules and regulations, control over hiring, or civil service status for administrative employees was contingent on his ability to coerce states through discretionary grants. But this was effectively countered by congressional policies that located administrative control, particularly over personnel policies, at the state level. Additional evidence can be found in the sister legislation to the ESA. FDR submitted another piece of relief legislation to Congress in the winter of 1935, the Emergency Relief Appropriations Act (ERAA) of 1935. Under the ERAA, Congress appropriated $4.8 billion for relief of the unemployed, to be spent through agencies that the president was authorized to create via executive order through methods not specified in the act. If the ESA was an attempt by Congress to secure state administrative control over the permanent part of the welfare system, the ERAA was an admission by Congress that for the remainder of the depression, the president would be allowed to exercise his discretionary powers to deal with the unemployment problem. Under the act’s authority, Roosevelt created the WPA, the Rural Electrification Administration, and several smaller agencies. Because of the unusual way it was created, the WPA could have been administered in any way that Roosevelt wanted. As we noted earlier, the fact that the WPA was a national program created some problems in the way its expenditures were treated, both in terms of grants and functions. In practice, however, and with a few important exceptions, the WPA was administered cooperatively with the states and came to be more cooperative as time went on. The two administrative mechanisms by which this cooperation was insured were project sponsorship and recipient selection. Hopkins became the WPA’s administrator, and he continued the FERA policy of encouraging state and local government participation in the program.

170

John Joseph Wallis and Wallace E. Oates

Although regulations regarding project sponsorship went through several changes after 1935, the basic structure was always the same. WPA projects were initiated by a sponsoring government, which could be local, state, or national, forwarded to the state WPA office for recommendation for approval, and then forwarded to the Washington office for final authorization. Between July 1935 and August 1937, 96 percent of all WPA projects were sponsored by local or state government^.'^ “Federal projects” were sponsored by the WPA itself, the most prominent of which were in the arts. After 1939 all federal projects were eliminated, and Congress insisted on a strict proportion of matching funds from sponsors of at least 25 percent of project ~ 0 s t . IThe ~ growing importance of sponsor contributions and the initiation of most WPA projects by local and state governments ensured a great deal of cooperation within the program. Cooperation was even more pronounced in the selection of WPA recipients. Although the WPA was free to hire its own administrative employees and supervisory employees (though sponsors also had some say in supervisory employees), in order for an unemployed worker to obtain WPA employment he first had to be certified as needy by the local relief agency. The WPA could not go out and provide relief to members of the community that the local government had not certified. Unlike the sponsorship policy, the WPA ultimately wished to control the certification process. Congress would not authorize money for the WPA to do this, however.I6 Project sponsorship and recipient selection demonstrate the extent to which intergovernmental cooperation pervaded the nominally “national” administration of the WPA. They also show that two forces were at work to produce such cooperation during the New Deal. First, the WPA itself sought cooperation from state and local governments to enable it to do its job more effectively. State and local governments were important sources of funds, administrative talent, and local expertise. By utilizing state and local cooperation, the WPA strengthened those governments even as it expanded its own role. Second, cooperation often gave state and local governments a measure of control over national administration. As such, cooperative policies were often mandated (or protected) by Congress even in the face of presidential opposition, as was the case with recipient certification. State and local independence was extremely valuable to congressmen. In summary, the New Deal initiated a pattern of cooperative intergovernmental activity with a distinctive bent: fiscal centralization and administrative decentralization. Not only were New Deal programs administered at the state 14. These figures are taken from Howard ([I9431 1973, 145). Howard provides an excellent history of the WPA. 15. Sponsor contributions to total WPA project costs were 10 percent in 1936, 14.7 percent in 1937, 21.4 percent in 1938, 19.3 percent in 1939, and 26 percent in 1940 (Howard [I9431 1973, 149). 16. The certification issue was quite complicated, see Howard ([1943] 1973, 356-79).

171

The Impact of the New Deal on American Federalism

level, but state governments in particular possessed real decision-making power. It was states, more than welfare recipients, who were entitled to categorical relief under the Social Security Act. State and local governments had their say in what projects the WPA would undertake, and by the end of the decade, the WPA could not build projects without state and local sponsorship and a 25 percent contribution. There were important differences between programs. The agricultural programs were cooperative, but state and local governments never became directly involved in their operation. The Civilian Conservation Corps, in contrast, was a national program, as were many public works projects. On the other hand, the WPA was a very important way that the national government financed education, water and sewers, parks, and highways. And the old highway programs continued their cooperative structure from 1916. 5.6 The New Deal and the States One other point needs to be addressed. Although the national government increased its share of total government activity during the New Deal, by no measure was the state share of government activity reduced. The state share of total government revenues from own sources rose from 16.4 percent in 1927, to 21.7 percent in 1934, to 28.2 percent in 1940.The state share of total government expenditures rose from 13 percent in 1927, to 16.8 percent in 1934, to 17.5 percent in 1940 (the own-expenditure shares are even higher). All of the growth in the national shares come at the expense of local governments. The reason for this is clear. National grants were given primarily to the states. Most of these grants offered incentives for state governments to increase their own spending. Whether these incentives were explicit, like the strict matching provisions in the categorical relief programs, or implicit as in Hopkins’s use of FERA grants, they were real. Wallis (1984) found that, in the late 1930s, every dollar of national grants increased state expenditures from own revenues by $0.3 1. At the same time combined state and national grants actually reduced local government expenditures. Similar effects are found in a longer analysis of grants and state and local fiscal activity (Wallis 1997). Where did these state revenues come from? In 1930, 16 states had individual income taxes, 17 had corporate income taxes, and none had a general sales tax. During the 1930s, 16 states added personal income taxes, 15 added corporate income taxes, and 24 created a sales tax. It is impossible to say that these taxes were the result of New Deal grant programs because the majority of new state taxes were put in place in 1933 at the same time that the New Deal grant programs were just getting under way. But one of the legacies of the New Deal was a much stronger state government sector with new and more flexible tax instruments. Another way to see this is in the structure of state government programs. In 1932, before FERA, only 7 states had spent money for unemployment relief

172

John Joseph Wallis and Wallace E. Oates

and had state relief agencies. By the end of 1933, all 48 states did. By 1939 all the states had approved UI schemes, and almost all had approved OAA, ADC, and aid to the blind programs in place. State highway boards were the result of the 1916 grants, but they too were still in place. The New Deal’s predilection for decentralized administration resulted in stronger and larger state governments.

5.7 Federal Fiscal Evolution Subsequent to the New Deal The New Deal created a new and major role for the federal government in domestic policy and set intergovernmental fiscal relations in the United States on a new course of joint responsibility for both the funding and the management of public programs. Since World War 11, this course has involved not only the expansion and extension of certain New Deal programs but also the creation of new forms of cooperative enterprise, some of which have flourished and others of which have proved much less successful. In this section, we explore briefly this later experience in order to provide some sense of what the New Deal might have bequeathed to us. The war, of course, had a dramatic impact on American government. During the fighting, the role of the national government expanded enormously, and both state and local government shrank in absolute terms. The war, and its employment policies, eliminated the need for the New Deal’s emergency relief programs, and the permanent programs became much less important. During the 1950s the national government returned to more active support for highway construction through the National Defense Highway Act in 1956, which began the interstate system. From 1952 to 1962, national grants for highway construction rose from $415 million to $2,748 million. The interstate system was built cooperatively, with states supervising the planning and construction in conjunction with national guidelines and direction. National grants for public welfare and education rose slowly in the 1950s, roughly in line with the overall rise of government. Social security, of course, became steadily larger as a growing portion of the labor force became eligible for benefits upon retirement. In the 1960s, the federal government took the design and operation of grant programs a step further. Responding to a sense of frustration with the lack of progress toward the attainment of certain basic social goals, including, for example, the elimination of poverty and the renewal of decaying center cities, it undertook the “War on Poverty.” The federal government enacted a whole series of grant programs that bypassed the states and, in some instances, even locally elected officials in an attempt to get resources “where they were needed.” Individually tailored project grants went to cities or special groups on the basis of approved grant applications. The emphasis of part of the War on Poverty was to encourage “maximum feasible participation.” It was intended to be empowering, and to the extent

173

The Impact of the New Deal on American Federalism

possible, the national government officials in charge of the programs were interested not in increasing the power and size of state and local welfare systems but rather in putting resources and control directly into the hands of the recipients. This brought federal officials into much closer contact with the “projects” that the money was intended to fund. Inevitably, it provided the national government much greater control over the specific ways in which grant funds were used. This was particularly true where project grants were specifically intended to circumvent the established political hierarchy. In this sense, these programs went well beyond the New Deal framework. At the same time, the national government redoubled its efforts to support the traditional, New Deal social welfare system. This effort continued well into the Nixon years. Part of this came through an expansion of the old programs: AFDC rolls exploded, supplemental social security expanded benefits for the aged, blind, and disabled. Part of this took the form of new programs: food stamps, Medicare, and Medicaid. Nixon’s Family Assistance Plan, which would have guaranteed all families with dependent children a minimum yearly income, was the most far-reaching proposal. It was not enacted by Congress, but it would have been a major change from the New Deal structure. At the same time the national government significantly increased its support for state and local government in a range of functions, including education, water and sewage, and natural resources. Not surprisingly (in retrospect at least), central intervention into the federal system eventually produced a strong reaction. Beginning in the late 1970s and especially under the new Republican administration in the 1980s, moves were initiated to cut back on the grant programs. This was typically described as an attempt to “return control of the programs to the states.” Narrowly targeted conditional grant programs were replaced by broad block grants that stipulated only very general areas in which the funds were to be used. Discretion in the employment of grant funds was supposed to devolve back to the states. Moreover, these reforms were accompanied by a general reduction in the relative size of grants; in the 1980s, federal grants as a fraction of state-local revenues and spending declined. In the early 1970s, another new intergovernmental structure was put in place. The Nixon administration and the U.S. Congress introduced a quite different sort of intergovernmental fiscal innovation: “general revenue sharing.” Under this program, the federal government distributed funds by formula to state and local governments with fairly loose restrictions on their use only at the local level. The objective of this program was to channel funds from the highly income-elastic federal revenue system to the state-local sector, where expenditure “needs” were growing rapidly-to link the most growthresponsive revenue sources to the most rapidly expanding forms of public expenditure. Revenue sharing, however, was a short-lived program. With the enormous federal deficits in the 1980s, the central government, not the states and localities, was under fiscal duress. The states (with their healthy treasuries)

174

John Joseph Wallis and Wallace E. Oates

were eliminated from the program in 1980, and with the support of the Reagan administration, the entire program was allowed to expire in 1986. This is curious in one respect. Revenue sharing in one form or another has played a major role in many federal countries such as Canada and Australia from very early days. Central government in these countries has been a basic and regular source of revenues for provincial, state, and local governments. But in the United States such programs have little history. The New Deal, while introducing large-scale transfers of funds from the center, certainly did not do so in a “hands-off‘’ fashion. In this sense, revenue sharing was not, in any direct way, a child of the Great Depression. With the massive deficits at the central level, there has been considerable pressure to cut back on federal transfers to the states and localities. But this has not proved easy. Intergovernmental transfers are now a basic part of the fiscal system; in 1992, for example, federal grants still accounted for over 20 percent of state and local revenues. Most recently, attention has focused on nonfiscal forms of central intervention, namely, various regulatory measures that impose responsibilities on state and local governments but do not provide the financial assistance to carry them out. These “unfunded mandates” have been the source of much contention and debate. This is a complicated issue. Many mandates seem well founded in terms of needed regulation of externalities ( e g , various environmental statutes). But others have a much less clear rationale. 5.8

Intergovernmental Grants and the Growth of the Public Sector

It is clear, in retrospect, that New Deal programs were the vehicle for the widespread introduction and heavy reliance on intergovernmental grants in the U S . federal fiscal system. In addition to transforming the character of American fiscal federalism, there is evidence in the public finance literature that this reliance on grants has made a significant contribution to the growth of the public sector in the American economy. As we have seen, many New Deal programs had as their very purpose the expansion of public sector spending to create new jobs and income and to stabilize the economy. To this end, they employed matching provisions that induced additional spending on the part of recipient state governments. The purpose of these grants was emphatically not to supplant spending by state and local governments with federal funds. From this perspective, New Deal grant programs clearly played an important role in extending the size and scope of the public sector as a whole. But at the end of this century, we find that federal grants (and state grants to local governments as well) no longer have economic expansion as their objective. Later grant programs are quite diverse in both form and purpose; most of them have aimed to provide fiscal support either for specific types of projects (e.g., highways, treatment plants, and retraining programs) or for income main-

175

The Impact of the New Deal on American Federalism

tenance. In the case of revenue sharing, the basic objective was simply to provide general purpose funds to state and local governments. Some of these programs contained matching requirements for recipients, but many others have not. Basic economic theory has some interesting implications for the effects of intergovernmental grants on the size of the public sector. In the case of most nonmatching grants, for example, grant funds simply augment the resources of the recipient community; they have no direct price effects, only income effects.” At the theoretical level, it has been established that such nonmatching grants have allocative and distributive effects that are no different than if these funds were distributed in a particular lump-sum fashion to the residents of the recipient jurisdiction (Bradford and Oates 1971). In short, nonmatching intergovernmental grants are formally equivalent in aZE their effects to a federal tax cut directly to the individuals in a community. In both cases, the measures simply increase the disposable income available to the community; in one case, the increased dollars show up in the public treasury and in the other in the pockets of the individual residents. But this should, in principle, make no difference to the ultimate outcome. This suggests that such grants should have no expansionary effect on the public sector. If the federal government collects taxes and redistributes the funds in the form of nonmatching grants to states and localities, then the negative income effects associated with the increased federal taxes should offset (approximately) the positive income effects from the rise in grants. Matching grants, in contrast, will have a net expansionary impact on the public sector since the income effect of the grants is accompanied by a price effect (i.e., the grant effectively reduces the price of services that the recipient purchases). But nonmatching grants should, in principle, be nonexpansionary. What is of interest here is the pervasive empirical finding in the public finance literature that this prediction of the theory is flat-out wrong. Dozens of studies of the stimulative effects of intergovernmental grants find that not only matching grants but nonmatching grants as well have a highly stimulative effect on the spending of recipients (Gramlich 1977; Hines and Thaler 1995). Based on income effects alone, we might expect nonmatching grants to state and local governments to induce an increase in spending on the order of 10 to 15 cents on the dollar (representing roughly their share in national income). But econometric (and other types of survey) studies find time and again that such nonmatching grants result in much more in the way of additional state17. This proposition is true so long as the grant does not exceed the amount in total that the recipient would have spent on the program in the absence of the grant. Moreover, even matching grants will have only income effects if they are closed ended (i.e., if matching stops at some prescribed level of spending by the recipient) and if recipients’ spending exceeds the sum for which they are eligible at the margin for additional matching funds. Many matching grant programs in the United States are, in fact, closed ended in structure, and the evidence suggests that in the great majority of the cases recipients spend more than the sum necessary to exhaust their grant entitlements. On the analytics of intergovernmental grants, see Oates (1972, chap. 3).

176

John Joseph Wallis and Wallace E. Oates

local spending, typically on the order of about 50 cents per dollar of additional grant monies. This anomaly has become known in the public finance literature as the “flypaper effect” (i.e., “money sticks where it hits”). And there is now a body of papers that seek in various ways to reconcile these empirical findings with the basic theory of intergovernmental grants (see Hines and Thaler 1995). For our purposes, this large empirical literature is important because it suggests that intergovernmental grant programs have had a substantial expansionary impact on the level of overall public spending. Additional federal taxes that are transformed into intergovernmental grants (even nonmatching grants) do not constitute a wash in terms of their impact on the size of the public sector. It is not easy to estimate the magnitude of this effect, as this would, in principle, require a detailed model of how these grants have influenced the growth over time of the state and local sectors. But we offer a simpler and admittedly crude calculation. If we take at face value the econometric estimates of the stimulative impact of intergovernmental grants, a conservative estimate would be that an average dollar of grants results in about 50 cents of additional statelocal expenditure and about 50 cents of tax relief at the state and local levels.’s Since federal grants account for about 20 percent of state-local revenues, this would suggest that state-local budgets in the aggregate are (at a minimum) about 10 percent larger than in the absence of these grant revenues. We do not intend this crude result to be taken as a serious effort to measure the effect of federal grants on state and local governments. Our intent is rather to make the point that in the process of transforming the character of the American federal system into one of cooperative federalism through an extensive reliance on intergovernmental grants, the New Deal also introduced a fiscal structure more conducive to overall growth of the public sector.

5.9 The Legacy of the New Deal The New Deal has had a lasting impact on the structure of American government. As we have seen, one dimension of this impact was to create a much larger and more active role for the central government. To get a clearer sense of this, we see in figure 5.1 the federal, state, and local shares in public expenditure excluding national defense and interest on the federal debt. This measure of fiscal shares (as discussed earlier) avoids some of the disruptions produced by wartime and required payments on the national debt. It is interesting to subject these data to some simple econometric scrutiny. It is clear from looking at the graph that the federal share has risen over the course of the century. In fact, if we simply regress the federal share of public expenditures on time, we find that

18. See Hines and Thaler (1995) for a summary of these estimates. This is a conservative estimate in that it is based on the measured impact of nonmatching grants: the stimulative effect of matching grant programs is substantially higher.

177

(1)

The Impact of the New Deal on American Federalism Federal share = - 4.5

+

(6.3)

.0025Time,

R2 = .75,

(6.8)

where the numbers in parentheses are absolute values of the t-statistics. The estimated equation confirms the growth in the federal share and indicates that on average the federal share of the public budget has increased over this period by about 1 percentage point every four years. But the path of the federal share in figure 5.1 also suggests that its growth has not been very regular. In fact, a large increase in this share seems to have occurred during the New Deal years. To account for this, we introduce into the equation a dummy variable that takes on a value of zero for those years before 1934 and a value of one for 1934 and thereafter. This yields

( 2 ) Federal share = -2.5

+

(2.7)

.0014Time (2.9)

+ .091Dummy,

R2 = .85.

(2.9)

We thus find that a “regime shift’’ took place during the New Deal period that increased the federal fiscal share by an estimated 9 percentage points. In addition, there remains a statistically significant trend over time toward greater fiscal centralization. But this effect is much smaller than in equation (1): the federal share now took about seven years to grow by 1 percentage point over the period under study. Moreover, if we return to figure 5.1, this process of continuing centralization appears to have its source in the New Deal years; there is no increase at all evident in the federal share prior to 1934. Thus, the data suggest that the New Deal programs both increased fiscal centralization and, at the same time, set in motion a process of further centralization. The basic theory of fiscal federalism can provide a partial explanation for this trend toward greater fiscal centralization. The motivation behind the assignment of functions to local, state, and national governments is different for the allocative function of providing basic public goods and services (such as education, roads, and police and fire protection) than it is for the redistributive function of providing support for low-income households (Oates 1972, chap. 1; 1994). Decentralized levels of government have as their primary economic role the provision of levels of “local” public goods that are tailored to the particular preferences and circumstances of their own jurisdictions. In contrast, providing assistance to the poor requires a more substantial central presence. A local government, for example, that attempts to redistribute income aggressively from wealthy to poor households immediately creates incentives for resident high-income families to move elsewhere and for poorer families to migrate into the jurisdiction. It is easy to show that such mobility can readily undermine the achievement of local redistributive objectives and will, in general, result overall in suboptimal levels of support for the poor (see, e.g., Brown and Oates 1987). The implication is that in a federal system, the central government must play a major role in the design and financing of assistance to the poor.

178

John Joseph Wallis and Wallace E. Oates

Since the New Deal had poor relief as one of its major objectives, it is not surprising to find that the introduction of needed redistributive programs involved an increased degree of centralization of the public sector. The New Deal, as we have seen, introduced a variety of major welfare programs including social security and categorical assistance programs such as OAA, aid to the blind, and ADC (the predecessor to AFDC). As programs like AFDC grew in importance in later years, we find that the increased public role in income redistribution manifests itself in part in an increased degree of fiscal centralization. Income maintenance programs could, in principle, be wholly centralized, with a centrally designed set of rules and benefits administered uniformly across jurisdictions. But, as we have seen, the New Deal took another course. The political setting for these programs was such that state and local officials demanded and got an active role in the specification of various key parameters for these programs (e.g., support levels) and for their administration, while the central government set up general guidelines and provided the bulk of the funds in the form of matching grants. The political economy of the new redistributive programs took the form of a cooperative enterprise between the various levels of government that has proved quite durable.19 The New Deal, however, was not the progenitor of all the changes in government structure since 1940. The New Deal was financed through national borrowing as much as through tax revenues. As a result, the changes in fiscal structure occurred largely on the expenditure side at the national level; it was at the state and local levels that new forms of taxation emerged in the 1930s. Not until World War I1 with the advent of income tax withholding did the national revenue structure take new shape. A whole range of grant programs was begun in the 1960sthat did not follow the New Deal pattern. These varied in their intent and structure, but typically the national government assumed much more direct control over the administration of the programs. The War on Poverty programs that cut state and even local governments out of the process are, by and large, gone now. The general revenue-sharing program is also gone. Although there was a move to consolidate conditional project grants into looser block grants for general purposes, block grants today are a relatively modest part of the fiscal system. For the most part, experiments in intergovernmental programs over the past 30 years that have not followed the New Deal pattern have not stood the test of time. We have argued that the New Deal pattern of intergovernmental relations was the result of the struggle between state and national governments, and also between the president and Congress, for control over these programs. Ever since the New Deal, Congress has shown less inclination to locate either ad19. However, certain recent changes in the structure of welfare programs aim at shifting more of the responsibility for poor relief back to the states and changing the form of federal support from matching to bloc grants. It remains to be seen how all this will work out.

179

The Impact of the New Deal on American Federalism

ministrative or financial control at the state level. As a result, the national government has continued to experiment with a wide range of intergovernmental forms. State and local governments, however, still want the substantial autonomy they were given in the New Deal programs. And they have effectively lobbied for a continuation of New Deal programs, albeit with growing complaints about unfunded mandates, and kept those programs in existence. A widely misunderstood result of all that happened during and since the New Deal is the notion that it was only the federal government that grew during the 1930s. In 1927, state own revenues were 2.1 percent of GNP, and local own revenues were 6 percent of GNP. In 1940, state own revenues were 5 percent of GNP, and local own revenues were 5.8 percent of GNP. During the 1930s neither state nor local governments became smaller relative to the economy, and state governments grew. In 1992, state revenues were 7 percent of GNP, and local revenues were 6.1 percent. Although the share of total fiscal activity at the local level declined sharply during the New Deal, the share at the state level actually rose. State expenditures were 13 percent of all government expenditures in 1927 and 17.5 percent in 1940. Rather than displacing state governments, the expansion of national government activity actually created incentives for state governments to expand their roles. The New Deal legacy has arguably been stronger governments at all levels, not just the central government. This is the result of national government programs that encourage, even demand, state and local participation in nationally funded programs. Even local governments, whose fiscal share has declined so sharply, continue to play the major role in the actual delivery of most major domestic public services (see figure 5.1). The New Deal was clearly the time during which the national government came into prominence in the domestic sphere, but the way in which this took place has led to a cooperative form of federalism involving active roles for all three levels of government. The relative importance of the three levels may have varied at times over the past 50 years, but as we approach the close of the century, all three seem well entrenched in an active, if sometimes contentious, public partnership.

References Advisory Council on Intergovernmental Relations. 1995. Signijicant features of jiscal jederalism, 1995. Washington, D.C.: Government Printing Office. Alston, L. J., and J. P. Ferrie. 1985. Labor costs, paternalism, and loyalty in Southern agriculture: A constraint on the growth of the welfare state. Journal of Economic History 45 (March): 95-118. Block, F., R. A. Cloward, B. Ehrenrich, and E F. Piven. 1987. The mean season: The attack on the welfare state. New York: Pantheon. Bradford, David F., and Wallace E. Oates. 1971. The analysis of revenue sharing in

180

John Joseph Wallis and Wallace E. Oates

a new approach to collective fiscal decisions. Quarterly Journal of Economics 85 (August): 4 16-39. Bremer, William W. 1984. Depression winters: New York social workers and the New Deal. Philadelphia: Temple University Press. Brock, William Ranulf. 1988. Welfare,democracy, and the New Deal. New York: Cambridge University Press. Brown, Charles C., and Wallace E. Oates. 1987. Assistance to the poor in a federal system. Journal of Public Economics 32 (April): 307-30. Brown, Josephine Chapin. 1940. Public reliej 1929-1939. New York: Holt. Bryce, James. (1888) 1901. The American commonwealth. London: Macmillan. Gramlich, Edward. 1977. Intergovernmental grants: A review of the empirical literature. In The political economy offiscal federalism, ed. Wallace E. Oates, 219-40. Lexington, Mass.: Lexington Books. Hines, James R., and Richard H. Thaler. 1995. Anomalies: The flypaper effect. Journal of Economic Perspectives 9 (fall): 217-26. Howard, Donald S. (1943) 1973. The WPA and federal relief policy. New York: Da Capo. Johnson, Ronald N., and Gary J. Libecap. 1994. Thefederal civil service system and the problem of bureaucracy. Chicago: University of Chicago Press. Oates, Wallace E. 1972. Fiscal federalism. New York: Harcourt Brace Jovanovich. . 1994. Federalism and government finance. In Modern public$nance, ed. John Quigley and Eugene Smolensky, 126-61. Cambridge, Mass.: Harvard University Press. Piven, E F., and R. A. Cloward. 1971. Regulating the poor: The functions of public welfare. New York Pantheon. Sawer, Geoffrey. 1969. Modemfederalism. London: Watts. Stewart, William H. 1984. Concepts of federalism. Lanham, Md.: University Press of America. Tocqueville, Alexis de. (1835) 1945. Democracy in America. New York: Vintage. US. Bureau of the Census. 1955. State and local governments. Special Studies, no. 38. Washington, D.C.: Government Printing Office. . 1975. Historical statistics of the United States. Washington, D.C.: Government Printing Office. US. Census of Governments. 1985. 1982 Census of governments: Historical statistics on government finance and employment. Washington, D.C.: Government Printing Office. Wallis, John Joseph. 1981. Public relief and unemployment in the 1930s. Unpublished Ph.D. diss., University of Washington, Seattle. . 1984. The birth of the old federalism: Financing the New Deal. Journal of Economic History 44 (March): 139-59. . 1991. The political economy of New Deal fiscal federalism. Economic Inquiry 29 (July): 5 10-24. . 1997. Form and function in the public sector: State and local government finances in the United States, 1902-1982. Manuscript. Williams, E. A. 1939. Federal aid for relief: New York: Columbia University Press.

6

The Great Depression and the Regulating State: Federal Government Regulation of Agriculture, 1884-1 970 Gary D. Libecap

Agriculture, noted Theodore Lowi, is “where the distinction between public and private has come closest to being completely eliminated” (1979,68). Agriculture is among the most regulated sectors of the American economy. The production and sale of almost all its commodities are affected by some government policy through a complex mix of programs. Leading students of agricultural regulation have attributed this regulatory regime to New Deal legislation, such as the Agricultural Adjustment Acts of 1933 and 1938, and to related laws, such as the Agricultural Act of 1949 and the Soil Bank Act of 1956.’ Yet, just how the New Deal changed the extent and nature of agricultural regulation in the United States remains to be demonstrated. This paper examines agricultural regulatory laws enacted by Congress between 1884 and 1970 and the corresponding budget expenditures between 1905 and 1970 to determine how the path of regulation was altered by New Gary D. Libecap is professor of economics and law at the University of Arizona and a research associate of the National Bureau of Economic Research. Revisions were completed while the author was an OIin Fellow in Law and Economics at the Cornell Law School, November 1996. Research assistance was provided by Joe Bial, Mike Miller, and Bart Wilson. Thorough comments were provided by Claudia Goldin, Bruce Gardner, and B. Delworth Gardner. Additional suggestions were provided by the participants at the preconference and by Sam Peltzman and conference participants. 1. Such students of agricultural regulation include Donald Blaisdell (1940, 39), Murray Benedict (1953.469-520). Murray Benedict and Oscar Stine (1956, xv), D. Gale Johnson (1973, I), Willard Cochrane and Mary Ryan (1976, 23, 132), Bruce Gardner (1981, 13; 1987b, 20), and B. Delworth Gardner (1995, 8). New Deal legislation includes the Agricultural Adjustment Act of 1933, P.L. 10; Agricultural Adjustment Act of 1938, P.L. 430; Soil Conservation and Domestic Allotment Act of 1937, P.L. 137; Sugar Control Act of 1937, P.L. 414; Commodity Credit Corporation, Executive Order 6359, 16 October 1933. Important post-New Deal legislation includes the Commodity Credit Corporation Charter Act of 1948, P.L. 806; the Agricultural Act of 1949, P.L. 439; the Agricultural Trade Development and Assistance Act of 1954, P.L. 480; the Soil Bank Act of 1956, P.L. 540; the School Milk Act of 1956, P.L. 465; and the Food Stamp Act of 1964, P.L. 525.

181

182

Gary D. Libecap

Deal programs. The start and end dates of the study are largely determined by the availability of data. The end date of 1970 was chosen because of two compilations of agricultural legislation assembled by Udell (1971, 1972), which contain entries for statutes enacted from 1884 through 1971. The nature of agricultural regulatory policy, however, is not sensitive to the end date chosen. Legislation enacted in the post-World War I1 period has continued policies prescribed by New Deal legislation, especially the Agricultural Adjustment Act of 1938 and the Agricultural Act of 1949, which itself extended New Deal regulations. This condition was true before 1970 and continued after that year.2 For example, the Freedom to Farm Act passed in 1996 ended New Dedbased acreage restrictions and price supports for grains and other commodities. Guaranteed-loan programs through the Commodity Credit Corporation (CCC), however, remained, as did regulations for peanuts, sugar, and dairy produck3 I contend that the New Deal increased the amount and breadth of agricultural regulation in the economy and, of more significance, shifted it from providing public goods and transfers to controlling supplies and directing government purchases to raise prices. This type of economic regulation was unprecedented. Government purchases during World War I were intended to aid the war effort, not raise prices. Prior to 1933, there is no record of government-imposed output controls on the scale that characterized New Deal programs. More important, I show that the New Deal created the institutional structure needed to continue the new regulation, which I contend, was the most consequential aspect of the agricultural legislation enacted between 1933 and 1939. Agricultural laws passed by Congress and the president from 1884 through 1970 are classified as to whether they provided public goods (controlled disease, fought insect pests, provided product quality standards), gave direct and indirect transfers (grants, subsidized loans and insurance, soil improvements), or engaged in economic regulation, where economic regulation includes demand enhancement through the government purchase of agricultural commodities and supply control through production and marketing limits.4 Addition2. For discussion of farm policies since 1970, see Bruce Gardner (1981), Pasour (1990), and Delworth Gardner (1995). 3. Although championed as a fundamental break from New Deal policies, the Freedom to Farm Act of 1996 in many ways appears to be a continuationof past programs. For example, grain prices in 1996 were at record levels, so that New Deal-based deficiency payments (difference between the target price and the market price) for farmers would have been zero in most cases. The legislation, however, provided for guaranteed payments to farmers unlinked to market prices. Further, under the law, should Congress fail to enact more permanent farm legislation by 2002, then farm policies will revert to those called for by the AgriculturalAdjustment Act of 1938 and the Agricultural Act of 1949. For discussion, see Wall Street Journal, 30 August 1995, 26 September 1995, 8 February 1996, and 1 April 1996. See also New York Times, 26 September 1995, 1 October 1995, and 29 February 1996. Articles on farm policy also appear in Economist, 7 October 1995 and 9 March 1996. 4. I recognize that distinctions between these categories are not always sharp. I categorized legislation based on the primary intent of the law as I interpreted it.

183

Federal Government Regulation of Agriculture, 1884-1 970

ally, laws enacted from 1940 through 1970 are classified as to whether they were linked to specific New Deal agricultural programs. The objective of this linkage is to see the degree to which post-New Deal regulation was directly tied to New Deal programs. The hypothesis is tested that absent the Great Depression and New Deal, the pattern of agricultural regulation with public goods and transfers that existed prior to 1933 would have continued through 1970. Additionally, budget appropriations for economic regulation of agricultural commodities are assembled and categorized as demand enhancement and supply control to analyze how the New Deal affected regulatory expenditures relative to what existed before 1933. The impact of the New Deal on agricultural regulation is clear. In terms of legislative activity, Congress from 1884 through 1970 enacted over 650 laws for the regulation of agriculture. Only 10 percent of the legislation was passed prior to the New Deal, and almost 80 percent of those laws were for the provision of public goods or transfers to agriculture. Intervention to fix commodity prices through economic regulation statutes was uncommon. In the New Deal period of 1933-39, however, 130 new laws were enacted (almost twice as many as in the previous 50 years), with more than 60 percent aimed at controlling supplies or increasing commodity demand. This legislative pattern continued after the New Deal, with over 450 regulatory laws passed between 1940 and 1970,60 percent of which were for economic regulation. Budget expenditures between 1905 and 1970 mirror the increase in regulation and the change in regulatory emphasis after 1932.5 Since 1933, commodity prices have been raised through demand enhancement and supply reduction. Supply control through the temporary or permanent removal of agricultural production from markets or through restrictions on the use of inputs, notably land, grew substantially during the New Deal compared with what previously existed. The most important demand enhancement policies were government purchases through the CCC and related agencies. These demand policies were implemented during the New Deal, and they, along with supply constraints, remain the post-New Deal centerpiece of federal agricultural regulation. Two indications of the long-term impact of New Deal programs are provided. One measure examines the impact of New Deal regulatory mandates on the staffing and budget of the U.S. Department of Agriculture (USDA). New Deal programs sharply increased the involvement of the USDA in American agriculture through new mandates and new programs, and the bureaucracy was active in drafting legislation that extended the department’s regulatory role. Indeed, the department became an important constituent in the development 5. See Johnson (1973, 22). The focus of New Deal agricultural regulation on price fixing to raise and stabilize farm incomes came after commodity and farm land prices had fallen sharply since 1929. Agriculture is particularly vulnerable to sharp cyclical swings due to low income elasticity and low price elasticity of demand for farm products. Under these conditions demand does not rise markedly with upswings in the business cycle, and price shifts often are sharp in response to supply changes.

184

Gary D. Libecap

and expansion of the institutional structure for agricultural regulation. To determine how the USDA fared, agency staffing (as a measure of agency size) relative to the total number of farms in the United States (as a measure of the magnitude of the principal clientele or political constituency) is presented from 1910 to 1970. The data reveal that while the number of farms declined, the number of USDA employees increased. This change also is largely a New Deal phenomenon. Throughout the 1920sthe USDA had approximately 20,000 employees, but by 1934, the agency had nearly three times that number, and 86,000 employees in 1939. By the late 1960%the USDA had over 120,000 employees. To gauge the USDA's position relative to other federal agencies, annual USDA expenditures as a share of total civilian federal government expenditures, from 1921 to 1970, are presented.6 During the 1920s, the department spent less than 6 percent of the federal civilian budget. But with new programs and regulatory mandates authorized by the New Deal, its share increased. By 1935, USDA expenditures were more than 18 percent of the federal civilian budget. Even after the depression subsided, the department's share of federal expenditures remained substantially above its pre-New Deal level. Another measure of the impact of New Deal regulation involves examination of domestic wheat prices. Wheat was one of the most important American agricultural commodities and the focus of various regulatory programs, during and after the New Deal. A domestic wheat price series is assembled from 1900 to 1970, with the post-1932 prices constructed using reported prices received by farmers plus price support payments. The series is compared to world prices as reflected in Australian wheat prices. If successful, New Deal programs should have raised (supported) U.S. domestic wheat prices relative to world prices after 1932. The price data reveal that they did. Commodity programs were also aimed at reducing the volatility of wheat prices, and that is another hypothesis tested in the paper. The role of the Great Depression in providing a crisis of sufficient magnitude to politically justify unprecedented new peacetime government intervention into agricultural markets conforms to the framework of Peacock and Wiseman (1961) and Higgs (1987), who argue that the growth of government obligations, regulations, and expenditures is discontinuous, with crises generating political support for government expansion. Agricultural regulation in other industrialized countries appears to have followed a path similar to that of the United States, generalizing the results of this study. In Australia, Canada, England, France, Germany, and Italy, the slump in commodity prices following World War I led to agricultural unrest and political agitation for government assistance. The general political response in the 1920swas to limit import competition through higher tariffs and quotas. But the Great Depression of the 6. The start date is 1921 because that is when the U.S. budget was first published as a single document and expenditure data can be obtained.

185

Federal Government Regulation of Agriculture, 1884-1970

1930s brought much greater and more direct government intervention into commodity markets to fix prices. Two-tier price policies were established, with domestic prices fixed above world prices. Chronic surpluses were controlled through government purchases and stockpiling and through output and marketing controls. Debt relief, lower interest rates, income deficiency payments, moratoria on mortgage foreclosures, and other programs were added. As in the United States, the legacy of depression-era intervention was an expanded role for the state in agricultural markets long after commodity prices had increased in the late 1930s.Across countries, agricultural programs enacted to meet the crisis were retained and broadened during World War I1 and in the postwar period.’ 6.1 Agricultural Regulation through 1970: An Overview

6.1.1 Pre-New Deal, 1884 to 1932 Compared to the period after 1933, the federal government played a limited role in regulating commodity markets through World War I. The reaction of Congress to late-nineteenth-century agrarian unrest was to provide indirect support-lower tariffs on manufactured products, railroad rate regulation, antitrust laws, food inspection to promote demand, restrictions against new competitors made possible by new technology (taxing oleomargarine in 1886 to make the product less competitive with butter, e.g.), increases in the money supply (through the coinage of silver), and promotion of marketing and buying cooperatives.8The demand for federal government assistance slackened after the turn of the century as commodity prices rose. By 1910,prices were so high and farm prosperity so great that the period 1910-14 has been described as “the golden age of agriculture” (Benedict 1953,115),and this became the “parity period” of later regulation. Rather than attempting to directly influence commodity prices between 1884 and 1917, Congress often addressed standard public goods problems, such as the investigation into and control of livestock and plant diseases and related measures for guaranteeing food quality through meat and drug inspection9 Increased settlement density and transportation improvements had pro7. The international comparison is discussed in more detail later in the text. Major sources include Corni (1990), Farquharson (1976), Hendricks (1991), Holt (1936), and Stuhler (1989) for Germany; Corni (1990) for Italy; Grigg (1989), Perren (1995). and Rooth (1993) for Britain; Moulin (1991) for France; Hefford (1985), Lloyd (1982), and Shaw (1982) for Australia; and Bothwell, Drummond, and English (1987) and Britnell and Fowke (1962) for Canada. A valuable crosscountry comparison of agricultural policies in Western Europe is provided by Tracy (1964). 8. For discussion ofrestrictions on oleomargarine,see Wood (1986). For discussion of the farm protest movement, see Benedict (1953,94-154) and Higgs (1987, 84-85). The peak of the farm protest movement was during the depression of 1893-96. 9. P.L. 41,29 May 1884, created the Bureau of Animal Industry for the inspection and quarantine of diseases among animals. P.L. 161.2 August 1886, and P.L. 110.9 May 1902, defined butter and taxed oleomargarine.P.L. 247.30 August 1890, was the first meat inspection law. See Libecap

186

Gary D. Libecap

moted the spread of diseases, such as hoof-and-mouth disease, Texas fever, and pleuropneumonia. Boll weevil and grasshopper infestations also threatened important export crops. Because these problems crossed state lines, the federal government was a natural focus of political demands for remedies. The response of Congress was to provide new mandates and appropriations to the USDA, which was given cabinet status in 1889. The Bureau of Animal Industry, the Bureau of Plant Industry, the Bureau of Chemistry, and the agricultural extension service, all within the USDA, carried out the new federal support of agriculture.1° The mobilization of the economy during World War I, however, brought greater federal intervention into commodity markets. The Lever Food Control Act of 10 August 1917 granted broad powers to the president for licensing the import, manufacturing, storing, and distribution of agricultural production, fuels, fertilizers, and equipment; regulating the prices of wheat, flour, meal, beans, and potatoes; and requisitioning food supplies for the war. The U.S. Food Administration, headed by Herbert Hoover, centralized the purchase and distribution of farm commodities.' I During the 1920s, additional regulatory legislation was enacted as a reaction to the sharp drop, in both real and nominal terms, in grain and other commodity prices in 1921. Farmers who had mortgaged their farms were unable to meet payments with current receipts, and a wave of bankruptcies and farm foreclosures ensued.'* The experience with the Lever Food Control Act of 1917 and the operation of the U.S. Food Administration during World War I had demonstrated the influence that government could have on prices, and farmers turned to the federal government for relief. The agricultural crisis of the 1920s brought a rise in militancy among farm organizations, such as the Farmers' National Relief Conference, the Farmers' Union, the National Grange, and the American Farm Bureau Federation. These groups lobbied Congress and the president for government intervention to raise prices and farm incomes. The first Farm Block lobby meeting was held in April 1921 in Washington, D.C., attended by representatives of the National Grange, American Farm Bureau Federation, National Milk Producers Federation, and Farmers' Union, as well (1992) for discussion. Other food inspection and livestock and plant disease laws included P.L. 49, 2 February 1902; P.L. 382, 30 June 1906; P.L. 384, 30 June 1906 (the Pure Food and Drug Act); P.L. 242, 4 March 1907; P.L. 275, 20 August 1912; P.L. 293, 4 March 1915; P.L. 390, 4 March 1917; and P.L. 4 0 , l O August 1917. 10. Benedict (1953, 111-18) provides discussion. See Young (1989) for discussion of the enactment of the Pure Food and Drug Act. 11. The agency remained active through January 1925 and later was revived as the Reconstruction Finance Corporation (Benedict 1953, 159-65). See Higgs (1987, 125-58) for discussion of federal government intervention into the economy during World War I through the Food Administration, the Fuel Administration, the Shipping Board, the Railroad Administration, and the War Industries Board. 12. The agricultural problem of the early 1920s is examined by James H. Shideler (1957). For additional discussion of the agricultural crisis of the 1920s, see Perkins (1969, 10-48) and Hoffman and Libecap (1991).

187

Federal Government Regulation of Agriculture, 1884-1970

by Secretary of Agriculture Henry C. Wallace and Secretary of Commerce Herbert Hoover. These farm organizations became increasingly adept in lobbying Congress to initiate farm programs in the late 1920s and 1930s and to extend the programs in the post-World War I1 period.I3 The early response of the federal government was to sanction and promote cooperative marketing associations to control market supplies to raise prices. This response was consistent with the dismantling of wartime programs because it relied on private organizations to address the problem of falling prices. The emphasis on cooperatives led to new congressional legislation to facilitate private cooperative efforts to control supplies and, thereby, raise prices.I4 For example, the 1926 Cooperative Marketing Act created the Division of Cooperative Marketing within the USDA to assist cooperatives in gathering and sharing data on output, prices, and demand. The 1929 Agricultural Marketing Act further promoted cooperatives and their joint efforts to control market supplies.15The law authorized the president to appoint a Federal Farm Board of eight members representing major agricultural commodities. Commodity advisory committees and price stabilization corporations, especially for wheat, cotton, wool, beans, corn, hogs, and cattle, were established to assist cooperatives in the enforcement of production and marketing rules and in promoting exports. The Federal Farm Board, drawing on a fund of $500 million, could make loans to cooperatives at subsidized interest rates to purchase and hold production temporarily off the market and to develop improved merchandising and distribution networks. Unpaid loans were to be absorbed by the Farm Board. Limited price insurance also was provided.I6But Hoffman and Libecap (1991) point out, given the number of farmers and cooperatives for major commodities such as wheat, cooperative efforts to control supply were unlikely to be successful. Because of problems of coordination among the many cooperatives and of 13. For discussion of early political activity, see Benedict (1953, 171-83, 277-80). For other discussions of the agricultural crisis of the 1920s and political lobbying by farm groups, see Perkins (1969, 10-48), Hoffman and Libecap (1991), and Saloutos (1982,28-43). 14. Even before 1920, agricultural cooperatives had received strong political support from Congress. Section 6 of the Clayton Act of 1914 specifically exempted certain agricultural cooperatives from the antitrust provisions of the 1890 Sherman Act. New statutes included the Capper-Volstead Act in 1922, P.L. 146; the Cooperative Marketing Act of 1926, P.L. 450; and the Agricultural Marketing Act of 1929, P.L. 10. Hoffman and Libecap (1991) examine the cooperative movement and efforts to obtain federal government assistance in raising farm prices. 15. The law declared that it was the policy of Congress “to promote the effective merchandising of agricultural commodities in interstate and foreign commerce, so that the industry of agriculture will be placed on a basis of equality with other industries, and to that end to protect, control, and stabilize the currents of interstate and foreign commerce in the marketing of agricultural commodities and their food products.” 16. The Farm Board had a short life; with falling farm prices after 1929, it paid more for the agricultural commodities it acquired than they could be sold for, and the board soon ran out of its appropriated capital. The effort was regarded largely as a failure because the various stabilization corporations could not purchase enough of commodities such as wheat and cotton to maintain prices. The Farm Board, however, set the precedent for later government programs to more directly influence prices. Further discussion is provided by Hamilton (1991).

188

Gary D. Libecap

policing compliance with marketing controls, there were efforts, especially among wheat growers, to obtain more direct federal intervention to raise prices. This lobby activity led to the McNary-Haugen bills considered by Congress between 1924 and 1928.” The initial McNary-Haugen bill, drafted by Charles J. Brand, former chief of the Bureau of Markets in the USDA, was introduced in Congress in January 1924. Under the McNary-Haugen bills, domestic and international markets for certain agricultural commodities, including wheat, cotton, wool, cattle, sheep, and hogs, were to be separated by a flexible tariff. In the domestic market, supply would be held to a level that would meet demand at a real price comparable to the commodity’s price during the period 1905-14. These prices were to be computed monthly by the Bureau of Labor Statistics. Production beyond that needed for domestic markets was to be purchased at the target price by an agricultural export corporation, composed of the secretary of agriculture and four presidential appointees. The corporation was to have a fund of $200 million to buy these “excess” supplies and sell them on the world market. The difference between the domestic price and the world price was to be absorbed by the corporation. In this way, the McNary-Haugen bill introduced domestic price supports for selected agricultural commodities nine years before the New Deal. But the legislation did not become law.’* Because of the unprecedented level of peacetime federal involvement required by the McNary-Haugen bills, they were controversial. Two versions were defeated in the House of Representatives in June 1924 and May 1926. Two other versions passed Congress in February 1927 and May 1928 but were vetoed by President Coolidge. The mobilization of political support in Congress for the McNary-Haugen legislation between 1924 and 1928 set precedents for subsequent lobbying activity during and after the New Deal. Farm organizations formed coalitions through the American Council of Agriculture, the Corn Belt Committee, and the National Producers’ Alliance. These groups in turn obtained the assistance of representatives of state agricultural colleges, USDA extension agents, producers of fertilizers and farm implements, millers, and meat packers. Further, because the legislation was seen as too wheat oriented, the coverage of the proposed statute was extended to include tobacco and rice, adding the political support of congressional representatives and senators from the South to those of the upper Midwest. Wheat prices had fallen earlier and more sharply than had other commodity prices, leading to initial advocacy by grain producers, 17. This legislation called for more direct federal intervention into agricultural markets, and it was most closely related to legislation later enacted during the New Deal. For discussion, see Benedict (1953,207-15), Fite (1954), and Hoffman and Libecap (1991). 18. The McNary-Haugen approach appealed to many farmers for two reasons. First, it involved direct federal intervention to raise commodity prices, with taxpayers providing funds to control surpluses. Second, the legislation did not require them to reduce output.

189

Federal Government Regulation of Agriculture, 1884-1970

but by the late 1920s cotton prices also had fallen so that cotton producers pushed for the McNary-Haugen bills as well.19 Although the McNary-Haugen bills were forerunners to subsequent New Deal legislation, there were two crucial differences: they were narrow in scope in comparison and had no supply controls. Limited supply controls, however, were introduced in the 1929 Agricultural Marketing Act, whereby the Federal Farm Board would purchase commodities from those farmers who complied with production targets.2O Despite the efforts of the Farm Board and its subsidiaries, commodity prices continued to fall in real terms between 1929 and 1932.*’About half of the 1930 U.S. wheat supply was held by the Grain Stabilization Corporation, created by the Agricultural Marketing Act, but even these purchases were insufficient to offset declining exports and falling domestic demand. The Cotton Stabilization Corporation, also created by the Agricultural Marketing Act, faced similar conditions, and the two organizations soon ran out of funds to purchase surpluses. The Farm Board was disbanded May 1933, and its assets transferred to the Farni Credit Administration. In addition to these actions designed to raise farm prices by promoting export demand and by limiting market supplies through organized cooperatives, Congress also expanded institutions that subsequently would be important for New Deal regulation. The Dairy Bureau was created within the USDA to provide marketing information support for dairy farmers. The Bureau of Agricultural Economics was set up within the USDA in 1922, combining the Office of Farm Management, the Bureau of Crop and Livestock Estimates, and the Bureau of Markets, to provide greater production and marketing cooperation with land grant colleges and to assemble market data. These data were essential for setting total production targets, assigning individual farm quotas, and directing government purchases of excess stocks that subsequently were part of the Agricultural Adjustment Administration (AAA) in 1933. 6.1.2 The New Deal, 1933-39

The federal government’s efforts to stem the fall in agricultural prices after 1929 through the Agricultural Marketing Act were ineffective, and farm groups lobbied for more direct economic regulation to reduce supplies and expand demand.22Farm organizations, such as the Farmers’ National Relief Conference, the Farmers’ Union, the National Grange, and the American Farm Bu19. For discussion, see Benedict (1953,207-30), Hoffman and Libecap (1991), and Fite (1954). 20. Discussion is given in Benedict (1953,216-39) and Perkins (1969,27-32). 21. For discussion, see Benedict (1953,205,262-67). 22. Agriculture was particularly hard hit by the Great Depression. The prices of many commodities had not recovered from their sharp fall in 1921, so that between 1919 and 1933, wholesale farm prices fell by 67 percent; whereas, over the same period nonagricultural wholesale prices fell by 45 percent. Moreover, the decline in agricultural prices was particularly severe after 1929. See U.S. Department of Commerce (1975, 199-200) and Perkins (1969, 11).

190

Gary D. Libecap

reau Federation lobbied Congress and the president for direct government intervention to control market supplies. Chronic overproduction was seen as the root of the problem, and federal government management of production was believed to be the only A policy of output control was strongly supported by the new secretary of agriculture in the Roosevelt administration, Henry A. Wallace.24The close ties between early New Deal agricultural and industrial policies were stressed by a leading historian of agricultural policy, Murray Benedict: “In both the NRA [National Recovery Administration] and the AAA emphasis was placed on the raising of prices through artificiallyinduced scarcity” (1953,294). Indeed, 450 agricultural codes were transferred from the NRA to the AAA between June and December 1933 for implementati~n.~~ The centerpiece of New Deal agricultural regulation, the Agricultural Adjustment Act of 1933 was the outcome of well-organized farm group lobbying, and the statute was drafted largely by Frederick P. Lee, legislative counsel for the American Farm Bureau Federation, once again showing the political influence of farm organizations.26The AAA was created within the USDA to implement the l e g i ~ l a t i o nThe . ~ ~ aim of the law was to raise agricultural prices to reestablish the relative purchasing power of farmers that had prevailed from 1909 to 1914. The statute called for farmers to enter into agreements with the secretary of agriculture to reduce their acreage in seven basic commoditieswheat, cotton, corn, rice, tobacco, hogs, and milk-in return for federal “benefit” payments to be derived from taxes levied at processing. These seven commodities later were augmented by beef, dairy cattle, peanuts, barley, flax, grain sorghum, sugar beets, sugar cane, and potatoes through legislation such as the Warren Potato Control Act of 1935 (P.L. 320). Output was to be reduced for these basic crops through acreage Each year, based on expected demand, the secretary of agriculture was to deter23. Cochrane and Ryan (1976, 12) describe the farm problem of the 1920s and early 1930s as one of chronic, excess productive capacity. 24. Perkins (1969,43, 81-86) discusses production control as the major tool for farm relief and describes Secretary Wallace’s strong commitment to it. See, also, Nourse, Davis, and Black (1937, 20). The literature of the time is clear on production control as the solution to the farm problem. E.g., the American Institute of Cooperation, which published American Cooperation, formed a roundtable committee in 1932 on production control, and the journal published articles in the early 1930s on the legality and necessity of production control (Hulbert 1932; Ezekiel 1934). Once the costs to individual members of supply controls became apparent, the emphasis of these organizations shifted from supply controls to demand enhancement. Tellingly, by 1938, the American Institute of Cooperation was publishing articles on government purchases and the problems of too much reliance on cooperative solutions (Brandt 1938; Stedman 1938). 25. For more discussion, see Nourse (1935, 24-49). 26. The politics underlying the enactment of the Agricultural Adjustment Act are described by Murphy (1955, 160), Shover (1965), and Perkins (1969.37-44). 27. For discussion of the early AAA, see Benedict (1953, 284), Irons (1982, 111-55). Schultz (1949,41), Higgs (1987, 175-78), and Gardner (1987b, 55). 28. Breimeyer (1983, 343) discusses paid acreage reductions under the AAA. Those farmers who did not take part would be ineligible for benefit payments.

191

Federal Government Regulation of Agriculture, 1884-1970

mine how much land should be removed from production for each commodity to raise prices to target levels. A base acreage was to be established for each grower, and production quotas were to be determined by percentage reductions in the base.29Each grower's base acreage, annual allotment, and compliance record were set and monitored by local, county commodity committees. These committees were made up of growers who coordinated with the relevant state allotment committee, a state crop and livestock statistician (at least partially paid by the USDA), USDA extension agents, and representatives of land grant c011eges.~"Farmers who complied with the acreage allotments were to receive rental payments for the land removed from production. Besides controlling supply, demand was to be promoted through nonrecourse loans (purchases) from the CCC.31 Financing for the AAA's regulatory policies was covered in Section 9(a) of the Agricultural Adjustment Act, which called for a tax to be levied on the first domestic processing of the commodity. The processing tax, which varied by commodity, was to be the difference between the market price and the commodity's "fair exchange value" or parity The tax revenues were to cover payments to farmers for reducing productive acreage and to cover the costs of subsidizing exports. Additionally, the secretary of agriculture was to receive 30 percent of customs revenues for use in agricultural price support and surplus removal activities. The same farm organizations that lobbied for the McNary-Haugen bills were active in the passage of the Agricultural Adjustment Act and in molding its provisions. Indeed, as noted above, the legislative counsel of the American Farm Bureau Federation helped to draft the legi~lation.~~ Two aspects of New Deal programs likely increased the political influence of farm groups in obtaining long-term regulatory policies in their behalf. One was the specialized, commodity nature of the legislation. In a practice that continues today, almost 29. Base acreage provisions are discussed by Benedict (1 953, 303). 30. For details on the administrative structure of the AAA, see Nourse et al. (1937,63-77). 3 1. For specialty crops, such as fruits and vegetables, the secretary of agriculture could issue a marketing agreement, if half of the shippers and two-thirds of the farmers in a region agreed to the provisions of the agreement. The marketing agreements authorized the secretary to limit interstate shipments through weekly allotments to shippers that were enforced through revokable shipping licenses and fines for violation. Marketing agreements also regulated shipments through flexible quality controls that could be tightened as production increased and shipping holidays that prohibited the movement of crops to market during specified periods. The original agreements for a variety of fruits, nuts, and vegetables were voluntary. In the face of noncompliance, they were supplemented with marketing orders issued by the secretary of agriculture as authorized by amendments to the Agricultural Adjustment Act (P.L. 340) on 24 August 1935. These marketing orders were binding on all growers and interstate shippers of the commodity covered by the agreement. Between 1933 and 1935,61 marketing agreements were approved by the secretary of agriculture for milk, oranges, grapefruit, dates, pecans, walnuts, olives, raisins, and asparagus, among other commodities. For discussion of marketing orders, see Nourse (1935,53) and Nourse et al. (1937,231-34). For analysis of their provisions, see Hoffman and Libecap (1994). 32. See Blaisdell (1940,45) for discussion of the tax. 33. For discussion of the political action of farm groups, see Blaisdell (1940,41) and Nourse et al. (1937, 268-79).

192

Gary D. Libecap

every statute mandating or extending federal agricultural policies was partitioned into commodity segments, such as wheat, cotton, and tobacco, with provisions and benefits varying across commodities. The bundling of commodity programs into single statutes no doubt helped agricultural groups collectively advance farm bills, and the range of commodities covered was extended as need be to obtain greater political support for the legislation. For example, the original 7 basic commodities covered by the Agricultural Adjustment Act of 1933 was extended to 16 through a series of amendments in 1934 and 1935 at the behest of various producer groups and their congressional representatives.34 With the commodity focus of the legislation, each producer group could mobilize member support for regulation by focusing on the specific benefits directed to their commodity. Further, the commodity-specific benefits were clear to members of Congress, who could monitor and expand them for their constituents. Given the committee structure of the Congress, the seniority vested in members from agricultural regions, and the overrepresentation of rural states in the Senate, farm representatives were in a position to deliver to their constituent^.^^ A second aspect of New Deal agricultural policies that likely increased constituent influence over the long term in devising and maintaining those policies was the delegation of administration of commodity programs to local and state grower committees and to USDA and land grant college extension agents. These groups in turn worked closely with the USDA Bureau of Agricultural Economics, Division of Crop and Livestock Estimates, and other USDA offices. Hence, there was a coalition between the administering agency and the constituent growers that extended from the grass roots up through the agency. They could jointly modify programs or, as necessary, draft new legislation and have it introduced by the relevant congressional representative^.^^ With the mandate of the Agricultural Adjustment Act of 1933, the federal government began to reduce supplies aggressively. In June 1933, between 25 and 50 percent of the sown cotton crop was plowed up, and wheat and tobacco acreage was red~ced.~’ A severe drought between 1933 and 1936 also helped to decrease supplies of cotton, corn, and wheat. To more strictly enforce farm 34. Expansion of commodities covered helped in obtaining final congressional passage of the McNary-Haugen legislation and was used to broaden political support of the Agricultural Adjustment Act (see Nourse et al. 1937, 42-43). The commodity nature of agricultural programs is emphasized in the analysis by Gardner (1987b). 35. Increasingly, logrolling has been involved between representatives of agricultural and urban regions with programs such as food stamps that benefited both. For discussion of farm group influence, see McCune (1956), and for analysis, see Gardner (1987a). 36. The interaction between farm groups, the USDA, and Congress in drafting legislation is illustrated with the Agricultural Adjustment Act of 1938 in Blaisdell(1940,55-60). Hoffman and Libecap (1994, 201-18) examine the negotiations between orange grower groups and the USDA in devising and modifying orange-marketing agreements. 37. Perkins (1969, 103, 124) discusses early reductions in supplies under the AAA.

193

Federal Government Regulation of Agriculture, 1884-1 970

quotas, Congress passed legislation to tax production that came from acreage beyond individual farmer quota allocation^.^^ Although the initial focus of New Deal farm policies was on supply controls as a means of raising prices, increasingly reliance was shifted to demand enhancement through the government purchase of surplus p r o d u ~ t i o nAs . ~ ~acreage and marketing limits were tightened in 1933 and 1934, farmer disenchantment with supply controls grew. Farmer participation in AAA programs declined after the first planting season, and farm groups lobbied the federal government to relax production quotas and to turn to alternative methods of raising agricultural prices.40Additionally, the dramatic actions taken by the AAA in 1933 to reduce supplies through the plow-down of cotton acreage and the “emergency” hog slaughter brought widespread criticism of the agency at a time when many Americans were having difficulty acquiring enough food. Supply controls of various types were maintained as part of agricultural regulation, but they were supplemented by government commodity purchases, funded by general tax revenues. Major government acquisitions to augment private demand reduced the supply reductions necessary to raise prices. Government purchases were also politically attractive because they avoided the distributional problems of assigning, enforcing, and reducing farm quotas. The CCC was the primary agency for buying agricultural commoditie~.~’ Substantial government purchases of agricultural production to support target prices was a new attribute of New Deal regulation that had not been relied upon extensively before 1933. In September 1933, the Federal Emergency Relief Administration announced that it would buy $75 million ($577 million in 1984 dollars) of surplus commodities. On October 1933, the Farm Credit Administration purchased 16 million bushels of wheat, and the CCC raised the loan rates (prices it paid) for 38. Bankhead Cotton Control Act of 1934, P.L. 169; the Kerr Tobacco Act of 1934, P.L. 483; and the Ken-Smith Tobacco Act of 1936, P.L. 534. 39. Nourse et al. (1937, 37, 186-91) discuss the initial focus on supply control and the shift to greater reliance on government purchases to raise farm prices. Benedict (1953, 304, 311, 313) notes that as opposition mounted to production quotas, yields were increased on quota acreage through the substitution of capital and labor for land, and farmer participation in AAA programs declined. For additional discussion and analysis, see Perkins (1969, 103, 140) and Hoffman and Libecap (1995). As argued by Hoffman and Libecap, tighter cartel controls were not the primary response of Congress and the Roosevelt administration to the political reaction to cartelization, and the literature is uniform in concluding that the output and market controls of the AAA were unsuccessful in raising agricultural prices relative to other goods. Schultz (1949, 143) points out that although corn acreage fell by 8 percent between 1937 and 1939, output grew by 17 percent. For additional assessments, see Nourse et al. (1937, 289-320). Stricter production controls were achieved only in tobacco and peanuts (Gardner 1987b, 21). 40. The incentive to exceed quota limits is a standard cartel problem that was not fully appreciated in congressional debate over the Agricultural Adjustment Act. For discussion, see Hoffman and Libecap (1995). 41. For discussion of the CCC, see Perkins (1969, 168,224).

194

Gary D. Libecap

hogs, wheat, and cotton. The Federal Emergency Relief Administration followed with a purchase of hogs for relief purposes. The October 1933 wheat purchases were about 3 percent of total U.S. wheat production that year. It was a small beginning that was to Nonrecourse loans of the CCC to farmers became a principal mechanism for government purchases of agricultural production that continued well beyond the New Deal. Farmers obtained nonrecourse loans from the CCC at the policy-determined price (the loan rate) and placed the crops covered by the loan in CCC-certified storage on their farms or in commercial storage to be held as collateral. If the market price exceeded the loan rate, farmers sold their crops and paid off the loan from the CCC. If the loan rate exceeded the market price, farmers defaulted on their loans, and the CCC, in effect, purchased their output. Supplemental or deficiency payments to farmers were made if the loan price were below the target parity price for the crop. With these price support purchases through the CCC, the federal government’s demand for commodities was perfectly elastic at the support price, which effectively became the domestic market price By the end of 1934,48 percent of U.S. cotton production that year was either purchased by the CCC or pledged to the agency as collateral for loans. Stocks held by the CCC declined as production fell during the drought years of 1935 and 1936, but in 1938 the agency once again held an equivalent of 38 percent of that year’s cotton crop and in 1939 had an equivalent of 12 percent of corn production and 23 percent of wheat production.- The 1938 Agricultural Adjustment Act specifically directed the CCC to make price-supporting loans (purchases) whenever certain specified conditions arose. The agency, for example, was to make such loans for cotton and wheat if market price fell below 52 percent of the parity price or if production was expected to exceed domestic consumption and export demand (apparently at the parity price). These loans could be defaulted on if the market price remained below the loan rate. Political pressure forced CCC loan rates (purchase prices) higher between 1934 and 1941, leading to the greater accumulation of wheat, cotton, and corn The value of CCC loans to farmers, with crops held as collateral, rose 42. Perkins (1965, 221-28) describes purchases by the Federal Emergency Relief Administration, the Farm Credit Administration, and the CCC. CCC purchases are from the U.S. Department of Commerce (1975,511). 43. As Gardner (1981, 22-23) points out, price support programs vary by commodity. He describes the actions of the Flue-Cured Tobacco Cooperative Stabilization Corporation to set price floors for different grades of tobacco and the actions of the federal government in setting dairy price supports for butter, various cheeses, and nonfat dry milk. In terms of quantities and expenditures, grains involve the most massive support programs. 44. CCC loans (inventories) as a percentage of annual production calculated from USDA, Annual Report (1941, 20) for CCC loans and purchases by commodity by year and USDA, Agricultural Statistics (1942, 65) for production data. 45. For discussion of political pressure to raise the CCC loan rate, see Benedict (1953, 333, 376-78).

195

Federal Government Regulation of Agriculture, 1884-1970

from $260 million in 1934 ($1.9 billion in 1984 dollars) to $457 million in 1939 ($3.3 billion in 1984 dollars).46Revenues from CCC loans and other government benefit payments became an important part of total farm income. Nourse et al. (1937, 285) suggest that one-fourth of the increase in farm income in 1933 was due to government transfer payments, two-thirds in 1934, and one-half in 1935.47These figures are consistent with Schultz’s (1949, 154) estimate that by 1939, loans and other benefit payments from the federal government provided as much as a quarter of total farm income. The CCC loan program became the center of the New Deal federal farm program, and Murray Benedict commented on the changing nature of its role: “The Commodity Credit Corporation’s activities came to have a second purpose which was not compatible with its [price] stabilization function. This was the function of maintaining prices continuously above the free-market levels, rather than merely that of ironing out the effects of ups and downs of production and control” (1953, 389). The accumulation of stocks through repeated purchases by the CCC was justified by the Ever-Normal Granary policy adopted by Secretary Wallace in June 1934. Stockpiles were distributed as relief through the Federal Emergency Relief Administration, the Surplus Commodities Corporation (later, the Federal Surplus Commodities Corporation), the Red Cross, and other agencies or sent overseas as subsidized exports.48 When the processing tax provision of the Agricultural Adjustment Act of 1933 was declared unconstitutional by the Supreme Court in January 1936 in United States v. Butler (297 US. l), production restrictions were reinstated within two months by Congress under the Soil Conservation and Domestic Allotment Act. The new law tied benefit payments to farmers to acreage reductions for “soil conservation.” Funds for benefit payments were provided by Congress from general tax revenues, rather than processing taxes. Crops were classified into soil depleting and soil conserving. Soil-depleting crops were basic cash crops-wheat, cotton, corn, tobacco, and sugar beets among others-whereas soil-conserving crops were forage crops that did not add to surpluses. The shift in orientation from price supports under the Agricultural Adjustment Act to soil conservation under the Soil Conservation and Domestic Allotment Act was an expedient for continuing programs that reduced productive acreage in exchange for benefit payments to farmers. Politicians apparently believed that output reductions as part of soil conservation programs were more popular with voters than were output reductions explicitly designed to 46. The quantities pledged to or purchased by the CCC are from USDA, Annual Report (1941, 20) and CCC loan amounts are from U.S. Department of Commerce (1975,488). 47. Income effects of farm programs are also discussed in Rucker and Alston (1987). 48. The Ever-Normal Granary policy and the distribution of government purchases are discussed in Breimeyer (1983,346) and Cochrane and Ryan (1976, 132-64).

196

Gary D. Libecap

raise farm (food) ~ r i c e s . Moreover, 4~ the emphasis on soil conservation and the maintenance of an Ever-Normal Granary reflected a shift in Congress from a short-term “emergency” farm program under the Agricultural Adjustment Act of 1933 to relieve agricultural distress to a long-term program of regulating commodity markets.50 Another major New Deal regulatory law was the Agricultural Adjustment Act passed in February 1938. The law expanded a number of previous regulatory and transfer policies. It provided for crop insurance, modified acreage restrictions and allotment features in the 1936 Soil Conservation and Domestic Allotment Act, outlined criteria for marketing quotas, redefined parity prices, defined parity income more clearly, and described policy regarding loans from the CCC. Parity prices were expanded to reflect interest payments on farm indebtedness, taxes, and freight rates. In terms of production quotas, if the current and prospective supply of a given commodity exceeded a “normal supply” the secretary of agriculture was directed to design quotas for the amount of the commodity that could be marketed during the year. The normal supply of basic commodities as corn, cotton, rice, and wheat was broadly defined in the law. The proposed quotas were to be voted on through secret ballot by those farmers who planned to participate in the program.*’ Once enacted, the quotas determined the amount of the commodity that farmers could sell. Amounts sold by farmers beyond their quotas were to be taxed. By 1938, under New Deal regulatory programs, there were four ways of affecting commodity markets: (1) Soil Conservation and Domestic Allotment Act conservation payments to encourage farmers to shift from soil-depleting (cash) crops to soil-conserving (forage) crops and benefit payments for adhering to acreage allotments for production controls, (2) marketing orders and marketing agreements under the Marketing Agreement Act of 1937 to limit weekly shipments of specialty crops and to set minimum prices for milk, (3) 49. For discussion of the Soil Conservation and Domestic Allotment Act, see Benedict (1953, 339-51). The American Farm Bureau Federation assumed leadership in drafting and securing enactment of both the 1933 Agricultural Adjustment Act and the Soil Conservation and Domestic Allotment Act of 29 February 1936. The Supreme Court left standing the provisions of the JonesCostigan Sugar Control Act of 9 May 1934, the marketing agreement provisions of the 1933 Agricultural Adjustment Act, and Section 32 of that law, which assigned 30 percent of customs revenue to the secretary of agriculture for administering agricultural programs. 50. Blaisdell (1940.40-54) and Nourse et al. (1937, 21-31) discuss the shift from emergency legislation to deal with immediate problems to longer term economic planning through government programs. Although the marketing agreement provisions of the Agricultural Adjustment Act were not affected directly by the Butler decision, Congress provided clear legislative authorization for their marketing control provisions with enactment of the Agricultural Marketing Agreement Act of 3 June 1937. The law called for the issuing of state marketing quotas for specific specialty commodities and the assignment of individual shipper quotas within those statewide allotments. Penalties for shipper violation of quota provisions were authorized. Marketing agreements were upheld subsequently in Edwards v. Unired States (14 F. Supp. 384) in July 1937. By this time, the Supreme Court’s ruling in the NLRB v. Jones and Laughlin Steel Co. (301 U.S. 1 [1937]) took a broader view of the power of the government in what could be regulated as interstate commerce. 5 1. If one-third of the growers of a crop voted against the quota for that year, it would not take effect; however, price support loans would be withheld from growers during that year.

197

Federal Government Regulation of Agriculture, 1884-1 970

production quotas under the Agricultural Adjustment Act of 1938 to control market supplies of basic crops, and (4) nonrecourse loans from the CCC to finance government purchases of commodities to support prices. 6.1.3 Agricultural Regulation, 1940-70 During World War 11, the regulatory problem in agriculture shifted from one of limiting supply and increasing demand to one of promoting supply, limiting civilian demand, and constraining agricultural price increases to support the war effort. New Deal institutions were in place, and in the case of parity pricing, the vested interest of agriculture in continued price increases remained. Despite the efforts of the Office of Price Administration to control agricultural prices, the farm lobby succeeded in obtaining congressional legislation that set a high floor for commodity prices at 110 percent of parity or higher. With various exemptions to price controls, agricultural prices rose relative to other prices throughout the war period.52 Between the end of World War II and 1970, the development of modern agricultural regulation continued through extensions of New Deal legislation with new price supports; deficiency, parity, or benefit payments to farmers equal to some portion of the difference between the income from the target price and the market price; acreage allotments and set-asides; marketing quotas; subsidized loans to farmers; marketing orders; CCC nonrecourse loans; and other government purchases of excess stocks at policy-determined prices.53 Additionally, export subsidies; tariffs and quotas on imports; and subsidies for inputs, such as irrigation water, electricity, and research and development were expanded. By 1970, commodities like cotton, wheat, rice, peanuts, tobacco, wool, mohair, honey, milk and other dairy products, corn, barley, oats, rye, sorghum, soybeans, and sugar, as well as specialty crops like fruits, nuts, and vegetables, were covered by separate programs that involved both supply controls and government purchases through the CCC.54 52. The political power of the farm lobby in gaining partial exemptions for agricultural commodities from wartime price controls is discussed by Benedict (1953, 403, 423-30). Rockoff (1984, 91-92), and Higgs (1987, 208-10). Higgs points out that price control exclusions were drafted by the American Farm Bureau Federation and other farm lobbies, the USDA, and the congressional farm block. With these exclusions farm prices could rise. By August 1942 food prices had risen by more than 1 percent a month. To obtain more effective control Congress passed the Economic Stabilization Act of 1942. Even so, agricultural prices still enjoyed a privileged position. Ceilings could not be set below either the parity level or the highest level that had prevailed during the period 1 January to 15 September 1942. 53. A variety of different income subsidies were adopted in the various laws passed over the period. Although they have different names-adjustment payments, benefit payments, conservation payments, parity payments, and so forth-they accomplish the same objective, paying farmers for some portion of the difference between the target price and the market price. A summary of agricultural regulation is provided by Gardner (1981, 21). Agricultural policy and some of the politics involved are discussed by Knutson, Penn, and Boehm (1983), Gardner (1987b. 1995). and Pasour (1990). 54. Gardner (1995, 148-205) argues that the various agricultural policies that emerged after 1933 came about through the lobby action of separate interest groups, responsive politicians and

198

Gary D. Libecap

Purchases of “excess” commodities by the CCC or other federal agencies led to the accumulation of stockpiles by the federal government that were kept off the market. Some of these stocks were distributed through the school lunch program, food for the poor and elderly through food stamps and related distributions, and foreign food aid.55Other stocks were stored and, occasionally, destroyed. Stockpiles were particularly large during the 1950s and 1960s, when support prices were well above prevailing market levels. For example, the support loan rate for wheat in 1954 reached $2.24 per bushel, when the average price received by farmers was $2. 12.56The reliance of farmers on CCC purchases through nonrecourse loans and the buildup of “surplus” stockpiles in the late 1950s is revealed in table 6.1 by inventories held by the agency as a share of that year’s production. With the accumulation of politically embarrassing government-owned stocks, loan rates gradually were lowered in the 1960s. But as loan rates fell below parity prices, deficiency payments increased, and Congress turned to greater emphasis on supply controls that varied by crop and by year. Controls were strictest on tobacco production, with more limited constraints on market supplies of peanuts, certain fruits and vegetables, and dairy products. From 1950 to 1970, the most important supply management involved acreage restrictions for grains and penalties for failure to comply (usually the denial of price support benefits). Generally, no cash crops were allowed on reserved acreage. To initiate supply controls each year, the secretary of agriculture set the allowable acreage for each crop and CCC loan rates based on expected prices, anticipated supply and demand conditions, and accumulated stocks. Production quotas as a share of base acreage for individual farmers were determined by the Agricultural Stabilization and Conservation Service (ASCS) through regional, state, and county offices.57In the 1960s, annual acreage reductions of 10-20 percent of base acreage were conditions for receiving federal price support benefits for farmers producing wheat, corn, barley, sorghum, and other crops. Even so, supply controls had little chance for success as the primary tool of agricultural regulation to raise commodity prices. As allowable acreage was reduced, farmers had incentives to farm remaining land more intensively, and bureaucrats, the congressionalcommittee structure, and a constitutional provision that gives undue representation to rural, agricultural states. 55. Indeed, Bruce Gardner argues that a motivation for enactment of food stamp and trade legislation was the desire to dispose of agricultural commodities in a politically acceptable way. In the case of the Agricultural Trade Development and AssistanceAct of 1954, he comments: “The political push behind P.L. 480 arises from its role, not in fighting hunger, but in the demand for U.S. farm products” (1981, 21). 56. Loan rate from USDA Agricultural Srurisrics (1952, 680) for 1933-50 and USDA, Wheat Yearbook (1996, app. table 9) for 1951-70. Average price received by fanners from USDA, Agricultural Statistics (1967, 1-2). 57. Discussion of supply management programs is provided by Gardner (1981, 26-34) and Pasour (1990,51,81-117).

199

Federal Government Regulation of Agriculture, 1884-1 970

Table 6.1

Inventories Held by the Commodity Credit Corporation (% of annual production) Year

Cotton

Corn

Wheat

1956 1957 1958 1959 1960

51 46 9 7 35

20 24 28 25 27

95 86 57 103 88

Sources: Inventory data are from USDA (1956, 3; 1957, 2, 3; 1958, 3, 4; 1959, 3, 4; 1960, 30). Annual production data are from U.S. Department of Commerce (1975,510-11).

there were few penalties for doing so. If commodity prices fell, diversion (or conservation) payments for reserved acreage would be increased by the secretary of agriculture as additional land was set aside. Further, benefit payments for differences between market prices and parity prices would be raised, following policies initiated in the Agricultural Adjustment Acts. Additionally, there was long-standing political opposition among farmers to tight production controls, dating from the beginning of the New Deal, and neither Congress nor the USDA was willing to confront that o p p o ~ i t i o nFinally, .~~ CCC loan rates were adjusted upward in the 1950s and 1970s, providing a rising price floor for farmers, and the provisions of the Agricultural Adjustment Act of 1938, which mandated CCC loans as market prices fell, provided farmers with guaranteed purchases. Gale Johnson (1973,22-31) concluded that through the early 1970s federal programs substantially increased the cost of regulated commodities to American consumers, and he argued that the programs achieved parity income targets for virtually every year since 1940.59Moreover, he determined that most of the success was through demand enhancement, rather than supply controls, which he concluded had not substantially reduced production.60The importance of government purchases is noted by Bruce Gardner: “The primary policy tool has been to have the government itself provide the demand that is necessary in order to achieve increases in farm prices. This approach in the past has 58. Benedict (1953,304,311,313). Perkins (1969, 103, 140). and Hoffman andLibecap (1995) discuss early farmer opposition to tighter production rules under the AAA. Schultz (1949, 143) argues that although corn acreage fell in the late 1930s, output grew. Only in tobacco and peanuts were output controls successful according to Gardner (1987b, 21). 59. Johnson (1973, 26) observed that there had been little political pressure to see whether parity income targets had been met since 1940, although he was confident that they had. Clear documentation of success would have caused a political backlash or justified cuts in programs. 60. Johnson (1973.3-4.23-41) argued that the reduction in farm output that can be attributed to supply management did not exceed 2.0-2.6 percent of total farm output. Most of the 55 to 60 million acres set aside or otherwise diverted from production either was land that would not have been planted or was poor-quality land. Further, farmers adjusted by adding other inputs to remaining land to raise production. Land diversion contributed to increased use of new seeds, fertilizers, and capital to increase production.

200

Gary D. Libecap

caused prices to be high enough to bring forth costly surpluses of production’’ (1981, 21). 6.2 Categorization and Analysis of Agricultural Legislation: The Impact of the New Deal Although the policy summary outlined above suggests a regime switch regarding agricultural regulation brought about by New Deal legislation, it is possible to address the issue more concretely. Following the industrial organization literature (Breyer 1982; Shepherd 1990; and others) it is possible to classify government regulatoq policy in commodity markets as providing: Public goods: Government provides consumer and producer information about products and production conditions, product quality assurance, industry standards, disease and insect control, and other actions that promote gains from trade.6* Transfers: Government provides grants, subsidized loans and insurance, research expenditures, infrastructure, labor and other input subsidies (soil improvement, windbreaks, watershed, irrigation), and marketing support. Economic regulation: Government intervenes directly into specific commodity markets to raise commodity prices and incomes of farmers through: 1. Demand enhancement: Government purchases agricultural commodities with the primary objective of increasing demand. 11. Supply control: Government limits total market production and supplies through input controls, marketing quotas, production allotments, entry limits (licensing, trade barriers), and regulation of shippers, handlers, and other middlemen.

These categories are used to classify 653 agricultural laws enacted by Congress between 1884 and 1970 to see more clearly the regime switch of the New Deal after 1932. Additionally, by classifying 586 laws passed between 1933 and 1970, it is possible to determine how major New Deal statutes or executive orders were linked to post-New Deal regulatory policies.62A finding of direct ties between legislation in both periods through amendments, modifications, and extended authorizations will more precisely reveal the influence of the New Deal on the pattern and scope of agricultural regulation in the United States. Udell (1971, 1972) provides a compendium of laws relating to agricultural 61. Another category would be to address natural monopoly markets. In agriculture this might apply to railroad rate regulation through the Interstate Commerce Commission. Railroad rate regulation broadly affected many sectors of the economy and is not included in the agricultural laws analyzed in this paper. 62. New Deal legislation considered includes the Agricultural Adjustment Acts of 1933 and 1938, the Taylor Grazing Act of 1934, the CCC, the Bankhead Jones Acts of 1935 and 1937, the Soil Conservation and Domestic Allotment Act of 1936, the Agricultural Marketing Act of 1937, the Sugar Act of 1937, and the Federal Crop Insurance Act of 1938.

201

Federal Government Regulation of Agriculture, 1884-1970

regulation between 1884 and 1972. Identification of the major objective of each law allows it to be placed into one of the three categories: public goods, transfers, and economic regulation. For example, P.L. 41, passed in May 1884, created the Bureau of Animal Industry “to prevent the exportation of diseased cattle, and to provide means for the suppression and extirpation of pleuropneumonia and other contagious diseases among domestic animals.” The primary goal of the law is to provide the public good of livestock disease control.63 For another example, P.L. 7, passed in January 1928, authorized $500,000 to fund actions by USDA county extension agents to rehabilitate certain farmlands damaged by floods. Because this law used government funds to improve private land, it is classified as providing an income transfer to farmers. Similarly, P.L. 10, enacted in May 1933, was the Agricultural Adjustment Act, and as a major piece of New Deal legislation, it had many complex provisions. Even so, the primary intent of the law was to control the supply of farm commodities through “the reduction in the acreage or reduction in the production for market, or both, of any basic agricultural commodity” (Pt. 2, Sec. 8), and hence it is classified as economic regulation of supply. Another example of economic regulation was P.L. 480, enacted in July 1954, to “increase the consumption of United States agricultural commodities in foreign countries,” largely through the granting or subsidized sale of surplus agricultural commodities held by the CCC for export. As such, this law is classified as economic regulation to promote demand. In addition to placing regulatory legislation enacted between 1884 and 1970 into one of three categories, those laws passed between 1940 and 1970 (the post-New Deal period) were classified as to whether they were directly linked to the major New Deal statutes listed above. In most cases the linkage is straightforward because much of the legislation enacted after 1939 makes explicit reference to the New Deal statutes being amended or extended. Table 6A. 1 illustrates the linkage identified in the analysis by examining agricultural regulation legislation passed in 1962. In 1962, 25 laws were passed, and of those, 18 were directly related to the Soil Conservation and Domestic Allotment Act of 1936 or the Agricultural Adjustment Act of 1938 through adjustments in parity prices and acreage allotments for various commodities, setting new times for marketing quota referendums, extending other provisions to stabilize prices, and broadening soil conservation provisions. The results of the classification of the agricultural regulation legislation into the three categories of public goods, transfers, and economic regulation and the linking of laws passed after 1939 to New Deal legislation are summarized in table 6.2. The table lists the number of laws by category for the entire period 1884-1970, the pre-New Deal years 1884-1932, the New Deal years 1933-39, and the post-New Deal years 1940-70. There were 653 regulatory laws enacted in the 87 years between 1884 and 1970. Approximately 10 percent were 63. The public good is due to the externalities associated with disease elimination.

202

Gary D. Libecap Agricultural Regulation by Category

Table 6.2

Economic Regulation Total Laws Passed (1 )

Public Goods Laws (2)

1884-1970

653

1884-1932

67 (10%) 130 (20%) 456 (70%)

115 (18%) 25 (37%) 19 (15%) 71 (16%)

Time Period

1933-39 1940-70

Demand Enhancement (3)

Supply Control (4)

Transfer Laws

69 (11%) 1

320 (49%) 15 (22%) 74 (57%) 23 1 (51%)

140 (21%) 26 (39%) 25 (19%) 89 (20%)

(1%)

8 (6%) 60 (13%)

(5)

Other (6)

New Deal-Based Legislation* (7)

9 (1%)

392 (60%)

4 (3%) 5 (1%)

81 (62%) 311 (68%)

Nore: Percentages (in parentheses) in col. (1) are of the total number of laws for the entire 18841970 period. Percentages in cols. (2) to (7) are of the total number of laws for the applicable subperiod. “Linked to major New Deal legislation as defined in the text and table 6A.1. The percentages in this column will not add to those to the left.

passed during the first 49 years through 1932. There were almost double that number enacted during the New Deal, and another 456 were passed after 1939. Not only did the number of regulatory statutes jump after 1932, but the emphasis of regulatory policy shifted. In the early period most regulation focused on the provision of public goods, such as plant and animal disease control and quality standards, and such legislation accounted for 37 percent of the total statutes enacted. Only subsidized loans to cooperatives and farmers, disaster relief, and other transfers were more common, with 39 percent of the legislation passed during the period. Economic regulation was comparatively less important, comprising about 23 percent of the laws passed. With the advent of the New Deal farm program in 1933, the proportions changed and remained different from what existed previously. During the New Deal and after, the relative emphasis on public goods provision declined, with 15 and 16 percent of the laws passed aimed at public goods in the two periods. With the New Deal the number of laws aimed at providing economic regulation and transfers increased. But transfer legislation became a comparatively less important part of total legislation, with only 19 percent of the statutes passed during the New Deal and 20 percent of the total laws passed from 1940 through 1970. The major growth was in economic regulation through supply control and demand enhancement as part of significant new intervention into commodity markets to raise prices to parity levels as defined in the Agricultural Adjustment Act of 1933. During the New Deal, 63 percent of all laws passed were directed at economic regulation, and after 1939, 64 percent were for supply control and demand enhancement. The importance of key New Deal laws in agricultural policy after 1939 is shown in column (7), where 68 percent of all laws passed

203

Federal Government Regulation of Agriculture, 1884-1970

Fig. 6.1 Agricultural regulation legislation, 1884-1970 Sources: Udell(1971, 1972).

between 1940 and 1970 were directly linked to New Deal economic regulation and transfer legislation. The impact of the New Deal can also be seen in figure 6.1, which shows annual legislation from 1884 and 1970. After 1939 the light-shaded bars show the total number of laws passed each year and the dark bars show annual legislation after removing New Deal-based statutes. Considering first the general pattern, the increase in legislative activity after 1932 is clearly demonstrated. The extraordinary legislative activity of the New Deal in establishing new regulatory policies, particularly in supply control and demand enhancement, is clear. Moreover, after 1939, the light-shaded bars show that the annual number of laws passed continued to be well above what existed prior to the New Deal. Robert Higgs argues that the growth of government occurs in stages, with crises leading to a flurry of new legislative activity that remains after the crisis ends (1987,57). The legislative pattern described in the figure supports this observation. The separation of the bars into light and dark divisions after 1939 shows the impact of the New Deal on subsequent regulatory policy. The dark bars from 1940 through 1970 show only public goods laws and non-New Deal transfers and economic regulation legislation.w Comparing the light and dark bars indicates the importance of the New Deal legacy for post-New Deal regulatory policy. Indeed, the period after 1940 without New Deal-based legislation appears quite similar to what existed from 1884 through 1932. 64. Examples include P.L. 645, the Potato Insect Control Act of 1948; P.L. 237, enacted 8 August 1953, authorizing the recruitment of Mexican labor; and P.L. 61, enacted 8 July 1963, to amend the Packers and Stockyard Act of 1921.

204

Gary D. Libecap

Table 6.3

Mean Percentage of Annual Laws Enacted by Category and Tests for Differences in the Means &Testsof Differences in the Means between 1884-1932 and 1940-70 for:

Mean Percentage of: Time Period

1884-1932 1933-39 1940-70 1940-70 (no New Deal)

Public Goods

Transfers

Economic Regulation

Public Goods

Transfers

Economic Regulation

46 14 16

31 21 20

23 63 62

3.52*

1.51

6.03*

49

26

20

0.27

0.61

0.43

*Significant at the 99 percent level

Table 6.3 provides a more explicit comparison of regulatory policy before and after the New Deal if New Deal-based legislation is subtracted. The table lists the mean annual percentage of statutes enacted providing public goods, economic regulation, and transfers for 1884-1932, 1933-39, and 1940-70, with and without New Deal-based legislation, as well as r-tests for the equivalence of the means.65As indicated in the table, the New Deal and post-New Deal periods are not different statistically from one another in terms of the aims of regulatory legislation. The periods 1933-39 and 1940-70 have comparable mean percentages of laws enacted annually for the provision of public goods, income transfers, and economic regulation, respectively. By contrast, however, the nature of agricultural regulation changed importantly after the New Deal from what had existed before 1933.As shown, the mean percentages of statutes providing public goods, income transfers, and economic regulation adopted annually during 1884-1932 are quite different from what was enacted either during the New Deal or after. Indeed, the difference in the mean percentages for the three types of laws between the pre-New Deal period, 1884-1932, and the post-New Deal period, 1940-1970, are statistically different at conventional levels, underscoring the shift in the regulatory regime. Subtracting New Deal-based legislation from the regulatory laws passed annually after 1939, however, changes the mean percentages for each regulatory category and the overall regulatory picture sharply. Absent the New Deal, the regulatory focus of the pre-and post-New Deal periods is close to the same, with the annual mean percentages for the three categories of regulation similar and statistically not different from one another. As a result, both the pattern shown in figure 6.1 and the statistical tests in table 6.3 support the hypothesis that had the Great Depression not occurred 65. The calculation of the means includes only those years in which legislation was enacted. Unequal variances across the periods is assumed.

205

Federal Government Regulation of Agriculture, 1884-1970

and New Deal regulatory policies not been enacted, the pattern of regulation through 1932 would have continued in the period 1940-70.@jThese data underscore the importance of the New Deal in changing the nature of government regulation of agriculture in the United States. 6.3 Measures of the Impact of New Deal Agricultural Policies

6.3.1 The Impact on Regulatory Budget Expenditures The data in tables 6.2 and 6.3 reveal that the major shift in regulation due to major New Deal legislation was in the growth of economic regulation through supply control and demand enhancement. It is possible to examine the impact of New Deal regulatory legislation on budget allocations for supply control and demand enhancement using data collected from U.S. Treasury Department, Digest ofAppropriations (1905-21) and U.S. Budget ( 1922-70).67From 1905 through 1949 actual expenditures by legislative act, agency, and program can be identified.68For the 20 years, 1950-70, budget data are grouped more broadly by expenditure categories, such as the CCC (part of the USDA budget from 1948 through 1970) and Soil Bank programs (1956-58).69 These expenditures can be classified by primary function and assigned to either demand 66. Another possibility for consideration is what would have happened to agricultural regulation after 1939 if no New Deal farm program had been enacted in response to the fall in farm incomes. The only way of addressing that scenario is to look at two earlier periods. Farm incomes fell sharply in the mid-1890s and in 1921 with the fall in commodity prices. But in the absence of major federal government intervention during both of those periods, no institutional structure was developed nor precedents established for the continuation of federal government regulation after the crisis passed on the scale that actually occurred after 1939. These outcomes suggest that the post-1939 period would have looked more like the regulatory pattern observed in the 1920s had there been no large-scale New Deal farm program. 67. Deering (1945) provides the organization of USDA and role of various agencies, branches, and bureaus within the department, as well as an appendix with budget appropriations by program for 1932-46. For budget data from the Digest ofAppropriations and the U.S.Budget, actual appropriations are used through 1920, actual expenditures are reported for 1921 and 1923-44, estimated expenditures are reported for 1922, and actual appropriations are shown for 1945-70. 68. E.g., expenditures are listed for the U.S. Food Administration (1918); the Bureau of Markets (1918-20); the Bureau of Markets within the Bureau of Agricultural Economics (1921-37) and within the Agricultural Marketing Service (1938-41); the Agricultural Marketing Administration (1942); the Agricultural Marketing Service (1943-44); the Production and Marketing Service (1945552). and the Agricultural Marketing Service (1952-70); the Filled Milk Act (1923); the Tobacco Inspection Act (1937-49); cotton regulation; wool regulation; the AAA (1933-44, and part of the Production and Marketing Administration for 1945-52); the Sugar Act of 1937; the Soil Conservation and Domestic Allotment Act; the Federal Farm Board (1929-33); and the CCC (not part of the USDA budget from 1934 through 1947). 69. Other budget categories include the ASCS (1961-70); Consumer and Marketing Services (after 1964); and the Production and Marketing Administration (1950-52), which includes all expenditures for acreage allotments and marketing quotas, the Sugar Act program, and AAA expenses. After 1952 the Production and Marketing Administration is divided into the Agricultural Marketing Service (which includes the school lunch program), the Agricultural Conservation Program Service (acreage reduction programs under the Soil Conservation and Domestic Allotment Act), and the Commodity Stabilization Service (AAA, Sugar Act programs).

206

Gary D. Libecap

enhancement or supply control. Table 6A.2 summarizes the programs included in each of the two categories. Figure 6.2 reveals the pattern of economic regulation between 1905 and 1970 through budget expenditures for supply control and demand enhancement regulation from 1905 to 1970.’O As indicated, there were virtually no expenditures for supply control until 1933, when they jumped with the advent of New Deal programs. Outlays remained high through the mid-1940s before declining for the next 10 years. But they did not return to their pre-1933 level. By the late 1950s, as stocks of wheat, corn, and cotton held by the CCC rose (table 6. l), expenses for supply control increased once again. Demand enhancement expenditures also were comparatively small prior to the New Deal. The early increases reflected purchases by the U.S. Food Administration during World War I and by the Federal Farm Board in 1930 and 1931. Most of the expenditures after 1933 were through the CCC. There were periodic drops in CCC expenditures that occurred during years when the market price exceeded the loan price, causing farmers to pay off their CCC loans in order to sell their collateral commodities, and in some years, CCC receipts exceeded expenditures. The budget expenditure pattern revealed in figure 6.2 corresponds with the legislative data shown in table 6.3 and figure 6.1. The statutes that implemented the economic regulation of agriculture came with the New Deal, and those same statutes provided the mandate for regulatory policies through 1970. Correspondingly, budget expenditures for economic regulation were not important until after 1933. The legacy of the New Deal in permanently inserting the federal government into commodity markets to reduce supplies and to increase demand is unmistakable. Additionally, figure 6.2 shows that, over time, the federal government shifted the relative emphasis of budget expenditures from supply control to demand enhancement, although the two policies remained intricately intertwined. Part of the reason for the shift was farmer political opposition to acreage and marketing controls. Moreover, supply controls alone were ineffective as farmers evaded acreage restrictions by substituting capital and labor for land and by removing only the least productive land from production. As output grew across most farm commodities, Congress was faced with enacting legislation authorizing the USDA to set ever stricter acreage and production limits if target prices were to be achieved. Government purchases to supplement supply management were an attractive solution. Resorting to general tax revenues to pay for agricultural surpluses reduced the pressure on production quotas and acreage allotments. In addition, demand enhancement purchases by the CCC could be directed to politically popular programs such as foreign food aid, 70. The data are shown in real terms (constant 1967 dollars using the All Item Index in U.S. Department of Commerce 1975, 199).

207

Federal Government Regulation of Agriculture, 1884-1970

7

---

Real Demand EnhancementAgricultural Regulation Real Supply Control Agricultural Regulation

Fig. 6.2 Agricultural regulation as demand enhancement or supply control, 1905-70 (billions of dollars)

Sources: U S . Department of the Treasury, Digest of Appropriations (various years) and U.S. Budget (various years).

trade promotion, disaster relief, school lunches, and food stamps. These programs in turn developed their own constituencies, who could be mobilized to join farm groups in promoting commodity purchases by the government. With agricultural surpluses directed to other worthy causes, the actions of the government could be seen less as efforts to raise farm prices and incomes. Moreover, the availability of CCC purchases as a last resort, together with various payments to farmers to cover the difference between parity or target prices and market prices, removed the incentive of farmers to be very concerned with adhering to supply management guideline^.^' 7 I . The nature of the payment varied by agricultural statute. Parity payments were authorized by the Agricultural Adjustment Act of 1938 to compensate farmers for the difference between market prices and 75 percent of parity (the percentage also varied according to subsequent legislation). Deficiency payments to cover the difference between market prices and “target” prices were introduced in 1973, where the target price was to be more flexible than the earlier parity price. The incentives put in place by the existence of demand enhancement programs helps explain why D. Gale Johnson (1973, 3-4, 23-41) found so little effective supply management. As stockpiles grew in the 1950s and 1970s, the federal government attempted new supply controls with little success. For example, the Soil Bank law was enacted in 1956 and the payments-in-kind (PIK) program was implemented in 1983.

208

Gary D. Libecap

6.3.2 The Impact on U.S. Department of Agriculture Staffing and Budget Another measure of the impact of the New Deal is its effect on the bureaucracy charged with implementing the expanded regulatory mandate. The USDA acquired new agencies, including among others the AAA, the CCC (after 1947), and the ASCS, as well as additional staffing and budget appropriations after 1932. Moreover, USDA personnel played an active role in drafting additional legislation to extend the department’s regulatory functions. Two measures are provided of the impact on the USDA of New Deal agricultural programs. One measure is agency staffing relative to the number of farms in the United States from 1910 to 1970.72Staffing is an indication of agency growth due to new mandates and programs, and the data are presented with respect to a measure of the size of the principal client group or department constituency, the number of farms in the United States. Figure 6.3 reveals the ratio of staffing to farms over the period. The ratio was fairly flat, rising slowly from 0.002 in 1910 through 0.004 in 1932. The ratio grew rapidly after the advent of the New Deal and the expansion of agricultural programs under the Agricultural Adjustment Act of 1933, jumping from 0.005 in 1933 to 0.012 in 1935. Through the next 20 years, USDA staffing relative to the number of farms remained in the range of 0.012-0.016. Importantly, the ratio never returned to pre-New Deal levels, even after the depression ended.73Further, beginning in 1955 as the number of farms declined while staffing levels continued to grow, the ratio rose from 0.018 in 1955 to a peak of 0.045 in 1969. Another measure of the impact of New Deal programs on the USDA relative other government agencies is the share of USDA expenditures (including the CCC) in total federal government civilian expenditures from 1921 through 1970.74The USDA share of the civilian federal budget is shown in figure 6.4. Throughout the 192Os, the USDA accounted for approximately 5 percent of total federal civilian expenditures. During the New Deal period, USDA plus CCC expenditures grew as a share of federal civilian expenditures, reaching 18 percent in 1935 and 19 percent in 1942. The USDA share of civilian expen72. USDA staffing is from the Statistical Abstract of the United States ( U S . Department of Commerce and Labor 1908-1912; U.S. Department of Commerce 1913-71). Data on the number of farms are from U.S. Department of Commerce (1975,457). 73. The number of farms peaked in 1935 at 6,812,000, whereas USDA staffing peaked in 1969 _ . at 125,034, 74. ExDenditures for total militaw functions are subtracted from total federal expenditures by year from 1921 to 1933. From 193ito 1941, War Department, Panama Canal, Veterans Administration, and total national defense expenditures are subtracted. From 1942 to 1949, war activities, national defense, and veterans’ services and benefits are subtracted. From 1950to 1970, the budget categories subtracted from total expenditures or outlays include military service and Department of Defense. USDA and federal expenditure data are from U.S. Department of the Treasury, Annual Reporf (1921-70). CCC data prior to transfer to the USDA are from U.S. Department of the Treasury, U.S. Budget (1936-49; for actual CCC expenditures during 1934-47).

Federal Government Regulation of Agriculture, 1884-1970

209 0 05

Fig. 6.3 Ratio of number of USDA employees to number of farms, 1910-70 Sources: Number of farms from U.S. Department of Commerce (1975,457). Number of USDA employees from U.S. Department of Commerce and Labor (1880-1912) and U.S. Department of Commerce (1913-71).

ditures declined in the early post-World War I1 period when commodity prices rose and farmers paid off their loans. During the 1950s and 1960s, however, the USDA share rebounded, reaching 13 percent in 1955 and 1956. The department's portion of federal expenditures did not return to predepression levels through 1970, mirroring the agency growth suggested by the staffing data shown in figure 6.3.75The expansion of the USDA, its regulatory mission, and its constituent ties are part of the regime shift that occurred after 1933. 6.3.3 The Impact on Domestic Wheat Prices Another way of gauging the effect of New Deal agricultural programs is to examine the impact on prices for a major commodity, wheat. Wheat was one of the most important U S . agricultural commodities. In 1935, it ranked third, after corn and cotton, in terms of the value of production. In 1949, wheat was the most valuable U.S. crop. In that year, wheat producers received 29 percent of all CCC loans.76Moreover, wheat producers were numerous, well organized, and politically influential, with wheat farmers in most western and northern states. Accordingly, it is no surprise that wheat was one of the initial 75. A t-test of the difference in the mean USDA expenditure percentages for 1921-32 and 1933-70 shows the means to be significantly different with a t-statistic of 6.16. The USDA share of federal expenditures fluctuates after 1970, dropping to 4 percent in 1975 and then rising to 8 percent by 1983. The fall in the USDA share largely reflects the rapid growth in overall nondefense federal expenditures after 1970. 76. Production data are from USDA, Agricultural Staristics (1937, 378-79; 1949, 592-94). CCC loan data are from USDA, Agricultural Statistics (1950, 706-7).

210

Gary D. Libecap

02 0 18 0 16

0 14 0 12 01

0 08

0 06 0 04

- m a + m N y u zN $ O * * O * * m o m mm a m z z z z z z z ? z m z m m m m z m ? ? ? ~ ? m

a

c

m

r

o

m

+

m

r

n

m

c

m

-

o

m

c

o

Fig. 6.4 USDA expenditures as a share of total federal civilian expenditures Source: U.S. Department of Treasury, Annual Report (1921-70). Note: USDA data include CCC expenditures.

basic commodities in the proposed McNary-Haugen legislation and the Agricultural Adjustment Act of 1933 and remained one of the key commodities targeted by New Deal farm programs. The objective of those farm programs was to raise commodity prices and farm incomes to parity with the levels of 1910-14. This section examines the success of those efforts with wheat by examining the path of domestic wheat prices relative to world prices between 1900 and 1970. To make these comparisons, domestic and world wheat price series were constructed from 1900 to 1970 and are reported in table 6A.3. From 1900 to 1932, the domestic wheat price is the “average price received by farmers” as reported in USDA, Agricultural Statistics (1967, 1). It is an average because it involves more than one variety and covers the entire crop year. From 1933 to 1970, the domestic wheat price is more complex because it includes either the “average price received by farmers” or the CCC loan rate, whichever was higher, plus various price support payments.77These price supports vary with legislation across the years, and only payments that can be tied to a bushel of wheat are included.78 77. If the CCC loan rate exceeded the average price received, the former would dominate because farmers would default on their CCC loans. For discussion of wheat price support payments, see Evans (1984), Hadwiger (1970), Heid (1980), Rasmussen and Baker (1979), and USDA, Economic Research Service (1996). 78. Some payments were not included because of missing years or because of uncertainty as to the effect they had on production choices. An example is storage payments. Producers participating in CCC loans and storing their crops were paid per bushel storage payments while their wheat

211

Federal Government Regulation of Agriculture, 1884-1970

For 1933 to 1935, the domestic wheat price series includes the average price received plus AAA price supports for those farmers who maintained their acreage within the authorized allotment^.'^ These payments were disrupted by the Supreme Court’s ruling in 1936 invalidating the Agricultural Adjustment Act of 1933. In 1936 and 1937, the average price received is used, but with the enactment of the Soil Conservation and Domestic Allotment Act and the Agricultural Adjustment Act of 1938, new per bushel price supports were available from 1938 through 1943 through parity and soil conservation payments.*O With high wheat prices between 1943 and 1962, there were no parity payments, and the price used is the higher of the average price received by farmers or the CCC loan rate. In 1963, new price support payments were inaugurated and added to the average price for wheat (see Evans 1984,19; Hadwiger 1970, 257; USDA, ASCS 1964, 1968). From 1964 through 1970, the wheat certificate program began whereby farmers received full price supports only for wheat covered by the certificates. These wheat-marketing certificates provided support payments, but they were not issued for all of the allotted acreage for wheat.81Along with Soil Bank payments, wheat certificates were an effort to reduce wheat production in the United States. The constructed domestic wheat price from 1964 through 1970 includes only wheat covered by certificates. Even so, the constructed domestic wheat price series likely is an understatement of the per bushel actual price received by farmers. Some subsidies were on a per acre basis or were lump-sum income payments with no obvious conversion to amounts per bushel of wheat. Moreover, there are so many subsidies, many of which were short lived, that some may have been missed.82 World wheat prices are represented by Australian wheat prices. Australia was a major wheat exporter, and the Australian wheat price series was constructed from three price series, for 1900-1931 from Shergold (1987,223), for 1932-38 from the Commodity Research Bureau (1939), and for 1939-70 from the Australian Wheat Board (1990). As necessary, Australian prices were converted to U.S. dollars using the exchange rate provided in Pope (1987, 244was held off the market. These payments are not available for all of the years in the study. Other payments not included include per bushel payments under various disaster relief programs, crop diversion programs, and per bushel export subsidies (except for 1964 and 1965 when export marketing certificates were widely dispersed). 79. These payments were funded by the processing tax under the Agricultural Adjustment Act of 1933. See USDA, Agricultural Statistics (1936, 361), Davis (1935), and Hadwiger (1970). 80. The parity payment is the difference between the average price received and the parity price, which in turn is based on a percentage of the prices received between 1910 and 1914 (Davis 1935). Parity prices were adjusted by the Agricultural Adjustment Act of 1938 and later legislation. The formulation of the parity price is described in USDA, ASCS (1985). Conservation and parity payments are from Davis (1942,366). 8 1. Wheat-marketing certificates covered approximately 40-45 percent of allotted acreage. For discussion, see USDA, ASCS (1964, 1968). 82. Green (1990) constructs a similar domestic wheat price series for 1961-90, and his series is comparable to the one presented in the appendix. See the notes to table 6A.3 for discussion of what price supports are included in constructing the domestic wheat price series.

212

Gary D. Libecap

2.5

2

1.5

1

0.5

0

Fig. 6.5 Ratio of U.S. supported wheat price to world wheat price

Source: See text and appendix. Note: World wheat price is the Australian export price.

45).83 The unsupported (prior to 1933) and supported (after 1933) U S . wheat price series is compared to the Australian wheat export price series. This comparison indicates the degree to which New Deal agricultural programs were successful in pushing domestic wheat prices above world prices. Figure 6.5plots the ratio of U.S. domestic (supported and unsupported) and Australian export wheat prices from 1900 to 1970. The figure reveals that during the pre-New Deal period, U.S. domestic wheat prices closely tracked world prices (except during World War I). From 1933 through 1970, other than the immediate post-World War I1 period when there was considerable world price volatility, the domestic price of wheat was held consistently above the world The mean difference between domestic and Australian wheat prices during 1900-1932 was $0.15 and during 1933-70 was $0.36; these are significantly 83. The series was converted from dollars per metric ton to dollars per bushel by multiplying by 0.0272. The constructed series was compared with selected Australian wheat export prices from other sources for verification: Commodity Research Bureau (1939, 114) and International Monetary Fund (1979,76-77; 1995, 180-81). The Australian series tracks wheat prices for Canada, which is a similar wheat export country. Price series for the two countries were assembled for 1950-68. 84. Between 1949 and 1953 there was a concerted effort to gain compliance with the International Wheat Agreement of 1933 to set world wheat prices. After 1953, the United Kingdom withdrew from the agreement, returning in 1959, but the United States withdrew in 1967, and for practical purposes the wheat agreement expired. The effort to fix high export prices between 1949 and 1953 may partly explain the spike in world prices during the postwar period. For discussion, see Hadwiger (1970, 70-72).

213

Federal Government Regulation of Agriculture, 1884-1 970

different from one another, with a t-statistic of 1.92. Moreover, the variation in domestic wheat prices declined after 1932, while world wheat prices became more volatile. The coefficient of variation for domestic wheat prices was 0.43 for 1900-1932 and 0.32 for 1933-70; whereas, for world prices, the coefficient of variation was 0.33 for 1900-1932 and 0.44 for 1933-70. The data indicate that the wheat commodity programs put into place by New Deal agricultural regulation generally succeeded in holding wheat prices above the world price and in providing more stability to domestic wheat prices.85 6.4

The Legacy of the Great Depression for Agricultural Regulation

This paper has measured the direct contribution of the Great Depression and associated New Deal policy to the development of agricultural regulation in the United States. By analyzing the approximately 653 laws enacted by Congress between 1884 and 1970 regarding agriculture and by linking over two-thirds of the 456 laws enacted between 1940 and 1970 to specific New Deal legislation, I have demonstrated how the New Deal affected the growth and nature of current-day agricultural regulation. Since 1933, Congress has enacted laws to modify and extend New Deal-based regulations to virtually every commodity and every agricultural market, making agriculture one of the most regulated sectors of the American economy. This legislative pattern is reflected in budget expenditure data reported from 1905 through 1970. Budget expenditures classified as demand enhancement or supply control were minimal until 1933. After that year, however, expenditures grew in real terms through 1970, with the greatest early growth in supply control efforts and later with more funds for government purchases to enhance commodity demand. In addition, the paper has examined the counterfactual of what agricultural regulation might have looked like had the Great Depression and the New Deal not occurred. Subtracting those laws passed after 1939 that were directly linked to key New Deal statutes leaves a pattern of regulation that is similar to what had existed prior to 1933-agricultural laws that primarily provided public goods and limited transfers to farmers. Prior to the New Deal, there was comparatively little emphasis on direct government intervention to fix prices and incomes through the economic regulation of supply and demand. This pattern of early regulation grew out of an apparent prevailing belief in small government and a corresponding lack of institutions and political constituents for the extensive regulation of commodity markets. Federal government intervention during World War I may have demonstrated its potential power to direct economic benefits to particular interest groups, but political resistance in Congress and from the president to the McNary-Haugen legislation of the 1920s indi85. After 1970 it becomes more difficult to compare U.S. domestic wheat prices with world prices because farm programs such as PIK, implemented in 1983, and other acreage-based or income payments are less easily converted to a per bushel figure.

214

Gary D. Libecap

cates that the time had not yet arrived for more substantial government involvement in agriculture. It took the crisis of the Great Depression to justify the creation of durable institutions for economic regulation.86 The Great Depression swept away much of the political opposition to direct federal government intervention in agricultural markets that had existed with the McNary-Haugen bills. With the AgriculturalAdjustment Act of 1933, both the Congress and the president supported government regulation of agriculture that went beyond what had been proposed with McNary-Haugen, allowing the federal government “to relieve the existing national economic emergency by increasing agricultural purchasing power” (Preamble of the Agricultural Adjustment Act, 12 May 1933). The paper has also presented data on staffing and budget for the USDA, which show that the department grew relative to the size of its constituency and to the size of federal domestic expenditures as its regulatory mission increased. These data reflect both the institutional regime shift after 1932 toward greater government intervention into agricultural markets and the results of that shift. That is, as the USDA was assigned a larger regulatory mandate and new administrative divisions were created, its role in the further expansion of regulation was enhanced. It collected the data on commodity demand and supply that were critical in administering (or justifying new) regulations. Further, it provided a broader array of services to farmers to promote constituent relations (extension, marketing, export promotion, and research and development). This relationship no doubt helped farm groups mobilize to protect and expand the benefits of agricultural regulation, even as the number of farmers fell.87 Finally, the paper has presented data on domestic wheat prices and compared them to world prices from 1900 to 1970. This comparison is more difficult than one might have expected because the many subsidy programs that have been put into place since 1933 have distorted domestic prices so that it is difficult to know just what a true domestic price is, let alone to calculate one that includes the various price supports. Similarly, World Wars I and I1 and the International Wheat Agreement inaugurated in 1933 have disrupted world wheat prices. Nevertheless, the data indicate that New Deal agricultural programs have succeeded generally in maintaining domestic prices above world prices and in making them more stable. Given how tenaciously farmers have worked to maintain and expand commodity programs it could hardly be otherwise. The role of the Great Depression in expanding agricultural regulation was to provide a crisis of sufficient magnitude to politically justify the establish86. This argument is made by Robert Higgs (1987.20-27, 192). 87. The political economy of agricultural regulation, including the roles of farm groups, politicians, and the USDA, requires thorough analysis. Parts of the farm program have been examined; e.g., see Gardner (1987a) for analysis of the characteristics of commodities receiving subsidies and Gardner (1995, 185-238) for discussion of special interest groups and PAC contributions in the late 1980s.

215

Federal Government Regulation of Agriculture, 1884-1 970

ment of regulatory policies on a scale that had not previously existed. The programs in agriculture were so extensive and the private benefits so great that they were not dismantled at the end of the depression. Rather, a coalition of farm groups, USDA officials, and congressional representatives from farm regions were able to expand farm programs and the benefits they provided through incremental adjustments and amendments to the original enabling legislation. This coalition was supported by logrolling with congressional representatives of urban areas, whose constituents benefited from food stamp, school lunch, and other agricultural programs. The pattern of agricultural regulation in the United States in the twentieth century appears to be remarkably similar to that found in other Western industrialized countries, such as Australia, Britain, Canada, France, Italy, and Germany. Agriculture has always been subject to sharp cyclical swings in commodity prices and incomes due to low income elasticity and low price elasticity of demand for farm products. With these conditions, demand does not rise markedly with upswings in the business cycle and prices fall as supplies increase (the problem of “chronic surpluses”). Across countries, relatively limited government regulation of agricultural markets occurred following World War I. The deterioration in many commodity prices after 1921 brought agricultural unrest and political pressure for government assistance. In Australia, Canada, England, France, Germany, and Italy the general political response in the 1920s was to limit import competition through higher tariffs and quotas. Germany, for example, raised tariffs on certain commodities in 1925 (Hendricks 1991, 31), as did other European countries (Corni 1990,275; Tracy 1964, 117-21).88 In the United States government policy promoted private cooperative controls on market supplies, and new restrictive trade policies were part of the McNary-Haugen legislation considered by Congress in the 1920s. But cooperatives in the United States and farmer syndicates in France, as well as higher tariffs and limited quotas on imports enacted in Western European countries, were insufficient to block the fall in prices in the 1930s. Much more aggressive policies were adopted for greater, direct government intervention into commodity markets to fix prices. Indeed, the plight of agriculture symbolized government response across the countries to the economic crisis of the 1930s. The focus on agriculture in the United States, for example, is indicated by the disproportionate share of federal government civilian expenditures made by the USDA, 18 percent in 1935, up from 6 percent or less in the 1920s. As in the United States, the scope and scale of new government intervention were beyond anything that had been implemented before. As noted by Perren: “In the 1930s British government assistance to farming came to exceed any88. As described by Tracy (1964, 120-28, 150-204) the reliance on tariffs varied across European countries, with tariff increases in Italy greater than in Britain, e.g.

216

Gary D. Libecap

thing that had been offered before in peacetime, even in the early nineteenth century” (1995,53). In Britain, government marketing boards were created for milk and bacon to manage the supplies placed on the market. Prices were fixed and production controls were installed for beef, pork, butter, cheese, wheat, barley, and oats. Deficiency income payments were added to cover the difference between the target price and the market price for grains (see Perren 1995, 7-60; Grigg 1989,21-24; Tracy 1964, 122-25, 150-69). In 1933 in Germany, the State Food Corporation was created to set prices, control production, and coordinate the sale of all agricultural commodities (see Hendricks 1991, 27-33; Stuhler 1989,443; Farquharson 1976,25-77; Corni 1990,43-95; Tracy 1964, 128-30). Similarly, in Italy, France, Australia, and Canada two-tier price policies were established, with domestic prices fixed above the world price (see Tracy 1964, 128, 170-84; Bothwell, Drummond, and English 1987, 254-59; Britnell and Fowke 1962; Shaw 1982, 14-16; Lloyd 1982, 359-61). Chronic surpluses were controlled through government purchases and stockpiling and through output and marketing controls. In Canada and Australia, wheat boards were created in the 1930s to purchase production and to manage the supplies placed on international markets. Other policies adopted in varying degrees across the countries included income subsidies, debt relief, lower interest rates, and moratoria on mortgage foreclosures (see Moulin 1991, 146-52; Corni 1990, 159,269-75; Farquharson 1976,62-66; Hefford 1985,64; Shaw 1982, 14-16; Lloyd 1982, 358-61; Tracy 1964, 122-31, 150-204; Bothwell et al. 1987,254; Rooth 1993,217-20). These programs typically were retained after commodity prices increased in the late 1930s and were expanded during World War I1 and in the postwar period.89 The crisis created by the Great Depression, the unprecedented, peacetime government intervention into agricultural markets it initiated, and the retention of those programs after it ended confirm the arguments made by Peacock and Wiseman (1961) and Higgs (1987), who argue that the growth of government obligations, regulations, and expenditures is discontinuous. Crises are necessary to change popular perceptions about the role of government and acceptable tax levels. Crises facilitate interest group formation and lobby activity for government relief or support. Taking advantage of the new political environment, politicians respond to lobby demands and general unrest by devising transfers and regulatory policies to address the crisis. These policies in turn create constituents who maneuver in the political arena to maintain and extend government intervention after the initiating crisis has passed. In this way, the Great Depression was the defining moment in the regulation of agriculture in the United States and elsewhere in Western industrial countries. 89. E.g., see Hefford (1985, 140-41) for discussion of the continuation of Australian policies and Grigg (1989,24) for discussion of the reaffirmation of guaranteed prices with the 1947 Agriculture Act in Britain. Other discussions of postdepression policies include Bothwell et al. (1989), Britnell and Fowke (1962,365-400), Keeler (1987), Muth (1970), and Tracy (1964,225-367).

Appendix Table 6A.1 Public Law

Linkage between New Deal Legislation and 1962 Agricultural Regulation Date

Link to New Deal Legislation

425

3130162

412

3/6/62

450

5/15/62

446

4/27/62

45 1

5/15/62

485

6/16/52

488

6/19/62

494

6/25/62

530

7110162

535

7/13/62

539

7/19/62

540

7/19/62

632

9/5/62

639 703

9/5/62 9/27/62

Ammended the Soil Conservation and Domestic Allotment Act and the AgriculturalAct of 1949 regarding participation in the 1962 feed grain support program. The AgriculturalAct of 1949 (P.L. 439), in turn, had modified the Agricultural Adjustment Act to define new parity prices, support levels, marketing quotas, and acreage allotments for tobacco, corn, wheat, rice, cotton, peanuts, wool, honey, potatoes, and other crops. Amended Sec. 353 of the Agricultural Adjustment Act regarding rice acreage history and allotments. Deferred proclamation under Sec. 332 of the Agricultural Adjustment Act of wheat-marketing quotas and acreage allotments for the 1963 crop. Extended coverage of Sec. 344 of the Agricultural Adjustment Act regarding cotton through 1962. Authorized planting of nonsurplus crops on diverted acreage as defined by Sec. 16(d)(l) of the Soil Conservation and Domestic Allotment Act. Deferred proclamation of a national wheat acreage allotment under Sec. 332 of the Agricultural Adjustment Act. Amended Sec. 204 of the AgriculturalAct of 1956 (P.L. 540) regarding restrictions on foreign imports of cotton or cotton products as outlined in the Agricultural Adjustment Act of 1933. Amended the Agricultural Marketing Act of 1937 to reduce the revolving fund. Amended the AgriculturalAdjustment Act of 1938 regarding tobacco acreage allotments. Amended and extended the provisions of the Sugar Act of 1948 (P.L. 388), which continued the sugar support program as defined in Sec. 301 of the Agricultural Adjustment Act of 1938 and the Sugar Control Act of 1937 (P.L. 414). Amended the Sugar Act of 1948 as authorized in the Agricultural Adjustment Acts and the Sugar Control Act of 1937. Amended the AgriculturalAdjustment Act of 1938 to extend the time for the referendum regarding the national marketing quota for wheat. Extended the International Wheat Agreement Act of 1949 (P.L. 421) to stabilize wheat prices as described in the amended Agricultural Adjustment Act of 1933 (P.L. 320, 1935). Extended water and soil conservation. Food and AgriculturalAct of 1962: amended extended provisions of Sec. 33 1-39 and others of the Agricultural Adjustment Act of 1938 and Sec. 7.8, 16 of the Soil Conservation and Domestic Allotment Act. Amended the Soil Conservation and Domestic Allotment Act

732 (continued)

10/2/62

Table 6A.1 Public Law

(continued) Date

80 1

10/11/62

824

10/15/62

Table 6A.2

regarding drainage of wetlands. Modified the wheat program as described in Sec. 309 of the Food and Agricultural Act of 1962 and the Agricultural Adjustment Act of 1938, Part 111. Amended Sec. 316(a) of the Agricultural Adjustment Act of 1938 regarding tobacco acreage allotments.

Demand Enhancement and Supply Control Programs as Listed in the Digest of Appropriations and U.S. Budget

Demand Enhancement

Supply Control

Link to New Deal Legislation

U.S. Food Administration, 1918-19; Federal Farm Board purchases via loans to cooperatives, 1929-33 (not part of USDA budget); CCC purchases, 1934-70; purchase of surplus sugar with AAA processing tax revenues, 1935-36; export and domestic consumption of surplus agricultural commodities, 1936-48, 1950-51; National School Lunch Act under the AAA, 1947-52; removal of surplus agricultural commodities 1940, 1949-52; surplus reserve for postwar price support via the AAA, 1946; foreign war relief via the AAA, 1942-47; foreign food programs via the AAA, 1946-47; Production and Marketing Administration (school lunch, removal of surplus agricultural commodities), 1950-52; Foreign Agricultural Service, 1953-70; and Agricultural Marketing Service, 1953-70 Milk Importation Act, 1928-41; Cooperative Marketing Act, 1927, 1930; Agricultural Adjustment Administration, 1934-49; Soil Conservation and Domestic Allotment Act, 1936-42; Filled Milk Act, 1937-41; Sugar Act, 1943, 1944; 1950-70: Production and Marketing Administration (administration of acreage allotments and marketing quotas of AAA, SCDA, and Sugar Act), 1950-52; Agricultural Conservation Program Service and Commodity Stabilization Service (AAA marketing quotas, acreage allotment, sugar quota program, Soil Conservation and Domestic Allotment Act), 1953-60; Soil Bank program, 1956-58); Agricultural Stabilization and Conservation Service (Agricultural Conservation Program Service merged with Commodity Stabilization Service), 1961-70

Table 6A.3

U.S. Domestic Wheat and World (Australian) Wheat Prices

Year

U.S. Base Price (larger of price received by farmers of CCC loan rate)

1990 1901 1902 1903 1904 1905 1906 1907 1908 I909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 I920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 I943 I944 1945 1946

$0.621 0.631 0.630 0.693 0.926 0.747 0.660 0.866 0.967 0.991 0.908 0.869 0.807 0.794 0.975 0.961 1.430 2.050 2.050 2.160 1.830 1.030 0.966 0.926 1.250 1.440 1.220 1.190 0.998 1.040 0.67 1 0.391 0.382 0.744 0.848 0.831 1.020 0.962 0.590b 0.691 0.682 0.980b 1.140b 1.360 1.410 1.490 1.900

(continued)

Price Supports”

0.290 0.290 0.330 0.120 0.280 0.181 0.180 0.234 0.234

Total US.Price (supported and unsupported)

Australian Export Price

$0.621 0.631 0.630 0.693 0.926 0.747 0.660 0.866 0.967 0.991 0.908 0.869 0.807 0.794 0.975 0.961 1.430 2.050 2.050 2.160 1.830 1.030 0.966 0.926 1.250 1.440 1.220 1.190 0.998 1.040 0.671 0.391 0.382 1.034 1.138 1.161 1.020 0.962 0.710 0.97 1 0.863 1.160 1.374 1.594 1.410 1.490 1.900

0.590 0.606 0.709 1.464 0.651 0.719 0.698 0.670 0.984 0.920 0.923 0.770 0.819 0.798 0.801 1.679 0.894 0.953 0.953 1.051 1.413 0.866 1.061 1.029 0.939 1.222 1.108 0.916 0.762 0.789 0.5.52 0.245 0.420 0.530 0.580 0.690 0.980 0.890 0.500 0.552 0.624 0.666 0.663 0.764 0.995 1.513 2.268

220

Gary D. Libecap

Table 6A.3

Year

(continued) U.S. Base Price (larger of price received by fanners of CCC loan rate)

1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970

2.290 2.000b 1.950b 2.000 2. 180b 2.200b 2.210b 2.240b 2.080b 2.000b 2.000b 1.820b 1.810b 1.780b 1.830 2.040 1.850 1.370 1.350 1.630 1.390 1.250b 1.250 1.330

Price Supports"

Total US.Price (supported and unsupported)

Australian Export Price

1.180 0.700 0.750 1.320 1.360 1.380 1.520 1.570

2.290 2.000 1.950 2.000 2.180 2.200 2.210 2.240 2.080 2.000 2.000 1.820 1.810 1.780 1.830 2.040 2.030 2.070 2.100 2.950 2.750 2.630 2.770 2.900

3.055 2.448 2.716 2.082 2.313 2.355 1.857 1.622 1.473 1.484 1.613 1.548 1.491 1.509 1.611 1.610 1.702 1.589 1.598 1.717 1.583 1.576 1.454 1.554

"Pricesupports for 1933-35 are benefit payments under the Agricultural Adjustment Act of 1933, for 1938-43 conservation payments under the Soil Conservation and Domestic Allotment Act, for 1963 Soil Bank payments, and for 1964-70 marketing certificate support payments. bCCCmarket market rate.

References American Institute of Cooperation.Various issues. American Cooperation. Washington, D.C.:American Institute of Cooperation. Australian Wheat Board. 1990. Export wheat prices, 1939-1989. Melbourne: Australian Wheat Board. Benedict, Murray R. 1953. Farm policies of the United States, 1790-1950: A study of their origins and development. New York: Twentieth Century Fund. . 1955. Can we solve the farmproblem: An analysis offederal aid to agriculture. New York: Twentieth Century Fund. Benedict, Murray R., and Oscar C. Stine. 1956. The agricultural commodity programs: Two decades of experience. New York: Twentieth Century Fund. Blaisdell, Donald C . 1940. Government and agriculture: The growth of federal farm aid. New York: Farrar and Rinehart. Bothwell, Robert, Ian Drummond, and John English. 1987. Canada 1900-1945. Toronto: University of Toronto Press.

221

Federal Government Regulation of Agriculture, 1884-1970

. 1989. Canada since 1947: Power; politics andprovincialism, rev. ed. Toronto: University of Toronto Press. Brandt, John. 1938. A national program to deal with surplus farm programs. In American Cooperation, 221-23. Washington, D.C.: American Institute of Cooperation. Breimeyer, Harold F. 1983. Agricultural philosophies and policies in the New Deal. Minnesota Law Review 68 (December): 333-52. Breyer, Stephen. 1982. Regulation and its reform. Cambridge, Mass.: Harvard University Press. Britnell, George Edwin, and V. C. Fowke. 1962. Canadian agriculture in war and peace, 1935-50. Stanford, Calif.: Stanford University Press. Cochrane, William, and Mary Ryan. 1976. American farm policy, 1948-1973. Minneapolis: University of Minnesota Press. Commodity Research Bureau. 1939. Commodity year book. New York: Commodity Research Bureau. Corni, Gustavo. 1990. Hitler and the peasants: Agrarian policy of the Third Reich, 1930-1939. New York Berg. Davis, Joseph S. 1935. Wheat and the AAA. Washington, D.C.: Brookings Institution. . 1942. World wheat survey and outlook, May 1942. Wheat Studies of the Food Research Institute. Stanford, Calif.: Stanford University Press. Deering, Ferdie. 1945. USDA: Manager of american agriculture. Norman: University of Oklahoma Press. Economist. 7 October 1995; 9 March 1996. Evans, Sam. 1984. Wheat background for I985 farm legislation. USDA, Economic Research Service. Washington, D.C.: Government Printing Office. Ezekiel, Mordecai. 1934. The cooperative approach to production control. In American Cooperation, 178-83. Washington, D.C.: American Institute of Cooperation. Farquharson, J. E. 1976. The plough and the swastika: The NSDAP and agriculture in Germany, 1928-1945. Beverly Hills, Calif.: Sage. Fite, Gilbert C. 1954. George N. Peek and thefightfor farm parity. Norman: University o f Oklahoma Press. Gardner, Bruce. 1981. The governing of agriculture. International Center for Economic Policy Studies and the Institute for the Study of Market Agriculture. Lawrence, Iowa: Regents Press. . 1987a. Causes of U S . farm commodity programs. Journal of Political Economy 95 (April): 290-310. . 1987b. The economics of agricultural policies. New York: Macmillan. Gardner, B. Delworth. 1995. Plowing ground in Washington. San Francisco: Pacific Research Institute for Public Policy Research. Green, Robert C. 1990. Program provisions for program crops: A database for 196190. USDA, Economic Research Service. Washington, D.C.: Government Printing Office. Grigg, David. 1989. English agriculture: An historical perspective. Oxford: Blackwell. Hadwiger, Don F. 1970. Federal wheat commodityprograms. Ames: Iowa State University Press. Hamilton, David E. 1991. From new day to New Deal: Americanfarm policyfrom Hoover to Roosevelt, 1928-1933. Chapel Hill: University of North Carolina Press. Hefford, R. K. 1985. Farm policy in Australia. St. Lucia: Queensland University Press. Heid, Walter G. 1980. U.S. wheat industry. USDA, Economic Research Service. Washington, D.C.: Government Printing Office. Hendricks, Gisela. 1991; Germany and European integration, a common agricultural policy: An area of conflict. New York: Berg. Higgs, Robert. 1987. Crisis and Leviathan: Critical episodes in the growth ofAmerican government. New York: Oxford University Press.

222

Gary D. Libecap

Hoffman, Elizabeth, and Gary D. Libecap. 1991. Institutional choice and the development of U.S. agricultural policy in the 1920s. Journal of Economic History 51 (June): 397-411. . 1994. Political bargaining and cartelization in the New Deal: Orange marketing orders. In The regulated economy: A historical approach to political economy, ed. Claudia Goldin and Gary D. Libecap, 189-221. Chicago: University of Chicago Press. . 1995. The failure of government-sponsored cartels and development of federal farm policy. Economic Inquiry 33 (July): 365-82. Holt, John Bradshaw. 1936. German agricultural policy, 1918-1934. Chapel Hill: University of North Carolina Press. Hulbert, L. S. 1932. Legal status of plans for production control. In American cooperation, 504-15. Washington, D.C.: American Institute of Cooperation. International Monetary Fund. 1979. International financial statistics. Washington, D.C.: International Monetary Fund. . 1995. Internationalfinancial statistics. Washington, D.C.: International Monetary Fund. Irons, Peter H. 1982. The New Deal lawyers. Princeton, N.J.: Princeton University Press. Johnson, D. Gale. 1973. Farm commodityprograms: An opportunityfor change. Washington, D.C.: American Enterprise Institute. Keeler, John T. S. The politics of neocorporatism in France: Framers, the state, and agriculture, policy-muking in the Fifth Republic. New York Oxford University Press. Knutson, R. D., J. B. Penn, and W. T. Boehm. 1983. Agriculture andfoodpolicy. Englewood Cliffs, N.J.: Prentice-Hall. Libecap, Gary D. 1992. The rise of the Chicago packers and the origins of meat inspection and antitrust. Economic Inquiry 30 (April): 242-62. Lloyd, A. G. 1982. Agricultural price policy. In Agriculture in the Australian economy, ed. D. B. Williams, 353-82. Sydney: University of Sydney Press. Lowi, Theodore J. 1979. The end of liberalism, 2d ed. New York: Norton. McCune, Wesley. 1956. Whok behind our farm policy. New York: Praeger. Moulin, Annie. 1991. Peasantry and society in France since 1789. New York: Cambridge University Press. Murphy, P. L. 1955. The New Deal agricultural program and the constitution. Agricultural History 29-30 (October): 160-68. Muth, Hans Peter. 1970. French agriculture and the political integration of Western Europe. Leyden: A. W. Sijthoff. New York Times. 1995. 26 September; 1 October. . 1996.29 February. Nourse, Edwin D. 1935. Marketing agreements under the AAA. Washington, D.C.: Brookings Institution. Nourse, Edwin D., Joseph S. Davis, and John D. Black. 1937. Three years of the Agricultural Adjustment Administration. Washington, D.C.: Brookings Institution. Pasour, E. C., Jr. 1990. Agriculture and the state: Market processes and bureaucracy. New York: Holmes and Meier. Peacock, Alan T., and Jack Wiseman. 1961. The growth ofpublic expenditures in the United Kingdom. Princeton, N.J.: Princeton University Press. Perkins, Van L. 1965. The AAA and politics of agriculture: Agricultural policy formation in the fall of 1933. Agricultural History 39-40 (October): 220-29. . 1969. Crisis in agriculture: The Agricultural Adjustment Administration and the New Deal, 1933. Berkeley: University of California Press. Perren, Richard. 1995. Agriculture in depression, 1870-1940. New York: Cambridge University Press.

223

Federal Government Regulation of Agriculture, 1884-1970

Pope, David. 1987. Private finance. In Australians, historical statistics, ed. Wray Vamplew, 238-63. Broadway, New South Wales: Fairfax, Syme, and Weldon Associates. Rasmussen, Wayne D., and Gladys L. Baker. 1979. Price support and adjustmentprograms from 1933 through 1978: A short history. U.S. Department of Agriculture. Washington, D.C.: Government Printing Office. Rockoff, Hugh. 1984. Drastic measures: A history of wage and price controls in the United States. New York Cambridge University Press. Rooth, Tim. 1993. British protectionism and the international economy: Overseas commercial policy in the 1930s. New York: Cambridge University Press. Rucker, Randy, and Lee J. Alston. 1987. Farm failures and government intervention: A case study of the 1930s. American Economic Review 77 (September): 724-30. Saloutos, Theodore. 1982. The American fanner and the New Deal. Ames: Iowa State University Press. Schultz, Ted W. 1949. Production and welfare of agriculture. New York: Macmillan. Shaw, A. G. L. 1982. History and development of Australian agriculture. In Agriculture in the Australian economy, ed. D. B. Williams, 1-29. Sydney: University of Sydney Press. Shepherd, William G. 1990. The economics of industrial organization, 3d ed. Englewood Cliffs, N.J.: Prentice-Hall. Shergold, Peter. 1987. Prices and consumption. In Australians, historical statistics, ed. Wray Vamplew, 210-26. Broadway, New South Wales: Fairfax, Syme, and Weldon Associates. Shideler, James H. 1957. Farm crisis, 1919-1923. Berkeley: University of California Press. Shover, John L. 1965. Populism in the nineteen-thirties: The battle for the AAA. Agricultural History 39-40 (January): 17-24. Stedman, Alfred D. 1938. The philosophy and practical operations of the new farm bill. In American Cooperation, 78-83. Washington, D.C.: American Institute of Cooperation. Stuhler, Elmar A. 1989. Changes in agricultural policy and in farm structures in West Germany in the past SO years. Munich: Hampp. Tracy, Michael. 1964.Agriculture in WesternEurope: Crisis and adaptation since 1880. London: Cape. Udell, Gillman G. 1971. Laws relating to agriculture. Washington, D.C.: Government Printing Office. . 1972. Farm relief and Agricultural Adjustment Acts. Washington, D.C.: Government Printing Office. U S . Department of Agriculture (USDA). Various years. Agricultural statistics. Washington, D.C.: Government Printing Office. . 1934. Yearbookof agriculture. Washington, D.C.: Government Printing Office. . 1941, 1956-60. Annual report of the Department of Agriculture. Report of the President of the Commodity Credit Corporation. Washington, D.C.: Government Printing Office. . 1996. Wheat yearbook. Washington, D.C.: Government Printing Office. . Agricultural Stabilization and Conservation Service (ASCS).1964. Wheatprogram 1965. Washington, D.C., Government Printing Office. . 1968. Wheatprogram 1968. Washington, D.C.: Government Printing Office. . 1985. Compilation of statutes relating to soil conservation, set aside, acreage diversion, marketing quotas and allotments, wheat certi$cates, CCC, price support, export and surplus removal, P.L. 480, crop insurance, and related statutes as of January I , 1985. Agricultural Handbook no. 476. Washington, D.C.: Government Printing Office.

224

Gary D. Libecap

. Economic Research Service. 1996. Wheat: Farm prices, support prices, and ending stocks, 19SO/51499S/96. Washington, D.C.: Government Printing Office. U.S. Department of Commerce. 1913-71. Statistical abstract of the United States. Washington, D.C.: Government Printing Office. . 1975. Historical statistics of the United States. Washington, D.C.: Government Printing Office. U.S. Department of Commerce and Labor. 1908-12. Statistical abstract of the United States. Washington, D.C.: Government Printing Office. U.S. Department of the Treasury. Various years. Annual report of the treasurer of the United States. Washington, D.C.: Government Printing Office. . Various years. Digest of appropriations. Washington, D.C.: Government Printing Office. . Various years. US.budget. Washington, D.C.: Government Printing Office. Wall Street Journal. 1995.30 August; 26 September. . 1996.8 February; 1 April. Wood, Donna. 1986. Strategic uses of public policy: Business and government in the Progressive Era. Marshfield, Mass.: Pitman. Young, James Harvey. 1989. Pure food: Securing the federal Food and Drugs Act of 1906. Princeton, N.J.: Princeton University Press.

7

A Distinctive System: Origins and Impact of U.S. Unemployment Compensation Katherine Baicker, Claudia Goldin, and Lawrence F. Katz

If Franklin D. Roosevelt had not been President and if the nation had not been plunged into the Great Depression of the 1930's, action might not have been taken at that time or might have taken a different form. Arthur Altmeyer (1966, 6)'

The unemployment compensation system of the United States is distinctive. Nowhere else does publicly mandated unemployment insurance (UI) contain the feature of experience rating, whereby firms are penalized for benefits paid out to their workers.2Unlike many European systems, unemployment compenKatherine Baicker is a doctoral candidate in economics at Harvard University. Claudia Goldin is professor of economics at Harvard University and director of the Development of the American Economy Program at the National Bureau of Economic Research. Lawrence F. Katz is professor of economics at Harvard University and a research associate of the National Bureau of Economic Research. The authors thank their discussant, Bruce Meyer, for his exceptionally detailed comments and others at the conference for general guidance. They are grateful to Stephen Wandner of the U.S. Department of Labor, Unemployment Insurance Service, for the updated data on UI finance and thank Donald Deere for making his UI data available to them. Many of their Canadian friends helped them locate historical data for Canada. They especially thank Thomas Lemieux, Frank Lewis, and Mary McKinnon. Bridget Chen and Mireille Jacobson served admirably as research assistants for this project. Goldin and Katz thank the National Science Foundation for funding a portion of this research. 1. Altmeyer was the chair of the Technical Board of the Committee on Economic Security, established by executive order in 1934, and was one of three members of the first Social Security Board, becoming its chair in 1936. 2. Both Great Britain and Austria had once allowed for some type of experience rating at the trade or industry level. In Britain the original act contained a section that has been interpreted as a form of experience rating. It was replaced in 1920 with a provision that entire industries could independently manage their own unemployment insurance (UI) plans. But the operation of the provision was suspended in 1921 and eliminated entirely in 1927 (Feldman and Smith 1939; Lester and Kidd 1939). Experience rating by industry was part of the original UI system passed in Austria in 1920, but it was later abandoned because labor felt it was too divisive an issue (Becker 1971, chap. 3). Even though there is no experience rating in European UI systems, many European countries have mandated severance pay for permanently separated workers. The system amounts to some experience rating, although employer taxes are not linked to unemployment benefits. They are, instead, a tax per worker laid off that is usually a function of years of employee service with the firm.

227

228

Katherine Baicker, Claudia Goldin, and Lawrence F. Katz

sation among the states has always limited duration of benefit^.^ And no national system in the industrialized world cedes to its states, provinces, or cantons as much autonomy to establish tax rates, set benefits, and determine eligibility as does that in the United States. Unemployment compensation in the United States is not one system; it is many. Yet, even though there are a multitude of systems-53 to be precise-there is unity in several features that are also distinct with respect to other countries. Experience rating, a federalstate structure, and the limitation on benefit duration are the most obvious. Others include less generous benefits, a tax base equal to a small fraction of payroll, restrictive eligibility, and currently low take-up rates. The three distinctive features of the U.S. unemployment insurance system were intentional. Even though UI was part of the omnibus Social Security Act (SSA) passed in August 1935, in a time of urgency and amidst a flurry of legislation, its details were well thought out. Two of the features-the state system and the potential for experience rating-were parts of the original federal law and were responses to past lessons fresh in the minds of the drafters. The limitation on duration of benefits, incorporated into all state laws, was also derived from historical lesson. The law was tailored to the times but not simply to the depression experience. Previous decades of experience with unemployment, in the United States and elsewhere, and with the legislative and judicial processes shaped the act. More so than any other part of the SSA, such as welfare and social security, the design of UI remains today that intended in the original document. Details have changed, to be sure. They had to, because UI is a state system and the states had to pass legislation to receive credit for the taxes charged their employers. Tax rates, benefits, the tax base, eligibility, maximum duration of benefits, and coverage have all changed countless times since the late 1930s. But the substance of the act remains very much the same. The act was intended to provide unemployment benefits to those having attachment to the labor force. Under the act, the unemployed are viewed as those temporarily out of work, not the permanently displaced. The law contains no provision for retraining. UI was not a handout or welfare, for benefits were not to be means tested. They depend on previous earnings and work experience and could be of limited duration.“ Employers were to be penalized for their contribution to (compensated) unemployment, and states were given an incentive to devise tax rates to accomplish that goal. Finally, the act encouraged 3. Federal-state extended benefits for periods of high unemployment, jointly financed by the federal government through the federal unemployment tax and the states, were proposed in the 1960s and eventually passed in 1970. Benefits, even when extended, are still limited, usually to 13 weeks beyond the exhaustion of regular state benefits. 4. The act barred means testing through its definition of “compensation” as “cash benefits payable to individuals with respect to their unemployment” (Sec. 907.b). Although the act has no provisions for retraining and prohibits the use of UI funds for any disbursements other than UI compensation and administrative costs (Sec. 903.a.4). prior legislation covered placement. Passed

229

Origins and Impact of U.S. Unemployment Compensation

states to tailor their systems to best meet the needs of their workers and firms. Because the substance of the UI system has changed little with time, many claim that it is antiquated, particularly with regard to its lack of retraining and the absence of benefit eligibility for many part-year and part-time employees who lose their jobs (see, e.g., Advisory Council on Unemployment Compensation [ACUC] 1996). Because, as we contend, today’s UI system has retained its original design, we can address part of the central question of the conference: How did the establishment of the UI system “define” or alter the economy in subsequent years? The answer involves observing how the UI system affected the economy through its particular design. We will emphasize the distinctive features of the system and not its existence because some UI system, we and most others believe, was destined to be establi~hed.~ But how would the UI system have differed had it been passed during an era other than the Great Depression? In what manner are its distinctive features today, and their evolution at the state level, products of the period of its passage, and what difference have they made to the economy? As Arthur Altmeyer conjectures in the epigraph, had there been no Great Depression social insurance might not have taken the form it did in 1935. Of the three distinctive features of the system, it is the federal-state structure that would probably be different had passage been delayed. And had there not been a federal-state structure, it is likely that the system would not have been experience rated, but we are less secure in this prediction. The most interesting issue concerns how the system functioned because of its federal, and not national, character. We explore various aspects of the system, two of which-the federal-state structure and experience rating-are related to the possible difference that passage during the 1930s meant. We look at how states tailored their systems from the late 1940s to the 1960s in response to the needs, objectives, or simple rentseeking activities of their citizens. We then ask whether firms in states with more experience rating responded to financial penalties by reducing employment fluctuations and whether seasonality in the United States evolved any differently from that in Canada, which does not have experience rating, after both passed UI.

in 1933, the Wagner-PeyserAct (Employment Exchange Bill) established a federal-state network of employment service offices to help workers in their job searches. The bill was not, however, intended to provide retraining. 5. Those who worked on the act believe that as well: “There is no question . . . that sooner or later the same world-wide economic, social, and political forces responsible for the emergence of social security programs in 125 nations would have resulted in the developmentof a social security program in the United States” (Altmeyer 1966,6).

230

Katherine Baicker, Claudia Goldin, and Lawrence F. Katz

7.1 Unemployment Insurance Today and Its Evolution 7.1.1 Coverage, Eligibility, and Generosity We step back, for a moment, from the questions at hand to briefly describe the UI system today and its evolution nationally and at the state level during the past 60 years. When UI was passed in 1935, the intent of the framers was to insure workers, with attachment to the labor force, against unemployment of short duration and to provide them with sufficient income to tide them over, but not so much as to generate serious work disincentives. That has remained the intent. Note, however, that the original law is extremely parsimonious in the details and that virtually all the features that realized the intent-those regarding eligibility, duration, and generosity-arose out of state law. Coverage is one of the few details in the 1935 act.6 At the time of its passage, UI covered nonagricultural, non-public-sector workers in firms with more than eight employees.’ In the six decades since its passage, coverage of those demonstrating “attachment” to the labor force has continued, but the act has become more inclusive in covering workers in virtually all firms and those in public sector employment. As shown in figure 7.1, covered employment was 60 percent in the early 1950s and it is now 90 percent8 Thus, U.S. unemployment insurance coverage today is nearly universal for wage and salary employees. Some of the time-series expansion reflects increases in the statutory number of employees per firm and in the distribution of employees across firms. The upper line in the graph (after 1971) shows the influence of other changes in the law. The various jumps in that line reflect the inclusion of employees of nonprofits and public sector workers. The original intent of UI-that it be for those attached to the labor force, of limited duration, and of moderate amount-was implemented by the various states and evolved over time. Attachment to the labor force is assured by eligibility requirements of two types: monetary and nonmonetary. Monetary eligibility requirements typically demand that the claimant demonstrate earnings in the recent past that exceed some minimum and have had employment that was sufficiently continuo~s.~ Claimants must also demonstrate availability for work and current search behavior. Unemployment must have arisen from eco6. The potential for experience rating was also part of the original law. We take up this subject below. 7. Also excluded were domestic servants, employees of nonprofits, casual laborers, those over age 65, and some others (Title VIII, Sec. 81 1). Railroad workers were excluded in 1938 when they became covered by their own statute. 8. In 1954 coverage was extended to firms with at least four employees. In 1970 the law was broadened to include employees of nonprofits, and in both 1970 and 1976 coverage was extended to all state and local government employees and to some agricultural workers. 9. Monetary eligibility requirements typically have three parts and take the following form: (1) earnings during some base period (sum of the four or the five most recent quarters) have to exceed some minimum level, (2) during at least one of the quarters they have to exceed a higher level, and (3) claimant had to have earnings during at least two of the four quarters.

Origins and Impact of US. Unemployment Compensation

231

0.9

1

?=--

I

0.8 I 0.7 0.6

0.5 0.4

0.3

'

CovEmpl CovEmp2 I I ' '

39

Fig. 7.1

"

I '

44

"

' I

" "

49

I ' '

54

I "

" '

59

"

I '

64

"

' I

"

69

' ' I

" "

74

I ' '

79

"

I '

84

"

' I '

89

"

"

94

Proportion of the civilian labor force covered by UI, 1939-94

Sources and Notes: U.S. Department of Labor, ETA (1996) for average monthly covered employment. After 1971, average monthly covered employment for the "reimbursable" sectors is included. "Reimbursable" includes the public sector and nonprofits that do not pay taxes but receive coverage. U.S. Council of Economic Advisers (CEA 1996, table B-31) for civilian employment. Civilian employment before 1947 includes those 14 years old or older; after 1946 it includes those 16 years old or older. CovEmp 1 = (taxable covered employment)/(civilian employment). CovEmp2 = [all (taxable + reimbursable) covered employment]/civilian employment. Note that CovEmpl and CovEmp2 are the same until 1971.

nomic reasons; quits and terminations for cause are typically invalid reasons. The maximum duration of benefits is typically 26 weeks in all states today. Potential duration increased somewhat in the early years of the system and rose to the current 26-week level by the mid-l950s.I0 The original act, although silent on most details, did set down a legal grievance procedure, and most grievances today concern the nonmonetary aspects of eligibility, such as whether an employee quit or was fired. States have considerable discretion to change eligibility. Yet it is important to realize that the eligibility requirements established at the very outset of the program are, by and large, those in place today. The requirements may have suited the labor force of the 1920s and 1930s, but with a larger share of the workforce employed part time and discontinuously,a lower fraction of the covered workforce is now eligible for benefits once unemployed. 10. Federal-state extended benefits, as part of the 1970 UI law, are triggered by a complicated formula concerning both the level and the rate of change of unemployment in a state. States can choose whether to use a trigger based on their insured unemployment rate or on their overall unemployment rate. More recently, as occurred in the 1991-93 recession, Congress has, by discretion and on an ad hoc basis, passed emergency extended benefits for the entire nation, thus bypassing the laws for the separate states.

232

Katherine Baicker, Claudia Goldin, and Lawrence F. Katz

0.6

0.5

0.4

0.3

,,,,,, , , ,

0.2"""' 40 45

UI claimdTot U benefiwage

I , , ,

50

55

60

65

70

75

80

85

, , , ,

90

95

Fig. 7.2 Evolution of (UI claims/total unemployed) and (benefitdwages), 1940-94

Sources and Notes: U.S. Department of Labor, ETA (1996) for the number of UI claims, average weekly benefits, and the average weekly wage. CEA (1996, table B-3 1) for civilian unemployment. Average weekly benefits and average weekly wages are entire US.averages.

There are various measures of the system's generosity. All begin with the legal replacement rate schedule. Most state laws, today and in the past, contain replacement rates in the 50-60 percent range." The ex post, actual rate can be computed only by using microlevel data because it requires knowing the earnings of benefit recipients.I2A more convenient measure for the past is the average weekly benefits of UI recipients as a fraction of the average weekly wage of all covered employees. As seen in figure 7.2 (open-dotline), this measure has been rather stable, nationally, from the early years of the program to the present.I3 Even though the ratio of benefits to average wages, for the entire United States, has been relatively stable, growing variation across the states has 11. Many are instead expressed as an equivalent fraction (e.g.. 1/26) of high-quarter wages. 12. Actual pretax replacement rates today have been computed in the 52-7 1 percent range, by state, with between 70 and 90 percent of recipients, by state, receiving more than 50 percent. See ACUC (1996,54), which calculates the rate for Illinois, Michigan, Pennsylvania, Texas, Washington, and Wisconsin using the Survey of Income and Program Participation. Gruber (1994) estimates a mean after-tax replacement rate of 58 percent for unemployed household heads during 1969-87, using the Panel Study of Income Dynamics. He finds, in most states, average after-tax replacement rates of below 50 percent in 1987, when UI benefits became fully taxable under the federal income tax. 13. UI benefits were not subject to federal income tax prior to 1979. The federal taxation of benefits was phased in from 1979 to 1987. Thus, the stability of the pretax replacement rate, as shown in figure 7.2, masks a decline in the after-tax replacement rate during the phase-in period.

Origins and Impact of U.S. Unemployment Compensation

233

0.1 0.08 0.06 0.04

0.02

!b

'45' 50 '

'

'

55 ' ' 6 0 ' 6 5 ' '

'

70

' '

7 5 ' ' 80 ' ' 85 ' ' 90

'

95

Fig. 7.3 Cross-state standard deviations of (benefitdwages)and (maximum benefitdwages), 194&94 Sources and Notes: U.S. Department of Labor, ETA (1996) for state-level data on average weekly UI benefit and average weekly wage. U S . Department of Labor, ETA (various years; 1996) for maximum weekly benefits in the case of no dependents. sd rrate = standard deviation of the log (average weekly benefit/average weekly wage) or the standard deviation of the (log) replacement rate. sd rrmax = standard deviation of the log (maximum weekly benefit, no dependents/average weekly wage) or the standard deviation of the (log) maximum replacement rate.

emerged during the past decades, as can be seen in figure 7.3. The bottom (open-dot)line shows the standard deviation of the logarithm of the replacement rate (benefit/wage) by state. After an initial decline until 1956, as the states tinkered with their new systems, this variance measure rose and then remained at that higher level until the late 1970s, when it rose again (although it has declined recently). A second measure (also in fig. 7.3, solid-dot line), the standard deviation of the log (maximum weekly benefit/average weekly wage), reveals considerably more upward drift. States have, over time, established maximum rates that differ more from each other, even though the ex post replacement rates exhibit less increasing variance. Generosity must be tempered by eligibility, and eligibility has fallen over the past 45 years.I4We graph a measure of eligibility in figure 7.2 (solid-dot line)-UI claims as a fraction of total national unemployment. The measure declines in two periods: during the early 1960s and the early 1980s. Another 14. Hamermesh and Slesnick (1995) demonstrate that consumption has been stable in recent years for those receiving benefits. Thus the consumption-smoothing mechanism of UI has not been altered for this group, but the group receiving benefits has changed.

234

Katherine Baicker, Claudia Goldin, and Lawrence F. Katz

Un Rate

UVGDP

0.12

0.012

0.1

0.01

0.08

0.008

0.06

0.006

0.04

0.004

0.02

0.002

‘45

50

55

60

65

70

75

80

85

90

0

95

Fig. 7.4 Unemployment benefits as a fraction of GDP and the unemployment rate, 1948-94 GDP. 1959-94, CEA (1996, table B-l), 1946-58, U.S. Department of Commerce (1993, table 1.1, 1946-58). UI Unemployment insurance benefits for 1963-94, CEA (1996, table B-41), 1946-62, CEA (1977, table B-31). Un Rate: Unemployment rate for 1948-94, CEA (1996, table B-39). UI does not include supplemental benefits paid by federal funds but does include extended benefits.

Suurces and Notes:

way of observing the same phenomenon is to express UI as a fraction of GDP and graph it against the unemployment rate, as has been done in figure 7.4.15 The two lines of figure 7.4 are very close until the early 1980s when they drift farther apart. Both figures 7.2 and 7.4 reveal the declining eligibility. One of the puzzles addressed in the recent literature on UI is why the fraction receiving benefits has decreased. Some emphasize long-term changes in the structure of the labor force and the erosion of union strength (Baldwin and McHugh 1992; Corson and Nicholson 1988). Others claim the change is part of a general “race to the bottom” in social insurance as well as a product of an interaction between UI and federally funded programs (ACUC 1996).16States have eroded UI eligibility to shift costs to programs, such as food stamps, that are more fully funded.I7Thus, it appears from various indicators that the frac15. UI in fig. 7.4 is total benefits paid by the states (also to federal employees, railroad workers, exservicemen, and veterans) not including federally funded emergency supplemental benefits. 16. Another possible factor is the decrease in net benefits with their inclusion in taxable income. Anderson and Meyer (1996) find that this reduction in net benefits reduced UI take-up rates and can explain some of the decline in UI receipts in the 1980s. 17. The regional distribution of the labor force matters to the aggregate take-up rate. States in the Northeast have high take-up rates but others, such as Texas, have low take-up rates. Blank and Card (1991) conclude that one-third to one-half of the decline in take-up is due to changes in the regional distribution of U.S. workers. Others argue that this factor is less important. Instead, they

235

Origins and Impact of U.S. Unemployment Compensation

tion of covered employment that is able to (or chooses to) collect UI, when unemployed, has declined over time. The size of UI in the national economy is a product of coverage, eligibility, generosity, and unemployment rate. Thus, it might not be an odd or unfortunate situation for UI to be a shrinking share of GDP. But it is curious and interesting that workers’ compensation (WC), the first social insurance program in the United States, is now considerably larger than UI. In 1960 UI was almost twice as large as WC, but in 1990, another nonrecessionary year, WC was almost twice the size of UI. Even during the 1991-93 recession, UI was a lower fraction of GDP than was WC. The WC explosion may, in part, be a product of the same forces that have increased all health care expenditures. Both UI and WC are fully funded at the state level, and thus both could have been subject to a race to the bottom. But if there has been such a race, it should have eroded expenditures on both.’* 7.1.2 Comparisons with Canada and Other OECD Countries Another means of judging the generosity, eligibility, and coverage of the U.S. system is by way of comparison. Our neighbor to the north passed a national UI law in 1940.19Through the early 1970s the general features of the U.S. and Canadian programs moved in tandem. But since then expenditures on the Canadian system have exploded in comparison with the United States. The reasons concern major legislative changes in Canada. Eligibility requirements were substantially reduced in 1972-just 8 weeks of work were required to qualify (increased to 10-14 weeks in 1978); benefits were expanded in numerous ways-maximum weeks and the replacement rate were increased (Green and Riddell 1993). In consequence, spending soared. Although further changes in the latter part of the 1970s reduced the generosity of the system, Canada had moved closer to the European UI model and farther away from that of the United States.20In 1976 the ratio of total UI benefits in the United States to that in Canada was 5; it was 1.3 in 1990 (Green and Riddell 1993, table 4). note that as states faced budgetary pressures, they got tougher administratively and encouraged employers to challenge claims. Further, as the fraction of workers with unstable workforce histories grows, so will the wedge between the total unemployment rate and the eligibility rate for UI. The decline of unions could also have contributed to the decrease in take-up rates if unions had supported workers whose claims were challenged and if they provided information on filing. 18. One reason there could be a race to the bottom in UI but less so in WC concerns asymmetric information. Employers have far more information on why workers are terminated than they do on workers’ health status, and thus they can more credibly threaten to contest UI claims. 19. Canada passed the Employment and Social Insurance Act in 1935, but it was deemed not within the realm of the federal government. A Royal Commission, set up in 1937, recommended a Canadian federal UI program and a constitutional amendment to allow for it. With the amendment, the passage and constitutionality of the Canadian Unemployment Insurance Act of 1940 was assured (Green and Riddell 1993). Interestingly, Canada has a national UI law and provincial minimum wages, whereas the United States has state UI laws and a national minimum wage. 20. A recent study shows that the United States has among the least generous UI systems in the Organization for Economic Cooperation and Development (OECD). Part of the lack of generosity comes from the lower replacement rate in the first half-year of unemployment, and part comes

236

Katherine Baicker, Claudia Goldin, and Lawrence F. Katz

7.2 Origins of the Distinctive Features How did the U.S. unemployment compensation system obtain its distinctive features, and in what sense were they the products of a distinctive decade? Were they designed to ensure passage of the act, take care of special problems of the day, or serve another function? Among the three features-experience rating, federal-state structure, and limited benefits-that which most shaped the system has been its federal-state structure. A discussion of all three takes us deep into the history of the subject at both the national and state levels and gives us a chance to explain some of the basic features of UI at its outset and today.21 7.2.1

State Action prior to 1935

Prior to the Great Depression, unemployment insurance was discussed at both the state and national levels, but there was little pressure to pass legislation. Massachusetts, in 1916, was the first state to introduce a UI bill, modeled after the 1911 British act that set up its social security and UI systems. Massachusetts was followed by New York and Wisconsin in 1921. The Wisconsin bill, originally drafted in large measure by John R. Commons, differed from the others by penalizing firms with higher unemployment rates, a system now known as experience rating and then termed “merit rating.” The Huber (or Groves) bill, as it was eventually called, underwent various modifications before becoming law in 1932. The most important change concerned the establishment of employer reserve accounts instead of a state-pooled fund, with employer liability limited to the amount of the fund.22The system of employer reserve accounts, although allowed by the federal act and adopted by some, was eventually dropped by all states. Seven states (Connecticut, Massachusetts, Minnesota, New York, Pennsylvania, South Carolina, and Wisconsin) introduced bills for compulsory systems of UI before 1929, and one was introduced in Congress in 1928 (Industrial from the absence of UI coverage after six months. The calculation also factors in assistance programs (e.g., food stamps) applying under particular family circumstances. In the overall generosity measure, covering five years, the United States is third from the bottom (out of 21 countries) and is third or fourth from the bottom for the first year. See OECD (1994, table 8.1. p. 175). 21. We have borrowed heavily from, among others, Altmeyer (1966), Becker (1971), Blaustein (1993), Douglas (1939), Haber and Murray (1966), Industrial Relations Counselors (1934), Nelson (1969). Perkins (1946), Schlesinger ([1958] 1988), Stewart (1930, 1938), and Witte (1962). Historians are fortunate that many who crafted and drafted the legislation wrote extensive memoirs, although they are not always in agreement. 22. Commons worked on the 1921 Huber bill, which contained a pooled state fund. According to Becker (1971), when Philip La Follette (the son of the senator) wanted to run for governor, he asked Paul Raushenbush and Harold Groves to help draft a UI bill. Groves, according to Becker, thought that employer funds, and thus full experience rating, would appear less like socialism to the voters. Some firms had already created their own UI funds, and with full experience rating there would be no cross-subsidization. The final Huber (Groves) bill is generally associated with Commons, although it, perhaps, should be associated more with the Commons school than with the man.

237

Origins and Impact of U.S. Unemployment Compensation

Relations Counselors 1934, 72). With the onset of the economic downturn, activity accelerated regarding UI legislation. Commissions to study UI were authorized in nine states and reported in six (California, Connecticut, Massachusetts, New York, Ohio, and Wisconsin). The governor of one convened a conference of seven contiguous states to explore the possibility of coordinated action. That governor, Franklin D. Roosevelt, assigned the task of organizing the conference to Frances Perkins, later to be his secretary of labor. Perkins, in turn, involved Paul Douglas in its planning. Douglas helped draft unemployment legislation for Ohio in 1932 that differed in two important ways from the Wisconsin model. The Ohio plan provided for pooled reserves at the state level, rather than employer accounts, and did not incorporate merit rating (although it left open the possibility of experience rating in the future). The two systems were similar in other respects. Benefits were limited to 16 weeks in Ohio and to 10 weeks in Wisconsin. Benefits equaled 50 percent of weekly wages in both states but with different minimum and maximum levels. The tax rate was 2 percent of covered wages to be paid by employers in Wisconsin but was 3 percent in Ohio, with 2 percent from employers and 1 percent from employees.23 The Commons and Douglas “model” plans-those of Wisconsin and Ohio-reflected opposing views of the causes of unemployment and the economic effects of the remedy. Commons, and his associates at the University of Wisconsin, believed employers had wide leeway to reduce seasonal and other layoffs, and thus they championed merit rating and employer reserve funds.24 UI would be to unemployment what WC was to industrial safety. Both laws would penalize employers who engaged in practices that, to Commons and others, injured labor. Douglas, and his associates at the University of Chicago, saw in unemployment compensation a means of lifting the economy out of its doldrums through increased purchasing power. Contrary to Commons, Douglas was unconvinced that a large part of unemployment was at the discretion of the employer. Employers already had incentives to reduce unemployment and seasonal layoffs. A sound system based on pooled accounts, he believed, was essential to the automatic stabilizing role he accorded unemployment compensation. Merit rating could unduly penalize some firms that experienced negative shocks and could even exacerbate a downturn by driving such firms into bankruptcy.25 The debates between the two schools took place mainly on substantive 23. See Industrial Relations Counselors (1934, table D) for state bills passed by either house after 1929. 24. It is not clear what market failure Commons believed was at work, other than the fact the firms did not directly take into consideration the value their workers placed on consumption smoothing. Of course, if workers were identical, those in more highly seasonal industries would receive a wage premium. 25. The Commons and Douglas models were not the only serious proposals. The American Association for Labor Legislation, a Progressive Era organization, drafted a compromise bill that retained merit rating but had some industry pooling of funds.

238

Katherine Baicker, Claudia Goldin, and Lawrence F. Katz

grounds, but ideological factors entered as Each group had the ear of an influential segment of the committees that framed the eventual document. Although these differences form an intellectual backdrop for the legislative history of UI, they also concern the essence of the bill-national versus state structure, pooled versus firm reserves, and merit rating versus none. Both Douglas and Commons played major roles in devising our current UI system, but the final bill would more clearly bear the Commons stamp. In 1933, 25 states together introduced 83 bills (39 of which were amendments). An equal number of the separate bills embraced the Douglas and Commons forms (16 each); fewer (6) espoused a compromise version. Although none passed any state legislature, bills in seven states passed at least one house (California, Connecticut, Maryland, Massachusetts, Minnesota, New York, and Ohio). Wisconsin had passed a UI bill the year before that, of course, embraced the Commons ~ystem.~’ The majority of the other seven state bills that passed one house were modeled after the Wisconsin bill. Of the eight state bills, all but two (Maryland and Ohio) called for individual employer reserves, the most extreme form of experience rating.z8All of these limited employer liability to their fund amount, thus placing the unemployed at risk of not receiving compensation. Five of the eight states limited benefits to 50 percent of the weekly wage with varying maximum benefit caps, one used a 75 percent replacement rate (California), one 40 percent (Minnesota), and one had a more complicated formula (Connecticut). Several offered benefits to those employed part time. There was considerable variation in the duration of benefits. At the lower end were Wisconsin and Massachusetts with 10 weeks and California with 13. In the middle were New York and Ohio with 16 weeks and Maryland with 20. At the upper end were Minnesota at 40 weeks and Connecticut at 46. As it became clear that an unemployment compensation act would be passed by Congress, states deliberated their own measures more seriously. Prior to passage of the SSA, signed into law on 14 August 1935, five states passed their own legislation. Apart from Wisconsin, the list includes, in chronological order of passage (all in 1935), New York (April), New Hampshire (May), California (June), and Massachusetts (August)-a Utah law was repealed and one in Washington was declared unconstitutional (Stewart 1938,28). New York‘s was the only one with a pooled fund.29 It should be transparent, by now, that there was considerable agreement and 26. See, e.g., Rubinow (1934) for a taste of the vitriol in the debate between Douglas and Commons. Rubinow and Douglas were closely allied. Some of the differences are comprehensible to us, while others are simply not. 27. See Blaustein (1993) and Industrial Relations Counselors (1934). 28. Of course, compulsory employer reserves does not constitute “social insurance” but rather is compulsory “self-insurance” by firms. 29. The California law, as passed, provided for a pooled fund but allowed firms with approved plans to opt out (Stewart 1938, 53).

239

Origins and Impact of U.S. Unemployment Compensation

uniformity in the state laws proposed and passed prior to the SSA. First, and foremost, states debated the legislation because it was believed, and with hindsight rightly so, that the legislation would be state based. Second, there was considerable agreement concerning the benefits of merit rating, and in many cases, the proposal was for the most extreme form-employer reserves. Third, there was considerable similarity in the details, such as limitation on benefits, no means testing, and the definition of eligibility. The content of these bills and the reports of the various commissions were important backdrops for discussions at the national level, which began in earnest in 1931. So too were the UI systems of the seven countries that had earlier passed national compulsory UI laws and the ten with subsidies to voluntary UI plans (Blaustein 1993, 82). The British UI system was passed along with other major social insurance in 1911, and Germany's has existed since 1927. British experience with soaring unemployment in the interwar period gave rise to a widely held view, in the United States and elsewhere, that unlimited benefits could have disastrous effects. Interwar unemployment in Britain heavily influenced the limitation on benefits adopted by every state. 7.2.2 Federal Action prior to Passage of the Social Security Act In the debate over UI at the federal level, there were many issues to be decided. Most basic was whether the system would be state or national. Next was whether there would be experience rating, and also whether the reserve fund would be state, national, employer, or industry based. Although there were many other parts of the system that varied from one draft bill to another, they were less basic and debated less intensely. A UI bill was introduced into Congress in 1934. Known as the WagnerLewis bill, much of it was modeled after the Wisconsin legislation. The proposed law provided for a state system with the states having the choice of a pooled or an employer reserve system. Portions were drafted by Paul Raushenbush, who with his wife, Elizabeth Brandeis, had been students of Commons at the University of Wisconsin. Raushenbush's father-in-law was Justice Brandeis and the justice's role in the legislation will soon be clear. The most difficult problem was how the federal government could entice the states to pass UI legislation without endangering the bill's constitutionality, for only one had done so by 1934. The answer was found in a precedent involving the federal estate tax. Justice Brandeis, or so a widely reported story maintains, devised the tax-offset plan (Blaustein 1993, 125; Schlesinger [ 19581 1988, 302). He conveyed it to his son-in-law Paul Raushenbush, who in turn incorporated it in the Wagner-Lewis bill. The tax-offset plan was ingenious and used a scheme the Court had already validated. In 1926 the federal government, in response to Florida's elimination of its estate tax through state constitutional amendment, imposed a tax on all states

240

Katherine Baicker, Claudia Goldin, and Lawrence F. Katz

but remitted credits if the state passed an estate tax. No state could then entice migration with the promise of no estate tax without paying a penalty. The law, later upheld by the Court in Florida v. Mellon (1927), provided the model for the federal unemployment compensation tax credit. Roosevelt had run in 1932 on a platform that included “unemployment and old-age insurance under State laws” and he did not wish to endanger the full legislation with the piecemeal approach represented by the proposed WagnerLewis legislation in the House and Senate. Unemployment was the gripping issue of the day, but Roosevelt had a wider vision of the future needs of the country. Passing UI, Roosevelt suspected, would be a far easier task than passing social security, old-age assistance, and other types of safety-net legislation. Thus, the answer was to bundle them in one bill. In June 1934 Roosevelt strategically asked that the Wagner-Lewis Bill be deferred and created the Committee on Economic Security (CES), with Frances Perkins as chair, to look into all aspects of social insurance. The CES staff largely consisted of those with expertise and interests in the Wisconsin system, but their ultimate decision was based more on practicality than ideology. The clearest statement of what Roosevelt wanted and how the CES made its ultimate decision comes from the recollections of Frances Perkins. The president, according to Perkins, had no favorite design for UI. As governor “he had not made up his mind” (Perkins 1946, 107), and as president he seemed open to most plans.3oDespite the fact that Roosevelt stated on 15 November 1934 that “this part of social insurance should be a cooperative federal-state undertaking,” Perkins takes the credit for writing these words that have been quoted so often since. According to her, Roosevelt had no independent view on the subject and the “question was by no means settled” in November (1946, 289).31 As the deliberations intensified, Perkins, and others on the CES, began to favor a strictly national UI system. But the constitutional matter was troubling and for good reason.32The National Industrial Recovery Act in May 1935 (Schechter Poultry Corp. v. United States) and the first Agricultural Adjustment Act (AAA) in January 1936 (United States v. Butler) would be overturned by the Supreme Court, The second child labor case (Bailey v. Drexel Furniture Co.) was lost in 1922 over the use of taxing powers that were not designed to raise revenue, the same reasoning that would shortly bring down the AAA. 30. Roosevelt was open to all except one. He opposed Harry Hopkins’s plan to institute a permanent work relief system, which the president regarded as constituting a permanent “dole” (Perkins 1946,205). 31. Altmeyer believed otherwise: “The President’s desire to rely upon the states as much as possible was based upon his lifelong belief in our federal form, rather than a national form, of government” (1966, 11). Perkins would have disagreed. 32. In Perkins’s words: “There was one outstanding, and in my mind, determining factor, at that time in favor of the federal-state system. Although we thought we were on the right track in using the taxing power of the Federal Government, we were never quite sure whether a federal system of unemployment insurance would be constitutional. The Federal Government could tax. . . . But could it distribute its funds on a basis of social benefit? The Attorney General’s office . . . repeatedly advised us that it was a doubtful constitutional point” (1946, 291).

241

Origins and Impact of U.S. Unemployment Compensation

Despite the concern, the CES, according to Perkins, decided in early December 1934 to recommend a national system.33Only after intense lobbying by the various cabinet agencies of CES members, did the CES decide, in late December 1935, to recommend the federal-state system still with us today. “The President knew of only part of this confusion,” recalled Perkins, “if we were agreed on a method, then he was for it.” “The Committee,” she believed, “must bear the responsibility for the pattern of unemployment insurance we have in this country today” (1946, 292). The CES was divided on experience rating, particularly whether firms could, by state law, have their own reserves (as they did in Wisconsin). It ultimately decided to allow states to have such reserves, and that became part of the act. But the most important issue was the federal-state structure. In the end, the pragmatists won and a state system was assured containing considerable leeway for state experimentation with experience rating, benefits, and e l i g i b i l i t ~ . ~ ~ To understand the long-run effects of UI legislation we must construct a counterfactual world. What changes would have brought about a different system in terms of the federal-state structure and experience rating? Although we cannot be certain, we suspect that passage just a year or two later, certainly a decade later, would have changed the bill considerably. The Court was a major hindrance but would not remain so for much longer. In 1937 the Court left intact the National Labor Relations Act (N.L.R.B. v. Jones and L a ~ g h l i n )By . ~ ~1941, even without packing the Court, Roosevelt had appointed seven justices, and by 1945 the entire Court had changed. A new era of “permissiveness” regarding government regulation began.36 33. The Technical Board of the CES independently decided on a national system, chiefly to ensure unimpeded geographic mobility of workers. Perkins does not mention how their deliherations affected her views. For an opposing view to Perkins’s regarding support for a national system, see the essay by Thomas H. Eliot (Eliot 1961). Eliot was a member and counsel of the CES Technical Board. 34. Directly following the CES report, a bill, which would become the SSA, was brought to congressional committee. It contained UI as a state system, with the tax offset, and with compulsory experience rating (for states to receive credit after reserves were sufficiently high). The House version of the bill removed the experience-rating feature on the grounds that it would differentiate among states and lead to the same problems that were feared before federal UI. That is, some states would have considerably lower taxes than others on particular industries or firms. The Senate version contained experience rating, as did the final bill. 35. The real break came with West Coast Hotel Co. v. Parrish (1937), which upheld a Progressive Era state minimum wage law. Although it did not directly concern New Deal legislation, West Coast Hotel has been interpreted as assuring the constitutionality of the Fair Labor Standards Act, yet to be passed. 36. According to Miller (1968) and many other constitutional historians, 1937 was a momentous turning point in Supreme Court history. The long period from Munn (1877) to Jones and Laughlin ( 1937) was one of laissez-faire conservatism and Court-dominated legislation. “Badly wounded in the [Court packing] struggle, the Court, without announcement and without fanfare, dropped its laissez-faire stop sign and erected in its place a placard of permissiveness” (Miller 1968, 79). More important than the Court-packing battle, however, was the overwhelming vote for Roosevelt in 1936. Though Roosevelt lost the Court-packing battle, an act of March 1937 eventually gave him what he wanted. It granted to Supreme Court justices over 70 years of age, with 10 years service, full

242

Katherine Baicker, Claudia Goldin, and Lawrence F. Katz

Assume for the moment that the CES had decided to back a national system, that the plan received overwhelming approval by the president, and that the Supreme Court was no threat to the legislation. Would Congress have passed it? Most histories of the period emphasize the constitutionality of the proposed legislation, but all mention, as well, the potential problems with Congress. When Perkins and others on the CES were lobbied by their colleagues after they decided in early December 1934 to back a national system, congressional approval was the major issue.37But congressional matters were always bound up with those of the Court. Several years later, when the Fair Labor Standards Act was introduced, there was never serious discussion of having the federal government compel more states to pass wage and hours laws. We will never know whether a national system could have passed Congress had the other stumbling blocks, mainly the Court, been removed, but we sense that it could have.38 Thus, it is our belief that the federal-state structure was a product of the times, particularly the threat of the Supreme But without the federalstate structure, would experience rating have been incorporated into a national law? We think not but acknowledge that this is the more difficult counterfactual. Those close to Roosevelt who preferred the national system were also individuals who did not favor experience rating, certainly not in the extreme employer reserve form. The coincidence of beliefs was no accident; it derived from economic reasoning. A state system, to many such as Douglas, would not be adequate during a major economic downturn and would hinder geographic mobility when employment shocks differed by state. Similarly, the penalties of experience rating could exacerbate unemployment if shocks differed substantially by industry or firm.Certain potential benefits from both the state ~~

pay upon retirement. Four conservative justices left by 1941 (Devanter 1937, Sutherland 1938, Butler 1939, and McReynolds 1940), although four liberal ones (Cardozo 1938, Brandeis 1939, Stone 1941, and Hughes 1941) did as well (see Eriksson 1941). 37. “There was grave doubt . . . that Congress would pass a law for a purely federal system” according to Perkins, who was recalling the results of some last-minute December interviews with members of Congress (Perkins 1946,292). 38. Another question is whether UI would have been different had the Wisconsin group supported a national system, rather than defending the components of the Wisconsin UI bill. Congress did vote on one national bill. The Lundeen bill, a Farmer-Labor proposal, received some 50 votes as an amendment to the SSA. The bill would have made UI a national program but would also have considerably broadened its scope, duration, and benefits. See Douglas (1939,82). 39. Why, then, was old-age insurance (OAI) passed at the national level? First of all, it should be remembered that old-age assistance (OAA), as well as various other SSA programs covering the poor and the infirm, were left to the states. The federal government provided incentives, but there was no coercion. OAI, however, would have been an extremely faulty system had it been left to the states, for it was a contributory program with a long time horizon for each individual. The movement of workers from state to state would have involved enormous complications for OAI (see, e.g., Eliot 1961). UI, on the other hand, was a short-run measure, as were welfare and OAA to some extent. The CES knew that the national character of OAI created potentially serious problems regarding the Supreme Court, and they crafted the tax and expenditure sections of the act with that in mind (see Miron and Weil, chap. 9 in this volume). The point is that the CES could advocate a state-level system for UI but simply could not for O M .

243

Origins and Impact of US. Unemployment Compensation

system and experience rating were not doubted by those who came to oppose them. It was the net effect that mattered. Among those who held these views, Frances Perkins was the closest to Roosevelt. She was his secretary of labor, had headed the New York State analogue when Roosevelt had been governor, and was chosen by him to head the CES. Her views carried considerable weight, and if not for the Court, she would have argued strenuously for the adoption of a national UI system without experience rating. She made forceful arguments to Roosevelt regarding experience rating when the position of the Court was still being assessed. Perkins presented her case to Roosevelt in a most personally persuasive manner. The comparison between merit rating in WC and UI was often made by the Commons group, and Perkins reminded Roosevelt of the dangers in WC incentives. “A [man] badly crippled . . . [by an industrial accident] might be put back to work on extremely low wages. He would get no compensation, and therefore the accident cost for the employer would go down. . . . The most serious defect of merit rating,” Perkins emphasized to the president, “was the refusal to employ slightly handicapped people” (Perkins 1946, 290). In summary, the UI portions of the SSA were well thought out and crafted, a triumph of pragmatism to many and of economic reason to others. Yet it can also be claimed that there is much in the final bill that was accidental, a product of the decade, of the people, of the Court, of the state bills that preceded it. That some of the features were eminently sensible for the time is clear from the historical record. In a 5-4 decision in 1937, the Supreme Court used the Florida v. Mellon precedent, just as Brandeis had predicted, to uphold the sections of the 1935 SSA dealing with unemployment compensation.40We know the act survived the Court, but how accurate were the framers concerning the impacts of the distinctive features? Merit rating, it was thought, would provide disincentives to unemployment, and the state system was touted as allowing UI to be tailored to the times and the place. 7.2.3

1935 Social Security Act and Unemployment Compensation

Better sense can be made of the impact of UI legislation if the details are spelled out. Our task is simplified because the original act left so much to the states. As passed in 1935 Titles I11 and IX of the SSA, those that dealt with unemployment compensation, called for a federal tax of 3 percent (by 1938, but beginning at 1 percent in 1936, rising to 2 percent in 1937) on all wages of employers of eight or more States that subsequently passed their 40. See National Conference of Social Welfare (1985) for the text of the Supreme Court opinions regarding the SSA, in particular Chas. C. StewardMachim Co. v. Davis (1937), which upheld the tax provisions for UI. 41. Employer size was later changed to a minimum of four and eventually to one. State law, however, often mandated fewer employees than the federal minimum. In 1945, e.g., 16 states required UI for firms with one or more employees, whereas 22 retained the federally mandated minimum of eight.

244

Katherine Baicker, Claudia Goldin, and Lawrence F. Katz

own UI legislation, approved by the Social Security Board, would receive a tax credit of 2.7 percent (0.9 of 3 percent), the 0.3 percent difference going to the federal government for overseeing the States not passing a UI law would be ineligible to draw from the fund. The tax provided a strong incentive to pass UI legislation. Benefits could be paid out beginning in 1938, when it was believed the coffers would be full enough to cope with requests for compensation. The federal act mandated few items. It was up to the states to determine benefits, duration, some aspects of eligibility, and, after some point, the degree of merit rating and the tax structure. Each state had its own fund, managed by the U.S. Treasury. When state reserves reached 7.5 percent of total covered wages the state could lower tax rates on some firms. Most important, the reduction in the tax rate would be acceptable only if the state devised an approved experience-rated system to penalize firms for unemployment. A fully experience-rated system would be akin to self-insurance by each of the firms, and incomplete experience-rated systems-even vastly incomplete onescould and did gain approval. Finally, the act implicitly disallowed the means testing of potential recipients.Although the act is not explicit on means testing, it has been interpreted in this manner:) By mid-1937 all states had passed an approved UI law. Each was guided by a set of “draft bills” containing variants of the central aspects of the legislation (U.S. Social Security Board 1937). Experience rating, for example, could be calculated numerous ways, and the draft bills outlined two possibilities for states adopting a pooled fund. One is now known as the reserve ratio approach, the most popular then and today, adopted by 29 states in 1948 and used by 33 states in 1990. (We discuss experience rating and the reserve ratio concept below.) The other possibility was general, giving states wide latitude. The benefit-wage method was adopted by nine states in 1948, although its popularity has since declined. The benefit-ratio method has become the second most popular system, although just three states chose it in 1948. Another draft bill was created for states that adopted individual employer reserve funds, as Wisconsin had done in 1932. Maximum and minimum benefit levels, eligibility, whether the accounts were pooled or firm-specific reserves, and other features 42. The use of the tax credit and the state-based system were not the only aspects of the law governed by historical precedent. The administration of the funds was located in the U.S. Treasury Department and paid for with funds collected from employers by the federal government because of a perception that WC was improperly handled by several states. The funds, moreover, are restricted to the payment of benefits and cannot be used for training, e.g. 43. As noted above, means testing is implicitly covered in Sec. 907.g of the SSA. Among other aspects of the law are (1) nine states had employee contributions at the outset (Blaustein 1993, 161) and (2) there was no federal taxable wage base stated in the act, although the base was $3,000 for social security. A federal taxable wage base for UI was added in 1940. But unlike the case of social security, the federal taxable wage base for UI has substantially eroded over time. Whereas the (Federal Unemployment Tax Act tax basdtotal covered wages) was more than 90 percent in 1940, it is less than 40 percent today (ACUC 1996). As a consequence, most states have taxable wage bases that exceed the federal level.

245

Origins and Impact of U.S. Unemployment Compensation

of the legislation differed from state to state in the initial legislation and evolved in ways that we have surveyed above. 7.3 7.3.1

The Distinctive Features: Origin, Evolution, and Impact Experience Rating: The Most Distinctive Feature of Unemployment Insurance

What Is Experience Rating and Why Was It Adopted?

Of the three distinctive features of UI, two are easily understood in historical context. The state system was largely designed to ensure passage of the act in Congress and to avoid its rejection by the Supreme Court. Some claimed that the state system would also allow much-needed innovation through state experimentation. But the same factors that demanded the federal government penalize states for not passing UI could lead them to race to the bottom when allowed to innovate. The limitation of benefits, adopted by all states, was prompted by the interwar lesson with unemployment in Britain. But what about the most distinctive aspect of U.S. unemployment insurance-its experience or merit rating by firm?44 Experience rating is defined as any tax system for UI in which firms encountering a greater amount of compensated unemployment are taxed more heavily. In practice, however, there are a large number of experience-rating systems that have been adopted by the various states. One of the most common is based on a construct called the “reserve ratio.” The reserve ratio is given by the difference between taxes paid by the firm and benefits disbursed to its employees, all divided by the payroll of a firm. The ratio is defined over some time period, often that since the firm began experience rating. All systems have a maximum tax and a minimum tax (which can be zero). When the reserve ratio of a firm changes, due say to an increase in its layoffs, it will often be subject to a different tax. A higher reserve ratio will bring about a lower tax rate, and a lower reserve ratio a higher tax rate. But there can be “flats” in the schedule, so that a larger movement in the reserve ratio would be needed to change the tax rate. A schedule with a high minimum and a low maximum will have less complete experience rating than one with the reverse. Flats, in addition, cause bunching between reserve ratios and zero experience rating in that range. A completely experience-rated system would have a minimum rate of zero and no maximum rate. It would penalize firms dollar for dollar with respect to benefits paid out to their workers. No systems have been completely experience rated, with the possible exception of Wisconsin’s employer fund system as passed in 1932. All, therefore, are incompletely experience rated to some degree. 44. It is interesting to speculate whether experience rating would have camed the day had Commons not supported it early in the history of UI and embedded it in the first state UI law.

246

Katherine Baicker, Claudia Goldin, and Lawrence F. Katz

The degree of experience rating also depends on rules concerning what changes the tax rate. States governed by reserve ratios are the simplest to describe and are the most numerous today and in the past. But there are other systems with considerably more complicated rules.45 Experience rating was championed for two reasons. One was equity. Firms that created more unemployment should pay for it. Sectors, such as construction, lumber, and food, that were inherently more seasonal should be taxed more heavily. The more complete the experience rating, the less would be the subsidy from low-unemployment to high-unemployment firms. The stabilizing effect of taxing layoffs furnished the other reason.46Experience rating would tax unemployment and thus provide an incentive for firms to smooth production over the season and the cycle. The belief that seasonality was responsible for substantial unemployment motivated both reasons for experience rating. If unemployment resulted instead from economic downturns, firms might have less control over employment. And if unemployment were more stochastic than predictable, equity considerations would argue against experience rating. Thus, seasonality played a crucial role in the formulation of experience rating in UI.47 In the discussion of experience rating, it was often claimed that because the 45. See Anderson and Meyer (1993, 1994), Card and Levine (1994), and Topel (1983) for fuller discussions of measuring experience rating across different state systems. 46. “Stabilization” can be interpreted two ways. First is the disincentive to lay off workers. Second is the fiscal effect of increasing consumer demand when there was unemployment. It was the latter, macroeconomic, reason that was foremost in the minds of the drafters of the SSA. According to Arthur Altmeyer, appointed chair of the Social Security Board in 1936, Roosevelt “stressed [the stabilization function of UI] in his initial meeting with Miss Perkins [secretary of labor], Dr. Witte [executive director of the Technical Board of the CES], and me in August 1934” (Altmeyer 1966, 25). Roosevelt also inserted a sentence in his message to Congress about the committee’s report, remarking that “federal legislation should not foreclose the States from establishing means for inducing industries to afford an even greater stabilization of employment” (Altmeyer 1966, 25). The message had been drafted by Altmeyer, who took notice of the insertion. The addition appears to refer to experience rating. 47. Douglas (Douglas and Director 1931) and Commons (Lewisohn et al. 1925) both contributed to volumes on unemployment that contain virtually identical discussions of the role of seasonality and the attempts by firms to smooth production, often through advertising campaigns to entice consumers to purchase goods more regularly over the year. Both groups of authors emphasized the role and ability of private enterprise to smooth employment over the season. Despite these similarities, their conclusions were often different. Commons and his coauthors argued for UI with employer-based experience rating to provide incentives for firms to further reduce seasonal unemployment. They recognized the role of cyclical downturns in causing unemployment, for they were writing in the wake of one of the worst economic downturns in American history. But, they noted, seasonal fluctuations were just as important, for virtually all industries “have busy and dull seasons which occur during practically the same month, year after year” (Lewisohn et al. 1925, 83). They argued against employers who contended “that they do not need additional incentive unemployment insurance to cause them to stabilize employment” (213). Douglas, mustering similar evidence and making similar points, concluded, contrary to Commons, that firms already had sufficient incentives to smooth employment and that the main function of unemployment compensation was its relief and stabilizing impact, which could be jeopardized by experience rating. Although, in his 1931 volume, Douglas did propose a UI plan with premium rates varying by unemployment, he downplayed the role of experience rating and opposed employer reserves.

247

Origins and Impact of US.Unemployment Compensation

United States was a more agriculturally based economy than most in Europe, its industry was more seasonal. Seasonality, according to many contemporaries, was at the root of the unemployment problem in the United States, apart from vastly aberrant periods such as the Great Depression. In the 1930s, the debate over experience rating and the related subject of pooled versus individual firm accounts was highly contentious. Commons and his influential associates viewed experience rating as the crowning glory of the legislation, with the capacity to reduce future unemployment. Although Douglas recognized the importance of seasonality, he believed, as we have already noted, that firms already had ample economic incentives to reduce its costs (Douglas and Director 1931)j8 But did states with more seasonality embrace merit rating to encourage firms to reduce unemployment? Or, on the other hand, did states have incomplete experience rating to subsidize certain seasonal i n d u s t r i e ~ ?The ~ ~ latter question is not a simple one because the ability of highly seasonal firms to extract a subsidy from less seasonal ones will depend on the lobbying strength of the more seasonal (and the lobbying weakness of the less). The first matter is to describe seasonality in the pre-UI period across the United States. Seasonality before and after the Adoption of Unemployment Insurance

There are many ways to capture the concept of seasonality. One is a simple peak-to-trough measure, by month or by quarter. The measure captures what is generally meant by seasonality, but it may miss fluctuations over the year. Another, and the one we will mainly rely on, captures the choppiness in the data and is the standard deviation of monthly employment growth rates (see Barsky and Miron 1989).50 We measure seasonality in manufacturing across states before and after UI. For the pre-UI period, we are able to use four years of manufacturing data (1909, 1914, 1919, and 1929), and for the post-UI period we use annual Bureau of Labor Statistics (BLS) establishment survey data for the 1949-69 peThe major difference between the pre- and post-UI data is that the pre48. Experience rating remains a controversial issue. The recent ACUC report (1 996) failed to reach a consensus on whether to weaken or strengthen experience rating. The argument for experience rating remains that of Commons. The argument against is that experience rating creates an adversarial system in which firms contest workers’ UI claims and that it creates an incentive for firms to use contingent workers who would not qualify for UI benefits. 49. See Adams (1986) for evidence to this effect for the 1970s. 50. More precisely, this measure of seasonality is the standard deviation of coefficient estimates on month dummies in a regression for each state, where the monthly change in log employment is regressed on a full set of month dummies. There are 11 monthly changes; that from December to January is not observable in the 1909-29 period because our data are for the census years 1909, 1914, 1919, and 1929. 5 1. The 1909-29 data are from the U S . census of manufactures (construction data were not collected until 1929) and give employment of wage earners (meaning production workers) by month (US. Bureau of the Census 1933b). The data for 1949-69 are from the U.S. Department of Labor, BLS LABSTAT database maintained on the Internet at http://stats.bls.gov and include both production and nonproduction workers.

248

Katherine Baicker, Claudia Goldin, and Lawrence F. Katz

04.8 - 12.7 \

I

Fig. 7.5

-

12.7 18.1

I ,

-

18.2 31.3 31.4 77.2

-

I-

Seasonality, 1909-29

Source: U.S. Bureau of the Census (1933b). Nores: The seasonality measure is the standard deviation of the monthly employment growth rate by state for the years 1904, 1909, 1919, and 1929. The numbers in the legend have been multiplied by 1,OOO. The ranges were chosen to divide the 48 states into four equal groups.

UI data contain only production workers whereas the post-U1 data contain all employees in manufacturing. To account for the difference in coverage we multiply the 1949-69 state seasonality estimates by a factor of l .31, the estimated ratio in those years of the seasonality of production to all manufacturing workers at the national level. The adjustment factor implicitly assumes that each state’s seasonality was affected to the same degree as was the nation’s.52If correct, we are comparing seasonality in the pre- and post-UI periods for production workers in manufacturing. The data for this measure of seasonality are given in the maps of figures 7.5 and 7.6, the ranges of which are the same to facilitate comparison. The greatest seasonality before 1929 was found in the Pacific, Mountain, and Plains states; the least in New England (except Maine), the Middle Atlantic, and parts of the South. States with more diversified manufacturing show less seasonality, possibly why much of Europe may have had less seasonal employment than did the American states. States with cold, harsh climates, not surprisingly, had 52. Another difference is that the pre-1930 manufacturing data include railroad repair shops. We can subtract this sector only for 1929. Railroad repair was less seasonal than the rest of manufacturing and will bias the results against a reduction in seasonality especially in the small westem states.

249

Origins and Impact of U.S. Unemployment Compensation

Fig. 7.6 Seasonality, 1949-69

Source: U.S. Department of Labor, BLS LABSTAT. Nures: The seasonality measure is the standard deviation of the monthly employment growth rate by state for all years between 1949 and 1969. The numbers in the legend have been multiplied by 1,OOO. The ranges were chosen to divide the 48 states into four equal groups in 1909-29 (see fig. 7.5).

more seasonality, as did those with greater dependence on the food (SIC 20) and timber (SIC 24) industries. In comparison with pre-1929 seasonality, that for the 1949-69 period is considerably lower.53The (unweighted) mean of our seasonality measure decreases from 0.0254 in 1909-29 to 0.0182 in 1949-69 (after applying our correction factor), about a one-third decline.54Not only did the mean of seasonality decline in the post-World War I1 era, its distribution changed. With few exceptions, seasonality in the post-World War I1 period increases uniformly when moving from east to west, whereas the geographic pattern in the pre- 1929 period is considerably more mixed. Some states shifted position between the periods. As seen in figure 7.7, Con53. For a discussion of seasonality around the turn of the century, see Engerman and Goldin (1993). 54. We compute similar changes for another measure, the log (maximum quarterlylminimum quarterly) employment. The (unweighted) mean for “quarterly” seasonality for 1909-29 is 8.65 log points but 5.49 for 1949-69. These figures are unadjusted for differences in employment coverage. National monthly data on production worker employment in manufacturing are available from the BLS starting in 1919. The standard deviation of monthly log employment growth rates, for the entire nation, declined from 0.0105 for 1919-29 to 0.0091 for 1949-69.

250

Katherine Baicker, Claudia Goldin, and Lawrence F. Katz

Change in Seasonality 1909l29 to 1949169

Fig. 7.7 Change in seasonality, 1909-29 to 1949-69 Sources: See figs. 7.5 and 7.6. Nores: The change in seasonality is the value for 1909-29 minus that for 1949-69; thus negative values are increases in seasonality, and positive values are decreases. The numbers in the legend have been multiplied by 1,OOO.

necticut, Florida, Minnesota, North Carolina, South Dakota, and Wyoming (darkest shading) actually became more seasonal. But considerably more states became less seasonal. Leading the list are Arkansas, California, Colorado, Delaware, Idaho, Iowa, Louisiana, Maryland, Mississippi, Utah, Virginia, and Washington (cross-hatching),all having among the greatest seasonality in the 1909-29 period. Fully 30 states had decreases in seasonality over the years we cover. Some of the declines, as in the South, stemmed from the growth of more diversified manufacturing. States with the most seasonality in the 1909-29 period generally had the smallest fraction of their employment in manufacturing (and construction). But there were exceptions, such as the states of the Pacific region, for which seasonality was extremely high and seasonal industries were a nontrivial fraction of total employment. Also Maryland, Delaware, Maine and some parts of the Midwest had considerable seasonality of employment and a high fraction of manufacturing and construction workers in total employment. Seasonality, therefore, was substantial prior to the passage of UI and declined subsequently. We now address whether pre-UI seasonality affected the evolution of state laws and also whether experience rating by state had an effect on seasonality.

251

Origins and Impact of U.S. Unemployment Compensation

7.3.2 Distinctive Features: Evolution and Impact Evolution of BeneJits The federal-state structure of UI allowed it to evolve differently across the various states. Such differences could emerge for reasons of efficiency or instead could reflect diverse lobbying interests and power. We investigate whether states with more seasonality in the predepression period were those that evolved relatively high benefits. Firms and workers in the more seasonal industries should have been the most interested in generous benefits (and low experience rating). But their ability to achieve high benefits will be greater when they are not too large a share of the labor force, because the more seasonal industries will want to use the UI system to transfer resources from less seasonal industries. As long as experience rating is not perfect, more seasonal industries are almost always subsidized by less seasonal ones.55 We regress various measures of the average level of state benefits from 1947 to 1969 on (1) seasonality in manufacturing employment for 1909-29 (also construction in 1929), (2) the proportion of the state labor force in manufacturing and construction in 1930, and (3) an interaction of the seasonality and labor force proportion measures.56The key independent variables were determined 30 to 50 years prior to the dependent variables. Six regressions are offered to demonstrate the range of results with different functional forms and specifications. Columns (1) and (5) use the log of the average replacement rate, whereas column (2) uses log (maximum benefits/ average weekly wage). An advantage of examining the determinants of maximum weekly benefits is that they are directly set by state legislatures. Columns (3), (4), and (6) loosen the constraint on the log (average weekly wage) coefficient and include it as a separate variable. Finally, in column (6), we add the logarithm of per capita income and the percentage unionized in the state. The results, given in table 7.1, are striking and suggestive. Benefits are higher the greater the seasonality in manufacturing employment in 1929, as long as the percentage in manufacturing and construction is below 30 percent (using col. [l]), just above the (unweighted) state mean in 1930 (25.5 percent).57And benefits are higher the greater the percentage in manufacturing and construction, as long as the seasonality measure is below 0.0213 (using 55. Many employees in highly seasonal industries are not subject to effective marginal experience rating in any state because their firms are already at the maximum tax rate. 56. We use the ratio of the average benefit to the average weekly wage in the state. The true after-tax replacement rate would be higher because UI benefits were not taxed during this time period. See ACUC (1996) for a discussion of alternative measures of replacement ratios. 57. If, e.g., seasonality increases by 1 standard deviation (a change of 0.0183) and if the proportion in manufacturing is 0.151 (1 standard deviation below the mean), then the impact will be to increase in the replacement rate by 9 percent (using the col. [1] results). Note that the construction seasonality results (col. [5]) also show that increasing seasonality in construction increases benefits but only if the share of the labor force in manufacturing and construction is below the mean in 1930.

Table 7.1

Explaining Unemployment Benefits by State, 194769

Variable Seasonality in manufacturing, 1909-29 Proportion in (manufacturing + construction), 1930 Seasonality in manufacturing X proportion in (manufacturing + construction), interaction Mean log (average weekly wage), 1947-69 Seasonality in construction, 1929 Seasonality in construction X proportion in (manufacturing + construction), interaction Log (per capita income), 1929 Percentage unionized, 1939

log (average weekly benefitdwage) (1)

log (maximum benefitdwage) (2)

log (average weekly benefits) (3)

log (maximum benefits) (4)

log (average weekly benefi ts/wage) (5)

log (average weekly benefits) (6)

10.1 (2.18) 0.725 (0.186)

7.45 (2.98) 0.499 (0.255)

10.0 (2.20) 0.75 1 (0.193)

6.18 (2.37) 0.729 (0.206)

8.12 (2.13) I .22 (0.342)

7.94 (2.03) 0.710 (0.373)

-34.0 (7.64)

-25.5 (10.4)

-33.3 (7.77) 0.933 (0.108)

- 18.1

(8.34) 0.378 (0.116)

-26.9 (7.43) 1.11 (0.433) -3.77 (2.33)

-27.0 (7.13) 0.724 (0.164) 0.701 (0.429) - 1.76

(2.26) 0.153 (0.0632) -0.226 (0.132)

Constant R2

Root MSE N

- 1.33

- 1.07

(0.0579)

(0.0791)

0.330 0.0776 48

0.125 0.106 48

-0.243 (0.0598)

-0.232 (0.0642)

0.741 0.0781 48

0.485 0.0838 48

- 1.48

(0.0784)

-1.19 (0.360)

0.453 0.0717 48

0.834 0.0656 48

Sources and Notes: Regressions are estimated using ordinary least squares. The continental 48 states are included. The seasonality measure is the standard deviation of monthly employment growth rates and is described in the text. Numbers in parentheses are standard errors. Dependent variables: Cols. (I) and ( 5 ) log (average benefitdwage): State mean of log (average state weekly benefitslaverage state weekly wage) for 1947-69; U.S. Department of Labor, Employment and Training Administration (ETA 1996). Col. (2) log (maximum benefitdwage): State mean of log (maximum state weekly benefits/average state weekly wage) for the 12 years from 1947-69 for which we have data; U.S. Department of Labor, Bureau of Economic Security (BES, various years). Maximum weekly benefits are for a worker with no dependents. Cols. (3) and (6)log (average weekly benefits): Coefficients on state dummy variables in a regression of log (average weekly benefits) on a full set of year and state dummies using state-by-year data covering 1947-69. Col. (4) log (maximum benefits): Coefficients on state dummy variables in a regression of log (maximum benefits) on a full set of year and state dummies for stateby-year data covering the 12 years from 1947-69 for which we have data. Independent variables: Seasonality in manufacturing, 1909-29: Monthly data for manufacturing during 1909-29 from U.S. Bureau of the Census (1933b). Proportion in (manufacturing + construction), 1930: Proportion of the state’s labor force in manufacturing and construction, U S . Bureau of the Census (1932). Seasonality in construction, 1929: U.S. Bureau of the Census (1933a). Mean log (average weekly wage), 1947-69: Coefficient on state dummies in a regression of log (average weekly wage) on a full set of year and state dummies for state-by-year data covering 1947-69. Average weekly wage data from U S . Department of Labor, ETA (1996). Log (per capita income), 1929: U.S. Department of Commerce, Bureau of Economic Analysis (BEA 1984, table 2). Percentage unionized, 1939: Troy and Sheflin (1985, table 7.2).

254

Katherine Baicker, Claudia Goldin, and Lawrence F. Katz

col. [ll), approximately the (unweighted) mean in 1909-29 (0.0253). States with more seasonal industries in the pre-UI period had higher benefits, as a fraction of wages, in the post-UI period but only if their manufacturing and construction sectors were not overly large in relation to total state employment. Seasonality in the construction industry, moreover, has an added effect and operates similar to the way overall seasonality in manufacturing does. When seasonality is high, benefits are higher, as long as the proportion of the state’s labor force in the combined manufacturing and construction sector is not too high. State per capita income enters positively, but unionization is weakly negative. Our findings suggest a political economy story of successful lobbying efforts by highly seasonal industries. Generosity has its limits, however. Lobbying is more effective when the costs can be spread over a sufficiently wide group of industries and workers. Impact of Experience Rating

The most distinctive feature of UI-experience rating-is embedded in various aspects of state UI laws. States were given an incentive by Title IX of the act to adopt some form of experience rating, although many had already incorporated merit rating in their UI laws. When reserves reached 7.5 percent of total payroll, a state could reduce the tax rates of some firms below the 2.7 percent rate set by law only if it adopted an approved form of experience rating.58 The tax schedules incorporating experience rating and the type of reserve system used evolved over time. The schedules are complex, often containing many flat portions, and it is difficult to construct useful summary measures of them. In assessing whether a firm will pay more if it lays off a worker, one must know the tax schedule facing the firm and where the firm is on it.59These statistics are difficult to obtain today, and we have nothing resembling them for the past. We use, instead, an admittedly crude state-level measure of experience rating to assess its long-run impact on seasonality. The measure is the “slope” of the schedule, which can be defined only for the 3 1 states adopting the reserve ratio basis. It is given by (maximum - minimum tax rate)/(reserve ratio at minimum - reserve ratio at maximum tax rate). It is a measure of the degree to which firms are taxed for an additional layoff when they are between the minimum and maximum tax rates. We use the measure to crudely assess whether the adoption of experience rating by states had some effect on the change in seasonality that we have just illustrated. The effect of UI on layoffs has been a subject of wide-ranging interest, for it addresses a central question about the economic effects of public policy. 58. This rapidly happened as unemployment plummeted in the early 1940s. 59. Anderson and Meyer (1993, 1994) show that experience-rating systems are extremely complicated and that there is considerable heterogeneity in the marginal tax cost of layoffs facing firms even within a single state system and a single industry.

255

Origins and Impact of U.S. Unemployment Compensation

Does UI encourage more layoffs because firms do not bear their full costs? Most researchers find, mainly for the post-1970s period, that incomplete experience rating increases layoffs, heightens seasonality, and shifts employment to more seasonal sectors. It is ironic, though, from the standpoint of our paper, that so much emphasis has been placed on incomplete experience rating when the most distinctive, and possibly most interesting, aspect of the U.S. unemployment insurance system is the presence of experience rating in any form. Anderson and Meyer (1994), Card and Levine (1994), Feldstein (1976, 1978),and Topel (1983,1985) all argue that imperfect experience rating leads to excessive cyclical and seasonal layoffs." In a study of the retail trade industry, Anderson (1993) finds that higher UI taxes reduce seasonal variation in employment, mainly during the Christmas boom, and raise the average level of employment in the retail trade industry.61Deere (1991) contends that the current imperfectly experience-rated UI system subsidizes unstable sectors and industries and thereby serves to expand their employment.62 Not only does UI involve cross-subsidization, but Anderson and Meyer ( 1993)also provide convincing evidence that the system has persistently subsidized the same firms and i n d ~ s t r i e sThus, . ~ ~ they argue that UI is less a system of insurance than of subsidies to unstable sectors and their workers. The subsidization, moreover, could originate in the political strength of unstable firms, industries, and their unionized workforces, as persuasively argued by Adams (1986). Rather than ask what impact incomplete experience rating has had, we inquire instead as to its impact on changes in seasonality. We employ two main tests, both admittedly imprecise due to data deficiencies for the long period we are exploring. One test is to see whether the marked change in seasonality we demonstrated above is related to the extent and completeness of experience rating across the various states. Our primary dependent variable is the difference, between 1909-29 and 1949-69, in one of our measures of seasonality (the standard deviation of monthly employment growth rates). The measure of experience rating is the slope of the tax function in reserve ratio states averaged over 12 years in the 1951-69 period. The slope is the change in the UI tax rate 60. Anderson and Meyer (1994) show that conclusions from analyses using industry-state data may be misleading but reach similar conclusions using firm-level data. Feldstein (1978), Topel (1983), and Card and Levine (1994), using Current Population Survey data matched to information on state UI laws and the individual's industry of employment, find that greater UI subsidies and lower experience rating are associated with more unemployment as the result of temporary layoffs. Temporary layoffs seem more sensitive to experience rating than are permanent layoffs. 61. Similarly, Halpin (1979) concludes that stronger experience rating reduced seasonal fluctuations in the employment of two highly seasonal industries during the 1960s and early 1970s (but not in a third). 62. But Iorio (1987) argues that, in a world of incomplete capital markets, imperfect experience rating provides insurance to riskier industries and allows their workers to smooth consumption (see Gruber 1994 for evidence on consumption smoothing). 63. Meyer and Rosenbaum (1996) find the same result for workers.

256

Katherine Baicker, Claudia Goldin, and Lawrence F. Katz

with a change in the reserve ratio and can be thought of as the degree to which an increase in unemployment increases a firm’s taxes, within the minimum and maximum rates.64We also use an experience-ratingmeasure devised by Deere (1991), the “subsidy” to firms in a state, measured as the present discounted value of the difference between UI benefits received and the increase in UI taxes owed as the result of an additional layoff. Because of the necessity of using only reserve ratio states, we have just 3 1 (sometimes 30) observations, and outliers can dominate the analysis. Figure 7.8 shows the scatterplotsfor manufacturing and construction. In the manufacturing sample California is a notable outlier, as is Nevada in the construction sample. Thus, we present both ordinary least squares (OLS) and median regression estimates in table 7.2. As is shown in table 7.2, the median regressions provide modest evidence of a negative impact of experience rating on long-run trends in ~easonality.~~ The high-low seasonality measure is more significant than is the measure employing the standard deviation of employment growth rates, and the coefficients using construction data are more significant than are those for manufacturing. In all cases, however, the coefficients are economically significant. An increase in experience rating of 1 standard deviation (0.165) decreases manufacturing seasonality by -0.00505, or about 0.5 standard deviations. Our evidence, therefore, is consistent with the microlevel data we cited above, but given the nonlinearity and complexity of the tax rate schedules, it must be interpreted with caution. A second test uses the natural experiment of two different UI systems just a border apart. Canada and the United States introduced UI systems at nearly the same time. That in the United States began paying out benefits in 1938; Canada’s system did in 1942. The systems were initially of similar generosity regarding benefits, duration, and coverage and remained so until the early 1970s.The major difference is that the U.S. system is experience rated whereas Canada’s is not. To the extent that experience rating was effective, the United States should have undergone more stabilization of seasonal employment fluctuations than did Canada from the pre-UI to the post-UI periods. To examine the hypothesis, we make some, admittedly crude, comparisons of changes in seasonal employment fluctuations in construction and manufacturing for Canada (as an aggregate) and for U.S. states on or close to the Canadian border. We use border states to control for other trends in seasonality that ought to be common to states and provinces with similar climates and, possibly, industrial structures. We aggregate monthly employment for 13 U.S. states (from west to east: 64. We ignore flats in the schedule and are aware of all the problems we previously mentioned that are raised in Anderson and Meyer (1994). Our slope measure is a simplification of the approach to measuring experience rating used by Brechling (1981). 65. Alternatively, OLS without the outliers provides similar results to the median regression.

0.02

A

Manufacturing RI

0.01

NY

gc 0.00

NC MA

NH

OR

ME

NJNM

9M

rn 0

WPA WI

NE

ND

m

a .-c Q)

d

-0.01

SD

MI

IN

W;ASCw

Q)

c

LA

IA

0,

6m

-0.02

co AR

-0.03 CA

-0.04

0.1

0.2

0.4

0.3

0.5

0.6

0.7

0.9

0.8

Experience Rating Slope

6

Construction

-0.02 -0.04

-0.06

a .- -0.08 m

g

-0.10

0

-0.12

m rn v)

LA

CA NJ MA

NY

M E RI

TN

NC

PA

NH OH

.r -0.14

KS

wv

co

IA

scn -0.16 Q)

NE

c

0 -0.16

SD

ND

-0.20

-0.22

NV

-0.24 ,

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.6

.

I

,

0.9

Experience Rating Slope

Fig. 7.8 Scatterplots of change in seasonality and experience rating slope by state: A, manufacturing, 1909-29 to 194949; B, construction, 1929 to 1949-69 Sources and Notes: The measure of seasonality is the standard deviation of the monthly employment growth rate. “Slope” is defined in the text and is computed for 1951-69. See table 7.2 notes for seasonality and “slope” sources.

Explaining Changes in Seasonality, 1920-29 to 1949-69

Table 7.2

Manufacturing

S.D. Measure OLS

Seasonality measure: Estimation technique:

(1) ~~~

Construction

S.D. Measure Median Regression (2)

High-Low Median Regression (3)

-0.0305 (0.0173)

-0.0875 (0.0167)

(5)

S.D. Measure Median Regression (6)

-0.0786 (0.0529)

-0.148 (0.0621)

0.0993 (0.0582) -0.0107 (0.00422)

-0.0693 (0.0200)

-0.0389 (0.0232)

0.0972 0.00867 29 -0.00402

0.0706 0.0479 31 -0.0960

0.09.56

S.D. Measure OLS (4)

S.D. Measure OLS

~

Experience-rating measure, slope, 1951-69 UI subsidy, 1962-69 ( 1967 dollars) X Constant

-0.0147 (0.0114)

R’ or pseudo R’ Root MSE N Mean of dependent variable

0.0547 0.0103 31 -0.0048 I

0.000 197 (0.000429)

0.0025 I (0.00359)

0.00399 (0.00694)

0.0274

0.126

31 -0.00481

31 - O.OOS06

31 -0.0960

Sources and Notes: Dependent variables: Cols. (I), (2), and (4) use the s.d. measure that is the change in the standard deviation in mean monthly employment growth rates for manufacturing from 1909-29 to 1949-69. Col. (3) uses the high-low measure, which is the difference between the log of employment in the highest and lowest quarters. Col. (5) and (6) uses the s.d. measure for construction from 1929 to 1949-69. Independent variables: Experience-rating measure, slope: US. Department of Labor, BES (various years): slope = (maxtax mintax)/(maxres - minres), where maxtax = maximum UI tax rate (including firms with negative balances); mintax minimum UI tax rate; rnaxres = maximum reserve ratio on “sloped’ part of tax schedule, which equals the lowest reserve ratio associated with mintax; minres = lowest reserve ratio on the “sloped’ part of the tax schedule, which equals highest reserve ratio associated with rnaxtax. UI subsidy, 1962-69: private communication from Donald Deere; see Deere (1991). ~

Col. (4) deletes California, which is an outlier. The coefficient of the U1 subsidy variable is 0.0537 with a standard error of 0.0694 if California is included. Numbers in parentheses are standard errors.

259

Origins and Impact of US. Unemployment Compensation

Table 7.3

Seasonality in U S . and Canadian Industry, Pre-UI to Post-UI Standard Deviation of Monthly Employment Growth Rates Pre-UI: 1929

Industry and Country Construction Canada United States Manufacturing Canada United States

(1)

Post-UI: 1947-63 (2)

0.170 0.21 1

0.085 0.063

0.0268 0.0237

0.0123 0.0079 [O.O I03J

Post

-

Pre, (2) (3)

-

(1)

-0.085 -0.148 -0.0145 -0.0158 [-0.01 341

Sources and Notes:

United States refers to the 13 states that are on or close to the Canadian border (from west to east: Washington, Idaho, Montana, North Dakota, Minnesota, Wisconsin, Michigan, Ohio, Pennsylvania, New York, Vermont, New Hampshire, and Maine). U.S. pre-UI manufacturing employment by month: U.S. Bureau of the Census (1933b). U.S. pre-UI construction employment by month: U.S. Bureau of the Census (1933a). U.S. post-UI manufacturing, construction employment by month: Department of Labor, BLS LABSTAT database maintained on the Internet at http://stats.bls.gov. Canadian pre-UI manufacturing, construction employment by month: Canada, Dominion Bureau of Statistics (1963). Canadian post-UI manufacturing, construction employment by month: Canada, Dominion Bureau of Statistics (1955-64). Standard deviation of monthly employment growth rates: see text. Figures in brackets adjust for differences in the coverage of the U S . data between in the pre- and post-UI periods. The pre-UI period covers production workers only, whereas the post-UI period covers all employees. The adjustment, derived from national data for the post-UI period on both groups, produces a ratio of seasonality for production workers to all employees equal to 1.31. Thus the figures in brackets adjust the post-UI data to approximate the coverage in the pre-UI period.

Washington, Idaho, Montana, North Dakota, Minnesota, Wisconsin, Michigan, Ohio, Pennsylvania, New York, Vermont, New Hampshire, and Maine) to create a comparison group and use two measures of employment. One compares construction in 1929 (the only pre-UI year for the U.S. data) with that in the 1947-63 post-UI period. The other looks at manufacturing in 1929 (the only pre-UI year for which we have both U.S. state-level and Canadian monthly employment data in manufacturing) versus that in 1947-63. The results of these comparisons are shown in table 7.3. A larger reduction in seasonality in construction is found for the U.S. border states than for Canada (col. [3]), consistent with there being an impact of experience rating. The difference in manufacturing is less pronounced.66Thus the bulk of our cross66. When we correct for differences in employment coverage in the pre- and post-UI U.S. data we get a smaller reduction in seasonality (in comparison with Canada) and a larger reduction if we do not use the correction factor. We are uncertain whether the correction factor is appropriate, given possible changes in employment coverage in the Canadian data.

260

Katherine Baicker, Claudia Goldin, and Lawrence F. Katz

state and cross-country analysis suggests that experience rating served to reduce employment fluctuations in highly seasonal industries. 7.4

Concluding Remarks

The UI system in the United States was established in a period of extreme and pervasive unemployment to be a program for ordinary times. The system could not have coped for long with the extent and duration of unemployment in the decade in which it was created. Its functions have changed little since its creation and an important question is whether UI is a program for the more ordinary times of today. In the 60 years since its creation, UI has served a multitude of functions. Its limited benefits have, or so many believe, shielded the United States from the high unemployment experienced in Europe in response to negative employment shocks. Although some emphasize the incomplete nature of experience rating in the U.S. system, it is far closer to a perfect merit system than is any other, and many have demonstrated the benefits of experience rating for lowering unemployment and seasonality. The main criticism of the current UI system is that it has not changed with the times. The federal-state system set up in 1935 to find a way around the Supreme Court was ingenious. But it may later have produced a race to the bottom. In comparison with other OECD countries, the U.S. unemployment insurance system has low generosity, and its eligibility rates have declined markedly since the 1960s.The original act stated that UI monies could be used only to fund benefits and program administration. But there are many who today believe that UI needs to add reemployment and retraining to its functions. We have described the origins of the distinctive features of the unemployment compensation system adopted in August 1935 as part of the omnibus SSA. Of greatest importance are its federal-state structure, experience rating, and the limitation of benefits imposed by all states. We addressed how the parameters of the UI system evolved at the state level and the impact experience rating has had on the seasonality of employment. Evidence was presented showing that more seasonality in manufacturing employment in 1909-29 is related to higher UI benefits from 1947 to 1969 if a state’s manufacturing employment share is below the national mean. Lobbying activities of seasonal industries appear important in the evolution of the parameters. We also presented suggestive evidence that there is a relationship between declining seasonality and experience rating. Had there been no Great Depression, it is likely that UI would have been passed at some future date. The bill, we believe, would have been closer to the one that Paul Douglas, Frances Perkins, and others in the Roosevelt administration favored-strictly national in scope. Without the federal-state structure, it likely that it would not have incorporated experience rating, although we are

261

Origins and Impact of U.S. Unemployment Compensation

less certain of that. Compromises had to be made by those who championed a national UI system in the exacting and uncertain 1930s. The federal-state structure and the manner in which the states were induced to adopt their own UI legislation assured passage of the act and guaranteed its constitutionality.As is the case with other pieces of New Deal legislation, the legacy of the 1930s lives on.

References Adams, James D. 1986. Equilibrium taxation and experience rating in a federal system of unemployment insurance. Journal of Public Economics 29 (February): 5 1-77. Advisory Council on Unemployment Compensation (ACUC). 1996. Dejining federal and state roles in unemployment insurance. Washington, D.C.: Advisory Council on Unemployment Compensation. Altmeyer, Arthur J. 1966. The formative years of social security. Madison: University of Wisconsin Press. Anderson, Patricia M. 1993. Linear adjustment costs and seasonal labor demand: Evidence from retail trade firms. Quarterly Journal of Economics 108 (November): 1015-42. Anderson, Patricia M., and Bruce D. Meyer. 1993. The unemployment insurance payroll tax and interindustry and interfirm subsidies. Tax Policy and the Economy 7: 111-44. . 1994. The effects of unemployment insurance taxes and benefits on layoffs using firm and individual data. NBER Working Paper no. 4960. Cambridge, Mass.: National Bureau of Economic Research. . 1996. Unemployment insurance takeup rates and the after-tax value of benefits. Hanover, N.H.: Dartmouth College, December. Unpublished paper. Baldwin, Marc, and Richard McHugh. 1992. Unprepared for recession: The erosion of state unemployment insurance coverage fostered by public policy in the 1980s. Economic Policy Institute Briefing Paper. Washington, D.C.: Economic Policy Institute. Barsky, Robert B., and Jeffrey A. Miron. 1989. The seasonal cycle and the business cycle. Journal of Political Economy 97 (June): 503-34. Becker, Joseph M. 1971. Experience rating in unemployment insurance: Competitive socialism. Baltimore: Johns Hopkins University Press. Blank, Rebecca M., and David E. Card. 1991. Recent trends in insured and uninsured unemployment: Is there an explanation? Quarterly Journal of Economics 106 (November): 1157-89. Blaustein, Saul J. 1993. Unemployment insurance in the United States: The.first half century. Kalamazoo, Mich.: Upjohn Institute. Brechling, Frank. 1981. Layoffs and unemployment insurance. In Studies in labor markets, ed. s. Rosen, 187-207. Chicago: University of Chicago Press. Canada. Dominion Bureau of Statistics. 1953. Employment, payrolls, and weekly earnings. Ottawa: Dominion Bureau of Statistics. . 1955-64. Review of employment and payrolls. Ottawa: Dominion Bureau of Statistics. Card, David, and Phillip B. Levine. 1994. Unemployment insurance taxes and the cyclical and seasonal properties of unemployment. Journal of Public Economics 53 (January): 1-29.

262

Katherine Baicker, Claudia Goldin, and Lawrence F. Katz

Corson, Walter, and Walter Nicholson. 1988. An examination of declining UI claims during the 1980s. Unemployment Insurance Occasional Paper no. 88-3. Washington, D.C.: U.S. Department of Labor. Deere, Donald R. 1991. Unemployment insurance and employment. Journal of Labor Economics 9 (October): 307-24. Douglas, Paul H. 1939. Social security in the United States: An analysis and appraisal of the Federal Social Security Act, 2d ed. New York: McGraw-Hill. Douglas, Paul H., and Aaron Director. 1931. The problem of unemployment. New York: Arno. Eliot, Thomas H. 1961. The legal background of the Social Security Act. Speech delivered at Social Security Administration Headquarters. Available at http://www.ssa. gov/history/eliot2.html. Engerman, Stanley, and Claudia Goldin. 1993. Seasonality in late nineteenth century labor markets. In Economic development in historical perspective, eds. D. Schaefer and T. Weiss. Palo Alto, Calif.: Stanford University Press. Eriksson, Erik McKinley. 1941. The Supreme Court and the New Deal: A study of recent constitutional interpretation. Los Angeles, Calif. : Lymanhouse. Feldman, Herman, and Donald M. Smith. 1939. The case for experience rating in unemployment compensation and a proposed method. New York: Industrial Relations Counselors. Feldstein, Martin S. 1976. Temporary layoffs in the theory o f unemployment. Journal of Political Economy 84 (October): 937-57. . 1978. The effect of unemployment insurance on temporary layoff unemployment. American Economic Review 68 (December): 834-46. Green, David A., and W. Craig Riddell. 1993. The economic effects of unemployment insurance in Canada: An empirical analysis of UI disentitlement. Journal of Labor Economics 11 (January, pt. 2): S96-S 147. Gruber, Jonathan. 1994. The consumption smoothing benefits of unemployment insurance. NBER Working Paper no. 4750. Cambridge, Mass.: National Bureau of Economic Research, May. Haber, William, and Merrill G. Murray. 1966. Unemployment insurance in the American economy: An historical review and analysis. Homewood, 111.: Irwin. Halpin, Terrence. 1979. The effect of unemployment insurance on seasonal fluctuations in employment. Industrial Labor Relations Review 32 (April): 353-62. Hamermesh, Daniel S., and Daniel T. Slesnick. 1995. Unemployment insurance and household welfare: Microeconomic evidence 1980-93. NBER Working Paper no. 53 15. Cambridge, Mass.: National Bureau of Economic Research, October. Industrial Relations Counselors. 1934. An historical basis for unemployment insurance. Minneapolis: University of Minnesota Press. 10150, Karl. 1987. A theory of quits and layoffs. Unpublished Ph.D. diss., University of California, Berkeley. Lester, Richard A,, and Charles V. Kidd. 1939. The case against experience rating in unemployment compensation. New York: Industrial Relations Counselors. Lewisohn, Sam A., Ernest G. Draper, John R. Commons, and Don D. Lescohier. 1925. Can business prevent unemployment? New York: Knopf. Meyer, Bruce D., and Dan T. Rosenbaum. 1996. Repeat use of unemployment insurance. NBER Working Paper no. 5423. Cambridge, Mass.: National Bureau of Economic Research, January. Miller, Arthur Selwyn. 1968. The Supreme Court and American capitalism. New York: Free Press. National Conference on Social Welfare. 1985. 50th anniversary edition, the report of the Committee on Economic Security of 1935 and other basic documents relating to

263

Origins and Impact of U.S. Unemployment Compensation

the development of the Social Security Act. Washington, D.C.: National Conference on Social Welfare. Nelson, Daniel. 1969. Unemployment insurance: The American experience, 19151935. Madison: University of Wisconsin Press. Organization for Economic Cooperation and Development (OECD). 1994. The OECD jobs study: Evidence and explanations. Pt. 2, The adjustment potential of the labour market. Paris: Organization for Economic Cooperation and Development. Perkins, Frances. 1946. The Roosevelt I knew. New York: Viking. Rubinow, I. M. 1934. The quest for security. New York: Holt. Schlesinger, Arthur M., Jr. (1958) 1988. The coming of the New Deal: The age of Roosevelt. New York Houghton Mifflin. Stewart, Bryce M. 1930. Unemployment benejits in the United States: The plans and their setting. New York: Industrial Relations Counselors. . 1938. Planning and administration of unemployment compensation in the United States: A sampling of beginnings. New York: Industrial Relations Counselors. Topel, Robert. 1983. On layoffs and unemployment insurance. American Economic Review 73 (September): 541-59. . 1985. Unemployment and unemployment insurance. Research in Labor Economics 7:91-135. Troy, Leo, and Neil Sheflin. 1985. U.S. union sourcebook: Membership,finances, structure, directory. West Orange, N.J.: Industrial Relations Data and Information Services. U.S. Bureau of the Census. 1932. Fifreenth census of the United States, 1930. Population. Occupation statistics. Washington, D.C.: Government Printing Office. , 1933a. Fifieenth census of the United States, 1930. Construction industry. Washington, D.C.: Government Printing Office. . 1933b. Fpeenth census of the United States, 1930. Manufactures, 1929. Washington, D.C.: Government Printing Office. U.S. Council of Economic Advisers (CEA). 1977. Economic report of the president: The annual economic report of the Council of Economic Advisers. Washington, D.C.: Government Printing Office. . 1996. Economic report of the president: The annual economic report of the Council of Economic Advisers. Washington, D.C.: Government Printing Office. U.S. Department of Commerce. 1993. National income and product accounts of the United States. Vol. 1, 1929-58. Washington, D.C.: Government Printing Office. . Bureau of Economic Analysis (BEA). 1984. State personal income by state: Estimates for 1929-1982. Washington, D.C.: Government Printing Office. U.S. Department of Labor, Bureau of Economic Security (BES). Various years. Comparison of state unemploymentinsurance laws. Washington, D.C.: Government Printing Office. . Bureau of Labor Statistics (BLS). LABSTAT database maintained on the Internet at http://stats.bls.gov. . Employment and Training Administration (ETA). Various years. SignGcant provisions of state unemployment insurance laws. Washington, D.C.: Government Printing Office. . 1996. ET handbook 394. Unemploymentinsurancejnancial data 1938-1994. Washington, D.C.: Government Printing Office. U.S. Social Security Board. 1937. Draft bills for state unemployment compensation. Washington, D.C.: Government Printing Office. Witte, Edwin E. 1962. The development of the Social Security Act. Madison: University of Wisconsin Press.

This Page Intentionally Left Blank

8

Spurts in Union Growth: Defining Moments and Social Processes Richard B. Freeman

The 1930s depression growth of unionism is the most studied and discussed period in U.S. labor history. Many analysts view the decade as a turning point in the development of the American labor relations system. Participation in unions exploded in the 1930s, with the sharpest increase in density occurring with the Supreme Court’s decision upholding the National Labor Relations (Wagner) Act in 1937 (see fig. 8.1, below). The Wagner Act, modified by the Taft-Hartley Act (1947) and ensuing legislation, continues to provide the legal framework for establishing unions in the United States. The 1930s growth of unionism was also associated with the development of industrial unions and the formation of a new national grouping of unions, the Congress of Industrial Organizations (CIO). In explaining the successful unionization of the bluecollar workforce in the 1930s, labor historians place great emphasis on the Wagner Act, the dramatic events surrounding the formation of the CIO, the development of the industrial union, and the specific battles and personalities of the era-John L. Lewis, Philip Murray, Walter Reuther, President Roosevelt, among others. It was an era of legends. Was the depression growth of U.S. unionism a unique event in trade union history or was it part and parcel of the normal pattern of union development? Is there an underlying social dynamic behind the growth in unionization? Did the 1930s leave an institutional legacy that makes this period a defining moment in the history of U S . labor relations; and, if so, what is that legacy? Because the expansion of unionism in a single country in a single period leaves much scope for interpretation and only limited possibility for testing interpretations (where do we find the appropriate counterfactual to any given Richard B. Freeman holds the Herbert Ascherman Chair in Economics at Harvard University. He is also director of the Labor Studies Program at the National Bureau of Economic Research and executive director of the Programme in Discontinuous Economics at the Centre for Economic Performance of the London School of Economics.

265

I I I

b I

= -w.

i

J

I

I

OE6L

EE61

0961

E961

OL61

! EP61

0961

G861 0861

OZ61

E68 L 068 1

E88 1 088 1

7

-$ %

e 3

8

0

267

Spurts in Union Growth: Defining Moments and Social Processes

interpretation?), I consider the U.S. depression-era experience in light of developments in other countries and in other time periods. I find that unionism generally grows in discontinuous spurts and that the period of the Great Depression was one of union growth in many countries. This leads me toward an explanation of the depression experience in terms of a general theory of employer-employee conflict over organization as opposed to an explanation rooted in specific events and personalities in the depression and United States. Specific events and people ignite processes and potentially impel those processes in particular directions at formative times, but the key to understanding union growth lies in the endogenous social process, not in the historical details, which vary from period to period and country to country. As to the legacy of the depression spurt, the 1960s-to-1990s decline of private sector union density to predepression levels belies the view, prevalent 20 or 30 years ago, that the New Deal established a stable system of collective bargaining. The legacy of the depression-era spurt is quite different: an institutional framework for establishing unions and collective bargaining that has become outmoded from the vantage point of workers, unions, and firms.

8.1 Spurts of Unionism: The Quantitative Record Statistics on union membership for the United States and other countries show that trade union growth takes the form of discontinuous “spurts” rather than gradual logistic growth to some equilibrium value. The expansion of unionism in the depression era was an exceptionally large spurt but one with parallels in U.S. history and in other countries as well.

8.1.1 The American Experience Figure 8.1 shows the pattern of union density in the United States from 1880 through 1995. For most of the period, unions were concentrated in the private sector, so the density figures largely represent that sector. In the 1960s, however, public sector unionism grew substantially, but the overall pattern is still dominated by developments in the private sector. “Optimetrics” shows five spurts in union density: 1880-86, 1897-1904, 1916-21, 1934-39, and 194245. Whether we should divide the depression-era growth of unionism into the 1934-39 and 1942-45 spurts is questionable. The Bureau of Labor Statistics (BLS) series on union membership, on which I rely, shows a break, but the series of union membership reported by Troy and Sheflin (1985) does not. Commons et al. (1966), Dunlop (1948), and most analysts differentiate between the two periods, and I shall do so also. In addition, there is a spurt in public sector unionism from about 1962 to 1972 or so. Going beyond the visual picture, the notion and treatment of spurts in union growth can be developed further. By a spurt I mean a sharp concentrated episode of union growth. In such an episode, membership should grow more rapidly in a few contiguous years than would be expected by any model of random

Richard B. Freeman

268 Table 8.1

Changes in Union Density during Spurts and Other Periods, 1880-1995 Years When Density Grew

Density

Year Spurts 1883-1 986 1899-1904 1916-21 1934-39 1942-45

All

No. of Years

Average Change in Density

0.4 0.3 0.4 0.4 1.o 0.5 0.5 0.5

6 6 7 2 40 17 23

-1.3 -0.8 -1.2 -1.8 -0.6 -0.7 -0.6

0.5

61

-0.8

Initial

Final

No. of Years

2.8 4.9 9.6 11.5 25

9.9 11.9 17.4 27.6 34.2

3 5 5 5 3

2.4 1.4 I .6 3.2 3.1

21

2.2

All Nonspurts 1880-83 1886-99 1904-16 192 1-34 1939-42 1945-95 1945-70 1970-95

Average Change in Density

1.7 9.9 11.9 17.4 27.6 34.2 34.2 26.4

2.8 4.9 9.6 11.5 25 14.9 26.4 14 33

Years When Density Fell

Source: Tabulated from union density statistics in table 8A.2. Nora: Average change in density in a spurt period is simply the average annual change in the specified years. Average change in density in a nonspurt period is the average annual change in the years when density grew and when density fell. The “all” figures are the averages of the change in density in the specified spurt and nonspurt years.

fluctuations in growth, even with some autocorrelation in the rate of growth. In addition, for spurts to be growth phenomena, the pattern of increases in union density should differ from the pattern of decreases in union density during periods of decline. The statistical measures of changes in U.S. union density in private sector spurts and in “nonspurt” periods shown in table 8.1 indicate that the designated periods meet these criteria. The change in density is larger in the spurt periods (an average annual gain of 2.2 percentage points) than in the nonspurt periods when density increased (an average annual gain of 0.5 percentage points) or than in nonspurt periods when density declined (an average annual loss of 0.8). The change is also highly concentrated: gains in union density are more likely to be clumped together than decreases in density and are more highly correlated than are decreases. The optimetrics reading of the data is not spurious: spurts in growth are real. By contrast, absent the periods of spurt, U.S. union history is characterized by gradual erosion of union density

269

Spurts in Union Growth: Defining Moments and Social Processes

How does the 1934-39 depression spurt look in this context? It had the largest growth of density and of density per year of any spurt. In contrast to the other spurts, it occurred in a peacetime period of high unemployment. It was associated with innovations in labor laws and in the nature of union organizations that created a “new unionism” in the form of industrial unions. Earlier spurts were associated with either substantive legal changes (during World Wars I and I1 the government encouraged settlement of disputes) or new forms of unionism (the 1897-1904 spurt was related to the formation of the American Federation of Labor [AFL]; the increase in unionism in the 1880s was due in large part to the growth of the Knights of Labor), but not with both. Finally, the 1934-39 spurt was followed by the 1942-45 spurt associated with World War 11. This arguably brought union strength to levels that would never have been achieved or maintained simply from the 1930s spurt. In fact, between 1939 and 1942 union density fell (though membership increased), suggesting that at least some of the 1930s gain in density would have eroded absent World war 11. There are two possible ways in which aggregate unionism can increase rapidly in a spurt. Existing unions with given jurisdictions could greatly increase their representation. The Carpenters’ Union might, for example, successfully organize carpenters in nonunion areas. Alternatively, new or existing unions could expand into previously nonunion sectors. Extant data on unionization by industry are limited before the 1950s, but figures on membership for particular unions provide a disaggregated picture of the spurt process by organization. Before the depression expansion, most unions were organized on a craft basis, with explicit jurisdictions; and the pre-CIO industrial unions, such as the Miners, also had relatively narrow jurisdictions. Thus, expansion of existing unions would indicate organization of sectors where unions were traditionally concentrated. Creation of new unions would indicate union growth in traditionally nonunion occupations or industries. Using data from Troy and Sheflin (1985), I have examined the growth of individual unions during the four spurts identified in figure 8.1 and table 8.1, and during relevant nonspurt periods. The results of this analysis are summarized in table 8.2. Because I do not have information on employment within each union’s jurisdiction, I report the growth in terms of percentage changes in absolute membership. While this overstates growth in union density in relevant jurisdictions during periods of economic upswing, the rates of change in union membership during the various spurts are so great that they invariably imply large changes in density. The table shows, in any case, that in a spurt nearly every union experiences membership gains. However, in the 1930s, a larger proportion of the growth occurred though the formation of new unions than in previous spurts. This helps explain the greater magnitude of that spurt than of previous spurts. The ubiquity of growth among existing unions and the formation of new unions during the depression suggests a sea change in labor-management relations that made this spurt something different. In fact, it is this sea change in the labor market that labor

Table 8.2

Percentage Growth of Union Membership in Selected Unions during Spurts

Union Automobile Boilermakers Bricklayers Carpenters ClothingRextile Communication UE (United Electrical Workers) IBEW (International Brotherhood of Electrical Workers) Operating Engineers Food and Commercial ILGWU (International Ladies' Garment Workers' Union) GlassK'otteryK'lastic Hotel and Restaurant Ironworkers Laborers (Hod Carriers) Longshore Machinists Maintenance of Way Musicians Operating Engineers Painters Plumbers Railway Clerks Retail and Wholesale Rubber Service Employees Sheetmetal Workers Steelworkers Teamsters Transit (Street and Motor) Transport Workers UFCW (United Food and Commercial Workers) UMW (United Mine Workers) Utility Workers Woodworkers

1891-1 904

1916-2 I

-

-

111 55 361

212 11

1934-39

1942-45

113 41 22 88

-45

51

-

0 162b 4 138'

294

122 21

13

900 949

52 31 345d

135 6

50 1,333 1O F

5 -3 55 44 131 132 385 24 50 34 -6 2,689

259 215 161 261 83 92 133 29 66 12 45 42

-

54 172

-

-

-

-

661

-

122 308 -

118 1,856 756 240 540' -

66

39 41

216 12

31

635

42 6 -9 -56 1 31 14 21 -2 9 65 44 11 71 7 25 65 14 81 88 51

20

-1

- 16

-

-

-

-

-

183 25

1,218 -

2,9419 934

-

1,329 1,370

4 51 21

40 -

-

-

-

-

-

-

-

-

-

25 25

Sources: Troy and Sheflin (1985, app. B) and Troy (1965, tables A1 and A2). Nutes: A dash means that the data were unavailable, usually because the union did not exist during

the observation period. The Troy and Sheflin (1985) data differ somewhat from Troy (1965) for the same time periods. When available, I used the Troy and Sheflin data. "Union was newly created during this time period. bNumbers reported are for the period 1940-46. 'Numbers reported are for the period 1939-46. dNumbers reported are for the period 1909-16. 'Numbers reported are for the period 1901-03. 'Numbers reported are for the period 1900-04. YNumbers reported are for the period 1899-1 904.

271

Spurts in Union Growth: Defining Moments and Social Processes

historians have stressed in their descriptions of the period (Bernstein 1971; Galenson 1960). Finally, note that the spurt not covered in the table, involving public sector workers in 1962-72, has properties similar to those of the spurts shown in table 8.2: rapid expansion of unionism in new areas with new or changed organizations (the conversion of employee associations such as the National Education Association [NEA] into unions) and growth of existing unions under the aegis of legal changes (Freeman 1986). 8.1.2 The Experience of Other Countries Are spurts in union growth unique to the United States, or do they characterize other countries as well? Given the historically confrontational labor relations of the United States, we might expect to see less jumpy patterns of union growth in other countries. In some countries, union growth has in fact been less discontinuous than in the United States, but in most advanced capitalist countries we find a pattern similar to the U.S. experience: union growth taking the form of sharp concentrated spurts in membership. Even more striking, union growth spurts have occurred during roughly the same periods in most countries. For instance, union density rose sharply in the United States, United Kingdom, Germany, Sweden, Denmark, Norway, and Australia around the early 1900s. Density also grew during the World War I period in all countries for which’data exist. Table 8.3 provides a rough picture of the time pattern of the growth or decline of union density across countries that highlights the similarity in the timing of spurts. The number of countries in the table differs from period to period due to differing availability of union membership data. There is information on 7 countries in the 1900s, 12 countries until the rnid-l930s/World War I1 period when Germany and Austria are excluded because of the Nazi suppression of free trade unions, and an increasing number of countries thereafter. The most striking pattern in the data is the similarity in most periods of union growth across countries that enables me to label the periods as times of spurt, stability, or decline. Union membership shows spurts in all of the countries for which I have data in the 1900s, in World War I, and in the mid-l930s/World War I1 period. For instance, in the United Kingdom, density jumped between 1910 and 1913 (15 percent to 23 percent), between 1916 and 1920 (26 percent to 45 percent), between 1939 and 1943 (32 percent to 40 percent), and between 1945 and 1948 (39 percent to 45 percent). It grew in the majority of countries in the 1970s, with the United States being an outstanding exception. In other periods, density was stable or falling. Declines in density occurred in nearly all of the countries in the 1920s and in a majority of countries in the 1980s to mid-1990s. Density was stable in 1910-16 and in the 1950s and 1960s. I have given broad time groupings in the table and have, in particular, compressed the mid-l930s/World War I1 period into one interval, because the depression and war occurred at different times among the countries. The table is

272

Richard B. Freeman

Table 8.3

A Century of Change in Unionization in the Developed World

Period 1900s 1910-1 6 World War I 1920s Mid-1930slWorld War I1 1950s 1960s 1970s 1980s/1990s

Qualitative Characterization Increasing union density Stability Spurt in density Fall in density Spurt in density Stability Stability Growth with diversity Fall in density

Mean Change in Density

No. of Countries That Fit Qualitative Characterization

9 0 19 -9 19 -2

7 of 7 rise“ 6 rise; 6 fallb 12 of 12 rise 10 of 12 falF 10 of 10 rised 10 of 14 modest changes‘ 12 of 14 modest changes‘ 13 of 18 rise 13 of 18 fall

0 5

-6

Sources: Visser (1989, 1992). Bain and Price (1980), Schneider (1991). and OECD (1991). Notes: Countries covered are Austria, Denmark, Finland, France, Germany, Ireland, Italy, Japan, Luxembourg, Netherlands, Norway, Sweden, Switzerland, United Kingdom, United States, Canada, Australia, and New Zealand. Mean change in density is the nnweighted average change in density for the countries over the relevant period. “1900s: The seven countries are United States, United Kingdom, Germany, Sweden, Denmark, Norway, and Australia. 1910s: Additional five countries are Canada, France, Italy, Netherlands, and Switzerland. ‘1920s: Exceptions are Sweden and Australia. d1930s/World War 11: Missing countries are Germany and Austria. ‘1950s: Norway rises; Italy, France, and Japan have sharp falls. ‘1 960s: Italy rises; Switzerland falls.

“stylized” because I have used different-sized time windows for the periods to summarize the pattern, but the mean change in density is a “hard’ statistic: the unweighted average of changes in density for the countries in the part of the period when they had their spurts. Table 8.4 focuses on union density in the depressionlworld War I1 period (with somewhat different years shown for the countries to capture the timing of their particular spurts). The table shows that union density rose prior to the war in several countries, as well as in the United States, though in some cases the increase was less concentrated in a few years (the depression experience of Denmark, in particular, is misleading as Danish density began trending up earlier). In the United Kingdom, density jumped from 22.6 percent in 1933 to 33.1 percent in 1940; in France, density went from 7.0 percent in 1935 to 35 percent in 1937; in Norway, density increased from 13 percent in 1927 to 43 percent in 1939; and so on. The similarity in the pattern suggests that unionization during the depression did not reflect country-specificpolitical events but rather a more fundamental response of employees to the depression situation. In sum, I conclude that data on union density for the United States and other countries show that unionism generally grows in sudden sharp spurts and that the depression spurt of unionism in the United States is not an isolated country phenomenon but rather exemplifies a general pattern in this seemingly inhospi-

273

Spurts in Union Growth: Defining Moments and Social Processes

Table 8.4

Union Density in the 1930s in Western Countries Density (Date)

Country

Before Spurt

At Peak Spurt

Average Annual Change per Year

Australia Canada Denmark France Netherlands Norway Sweden Switzerland United Kingdom United States

34.9 (1933) 17.6 (1936) 36.2 (1929) 7.0 (1935) 26.3 (1928) 13.2 (1927) 37.9 (1933) 21.1 (1928) 22.6 (1933) 11.9 (1934)

38.8 (1939) 22.2 (1938) 46.6 (1939) 25.4 (1937) 43.0 (1932) 42.9 (1939) 53.7 (1939) 27.7 (1932) 33.1 (1940) 28.6 (1939)

0.7 2.3 1 .o 9.2 4.2 2.5 2.6 1.7 1.5 3.3

Sources: Bain and Price (1980) and Visser (1989). aDenmark had fairly steady growth before and after the depression.

table period for unionism. This regularity raises two questions about the growth of unionization. Is there is a single social process underlying growth by spurts, and if so, what is that process? Why does union growth occur in roughly similar periods across countries, including the depression when one might have expected unions to be particularly weak? The former question relates to the dynamics of growth. The latter relates to its timing.

8.2 Why Spurts? Two types of models can generate spurts in union growth. The first are models in which the process of growth creates nonlinearities that produce “phase transitions” when certain conditions are met-models of tipping, contagion, self-organized complexity. The second are standard comparative statics linear models in which massive shocks or environmental changes generate commensurately large responses in otherwise stable union membership. The former models stress the underlying process by which organization occurs and the cumulative behavior of individual workers, unions, and firms. The focus is on the behavior of thousands or millions of individuals acting in response to one another. The latter models stress the exogenous shock, usually generated by political forces. Historians of unionism in particular countries generally interpret the growth of unionism in their country as resulting from political “topdown” changes: unions grew in the United States because of the Wagner Act, in France because of the Blum government, in Canada because of the enactment of PC 1003, and so on. Without denying the importance of particular laws or events as catalysts or triggers for the growth process, I lay out in this section a model in which sudden sharp changes in union density-discontinuitiesarise from the process of organization. My model concentrates on two aspects of unionization: the conflict between

274

Richard B. Freeman

management and labor in the formation of unions; and nonlinearities in the benefits, costs, and strengths that accrue to the two sides from different levels of unionization. The model makes union formation an outcome of a battle between management and labor, rather than of the “laboratory voting” procedure the National Labor Relations Board (NLRB) envisages. Prior to the Wagner Act, organization was largely a matter of strategic workers using economic muscle to force employers to accept a union at their workplaces (Dunlop 1948) per the battle motif. Absent the strike threat by workers to impose economic costs on firms, management might refuse to recognize a union even if virtually all workers supported it. The result was a large number of often bitter recognition strikes during many union spurts. A major goal of the Wagner Act was to transform this process into a secret ballot election campaign. The model generates nonlinearity in organizing because the resources unions or employers bring to the campaign and the incentives that they have to use their resources to organize or oppose organization depend critically on the percentage already organized or covered by collective bargaining in the relevant market. When union density is low, unions have little strength to organize new firms and firms have a large incentive to oppose organization. As union density rises, unions have increased resources to use for organizing and firms see less competitive disadvantage in being organized. Thus, at some range of union density, union organizing strength rises and employer opposition falls, potentially producing a spurt in membership. The backbone of this model is an accounting identity for union density:

(1)

UNION, = (1 - r)*UNION,-,

+

NEW,,

where UNION is union density in a given product market in a particular year, defined as the ratio of union members to employment in that sector; r is the “normal” rate of depreciation of that density due, say, to attrition of members or firms or the growth of employment in new firms that have not yet been organized; and NEW is the ratio of new members organized to employment in the given year. I define UNION as union density in a product market because the model focuses on the effects of unionization on firms that compete in the same area, and on the ability of collective bargaining to create a “level playing field” in terms of a single package of wage, benefit, and conditions in that sector that can reduce initial employer opposition to unions. The key to the model is the rate of new organization, which depends on the resources that labor and management devote to organizing or opposing organization:

(2)

NEW = f(ORG, OPP) ,

where ORG is the resources unions give to organizing and OPP is the resources management gives to opposing union drives.

275

Spurts in Union Growth: Defining Moments and Social Processes

The functionfis the key element of the model. As written, it is a production function of sorts, with df/dORG > 0 and af/aOPP < 0. Beyond this, it is difficult to specify functional form. One possibility is that f has diminishing marginal productivity in both ORG and OPP, so that, conditional on employer opposition, more organizing effort will yield diminished growth of membership and, conditional on organizing effort, employer opposition will also have diminishing returns. But it is also possible that ORG or OPP has increasing marginal productivity over some range: a massive union drive might pay off much more than a smaller drive; employer opposition around a single banner-the American plan-might be more successful in defeating union organizing than efforts by individual employers. To close the model requires equations for ORG and OPP. ORG presumably depends on existing union resources, the potential benefits to existing members from spending their dues money on organizing new members, the benefits to nonunion workers of unionization, and so on. One could readily imagine ORG resulting from some form of union optimizing behavior. What my model requires is that at very low levels of unionization, ORG will be small: the union simply does not have the resources to devote to organizing campaigns. I expect that ORG is also low at high levels of unionization: at high density the union will have effective control over the market for labor, and existing members will gain little from expanding membership. If union benefits spill over to nonunion workers through “threat effects,” new workers may also gain little. For simplicity, I postulate that ORG rises, then falls more or less parabolically, with UNION. The OPP relation depends on firms’ estimates of the benefits and costs of operating union or nonunion, their assessment of the likely success of efforts to defeat union organizing drives, and the resources they have to combat the union. OPP is likely to be high when UNION is low: if the union is too weak to establish a level playing field in a sector, firms that are organized risk competitive disadvantage by paying higher wages or benefits. OPP may or may not be high when UNION is high: a nonunion firm in a primarily union market may find that to maintain this status, it must operate by union rules and pay union or higher wages or spend considerable resources fighting organizing drives, reducing the incentive to oppose union organizing efforts. On the other hand, if the firm can remain nonunion and avoid sharing economic rents with workers, it may be able to earn exceptionally high profits. As the analytics require only a single nonlinearity in the relation between union density and the resources devoted to the process of gaining new members, I assume for simplicity that employer opposition is simply a declining function of UNION. The result is a nonlinear difference equation that readily generates sharp jumps in union density.’ Figure 8.2 captures the essential nonlinearity. In panel A of figure 8.2 there are two stable union equilibria, 0 and U*, and one unstable 1 . In the simple case where ORG= a . UNION-, - b . (UNION_,)ZandOPP= 6 - c . UNION-,, the difference equation is second order.

A

Organizing Resources (ORG) and Opposing Resources (OPP)

I I

I I I

I I I I

OPP

I I

I

I

0

(stable equilibrium) B

U' (unstable equilibrium)

Union Density

I I

U* (stable equilibrium)

Organizing Resources (ORG) and Opposing Resources (OPP)

Density (initial density)

(new density when ORG rises to ORG*)

Fig. 8.2 Nonlinearity and spurts: A , stable and unstable equilibria in union spurt model; B, discontinuousjump in density when organizing resources rise

277

Spurts in Union Growth: Defining Moments and Social Processes

equilibrium, U‘. Union density grows whenever ORG exceeds OPP (more properly, when NEW =f(ORG, OPP) > r * UNION-,). In panel B of the figure I assume that the initial unionization point is 0 (or some other low level). Beginning at 0, a gradual increase in the union-employee desire for organization, which shifts upward the ORG curve, or a gradual decrease in employer opposition to unionization, which shifts downward the OPP curve, has no effect on union density until a critical point is reached: the point where ORG(0) > OPP(0). In panel B the increase in the organizing function to ORG* creates this situation. The result is a sudden spurt in density to the equilibrium U*. As it stands, this model predicts both sudden spurts and sharp declines in union density. But in the United States (and many other countries), it is difficult to displace unions in workplaces where they exist. The Wagner Act made that particularly difficult by requiring a decertification process so that a firm could end a union relationship only when workers voted to decertify the union, though a firm with strong bargaining power could effectively eliminate unions by refusing to come to a collective bargaining agreement and bringing in replacement workers to take the jobs of union members if the union struck. For the United States, though, I assume that shifts in ORG or OPP that reduce density do not have such a discontinuous effect and make the maximum loss of density in a given period r * UNION-,. Thus, the nonlinearity of the model transforms gradual changes in the underlying desire for or opposition to unionism into jumps to a new equilibrium on the growth side but produces gradual drops in density on the decline side. The point of this exercise is not to derive “the” union growth or decline equations applicable to all institutional settings or time periods but rather to highlight the potential for producing endogenous spurts when union and firms battle over organizing drives, given the likely relation between their allocation of resources to organizing or opposing activity and extant density. The key condition for growth spurts rather than gradual growth of unionization is confrontation over the union institution. The logic of the model suggests that in the early phase of a spurt there will be considerable conflict but that this conflict will diminish as the incentive for firms to oppose unions falls with higher levels of density, and that unionism will be concentrated in selected sectors as opposed to being evenly spread across sectors in the workforce. It makes product markets critical in analyzing union growth. To go further, it is necessary to specify the factors likely to shift the ORG and OPP schedules over time in a particular setting. For the United States in the depression period, these factors include the New Deal labor policies, notably the Wagner Act that most labor historians stress; the attitudes of workers, dependent on their assessment of the likely benefits and costs of unionization; the resources and incentives facing firms; and the resources and incentives of unions, including innovations in union structure, such as the formation of the CIO under John L. Lewis. The magnitude of the depression and consequent loss of belief in business leadership offers a potentially strong candidate for

278

Richard B. Freeman

raising worker desire for unionization, which in the framework given above shifts the ORG curve. In other time periods, such as the World War I or World War I1 spurts, economic booms might have reduced employer opposition, making unionization easier, even absent a change in ORG. For other countries, one may come up with a similar list of incentives to governments, employees, and firms and unions that arguably shifted over time. Since institutions affect compensation and benefit packages, moreover, the same economic changes might alter the benefits and costs to different parties differently in different countries. For example, in a European country with extension of collective bargaining contracts (from the organized sector to nonunionized firms), the incentive of employers to oppose unions will be much less than in a country with plant- or firm-level bargaining. But the task of this paper is not to explain unionization around the world but to use that and other evidence to cast light on the most dramatic period of U.S. union history.

8.3 The Depression and World War I1 Spurts in the United States The 1934-39 depression spurt and the World War I1 spurt raised U.S. union density to unprecedented levels and seemed, through the 1960s at least, to make unions a permanent and accepted part of American society. During both spurts there was a substantial increase in density in private sector industries outside of the services but little growth of density in the service sector or government (table 8.5). The World War I1 spurt fits an overall pattern in which unionism grows in war periods (recall the growth throughout the advanced world in World War I), as governments seek to maximize production by minimizing labor disputes. Most labor experts would predict a growth of unionism in such a time period. But the depression-era spurt came as a shock to the experts of its period. In 1932 George Barnett, president of the American Economic Association (AEA) and an expert in industrial relations, declared: “American trade unionism is slowly being limited in influence by changes which destroy the basis on which it is erected. . . . I see no reason to believe that American trade unionism will . . . become in the next decade a more potent social influence” (1933,6). Perhaps in no other period of American history has the growth of unionism seemed so surprising. There are two interpretations of the causes of the 1930s spurt in the United States, each focused on the possible catalytic role of particular social actors. The first interpretation, which I will call the top-down hypofhesis, is that the 1930s growth resulted from the decisions of the Roosevelt administration, the passage of prolabor New Deal legislation, and the formation of the CIO by a group of innovative union leaders led by John L. Lewis. Many analysts adhere to the view that a sympathetic administration in Washington was necessary, and perhaps even sufficient, to spark the growth of unions. They place great weight on the enactment of the National Industrial Relations Act (NIRA) and the Wagner Act. Some have gone so far as to call the Wagner Act “the most

279 Table 8.5

Spurts in Union Growth: Defining Moments and Social Processes Union Density of Industries during the Depression and World War I1 spurts Year

Industry Manufacturing= Manufacturingb Metals and machinery Clothing Food, drink, and tobacco Paper, printing, and publishing Leather and leather products Chemicals, rubber, clay, glass, and stone Textiles Lumber and lumber products Transportation, communication, and utilities Railway transportation Building and construction Mining, quarrying, and oil Government Services

1933

1935

1939

1947

13.6 11.0 51.6 5.5

22.8 23.1 24.5 53.8 24.1

40.1 40.2 49.6 60.9 29.6

20.5 23.1

28.7 16.7

38.2 41.4

4.4 1.5 10.0

13.2 7.1 11.1

26.3 30.2 16.0

50.0 58.3 77.3 65.4 10.5 6.0

70.8

-

27.9 55.3 71.7 8.8 2.8

89.0 69.3 11.6 9.2

Sources: Bain and Price (1980) and U.S. Department of Commerce (1966). ”This line is from Bain and Price (1980, table 3.3). hTheselines are obtained by dividing union membership from Bain and Price (1980, table 3.4) by employment data from the U S . Department of Commerce (1966).

radical piece of legislation ever enacted” (Klare 1978, 265). Bemstein (1971) devoted much of his prologue chapter to Roosevelt and his administration because he believed that the coming of the New Deal was the key event in American labor history. Taft claimed that “the change in the future of American labor which took place in 1933 was almost entirely due to the legislative measures . . . of FDR’ (1964, 416). Goldfield (1989) attributes similar views to many other analysts: Bums (1956), Derber and Young (19S7), Schlesinger (19S8), Leuchtenberg (1963), and Freidel (1952). The act aside, when unions can declare that “the President wants you to join” (as in one CIO organizing poster), surely it had to matter somewhat. A variant of the top-down hypothesis is that the leadership provided by John L. Lewis in forming the CIO was also a critical component in growth. According to Taft, “The CIO was largely the creation of John L. Lewis without whose leadership and financing the movement would have foundered and expired” (1964, xx). Taft’s story is basically that if you removed FDR and Lewis, nothing much would have happened on the union front during the depression. The second interpretation is quite different. It is that the catalyst for the

280

Richard B. Freeman

1930s spurt was the grievances of employees and their loss of faith in business leadershlp. Government policies and union leadership were endogenous responses to the changed views of employees. Since the spark for union growth emanates from workers, I call this the bottom-up hypothesis. Horace Davis’s theory of union growth exemplifies this view: “When labor has major grievances and an improving position in the labor market, unions tend to grow” (1941,623). He argues that the grievance-loss of jobs or risk thereof-was exceptional during the depression and that the direction of movement in the economy, rather than the level, was critical in allowing unions to develop. In this view, unionism can grow during an economic recovery, even if the recovery is still associated with high joblessness. This hypothesis stresses the activities of workers, firms, and unions operating as individual agents, whose collective action determines unionization and influences governmental policy. Absent FDR and the Wagner Act and John L. Lewis, there still would have been substantial growth of unionization in the late 1930s, according to this hypothesis. There are problems in differentiatingbetween these explanations.The 1930s spurt occurred during the New Deal, was accompanied by the CIO split, and occurred during a period of recovery from economic recession. One could legitimately argue that the spurt began when a weighted sum of contributing factors, Cw,X,, exceeded some critical value, where Ximeasures one of the factors and w,is its effect on unionization. In figure 8.2, ORG - OPP would depend on this sum. In this case, the issue of deciding between the two interpretations is not one of accepting either the top-down story or the bottom-up story but of partitioning the weighted sum to determine which set of factors was more important. Such a calculation would presumably give some weight to all factors and thus lead to a multicausal explanation of the spurt that would not reject either hypothesis. But the counterfactualfor determining what would happen absent an observed X should not be a simple ceteris paribus thought experiment in which one removes X and holds everything else the same. Rather, the counterfactual requires a more complex thought experiment that assesses how the remaining factors might adjust to the change. Perhaps absent one factor, other causal factors would have increased sufficiently to keep the weighted sum above the critical value. In the case of the 1930s depression spurt, had Congress failed to enact the Wagner Act, might unions have expanded more through traditional recognition strikes? Absent John L. Lewis and the CIO, would AFL unions have eventually responded to the opportunities afforded by worker discontent in mass production industries, or might some other entrepreneurial union leader have stepped forward to take greater initiative than the AFL had shown up to that point? Or might another form of worker organization have come forth to unionize the mass production industries? To assess the top-down and bottom-up interpretationsof the depression -era spurt, I have examined three aspects of the unionization drive of the period: the extent to which unions formed through NLRB elections versus the recognition

281

Spurts in Union Growth: Defining Moments and Social Processes

strikes the legislation was designed to supplant, the extent to which the new CIO unions were in fact aided by the CIO, and the extent to which the old AFL unions responded to the new situation. If the top-down analysis is correct, I would expect (1) that recognition strikes would not continue to be a major mode of unionization once the benefits of the secret ballot process were established, (2) that the bulk of the new unions in the CIO would have relied on central funds or organizing assistance, and (3) that the AFL unions, which lacked the dynamic leadership of Lewis, and which were less favored by the New Deal than the CJO, would gain fewer members than the CIO unions. Data relating to all three of these “tests” reject the top-down hypothesis in favor of an explanation of the depression-era spurt that places greater weight on the independent activities of workers and local unions in a bottom-up organizing effort. From this perspective the Wagner Act was less an exogenous change in regime that caused the union spurt and more an endogenous outcome of unionizing pressures-a government means for channeling worker desires for unionism and employer opposition into a less violent and confrontational mode for determining organization and collective bargaining arrangements. This does not mean that the act had no effect on the events of the period, but that absent the act or in the presence of, say, a weaker labor law, there still would have been a major union spurt in the depression period. Consider first the evidence on recognition strikes. Despite the creation of the NLRB and “laboratory” elections for workers to choose whether or not they wanted to unionize, a huge number of workers were organized in the 1930s and through the war years by recognition strikes. In 1934, when Roosevelt ordered the pre-Wagner Act Labor Board to conduct bargaining elections, there were 562 recognition strikes involving over 700,000 workers. Table 8.6 shows that the number of workers involved in recognition strikes varied in ensuing years, trending downward as the National Labor Relations Act procedures become increasingly accepted but still remaining high as late as 1937, when the Supreme Court declared the Wagner Act constitutional.2 More surprising is the fact that recognition strikes were also an important method of organizing in 1941, 1944, and 1945. Overall, more workers were organized through recognition strikes during the 1934-39 spurt than were organized through NLRB elections: 1.8 million through recognition strikes versus 1.0 million via NLRB elections. This suggests that the act may have changed the nature or process of union organization after 1937 more than it changed the actual number organized.A legitimate interpretation of the data is that the election procedure largely substituted for recognition strikes that would have cre2. The timing of the legal changes deserves some attention. On 16 June 1933, the NIRA was signed into law. The first code, with a statement on unionization, was given in July. In 1934 Roosevelt ordered the Labor Board to conduct bargaining elections, and they held 528 elections with approximately 30,000 votes. On 27 May 1935, the Supreme Court declared the NIRA unconstitutional in the Schechter Poultry Co. decision. The Congress passed the WagnerAct on 27 July 1935. The act was upheld by the U.S. Supreme Court in N.L.R.B. v. Jones and Loughlin (April 1937).

282

Richard B. Freeman

Table 8.6

Number of Workers Organized by Way of Recognition Strikes and by

NRLB Elections

Year

No. of Recognition Strikes

No. of Workers in Recognition Strikes

1934 1935 1936 1937 1938 1939 1940 1941 1942 I943 1944 1945

562 560 809 2,200 867 885 767 1,466 684 244 389 592

701,101 202.118 272,O13 941,802 110,398 132,034 71,054 444,551 59,876 7 1,764 213,387 436,500

~

No. of Workers Organized by Recognition Strikes (A)

No. of Recognition Elections

554,755 163,513 225,498 71 1,060 86,104 99,817 58,051 337,868 47,906 57,411 29,875 103,320

528 163 708 949 69 1,880 3,390 4,182 4,432 4,815 5,253

No. of Workers in Election Units Voting Union (B)

~~~

Percentage of Workers Unionized by Recognition Strikes (AMA + B)

30,000

95

74,000 262,000 286,000 373,000 657,000 1,001,000 1.281,000 1,253,000 1,084,000 773,000

75 73 24 21 9 25 4 4 3 12

Source: Recognition strike data from the U S . Department of Labor, BLS, Monthly Labor Review (Washington, D.C., May 1936-May 1946,January 1986). Notes: 1 estimated the number organized by recognition strikes by multiplying the number reported involved in those strikes by the percentage of workers the BLS reported as being involved in recognition strikes where the outcome was substantial or partial gains to the union. Election statistics are from NLRB, Annual Report (Washington, D.C., June 1936-June 1946), with the monthly NLRB data adjusted from a fiscal year basis to be on a calendar year comparable to the strike data. From 1936 to 1945 the NLRB reported the number of workers eligible to vote and the number who voted union but not the number in elections where the union won. In a term paper (Giebisch 1979). Robert Giebisch estimated that the number of workers in units won by the union was 15 percent higher than the number who voted union. The number of workers in election units voting union is based on this adjustment.

ated a comparable growth of unionism in a more confrontational way. If this is the case, the claim that the depression-based spurt was bottom-up driven rather than top-down driven is enhanced. How many workers might have successfully unionized through 1939 in the absence of an election procedure? A minimal estimate would be the number who in fact organized through recognition strikes: the 1.8 million workers organized through recognition strikes is 33 percent of the 1934-39 growth in union membership of 5.5 million reported in table 8A.2. Many workers, of course, joined existing unions so that membership rose absent either strikes or NLRB elections. If this growth was simply proportionate to the growth of nonagricultural employment, membership would have risen by 18 percent from 1934 to 1939, or by 900,000 persons. I will not try to predict how many persons would have joined unions over the 1937-45 period absent the Wagner Act but will simply note that the growth in the latter period was comparable to that during World War I, with no New Deal legislation.

283

Spurts in Union Growth: Defining Moments and Social Processes

Consider next the evidence on the growth of CIO unions. How much of CIO membership growth resulted largely from the activities of workers and autonomous unions and how much from top-down central CIO assistance? Two of the most important organizing drives in the 1930s and 1940s were in automobiles and steel, both highly oligopolized sectors with a large dominant firm whose unionization virtually guaranteed success in the sector. The automobile workers, employing the sit-down strike, forced General Motors to accept their organization after numerous abortive efforts. But the United Auto Workers (UAW) did not rely intensively on CIO monies for its organizing success or on Lewis or other CIO leadership for its operations. The history of the union is a story of workers trying to organize, subject to generally bungled efforts by the AFL to provide leadership and aided only modestly by the CIO leadership. The organization of the steelworkers was quite different. In steel, the CIO set up and funded the Steelworkers Organizing Committee (SWOC), and the United Mine Workers provided organizers and money as well as Philip Murray, the Steelworkers Union’s first president. Much of the growth, however, came by gaining the support of the company unions that the steel firms had set up to buffer themselves from independent unions, which makes it clear that there was an important bottom-up character to this drive as well. Still, the great success of the SWOC came with the negotiations between Lewis and the head of the U.S. Steel Corporation, Myron Taylor-a closed-door meeting that organized the largest company without a strike. The Steelworkers is the prime example of a union that fits the top-down model. To what extent did the organization of workers in CIO unions follow the UAW bottom-up pattern as opposed to the Steelworkers top-down pattern? One way to answer this question is to categorize CIO unions according to the financial and organizing resources that the federation gave them and to contrast the membership or growth of membership in those unions over the relevant time period. Table 8.7 presents data on the amount of money the CIO gave to various unions from 1935 to 1941 and gives the dollars per member of each union in 1937. While financial support is not the sole indicator of CIO effort, it is an important measure of how involved the central federation was in particular organizing campaigns. The table shows wide variation in the absolute amount of money given and in the amount given per member. Consistent with the history of unionization in autos and steel, the UAW received relatively little money while the Steelworkers obtained a lot. Table 8.8 contrasts the 1942 membership in the various CIO unions according to the amount of CIO support given per member in 1937. I treat the unions receiving less than $2.00 per 1937 member as organized independently of the CIO, those receiving more than $5.00 per member as being dependent on the CIO, and unions receiving between $2.00 and $5.00 as being in an intermediate group. A classification of this type, based on a single indicator, is rough but, as noted in the auto and steel cases, my classification is generally consistent with the histories of particular

Table 8.7

Assistance Given by the CIO to National Affiliates, 1935-41, and Estimated Assistance per Member in the Affiliates as of 1937

Union Aluminum Workers Architects Automobile Barbers Cannery Communication Construction Die Casters Distillers Electrical and Radio Farm Equipment Federal Flat Glassblowers Furniture Inland Boatmen Iron, Steel, and Tin Longshore Marine Engineers Marine and Shipbuilding Maritime Mine, Mill, and Smelter Newspaper Guild Office and Professional Oil Optical Packinghouse Workers Stone Radio and Telegraph Retail and Department Rubber Shoe State,Municipal, and City Steel Studio Technicians Textiles Toy and Novelty UMW, District 50 Utility Woodworkers

Total Assistance (thousand $)

30 44

60

3 88 68 313 13 30 52 72 108 14 54 9 20 29 1 52

49 62 41 66 117 4 93 1

33 121 23 82 134 1,019 4 133 53 110 76 85

Assistance per Member

($1 1.17 12.09 .31 6.00 6.18 125.20 2.60 1.68 1.45 33.75 .82 4.25 2.90 1.03 1.16

.80 2.67 1.67 3.18 3.59 7.59 6.92 2.50

2.94 .20 66.00 3.02 .74 4.77 23.10 8.15 1.53 1.97 11.04 12.09 9.50 4.36

Sources: Total assistance from Galenson (1960, table 26); for assistance per member, I divided

total assistance by union membership in 1937 from Troy (1965, table A-2). When Troy reported no membership in a given year, I have taken the membership in the nearest available year.

285

Spurts in Union Growth: Defining Moments and Social Processes

Table 8.8 Dollars of Aid per 1937 Member, 1935-41 Less than $2 $2-$5 More than $5

Membership in CIO Unions in 1942, According to Dependence on CIO for Financial Assistance No. of Unions

Membership in 1942

Major Unions in Group

15 10 12

1,055,000 379,000 584,600

UAW, Electrical, Rubber Textile, Marine and Shipbuilding Steel, Construction District 50

Notes: Calculated from table 8.7 and membership data from Troy (1965, table A-2). Includes only unions given some central supprt. Thus, ACW, UMW, and ILGWU, among others are excluded from the data. When a union was not affiliated with the CIO (UMW, District 50, left with the UMW), I used the most recent CIO affiliation year.

unions. The striking fact is that in 1942 the CIO was not dominated by unions whose formation and growth had depended greatly on CIO financial and organizing support. There are more than twice as many members in the unions that received relatively little CIO financial support than in those that received considerable support. One reason is that the CIO gave considerable money to some organizing drives, such as for construction workers, that had relatively little success (and were motivated less by the desire of workers to join CIO unions than by John L. Lewis’s desire to create trouble for the AFL craft unions). The most dramatic histories of the depression spurt focus on the growth of the new CIO industrial unions, but a major component of union expansion in the period was the AFL response to the new competing federation (Galenson 1960). Indeed, the absolute growth of membership in AFL unions exceeded the growth in CIO unions, even if we count the CIO as having no members prior to 1937 (although some founding organizations like the Miners were already large). Between 1935 and 1942 the AFL gained 2.7 million members compared with a gain of 2.5 million for the CIO (see table 8.9). Several AFL unions made particularly large gains in the period: the Machinists, which had opposed industrial unionism in internal AFX debates, but which transformed itself into “one of the great mass production unions in the country” (Galenson 1960, 141), in part to c y p e t e with the UAW in areas like airplane production; the Teamsters, which debeloped regional conferences to create a multiindustry general transport workers’ union; the Carpenters (whose head, Bill Hutcheson, was the man Lewis punched when he quit the AFL to form the CIO); the Railway Clerks; and the Building Laborers; though membership grew in other unions as well. To some extent, AFL unions were galvanized to fight for membership by rival CIO unions that threatened their jurisdictions rather than by the opportunities created by the Wagner Act election procedure per se. But they were also galvanized by worker desires to unionize in response to the depression conditions. The problem for the AFL unions was to get their “act in

286 Table 8.9

Richard B. Freeman Membership in the AFL and CIO, 193542 (thousands)

Union

1935

1936

1937

1938

1939

1940

1941

1942

AFL Machinists Teamsters Retail clerks Bakers Carpenters Electrical Hotel Other building trades Boilermakers Bridge and iron Operating engineers Painters Plumbers

3,218 98 162 12 21 129 57 82

3,516 116 188 18 25 150 67 110

3,180 167 353 24 30 209 94 194

3,547 171 394 30 56 215 107 187

3,878 178 442 51 69 215 125 211

4,343 207 47 8 74 81 233 146 226

5,179 313 599 83 83 357 202 269

6,073 489 566 80 91 517 277 237

15 13 35 64 35

16 17 35 75 37

18 32 42 93 44

28 42 42 101 54

29 39 58 103 59

33 41 64 114 62

43 63 97 131 81

90 94 138 127 114

1,991

1,958

1,838

2,154

2,654

2,493

CIO

-

-

Source: Troy (1965, table A-I).

gear” to take advantage of a market opportunity created more by worker desires than by government legislation. Only three AFL (later CIO) unions exploited the opportunity first created by the Roosevelt administration’s NIRA to build or rebuild their membership base: the United Mine Workers (UMW), the Amalgamated Clothing Workers (ACW), and the International Ladies’ Garment Workers’ Union (ILGWU). All three unions operated in competitive industries. All had experienced periods of great growth followed by collapses throughout their histories. From 1927 to 1933 in particular, the UMW, then the largest exemplar of industrial unionism in the United States, had suffered massive losses of membership. The success of each of these unions depended critically on establishing some form of national or at least regional wage pattern, so that firms that signed with the union were not driven out of business by nonunion competitors. Thus, the NIRA gave them a particular institutional setting through which to develop sectoral agreements. The UMW accomplished this in part through bargaining with the bituminous coal manufacturers’ federation and effectively helping them to oligopolize the industry. The ACW and ILGWU did the same in the major apparel-producing areas in the North but failed to extend their organizations to the South. Most of the successful new CIO unions, by contrast, were in industries dominated by a few large employers, often located in a single area: steel, autos, rubber. The great success of the AFL construction unions in the late 1930sreflected their monopoly of skilled crafts workers in particular localities, where organization depended largely on the activities of local unionists. In sum, to understand the development of unionism during the Great Depression, it is more useful to think of an endogenous bottom-up response op-

287

Spurts in Union Growth: Defining Moments and Social Processes

erating during a period of great change in worker desire for unionism than of a legalistic top-down unionization drive. 8.4 Conclusion: What Does a Defining Period Define? For a particular event to be “defining,” it must lock in certain outcomes that persist into some future period when, given a blank slate, the society could have developed something very different. Historians and labor analysts writing in 1950s and 1960s believed that the depression spurt and the Wagner Act were such events, producing a substantial and stable collective bargaining system for the United States. According to Fleming, the Wagner Act “unquestionably contributed enormously to the growth of a large and independent labor movement . . . and to the . . . acceptance of that movement as a desirable part of a modern American society” (1957, 149). Bernstein wrote that “American labor history took an eventful turn with the coming of the New Deal” (197 1, ix). In 1964 Taft contrasted the labor movement “of today” with the “one that existed in the early 1930s:’ noting that “its numbers are about six times as great, and the level upon which it operates its legal, political, and research activities has been greatly expanded” (1964,708). In 1997 the effect of the depression and World War I1 spurts on union density and the U.S. labor relations system appears quite different. Private sector union density has dropped to the levels of the mid- 1900s. The social accord that made unions a part of national decision making has broken down. An increasing number of employers seek the “union-free’’ environment that only retrograde right-wing ideologues once sought. Many firms that do not espouse the unionfree world act as if that is what they truly want when their own workforce seeks to organize. The depression and World War I1 growth of unionism thus looks more like a diversion from American “exceptionalism”-a long and important diversion, but a diversion nonetheless-rather than a critical turning point in labor relations. This does not, however, mean that the period did not leave a lasting legacy. What remains to this day is the nation’s legal framework for conducting private sector labor relations: a national labor code, based on the Wagner Act as amended in ensuing years. This is a framework that is arguably outmoded. One major purpose of the act-to encourage collective bargaining-has failed: the proportion of nonagricultural workers in unions is below what it was in the five depression years prior to its enactment. From the union side, the Wagner Act turned the process of unionization into a legalistic business, in which fimns and union organizers battle before the NLRB and courts as part of the election process. The protections that the act gave to workers who want to unionize have de facto been eliminated as the number of unfair practices committed by employers per election has risen sharply. Never envisaging a world in which upward of 35 percent of nonmanagerial workers would supervise others, where white-collar managerial and professional jobs would constitute the dominant

288

Richard B. Freeman

occupations, the act, with its Taft-Hartley amendments, fails to provide any place for intermediate organizations-staff associations, works councils, or more pejoratively company unions-that some of these workers may prefer to the stark “collective bargaining or nothing” choice that the Wagner Act offers. The labor relations code provides no place for unions of supervisors, professionals, or managers. From the employer side, the act makes it illegal for firms to set up and support worker organizations that might give some legitimate voice to workers who want some organization at their workplace short of an independent union that bargains collectively (the “company union”). Perhaps most important, the NLRB representation procedure has not reduced the confrontation between management and employees over unionization but has simply transferred it from one setting to another. By contrast, state regulation of public sector labor relations has produced what the Wagner Act has not-a stable collective bargaining system in that part of the economy with much less confrontation in the organizing process. As the laws governing unionization in the public and private sectors are reasonably similar, this difference in outcomes is probably due to the smaller incentive and willingness of managers in the public sector than of managers in the private sector to fight union drives. Public sector managers are not motivated by corporate profit to oppose unionism. My view that the Wagner Act has locked the United States into an outmoded labor relations framework that does not fit labor market realities as the country moves into the twenty-first century is not an isolated one. Many labor experts concur that the Wagner Act framework no longer fits the U.S. economy, though they often disagree about the specifics of labor law reform, as can be seen in the differing views given by labor, management, and independent scholars before the 1993-94 Commission on the Future of Worker-ManagementRelations (U.S. Department of Labor 1994a, 1994b). Workers, firms, union membership, and the economy have changed greatly since the 1930s. But, despite several amendments to the Wagner Act, the basic structure of the law has not changed, creating an institutional straitjacket that helps neither U.S. workers, nor firms, nor unions, but one that has proved difficult to change, given the fears of labor and management that any shifts in the law will tilt the balance of power against their side. The lesson I draw from the depression spurt is that these fears are probably ill placed. No plausible “labor law reform” is likely to induce a burst of unionism in the United States. The lesson from the depression experience is that bottom-up employee-driven bursts of union activity rather than particular laws are necessary for any resurgence of union density. Another lesson is that any such resurgence of unionism will come suddenly, probably surprising the current crop of experts and labor historians as much as the depression spurt surprised Barnett and other observers of the period.

289

Spurts in Union Growth: Defining Moments and Social Processes

Appendix Union Membership and Density Estimates There are alternative estimates of union membership for the United States that cover the depression era (and earlier) through the 1990s. These include Troy’s (1965), Troy and Sheflin’s (1985), and those of the BLS (reported in diverse publications such as various statistical abstracts and BLS bulletins, such as U S . Department of Labor, BLS 1980). Galenson (1960,584-87) reports different numbers for various unions in the depression period and describes some of the problems of determining membership at a time when unions were competing. He also contrasts “his” estimates with those of Wolman. There are differences between convention strength and membership claims, differences between members in good standing and dues-paying members, that produce wide variation. Even the head of the CIO, Philip Murray, was uncertain about membership in 1939, when per capita payments gave a membership estimate of 1.7 million: “He indicated that many affiliates were not paying their per capita fees to the CIO, and that actual membership might be as high as 3,000,000’ (Galenson 1960, 585). The contemporaneous BLS estimate for 1939 was 4 million. Given these differences, it is not surprising that while all extant estimated series show a sharp increase in membership from 1935 through 1939 and from 1941 or 1942 through 1945, the timing and magnitude of the changes does differ, in some cases for reasons that are unclear. In table 8A.1, I report five different union membership series for 1933-48 that show the range of variation in these estimates. The differences among the series are largest in 1936 Union Membership (thousands)

Table 8A.1

Year

BLS

Troy

Troy and Sheflin

1933 1934 1935 1936 1937 1938 I939 1940 1941 1942 I943 1944 1945 1946 1947 1948

2,857 3,249 3,728 4,164 7,218 8,265 8,980 8,944 10,489 10,762 13,642 14,621 14,796 14,974 15,414 15,000

2,973 3,609 3,753 4,107 5,780 6,080 6,556 7,282 8,698 10,200 11,812 12,628 12,562 13,263 14,595 I 5,020

3,659 4,164 3,794 4,3 I6 5,923 6,193 6,708 7,524 9,017 10,569 12,103 12,605 12,728 13,515 14,694 14,953

Galenson

Wolman

4,164 5,080 5,944 6,680 6,669 8,339

4,075 6,334 7,342 7,735 8,101 8.614

290

Richard B. Freeman

and 1937, when the BLS data show a much greater spurt in membership than the Troy, Troy and Sheflin, or Galenson series does. The Wolman series shows a large increase in membership for 1936-37 but still falls short of the change in membership in the BLS data. The differences in the series imply somewhat different timing and magnitudes for the depression-era spurt, but all still show sizable gains in membership. The BLS series also shows a larger increase in 1942-43 than the Troy or Troy and Sheflin series. But since the Troy and Troy and Sheflin data are roughly comparable to the BLS data in 1948, they show larger estimated increases in union membership from 1944 to 1948 than does the BLS series. Why do the series differ? The BLS and Troy data differ presumably because the Troy figures are based on dues-paying membership while the BLS data are derived from union reports, which may include persons who do not pay dues, particularly in 1937. Troy and Sheflin are based on the earlier Troy figures, with adjustments from fiscal year to calendar year, but this does not readily explain some of the differences between these series, particularly in 1933 and 1934. The Galenson series uses AFL and independent union data together with Philip Murray’s estimates for the CIO, which may have understated even duespaying membership, as individual unions sought to keep more of the dues for themselves. There is no strong reason to prefer the BLS or Troy or Troy and Sheflin data for analyzing union growth in the United States. But because I am uneasy about the unexplained changes between the Troy estimates and the Troy and Sheflin estimates and have a mild preference for measures that reflect the broadest possible membership in unions to those limited to annual dues-paying members, as the former may give a better indication of changes in periods of rapid growth, I base my estimates on the BLS series spliced to other series. Use of any of the other series or variants of the particular splicing that I chose will not, however, affect the basic findings or analysis. The estimates reported in table 8A.2 provide one continuous union membership series from 1880 to 1995, together with a single nonagricultural employment series, from which I derive a density series. The union membership figures refer to U.S. union members only. To obtain the figures, I spliced together three different series. The 1995 numbers come from U.S. Department of Labor, BLS (1996). For the period 1983-94, I use the numbers reported from the Current Population Survey (CPS) in Bureau of National Affairs (1995, table I), with the 1981 number based on the reported percentage of wage and salary workers who were union members. As there are no numbers for 1982, I estimated membership in that year by assuming the change between 1981 and 1983 was proportionate to the change given in Troy and Sheflin (1985, table 3-10). For 1978-80, I use the CPS numbers reported in Bureau of National Affairs (1983, table 1). For 1930-77, I use the series reported by the U.S. Department of Labor, BLS (1980), spliced to be consistent with the CPS at the 1978 overlap year between the BLS and CPS series. For the period 1897-1929

New Estimates of Union Membership and Union Density in the United States, 1880-1995

Table 8A.2

Year

Nonagricultural Employment (thousands)

Union Density

(%I

1912 1913 1914 1915 1916 1917 1918 1919

149 167 253 278 343 427 1,060 828 612 627 722 753 697 703 762 614 539 544 624 730 1,028 1,265 1,477 1,982 2,094 2,001 1,895 2,02 1 2,000 1,895 1,993 2,153 2,213 2,436 2,393 2,315 2.46 1 2,69 1 3,045 3,658

9,284 9,520 9,757 9,993 10,229 10,465 10,702 10,938 11,174 11,411 11,980 12,386 12,956 12,684 11,926 13,010 12,956 13,498 13,552 14,988 15,178 16,294 17,395 17,858 17,640 18,707 20,069 20,523 19,259 2 1,203 2 1,697 22,093 23,191 24,143 23,190 23,149 25,510 25,802 26,432 27,270

1.61 1.76 2.59 2.79 3.36 4.08 9.90 7.57 5.48 5.49 6.03 6.08 5.38 5.55 6.39 4.72 4.16 4.03 4.60 4.87 6.78 7.77 8.49 11.10 11.87 10.70 9.44 9.85 10.38 8.94 9.18 9.75 9.54 10.09 10.32 10.00 9.65 10.43 11.52 13.41

1920 1921 1922 1923 1924

4551 4,269 3,571 3,28 I 3,209

27,434 24,542 26,616 29.23 1 28,577

16.59 17.40 13.42 11.22 11.23

1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 I893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 I907 1908 1909 1910 1911

(contiiiued)

Estimated Membership (thousands)

Table SA.2

(continued) Nonagricultural Employment (thousands)

Union Density

Year

Estimated Membership (thousands)

1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 I944 1945 1946 1947 1948 1949 1950 195 1 1952 1953 1954 1955 1956 1957 1958 1959

3,224 3,248 3,255 3,225 3,277 3,284 3,196 2,945 2,596 2,982 3,460 3,85 1 6,760 7,757 8,461 8,416 9,849 10,022 12,757 13,658 13,828 13,899 14,277 13,825 13,790 13,715 14,962 15,344 16,364 16,435 16,223 16,887 16,770 16,442 16,527

29,75 1 30,599 30,48 1 30,539 31,339 29,424 26,649 23,628 23,711 25,953 27,053 29,082 3 1,026 29,209 30,618 32,376 36,554 40,125 42,452 41,883 40,394 41,674 43,881 44,891 43,778 45,222 47,849 48,825 50,232 49,022 50,675 52,408 52,894 51,368 53,297

10.84 10.61 10.68 10.56 10.46 11.16 11.99 12.46 10.95 11.49 12.79 13.24 21.79 26.56 27.63 26.00 26.94 24.98 30.05 32.61 34.23 33.35 32.54 30.80 3 1.50 30.46 31.27 3 1.43 32.58 33.53 32.01 32.22 31.71 32.01 31.01

1960 1961 I962 1963 1964 1965 1966 1967 1968 1969 1970

16,46 1

54,203 53,989 55,515 56,602 58,156 60,444 63,901 65,803 67,897 70,384 70,880

30.37 29.16 28.85 28.19 27.96 27.63 27.11 26.95 26.90 26.11 26.40

15,741 16,014 15,954 16,260 16,703 17,322 17,734 18,264 18,380 18,713

(%)

293

Spurts in Union Growth: Defining Moments and Social Processes

Table 8A.2

(continued) Nonagricultural Employment (thousands)

Union Density

Year

Estimated Membership (thousands)

1971 1972 1973 1974 1975 I976 1977 1978 1979 1980 1981 I982 1983 1984 1985 1986 1987 I988 1989 I990 1991 1992 1993 1994 1995

18,549 18,765 19,167 19,503 18,935 18,957 19,016 19,548 20,986 20,095 19,507 18,558 17,717 17,340 16,996 16,975 16,913 17,002 16,961 16,740 16,568 16,390 16,598 16,740 16,360

71,211 73,675 76,790 78,265 76,945 79,382 82,471 86,697 89,823 90,406 91,152 89,544 90,152 94,408 97,387 99,344 101,958 105,210 107,895 109,419 108,256 108,604 110,730 1 14,034 116,609

26.05 25.47 24.96 24.92 24.6 1 23.88 23.06 22.55 23.36 22.23 21.40 20.73 19.65 18.37 17.45 17.09 16.59 16.16 15.72 15.30 15.30 15.09 14.99 14.68 14.03

(%I

I use the BLS series reported in U.S. Bureau of the Census (1960, series D735), adjusted so that the data refer to U.S. members of unions on the basis of the U.S. proportion of labor union membership in 1930 in series D-741 and D742. This eliminates Canadian members of U.S. unions. For the period 18801913 I used the series of union membership reported by Gerald Friedman (1995), spliced to the BLS series (itself adjusted to be on a CPS-comparable basis) at the overlap year 1914. The nonagricultural employment series is derived from several sources as well. For 1966-95, I used the Economic Report of the President (Council of Economic Advisers 1996, table B-42). For 1930-65, I used the data in U.S. Department of Commerce (1966, series A-88). For 1900-1929, I used the data in the U.S. Department of Commerce (1966, series A-87). For 1889-99, I used series A70 from the same volume; those data are an index of man-hours in nonagricultural industries, and I applied the index numbers to the 1900 nonagricultural employment data in series A-87. Finally, as there are no annual nonagricultural employment series prior to 1889, I estimated employment by assuming that it grew proportionate to population, as reported in series A-106.

294

Richard B. Freeman

References Bain, G. S., and Robert Price. 1980. Projles of union growth. Oxford: Blackwell. Bamett, George E. 1933. Presidential address to the 45th annual meeting of the American Economic Association, Cincinnati, Ohio, 29 December 1932. Published as “American trade unionism and social insurance,” in American Economic Review 23, no. 1 (March): 1-15. Bemstein, Irving. 1971. The turbulent years. New York: Houghton-Mifflin. Bureau of National Affairs. 1983. Directory of U.S. labor organizations, 1982-83 edition. Washington, D.C.: Bureau of National Affairs. . 1995. Union membership and earnings data book 1994. Washington, D.C.: Bureau of National Affairs. Bums, James MacGregor. 1956. Roosevelt: The lion and the fox. New York: Harcourt Brace Jovanovich. Commons, John R., David J. Saposs, Helen L. Sumner, E. B. Mittelman, H. E. Hoagland, John B. Andrews, and Selig Perlman. 1966. History of labour in the United States. New York: A. M. Kelley. Council of Economic Advisers. 1996. Economic report of the president 1996. Washington, D.C.: Government Printing Office. Davis, Horace B. 1941. The theory of union growth. Quarterly Journal of Economics 55 (August): 611-37. Derber, Milton, and Edwin Young, eds. 1957. Labor and the New Deal. Madison: University of Wisconsin Press. Dunlop, John T. 1948. The development of labor organization: A theoretical framework. In Insights into labor issues, ed. Richard A. Lester and Joseph Shister, 163-93. New York: Macmillan. Fleming, Robben Wright. 1957. The significance of the Wagner Act. In Labor and the New Deal, ed. Milton Derber and Edwin Young. Madison: University of Wisconsin Press. Freeman, Richard B. 1986. Unionism comes to the public sector. Journal ofEconomic Literature 24 (March): 41-86. Freidel, Frank. 1952. Franklin D. Roosevelt. Boston: Little, Brown. Friedman, Gerald. 1995. New estimates of union membership, the United States, 18801914. University of Massachusetts at Amherst, October. Mimeograph. Galenson, Walter. 1960. The CIO challenge to the AFL. Cambridge, Mass.: Harvard University Press. Giebisch, Robert. 1979. The impact of the Wagner Act on trade union growth. Paper prepared for Economics 1650, Harvard University, Cambridge, Mass., 1 January. Goldfield, M. 1987. The decline of organized labor in the United States. Chicago: University of Chicago Press. . 1989. Worker insurgency, labor, radical organization, and New Deal labor legislation. American Political Science Review 83, no. 4 (December): 1257-82. Klare, Karl. 1978. Judicial deradicalization of the Wagner Act. Minnesota Law Review 62 (March): 265-339. Leuchtenberg, William. 1963. Franklin D. Roosevelt and the New Deal. New York: Harper and Row. Organization for Economic Cooperation and Development (OECD). 1991. Employment outlook. Paris: Organization for Economic Cooperation and Development, July. Schlesinger, Arthur M . 1958. The coming of the New Deal. Boston: Houghton Mifflin. Schneider, Michael. 1991. A brief history of German trade unions. Bonn: Dietz. Taft, P. 1964. Organized labor in American history. New York: Harper and Row.

295

Spurts in Union Growth: Defining Moments and Social Processes

Troy, Leo. 1965. Trade union membership, 1897-1962. New York: Columbia University Press. Troy, Leo, and Neil Sheflin. 1985. U S . union sourcebook. West Orange, N.J.: IRDIS (Industrial Relations Data and Information Services). US. Bureau of the Census. 1960. Historical statistics of the United States: Colonial times to 1957. Washington, D.C.: Government Printing Office. . 1966. Long-term economic growth, 1860-1965. Washington, D.C.: Govemment Printing Office. U.S. Department of Commerce. 1966. The national income and product accounts of the United States, 1929-1965: Statistical tables. Washington, D.C.: Government Printing Office. U.S. Department of Labor. 1994a. Factjinding report: Commission on the Future of Worker-Management Relations. Washington, D.C.: Government Printing Office, May. . 1994b. Report and recommendations: Commission on the Future of WorkerManagement Relations. Washington, D.C.: Government Printing Office, December. . Bureau of Labor Statistics (BLS). 1980. Directory of national unions and employee associations 1979. Bulletin no. 2079. Washington, D.C.: Government Printing Office. . 1986-96. Employment and earnings: January issues. Washington, D.C.: Government Printing Office. Visser, Jelle. 1989. European trade unions injgures. Netherlands: Deventer. . 1992. Trade union membership database. Computer files. University of Amsterdam, March. Wolman, Leo. 1924. The growth of American trade unions, 1880-1923. New York: National Bureau of Economic Research.

This Page Intentionally Left Blank

9

The Genesis and Evolution of Social Security Jeffrey A. Miron and David N. Weil

Social security is the largest transfer program in the United States. In 1994, total tax collections under old-age and survivors insurance (OASI, the largest part of social security, and the part on which we focus in this paper) were $293 billion, or 4.2 percent of GDP. Benefit payments were $279 billion. Thirtyseven million people received benefits, and 139 million workers paid taxes on their wages or self-employment income. OASI accounted for 40 percent of the income of the aged in 1992, paying benefits to more than 90 percent of those aged 65 and over and providing more than half of total income to 63 percent of beneficiary units (U.S. Department of Health and Human Services [HHS] 1994b, 1995). In this paper we examine the genesis and evolution of social security in the United States, with special attention to the role of the Great Depression. We ask to what extent the program as it exists today is the same as that created during the depression. Has social security developed in the past six decades according to an original plan, or have subsequent legislation and changes in the environment in which the program operates fundamentally altered it? As a subsidiary question, we ask to what extent social security as it exists today bears the imprint of the Great Depression. We identify the features of the program that were most influenced by the depression and ask whether these features have endured. We begin in section 9.1 by describing social security as created during the Great Depression. We highlight the key features of the program and ask how its initial design was likely to affect its subsequent evolution. In section 9.2 we discuss the economic and political factors that determined Jeffrey A. Miron is professor of economics at Boston University and a research associate of the National Bureau of Economic Research. David N. Weil is professor of economics at Brown University and a faculty research fellow of the National Bureau of Economic Research. The authors are grateful to Dora Costa and Claudia Goldin for helpful comments.

297

298

Jeffrey A. Miron and David N. Weil

the specific design chosen for social security. Perhaps surprisingly, our conclusion is that economic factors played a relatively small role, while both precedent and political factors, which were not obviously related to the Great Depression, were far more important. This conclusion does not mean economic factors were irrelevant; indeed, absent the depression, social security might not have been created at all. But, overall, the type of program adopted appears to have been remarkably unaffected by the economic circumstances in which it was created. Section 9.3 examines the evolution of the program in the six decades since the Great Depression. By many measures, social security has grown enormously during this period, yet there has also been surprising continuity. Much of the growth was built into the original legislation, or else resulted from extensions that were consistent with the plans of the program’s architects. While there have been numerous substantive modifications in the workings of social security, we show that the most significant changes in the program’s operation have resulted from unanticipated changes in the conditions under which it operates, rather than from legislation.

9.1 Social Security as Designed in 1935-39 The social security system that emerged from the Great Depression was created in two steps. The Social Security Act of 1935 established much of the basic structure, but a substantial amendment to the act in 1939 produced a system that was materially different in several ways. 9.1.1

Old-Age Insurance and Old-Age Assistance

In his message to Congress on social security (17 January 1935), Roosevelt called for a multipronged assault on poverty in old age: “First, noncontributory old-age pensions for those who are now too old to build up their own insurance. It is, of course, clear that for perhaps thirty years to come funds will have to be provided by the States and the Federal Government to meet these pensions. Second, compulsory contributory annuities which in time will establish a self-supporting system for those now young and for future generations” (Zevin 1947,44).’ Three titles of the Social Security Act, signed into law on 14 August 1935, dealt with the aged. Title I made provision for federal subsidies to old-age assistance (OAA) programs administered by the states. The program that is now called social security was originally called old-age insurance (OAI) and was created by Titles I1 (benefits) and VIII (payroll taxes) of the act. The I . Roosevelt called for a third program, voluntary contributory annuities that would supplement the compulsory program and provide for individuals not covered by the compulsory program. This program, which had also been recommended by the Committee on Economic Security, was never enacted. “Pensions” in the usage of the time generally referred to noncontributory payments to the elderly.

299

The Genesis and Evolution of Social Security

division of OAI between two titles was motivated by the fear of legal challenges.z OAA was a program of grants-in-aid to states to provide cash payments to destitute aged. In states with plans that met the act’s criteria, the federal government agreed to pay half of the cost of assistance, up to a federal contribution of $15 per month per beneficiary (with no minimum benefit). States could not impose a minimum age above 65 (70 until 1940), and limits were placed on the stringency of residency requirements. The program was targeted at “aged needy individuals,” but no definition of need was provided in the federal act (Committee on Economic Security [CES] 1937, chap. 11). By the end of 1938, OAA was in operation in all 48 states, the District of Columbia, Alaska, and Hawaii. Expenditures on OAA by federal, state, and local governments in 1938 were estimated to be $380 million. During September 1938, 21.6 percent of those aged 65 and over received assistance; the fraction ranged from 7.2 percent in New Hampshire to 54.5 percent in Oklahoma. The average grant was $19.21 per month. Of all married people receiving a grant, over 40 percent had a spouse receiving a separate grant (Advisory Council on Social Security [ACSS] [1938] 1977, 16). OAI created a system of contributory old-age insurance. Unlike OAA, unemployment insurance, and aid to dependent children, it was a purely federal program. Retiring workers would receive lifetime benefit payments based on the total value of mandatory “contributions” made during working years. Contributions were to begin in 1937, with the first benefits to be paid out in 1942. The program covered almost all employees in industry and commerce but excluded domestic and farm workers, railway workers (who were covered under the railroad retirement program), the self-employed, and employees of federal, state, and local governments. 9.1.2 Taxes and Trust Funds OAI was financed by equal taxes on employers and employees, and subsequent increases have maintained this equality of tax rates. Although economists now agree that it is irrelevant on which side of market a tax is imposed, the symbolism of the “fifty-fifty” split between workers and their employers has been important in the development of the program (Cates 1983). Taxes applied to the first $3,000 of earnings, covering 92 percent of wages earned in covered employment in 1937. Of covered workers, only 3 percent had earnings that equaled or exceeded the ceiling (Myers 1993, table 3.4; HHS 1994b, 9). 2. In early 1936, the Supreme Court struck down the Agricultural Adjustment Act, which earmarked the revenues of a specific tax on processors to be paid out to farmers who agreed to reduce production. Drafters of OAI feared that an explicit insurance program, in which taxes were earmarked to be paid out to a specific group (former contributors), would be subject to a similar challenge. They thus tried to make OAI look like separate taxation and spending schemes. Even so, Thomas H. Eliot, counsel to the CES, had “grave constitutional doubts” about the program. It was not until 1937 that the legal threat to OAI was eliminated. See Eliot (1961).

300

Jeffrey A. Miron and David N. Weil

OAI began with only contributors and no beneficiaries. Thus, if the system had been funded on a pure pay-as-you-go basis, the initial tax rate would have been zero. By contrast, if the planners had desired a constant tax rate through time, the system would have built up a large trust fund in its early years, when there were few beneficiaries. The course chosen was between the two: tax rates were initially set low but scheduled to rise over time as the number of beneficiaries increased. At the same time, a trust fund was to be accumulated, with trust fund interest helping to cover benefit payments as the number of beneficiaries grew. The 1935 act established a combined employer-employeetax rate of 2 percent on covered workers, to be collected beginning in 1937.3It called for increases in the combined tax rate in 1 percentage point increments every three years starting in 1940, to reach a maximum of 6 percent in 1949. The 1939 act canceled the tax increase scheduled to have taken place in 1940 but called for a 2 percentage point increase in 1943, following which tax rates would follow the path specified in the 1935 act. The 1939 act envisioned the accumulation of a trust fund that would reach $50 billion in 1980 (Parker 1942, 39). Interest on the trust fund would allow benefit payments to exceed revenue collected through payroll taxes. It was the view of the ACSS that the size of total benefit payments would be much larger than 6 percent of payrolls: “Information now available indicates that the benefit structure under Title 11 of the present Act will involve financing from all sources of an annual disbursement equivalent to ten to twelve per cent of covered payroll by 1980 when persons now in their twenties will be at retirement age” (ACSS [1938] 1977,23).4 Recognizing that a trust fund might not be accumulated, both the CES and the ACSS envisioned that general revenues would have to supplement payroll taxes: “If the method of accumulating a relatively large reserve is eliminated, there must be, instead, the definite assurance that the program will be financed not by payroll taxes alone but, in addition, by governmental contributions from other sources. Without interest returns on a relatively large fund, payroll taxes alone would prove insufficient to meet the current disbursements necessary as the system matures” (ACSS [1938] 1977,46). 9.1.3 Benefits Receipt of benefits under OAI was conditional on complete withdrawal from the labor force. Under the 1935 act, no benefits were to be paid in months during which any wages were received from “regular employment,” although 3. We maintain the convention of expressing tax rates as a fraction of the wage inclusive of the employer’s tax payment. Thus, if a worker receives $97 after taxes, with $3 paid to the government as “employee” taxes and $3 paid as “employer” taxes, we call the payroll tax rate 6 percent. Alternatively, one could calculate the tax rate as $6/$97, or 6.19 percent. 4. The text refers to benefits payable under the 1935 act; however, the recommendations of the ACSS were designed to keep total disbursements in the 1939 act equal to those under the 1935 act. See also Derthick (1979, 233).

301

The Genesis and Evolution of Social Security

this term was not defined. Under the 1939 act, workers received no benefits in any month in which they earned $15 or more. The earnings test represents an area of confrontation between the different conceptions of social security, usually summarized by the terms “equity” and “adequacy.” If workers earn their entitlement to benefits through their contributions, then a person who is able to work beyond age 65 should be just as entitled to get his benefits as a person who is unable to work or who chooses not to-this is the equity view. The adequacy view is that social security is a form of “social insurance” against the misfortune of forced retirement, and so people who are able to work do not deserve benefits. Conditional on eligibility for benefits, the monthly benefit formula in the 1935 act was 1/2 percent of total lifetime covered wages up to $3,000, 1/12 percent of the next $42,000, and 1/24 percent of wages above $45,000. For a worker with annual wages of $1,000 (the average annual wage in 1937 was $979) who retired in 1942 with five years of contributions, the replacement rate would have been 20 percent. For a worker with 45 years of contributions, the replacement rate would have been 60 percent. There were no benefits for spouses or widows. The benefit formula was changed significantly in the 1939 act, h ~ w e v e r . ~ First, while the original act had promised benefits only to former workers themselves, the 1939 act extended benefits to both spouses (50 percent of the benefit to the primary worker) and widows (75 percent of the benefit to the primary worker) and correspondingly reduced benefits to unmamed workers. The program’s name was thus changed to old-age and survivors insurance (OASI). The 1935 act originally included a provision for a death benefit that would return to the estate of a deceased participant any employee contributions that had not been paid out as benefits. This was eliminated in 1939. The most important change was a shift in the benefit formula. Under the 1935 act, benefits had been based solely on the total value of contributions. Under the 1939 act, benefits were based on both the period of time over which contributions were made, and the average wage during that period.6Although benefits would still rise over time as the system matured, the new formula awarded much higher benefits to those retiring in the early years of the system 5. Why was the benefit formula altered? J. Douglas Brown, who chaired the 1937-38 ACSS, which designed the 1939 act, argued: “In the pressure to design a working model of an old-age insurance system that could be described in detailed legislative language, a benefits schedule was needed to complete the blueprint. The last-minute invention inserted in the bill, though simple to explain, was bound to need revision before any benefits were paid” (Brown 1977, 3; see also Brown 1969). As mentioned above, it had not been clear in 1935 that the OAI program would be constitutional, and this uncertainty had inclined the program’s designers not to worry about the details of the benefit formula. The political power of the Townsend movement also was clearly a factor in the redesign. Our conclusion that social security was not radically altered after its creation would be very different if we had viewed the 1935 plan as the original and the 1939 act as the subsequent amendment. 6 . The formula for monthly benefits was L.3 x min (AMW, 50) + .1 x min (AMW, 200)] x (1 + .01 x Y), where AMW is average monthly wages and Y is years of contribution.

302

Jeffrey A. Miron and David N. Weil

relative to those retiring when the system was mature than did the 1935 act. Finally, the year in which benefit payments were to begin was moved from 1942 to 1940. The 1939 act also raised the maximum federal share of OAA benefits from $15 to $20. 9.1.4 Illustrative Calculations Table 9.1 presents calculations designed to illustrate the provisions of the 1939 version of the program. An obvious difficulty in making these calculations is that the original design of the system did not explicitly take into account either inflation or real wage growth. Rather than try to anticipate what changes could have been expected in 1939, we simply hold prices and real wages constant. We use the scheduled phase-in of tax rates and the benefit formula from the 1939 act. We examine a married couple with a nonworking wife who is the same age as her husband. The husband is assumed to have a continuous earnings history from 1937 through retirement at age 65 at a wage of $1,000 per year.7 Column 2 of the table calculates the replacement rate: the fraction of the husband’s preretirement income that is accounted for by his retirement benefit plus the wife’s spousal benefit. To examine the size of the intergenerational transfers contemplated in the original system, we calculate two measures: the rate of return to contributions and the present value at age 65 of the transfer received by different cohorts. All three of these measures can be compared with the actual evolution of the system. The system was designed to phase in gradually. Tax payments as a fraction of lifetime wages phased in gradually both because workers who retired in the early years of the system would have short histories of contributions and because tax rates were to rise over the first decade of the system’s operation. Benefits also phased in gradually since the benefit formula rewarded extra years of contributions. Thus, early retirees had lower replacement rates than those retiring when the system was mature. The reduction in benefits for early retires was much lower than the reduction in contributions, however. Workers retiring in the early years of the program were to receive benefits more than half as large as those retiring when the system was mature, even though their histories of contributions were as little as one-tenth as long and the tax rate they paid was as little as one-third as high as those of later retirees. As a result, retirees in the early years were to receive large transfers.8 The system matured very slowly. Only after 1990 would workers with a full history of contributions at the maximum tax rate begin retiring. By that point, the table shows, the internal rate of return to participation in the program would be close to the 3 percent interest rate that the planners assumed would be 7. Rather than integrate over different possible dates of death, we simply assume that each member of the household lives to his or her life expectancy at age 65. 8. The calculation here does not take into account workers who died before they reached age 65 or single workers, both of which groups received much lower returns on their contributions.

The Genesis and Evolution of Social Security

303 Table 9.1

Financial Characteristicsof the 1939 Act First Year Receiving Benefits

1940 1950 I960 1970 1980 1990

Replacement Rate

.43 .47 .52 .56 .60 .60

Real Rate of Return (%)

Value of Transfer at Age 65

183 25 12 7.0 4.7 3.9

4,s 17 4,835 4,442 3,753 2,662 1,712 ~

Source: Authors’ calculations. Calculations assume a same-age married couple in which the husband has annual wages of $1,000 per year and a continuous work history from 1937 through age 65.

earned on trust fund reserves. The designers of social security recognized that the benefit formula involved significant transfers to cohorts who retired early in the program, but they argued that for later cohorts it was actuarially fair: “The plan outlined above contemplates that workers who enter the system after the maximum contribution rate has become effective will receive annuities which have been paid for entirely by their own contributions and the matching contributions of their employers. Workers now middle aged or older will receive annuities which are substantially larger than could be purchased by their own and the matching contributions, although considerably less than the annuities which will be paid to workers who contribute for longer periods” (CES 1935, 31).9 9.2 Factors in the Design of Social Security 9.2. I

The Precedents for Social Security

One set of factors determining the structure of social security was the precedent provided by preexisting programs. In the case of OAA, the precedents were state-level programs. All were of the assistance rather than the insurance variety; that is, they provided payments to the aged out of general revenues, with no mandatory or voluntary contributions by beneficiaries. Approximately 30 states had such programs by the beginning of 1935, and about 8 more adopted them in the first half of 1935, likely in anticipation of OAA. The state programs were extremely limited in the level of old-age income support they provided. Eligibility was means tested using very low income and wealth cutoffs, and virtually all programs barred beneficiaries with financially 9. It should be noted that in a dynamically efficient economy (in which the growth rate of output is less than the interest rate), such a combination of transfers to early cohorts and actuarial fairness for all later cohorts is not possible.

304

Jeffrey A. Miron and David N. Weil

competent children or other family members. The programs also typically required a minimum of 15 years U.S. citizenship and residency in the state. Yearly expenditure of all state old-age pension plans (noncontributory meanstested assistance) in 1934 was $31.2 million, and the total number of pensioners was 180,003 (CES 1935, table 14). By 1938, nearly 10 times as many people would be receiving OAA. The precedents for OAI consisted of the preexisting social security systems in other countries. According to the CES, 20 countries had adopted compulsory, contributory old-age insurance laws by 1935, and these programs provided precedent for virtually every feature of the U S . system. There were of course differences between the U.S. system and any given system elsewhere. Some of the insurance programs applied only to specific groups or industries, while others included groups that the initial U.S. system excluded (e.g., agricultural workers). Several of these systems made use of general tax revenues in addition to employer and employee taxes. At least one country, France, included benefits for workers who had already retired when the program was created as part of its contributory pension program. Nevertheless, the basic similarities between these systems and the 1935-39 version of OASI were great. They were compulsory, contributory old-age insurance systems, not just old-age assistance systems. In particular, they were not means tested. The financing generally involved taxes on both employers and employees, usually at the same rate. The systems typically excluded selfemployed workers, likely because of administrative difficulties. Thus, it is tempting to conclude that the U.S. system was modeled on existing systems, rather than designed from scratch or tailored in any obvious way to the U.S. experience. Moreover, the fact that the report of the CES contained detailed and explicit material on all these systems makes clear that the planners had all the relevant models in front of them when they designed OAI, and the comments written about these other systems were generally highly approving. For example, Frances Perkins, secretary of labor and chair of the CES, stated that “we shall have to establish in this country substantially all of the social-insurance measures which the western European countries have set up in the last generation” (Schlesinger 1958, 304). 9.2.2 The Role of Political Factors

Financing the System The first area in which political constraints seem to have been key is in the reliance on payroll taxes to finance the system. The choice was a conservative one in two senses. First, social reformers saw the payroll tax as regressive (although the benefit structure was progressive). And second, the direct link between benefits and an earmarked tax was seen as a natural way of limiting demands for benefit increases (CES 1937, 205).

305

The Genesis and Evolution of Social Security

The original proposal of the CES would have supplemented payroll taxes with contributions from general revenue. Supplements would not have begun until 1965 and would have been equal to roughly one-third of total benefit payments in 1980, when the system had matured (CES 1937, table 45). The idea of a federal contribution was not based on a misperception that federal revenues came from a source other than taxes. Rather, it was intended to enhance the progressivity of the system. Liberals were not shy about making this argument: “There is one defect in this old-age insurance system as set up in the act of Congress, a vital, fundamental defect. The Federal Government does not contribute to it. The Federal Government, as in many European countries, should contribute one-third of the total fund. Where will it get the money? I do not want to use this speech as a springboard for a dissertation on the maldistribution of wealth and income in this country, but I will venture to state that in a country where 87 percent of its wealth is owned by 4 percent of its population, inheritance and income taxes could well be increased for this purpose.” l o Another argument offered for federal financial participation in the program was that the existence of an OAI program would reduce the cost of OAA. The CES estimated that, by 1980, the existence of OAI would reduce OAA expenditures by over 60 percent compared with the case in which OAI was not implemented (CES 1935, 28). More generally, it was argued that since “the nation as a whole will materially and socially benefit by such a program, it is highly appropriate that the Federal government should participate in the financing of the system” (ACSS [1938] 1977,44). The House Ways and Means Committee rejected the use of general revenues to supplement OAI and instead adopted a Treasury plan with a higher maximum tax rate (6 percent vs. 5 percent, with the maximum reached in 1949 rather than 1957). The idea that the system should be self-financing out of payroll taxes, which had not been part of the conception of the CES or ACSS, became over time one of the core principles of OASI. The failure to secure a promise for a contribution out of general revenues turned out to be a dubious defeat for the liberals. By not relying on future supplements, the system would not be at the mercy of future Congresses. More significantly, reliance on payroll tax financing strengthened the political perception that benefits were paid as a matter of right-that workers who received benefits had “bought” them with their payroll taxes. Roosevelt most famously stated this view: “We put those payroll contributions there so as to give the contributors a legal, moral, and political right to collect their pensions. . . . With those taxes in there, no damn politician can ever scrap my social security program” (Schlesinger 1958, 308-9; see also Romer 1995). 10. From speech on floor of U.S. House by Representative Henry Ellenbogen of Pennsylvania, 19 August 1935 (Zinn 1966,285).

306

Jeffrey A. Miron and David N. Weil

Accumulating a Trust Fund

Political considerations played a key role in the debate over the accumulation of a trust fund. But the manner in which traditional divisions were reflected in the debate was complex. Many fiscal conservatives had argued for extending benefits to the elderly who were already retired at the program’s inception and for financing the system on a pay-as-you-go basis. In such a case, generous benefits would have to be imhediately matched by higher taxes, thus creating a countervailing force to politicians’ natural largesse. In their efforts, conservatives found allies among the Townsendites, who consistently opposed the restriction that benefits be paid only to individuals who had made contributions.” At the opposite extreme, running social security on a fully funded basis struck many as the more conservative approach. Alanson Willcox, counsel for the Social Security Board, argued that “the most effective protection” against Congress’s tendency to liberalize benefits “is a rigid adherence to the principle of a ‘fully financed‘ system, and the adoption by Congress of a ‘self-denying ordinance’ by which-though it could have no legal force-Congress abjures forever the right to increase the benefits without simultaneous imposition of sufficient additional taxes. The unpopularity of new taxes seems to me the strongest political antidote to the popularity of increased benefits” (Willcox 1938, 300). Many conservatives, however, feared that the accrual of a large trust fund would in itself lead to the expansion of benefits, as well as to two other evils: the expansion of other spending (since the government would have an easy market for its debt) and possible government ownership of a large chunk of the capital stock. Brown recalls that if the program had been fully funded “in 1934, there was no prospect that there would be enough federal securities in which to invest it” (1969, 12).12“It is scarcely conceivable,” argued one Republican senator, “that rational men should propose such an unmanageable accumulation of funds in one place in a democracy” (quoted in Derthick 1979, 232; see Derthick 1979, chap. 11, for more on this subject). Liberals, too, were divided about the wisdom of collecting a large trust fund. To some, the accumulation of a trust fund was the only way to guarantee the long-term viability of the program. But Leff (1988) argues that allowing the 1I . See Derthick (1979, chap. 11). In 1953 the Chamber of Commerce proposed a reform of social security that would have eliminated OAA and extended OAI benefits to all the elderly, whether or not they had made contributions. Financing would be on a pay-as-you-go basis via a payroll tax. The measure was designed to control the natural tendency toward expansion by making it clear that workers were not putting aside money for their own retirements but rather supporting the current elderly. It was introduced into Congress in 19.53 but died there. See Derthick (1979, chap. 6 ) and Romer (199.5). 12. Federal debt held by the public at the end of fiscal 1935, $28.7 billion, was roughly half as large as the eventual size of the social security trust fund under the scheme that was finally adopted. But this was itself only a partial reserve, which was to be much smaller than the full reserve called for by Willcox.

307

The Genesis and Evolution of Social Security

trust fund not to materialize made it more likely that future shortfalls in financing would be made up from general revenues rather than by payroll taxes, a plan that had always had liberal support. Debates about implementing the scheduled tax increases in the early years of the program centered on the ratio of the trust fund to annual benefit payments, which would have been enormous in the early years under the original schedule of tax increases. Of course, given the structure of the program this was inevitable-but it nonetheless served as a powerful argument against the increases (see Leff 1988). The Division between Old-Age Assistance and Old-Age Insurance

While OAA was simply an extension of existing state programs, OAI was completely new. To politicians and reformers, the contrast between the two programs was stark: “Contributory annuities are unquestionably preferable to noncontributory pensions. They come to the workers as a right, whereas the noncontributory pensions must be conditioned upon a ‘means’ test. Annuities, moreover, can be ample for a comfortable existence, bearing some relation to customary wage standards, while gratuitous pensions can provide only a decent subsistence” (CES 1935, 25). Under OAA, benefits would be “conditional on [the retiree’s] accepting either scrutiny of his personal affairs or restrictions from which others are free” (Ball 1988, quoting from the report of the 1949 ACSS). In contrast, it was argued of OAI that “such a method of encouragement of self-help and self-reliance in security protection in old age is essentially in harmony with individual incentive within a democratic society” (ACSS [I9381 1977, 16-17). Freedom from reliance on means-tested pensions, such as OAA, was a matter of dignity. The restriction of OAI benefits to workers who had made contributions during their working years was a crucial feature of the program-perhaps the most important one for its later development. Although the possibility of “blanketing in” the currently elderly (i.e., paying OAI benefits to workers who had not made contributions) was frequently raised, it was never implemented. The original design of the system made large transfers to workers who had contributed for only short periods, and subsequent changes increased the size of these transfers, but the rule that retirees who never made contributions (and were not the dependents of a contributor) could not receive benefits was always maintained. The design of OAI, in which benefits were conditioned on previous contributions, meant that the program could do nothing to help the currently retired. Indeed, under the 1935 act, even those retiring soon after 1942 would receive relatively small benefits, since they would have contributed for only a short period of time. (The 1939 act raised the replacement rate for workers reaching retirement age in the near future and also moved forward the starting date for benefits.) Thus, OAA was the only program addressed to the immediate burden of old-age poverty. Current retirees who were not poor would receive nothing. Many observers argued that the division between OAA and

308

Jeffrey A. Miron and David N. Weil

OAI, and in particular OAI’s exclusion of those already retired at the program’s inception, was conducive to the expansion of OAI benefits. Since there were few beneficiaries in the program’s early years, increases in benefits could be paid for by reducing the accumulation of the trust fund, rather than by explicit tax increases. Contributors would be in favor of high benefits, since they would think that they were “earning” similar benefits for themselves. Derthick summarizes this aspect of the design: “As did their conservative critics, [program executives] recognized that its initial incompleteness could in the long run serve their goals. By doing less initially, insurance would do more when mature, because artificially low costs would foster liberalizations the full burden of which would not be felt for many years” (1979, 223). 9.2.3 The Role of Economic Factors The Economic Situation of the Elderly Perhaps the most obvious economic factor that should have played a role in the design of social security was the economic situation of the elderly both before and during the Great Depression. The trend in retirement in the period before the Great Depression has been the subject of recent debate. Ransom and Sutch (1986, 1988) argue that the labor force participation rate for men over age 60 was roughly constant between 1870 and 1930. Costa (1996) and Moen (1994) take issue with this view, arguing that the decline in the labor force participation of older men predates social security by half a century. Estimates of the level of participation on the eve of the depression are less controversial: in 1930, labor force participation for men aged 65 and over was 58 percent (Costa 1996). In addition to high labor force participation (by today’s standards), the period before the depression was characterized by lower life expectancy. The probability that a 40-year-old man would live to reach age 65, computed using cross-sectional mortality probabilities from 1930, was only .61. High labor force participation combined with high mortality meant that few people would actually experience “retirement.” Carter and Sutch (1996), using data from 1900 and 1910, estimate that a 55-year-old man had a 21.5 percent probability of retiring before he died, excluding “death-bed retirement.” A large fraction of the retirement that did take place in the period before social security was due to ill health. In the 1940 census, of men aged 65-74 who were not in the labor force, 71.7 percent reported themselves unable to work (Ransom and Sutch 1988, table 1.4). Much of the discussion of old-age security stressed that secular changes were underway that made such a program necessary. There was a common view that machine production burned out workers at a younger age than previous technology. Similarly, large, modem organizations did not allow for a gradual reduction of work, and scientific management did not allow for the informal on-the-job pensions that had been common (Graebner 1980). “The small

309

The Genesis and Evolution of Social Security

shop and business, moreover, which formerly absorbed a considerable percentage of older workers, who dropped out of more strenuous industrial pursuits, are becoming less and less prevalent. This trend reduces the economic opportunities previously open to men and women after their peak of physical activity was passed” (CES 1937, 148). The CES reported that the fraction of the population aged 65 and over had risen from 4.1 percent in 1900 to 5.4 percent in 1930 and was forecast to reach 7.7 percent by 1950 (CES 1935, table 13). This rapid increase put further pressure on the informal support system for the aged: “This increase of older persons in the group which is employed or seeking work means that, unless the number of employment opportunities increases at the same rate, the competition for employment will become keener and more difficult as time goes on. In such competition the position of older persons tends to become more and more unfavorable” (CES 1937, 143). Private pensions did exist in the United States in the period before the Great Depression, although they were not very widespread. The CES estimated that 3.5 million workers (roughly 7 percent of the labor force) were covered by pensions in 1930 (CES 1937, 172). To a large extent, these pension plans seem to have been worker discipline devices. The pensions typically had long minimum service requirements, and employees lost all accrued pension eligibility if they quit or went on strike (CES 1937. 172-80; Lubove 1968,132). In 1935, 75 percent of employees of federal, state, and local governments were enrolled in retirement systems. The program for federal workers, established in 1920, required workers to contribute 3.5 percent of their wages to a retirement fund, with varying contributions from their employer (CES 1937, 179). The CES argued that the Great Depression had greatly exacerbated the plight of the elderly. The committee claimed that the elderly were among the first to lose their jobs. A May 1934 survey of people receiving relief and looking for employment showed that of those aged 65 and over, 30.8 percent had been unemployed for over two years, compared with 17.4 for people of all ages (CES 1937, table 3 1).l3Since they were more dependent on savings than young workers, the elderly were thought to be especially vulnerable in the financial collapses that accompanied the downturn in output. Finally, the depression put strain on the traditional family support system. The CES estimated that between 30 and 50 percent of the aged were supported by children, friends, and relatives. “During the present depression, this burden has become unbearable for many of the children, with the result that the number of old people dependent upon public or private charity has greatly increased” (CES 1935,24). The 13. Similar evidence comes from Ransom and Sutch’s calculations of labor force participation rates for men from the 1930 population census and the 1937 enumerative check census, adjusted for comparability. The participation rate for men aged 55-59 rose from 89.5 to 90.9, and for men aged 60-64 rose from 82.6 to 83.7, while the rate for men aged 65-74 fell from 63.1 to 57.7 (Ransom and Sutch 1988, table 1.1). By 1937 OAA payments had begun, so some of this decline in labor force participation could have been a response to this program.

310

Jeffrey A. Miron and David N. Weil

CES also argued that the effects of the depression would be felt for a long time to come, since many families had seen their lifetime savings wiped out. Recent literature has questioned the accuracy of this view as well as the motives of the social reformers who promulgated it. Haber and Gratton (1994) argue that one of the most powerful facts cited by the reformers-the rise in the fraction of the poorhouse population made up of the aged, from 25.6 percent in 1880 to 53.8 percent in 1923-was actually a statistical artifact: the fraction of the population that was aged had grown, while the fraction of the aged in poorhouses remained constant. Weaver (1987) makes the case that prior to the Great Depression, the economic position of the elderly was fairly secure: most owned their own homes and lived off labor income, which was supplemented by emerging private pension plans as well as life insurance and family support. Gratton (1996) shows that the cross-sectional age-earnings profile for men had the same shape in 1935 as it had in 1890. Though there was a decline in earnings at older ages, it was modest in size and had not changed over time. Gratton’s methodology is subject to the major drawback that his data set covers only households with economically active male heads, however; retired or unemployed older workers are excluded. Thus, to the extent that the impoverishment of the elderly took the form of the elimination of jobs, rather than wage reductions, he will be unable to detect it. This point is amplified in examining the effects of the depression, during which low employment, rather than low wages, was the problem facing the economy. Gratton also questions the conclusion that the depression hit the elderly especially hard. By examining the welfare of the elderly outside the family context, reformers such as those staffing the CES drew an exaggerated picture of their plight. Gratton quotes internal analyses by the Social Security Board of a survey from 1935-36, showing, for example, that per capita household income peaked for persons aged 60-64, at $627. For people aged 65 and over, average per capita income was only slightly lower, $601. This finding is not as inconsistent with the view of the CES as it might at first appear, however, since the reformers were not concerned with the status of the average elderly person, but with an unlucky fraction of that group. The High Level of Unemployment Economic factors appear to have played a more important role along a different dimension. Social security was created during a time of high unemployment, and it is often argued that one purpose of the program was to reduce unemployment by encouraging elderly workers to leave the labor force. The main way in which social security encourages workers to exit the labor force is through the earnings test: since benefits cannot be collected while one continues to work, the price of retirement is reduced. To the extent that social security represents forced saving during working years, there is also an income effect that encourages retirement.

311

The Genesis and Evolution of Social Security

Economists generally reject the “lump of labor” view under which such a policy would reduce unemployment in the long run, and several contemporaries denied that this view was the justification for the earnings test. Robert Myers, who began working for the CES in 1934 and served as chief actuary for the Social Security Administration from 1947 to 1970, argued: “Many people assert that the main purpose of establishing the Social Security program was to get older workers to retire and make jobs available for younger workers. Again, this was at most a minor factor. Actually not many workers aged 65 and over would be eager to retire on a benefit of about $15-$20 a month (when the average wage was about $90). Moreover, the Social Security program would not be very expeditiously effective in this manner because, under the original 1935 law, monthly benefits would first be payable so long afterward-in 1942!” (Myers 1993, 232). The lump-of-labor view seems to have been more influential than academically oriented participants in the process let on, however. For example, the CES wrote that “under the method of old-age assistance superannuated persons are encouraged to compete in the labor market as long as possible to eke out a more satisfactory existence. . . . The effect on the employment opportunities of younger men and upon wage rates might prove most unfortunate. Under a system of old-age insurance, the opposite effect would ensue. Since retirement from regular employment could be made a condition for receipt of benefits, the displacement of superannuated workers from the labor market would be accelerated” (CES 1937, 198). Similarly, debates in Congress stressed the role of the program in reducing unemployment. Barbara Armstrong, one of the small group within the CES involved in designing OAI, recalled that “the interest of Mr. Roosevelt was with the younger man” (quoted in Graebner 1980, 186) and that one of the intended functions of the program was to create job opportunities for the young. As to Myers’s argument that the earnings test could not have been targeted at unemployment because it would have no effect until 1942, it should be recalled that in the mid-l930s, it was not obvious that unemployment was a temporary problem. Robert Wagner, the Senate sponsor of the Social Security Act, argued that even in normal times technology would create a large group of unemployed workers, which social security could encourage to leave the labor force (Graebner 1980, 185; for more on this point, see 184-90). 9.3 The Subsequent Development of Social Security

9.3.1

Changes in the Program

Extensions of Coverage

One important change that has occurred in social security has been extensions of coverage to groups of workers not initially included in the system. The fraction of the labor force covered rose from 63.7 percent in 1940 to 77.8

312

Jeffrey A. Miron and David N. Weil

11 -

10 9 8 -

7-

1B5OAd

6-

1986ACt

5,......

4-

3 -

......

2-

1930

1940

1950

1960

1970

1980

1990

2000

Fig. 9.1 Scheduled and actual OASI tax rates

percent in 1950, due largely to the decline in unemployment. The 1950 act extended coverage to nearly all private employees, including regularly employed farm and domestic workers, government employees not covered by existing retirement systems, and some nonfarm, self-employed workers. A total of 10 million new workers were brought into the system, bringing the fraction of the labor force covered to 93.7 percent in 1951. The 1954 act brought in self-employed farmers, and the 1956 act brought in all self-employed professionals other than physicians (who were covered in 1965). The 1983 act extended coverage to federal civilian employees hired after that year (Myers 1993,232-38). By 1993,97.6 percent of the labor force was covered. These extensions in coverage were unquestionably expansions in the scope of social security. Nevertheless, it is natural to view these as consistent with the plans of social security's architects. The expansion of coverage to all workers was a goal espoused by the 1937-38 ACSS as well as all those that followed it (see ACSS [1938] 1977, 39-41; Ball 1988; Myers 1993, 234, 478). Taxes and Trust Fund

Figure 9.1 shows actual OASI tax rates as well as the scheduled paths of future tax rates incorporated into the 1935, 1950, and 1965 acts.14 Politicians persistently failed to anticipate how high tax rates would have to rise in the 14. Schedules for future tax rates were included in all of the acts but are not shown here for clarity. See HHS (1994b, table E) for the full set.

313

The Genesis and Evolution of Social Security

long run. But in the early years of the program there was an equally significant mismatch between expectations and reality: Congress repeatedly refused to implement scheduled tax increases, despite strong objections from the Social Security Administration. They refused both because of already high taxes during the war and because of conservative fears of the consequences of building up a large trust fund.I5 Not surprisingly, given the frequent adjustments of benefit formulas and tax rates, the social security trust fund has not followed the path envisioned by the system’s planners. Indeed, no large trust fund ever appeared. Under the original plan, interest on the trust fund should have covered roughly one-third of benefits when the system was mature. In fact, interest on the trust fund covered 27 percent of benefits in 1950 (when the system still had few beneficiaries), 4.8 percent in 1960, 5.3 percent in 1970, 1.8 percent in 1980, and 7.3 percent in 1990 (the last figure reflecting the buildup of the trust fund scheduled under the 1983 act; calculations based on Myers 1993, table 10.29). Ben& A second major class of legislative changes in social security has been to the formula for paying benefits. The 10 million new workers who were brought under social security in the 1950 act, as well as those already covered, were given a “new start” by basing the calculation of benefits on average wages after 1950. Thus, members of the newly covered groups who retired soon after 1950 (having contributed for at least six quarters) received full benefits, and a large transfer. The 1950 act also eliminated the increment for additional years of contributions that had been present in the 1939 system (and in turn had been a watered down remainder of the 1935 act, which based benefits on cumulative wage credits). The effect of this change was to eliminate the gradual rise in the replacement rate discussed above. Under the 1950 act, the replacement rate would be the same for workers retiring soon, with short contribution histories, as it would for workers retiring later, with full histories of contributions. To offset this, the level of benefits was lowered. The period since the enactment of social security has seen both rapid real wage growth and high inflation. The Social Security Act did not anticipate either of these eventualities. It may be that this was because they were not 15. The ceiling for wages on which taxes were assessed, set at $3,000 by the 1935 act, was not adjusted until 1950, with the result that the fraction of wages in covered employment subject to the tax fell from 92 percent to 79.7 percent. Between 1950 and 1968, the ceiling was adjusted on an ad hoc basis-the fraction of wages subject to the tax was set at roughly 80 percent by each new act but fell between acts as wages rose. At the most extreme, in 1965 only 71.3 percent of wages were subject to tax, and in that year 36 percent of workers had earnings that exceeded the ceiling (compared to 3 percent in 1939). Under the 1972 act, the ceiling was indexed to the average level of wages, but legislation in 1973 and 1977 raised the ceiling beyond the increase due to wage growth. By 1983, 90 percent of wages were subject to tax. Since then the fraction has fluctuated between 85 and 90 percent. In 1993.6 percent of workers had earnings that exceeded the ceiling. See Myers (1993, table 3.4; HHS 1994b, 9).

314

Jeffrey A. Miron and David N. Weil

Table 9.2 Year Cohort Turns 65

Actual Characteristicsof OASI Replacement Rate in First Year of Retirement

Real Rate of Return (%)

Value of Transfer at Age 65

Transfer as Percentage of Lifetime Earnings

135.1 37.4 23.5 18.8 14.6 12.1 10.3 8.85 7.66 6.20 5.66 4.79

2 1,800 34,087 48,120 72,884 89,895 105,278 122,907 137,437 158,906 125,576 135,466 122,512

4.45 6.29 8.16 11.24 12.47 12.98 13.59 13.67 14.24 10.27 10.34 8.74

~

1940 1945 1950 1955 1960 I965 I970 1975 1980 1985 I990 I995

39.1 25.9 23.5 52.1 49.0 46.4 47.9 58.0 66.7 59.1 61.2 61.7

Source: Steuerle and Bakija (1994, tables A.2, A.9, A.6, and A.7). Calculations assume a sameage couple with nonworking wife and husband earning the average wage with a continuous work history retiring at age 65. The transfer is in 1993 dollars, calculated using a 2 percent real interest rate.

viewed as likely occurrences; alternatively, it may be taken as evidence that revisions of the program were expected and that therefore there was no need to put in place any mechanism for automatic adjustment.I6 Prior to the 1977 act, adjustments for real wage growth and inflation were made on an ad hoc basis and were indistinguishable in structure from changes in the real level of benefits. The formula for determining benefits remained based on average monthly wages (up to a maximum), unadjusted for changes in price level. Had the benefit formula remained unchanged, its progressivity would have resulted in falling replacement rates. Indeed, over the period 193950, in which the benefit formula remained static, this is exactly what happened. Between 1950 and 1975 there were 11 different benefit formulas. Table 9.2 shows the replacement rate for newly retiring workers with average wages. The rise in the replacement rate reflected changes in benefits that intentionally went beyond adjustments for inflation. Changes in the benefit formula affected both new retirees and those who were already receiving benefits. In the context of rising wages, changes in the formula that kept the replacement rate for new retirees constant would at the same time raise real benefits for those already retired because of the progressivity of the benefit formula. Under the 1977 act (taking effect in 1979), benefits were based on average 16. As some evidence for the former view, the CES wrote in a different context (specifically, arguing that shorter working lives due to strenuous industrial jobs were not accompanied by higher wages that would allow saving for retirement) that over the period 1890-1928, “when real rather than nominal earnings are analyzed, popular impressions of greatly increased earnings in this country during the last four decades prove to be based more upon fiction than upon fact” (CES 1935, 148). Nonetheless, this argument does acknowledge that nominal earnings had risen.

315

The Genesis and Evolution of Social Security

lifetime wages deflated by the average level of wages in the economy, and the benefit formula was frozen. Further, benefits for new retirees and for existing retirees were “decoupled.” Benefits, once started, were thereafter held constant in real terms. The widow’s benefit was increased from 75 percent of the primary worker’s benefit under the 1939 act to 82.5 percent in 1961 and to 100 percent in 1972. Several of the changes discussed above served to increase the size of the intergenerational transfer received by early cohorts of retirees as well as to increase the number of cohorts that received such transfers. The failure to enact tax increases in the early years of the program reduced lifetime contributions for early cohorts. Changes in the benefit formula (in particular, the moving forward in 1950 of the time at which the replacement rate would reach its steady state level) increased its generosity. And changes in life expectancy increased the actuarial value of the stream of annuity payments. Examining the effects of each of these changes individually would be beyond the scope of this paper. As a measure of their combined effect, table 9.2 presents calculations from Steuerle and Bakija (1994). The table shows that the rate of return implicitly earned on social security contributions fell over time but did so much more slowly than had been scheduled under the 1939 act (shown in table 9.1). The cohort retiring in 1970, for example, received an average rate of return of 10.3 percent, compared with a rate of 7.0 percent planned in the 1939 act.17 Although the rate of return fell monotonically, the same was not true of the size of the total transfer received (present value of benefits less present value of contributions). The cohorts retiring around the decade of the 1970s made larger “investments” (in the form of contributions) than those that came before them and earned higher rates of return than those that followed-the combination meant that these cohorts received the largest transfers, either in absolute terms or as a fraction of lifetime earnings.

Earnings Test The earnings test is probably the area in which the original design of social security has been most significantly altered. Successive social security acts liberalized the test along multiple dimensions. Recipients aged 75 and over were exempted from earnings limits in 1950, and this limiting age was lowered to 72 in 1954 and 70 in 1970. The original “all or nothing” approach, under which a recipient lost his entire benefit for any month during which wages exceeded a cutoff, was replaced by a system in which benefits were reduced 1-for-1 with wages over an “annual exempt amount” in 1954. In 1972, the offset was lowered to 1-for-2 and under the 1983 act was lowered to 1-for-3 17. Boskin and F’uffert (1988) conduct a similar analysis and conclude that the real rate of return on social security contributions for cohorts that reached age 65 before 1977 averaged 11.2 percent. The cohort retiring in 1978-87 received 5.7 percent; the 1988-97 cohort received 3.7 percent; and the 1998-2007 cohort will receive 2.75 percent. Subsequent cohorts will receive between 2.0 and 2.3 percent real returns, assuming no future changes in the program.

316

Jeffrey A. Miron and David N. Weil

for recipients aged 65 and over. At the same time, annual exempt wages have risen as a fraction of the average wage: in 1935 the cutoff was equal to roughly 15 percent of average wages; in 1994 it was 34 percent for recipients below age 65 and 47 percent for those above. The annual exempt wage for recipients aged 65-69 is currently scheduled to rise from $11,160 in 1994 to $30,000 in 2002. A “delayed retirement credit” was introduced in 1972 (and subsequently expanded), which increases benefits to partially make up for delayed retirement and benefit reductions caused by the earnings test.I8The only major reversal in this trend came with the 1983 act, which provided for the inclusion in gross income for income tax purposes of up to one-half of social security benefits for households with sufficiently high wealth or income (raised to 85 percent in 1993).19 Why was the earnings test so thoroughly weakened? Brown argues that “the lack of understanding of those who believe that they have some ‘equity’ in full benefits at 65, regardless of the purpose of the system, may have been a political factor in accelerating the liberalization of the earnings test” (1977, 12-13). What Brown interprets as a lack of understanding of the equity-adequacy trade-off might also have been a simple downweighting of adequacy as a consideration. Another possibility is that, with the end of the Great Depression, the motive of reducing unemployment by removing older workers from the labor force was less salient. As we discuss further below, a final reason for the weakening of the earnings test was a change in the environment in which social security functioned. Overall Size of the Program

At first glance, one might be tempted to conclude that the most important change in social security has been its growth. Total OASI benefit payments were 0.33 percent of GDP in 1950, 2.8 percent of GDP in 1970, and 4.0 percent of GDP in 1990. Of course, much of this growth had been planned from the start. OAI had been designed gradually to replace OAA as covered workers retired. Table 9.3 shows that this is what happened. (OAA was replaced by a purely federal program, supplemental security income [SSI], in 1974.) Similarly, the extension of coverage to additional workers had been planned from the outset. Nevertheless, many of the legislative changes discussed above served to expand the program’s scope. Is social security much larger than it was originally designed to be? Different measures would seem to tell different stones. The fraction of the average worker’s preretirement income replaced by social security is today almost exactly the level designed in the 1939 program. But the payroll tax rate, originally 18. For a more detailed discussion, see Myers (1993, 270-78). 19. The additional tax revenue generated was transferred back to the social security trust fund. Since the thresholds for inclusion in taxable income were not indexed, the fraction of beneficiaries would grow over time. The value of these receipts was 1.2 percent of total benefits in 1984 and 1.5 percent of total benefits in 1991 (HHS 1992, tables 4.A3 and 4.A4; Myers 1993, 278-79; Steuerle and Bakija 1994, 217-29).

317

The Genesis and Evolution of Social Security Percentage of the Population Aged 65 and Over Receiving Benefits

Table 9.3

~

Year

OASI

OANSSI

1940 I945 1950 1955 1960 1965 1970 1975 1980 1985 1990

.7 6.2 16.4 39.4 61.6 75.2 85.5 90.4 91.4 91.7 92.4

21.7 19.4 22.4 17.9 14.1 11.7 10.4 11.1 8.7 7.1 6.6

~~

Both OAA/SSI and OASI .1 .5

2.2 3.4 4.1 5.2 6.3 7.8 6.1 5. I 4.6

~

Source; HHS (1992, table 3C5).

scheduled to rise to 6 percent, is now 11.2 percent.*O If one uses benefits as a measure, OASI is as large as it was scheduled to be. If one uses tax rates, it is twice its intended size. The key to reconciling these two measures is to recall that under the original plan, interest on a large trust fund was to supplement contributions. Since the trust fund was never accumulated, taxes today are higher than initially planned, while benefits are not. As we showed above, the original plan called for benefits to be in the neighborhood of 10 percent of payrolls. Since both replacement rates and the ratio of workers to beneficiaries are roughly at their forecast levels, it should be the case that the size of benefits relative to total payrolls is also at the level that would have been expected.*’In fact, benefits in 1994 were 10.7 percent of taxable payrolls (contributions were higher since the program is belatedly building up a trust fund). Does this mean that benefits are a better measure of the program’s size than tax collections? We think that the answer is yes, for a reason that has been stressed most forcefully by Kotlikoff (1986): the collection of a “trust fund” is largely an accounting exercise. Social security’s trust fund has never been invested in anything other than debt of the federal government. Thus, had the system built up a large trust fund on which it was earning interest, current OASI contributions would be lower, but the government would have an additional large item for “interest on the social security trust fund” in its spending. And this interest would have to be funded out of other tax revenue. From the point of view of flows of money between the government and the public, a policy in which social security builds up a trust fund holding government debt 20. In addition to OASI, payroll taxes are also collected for disability insurance, with a current tax rate of 1.2 percent, and hospital insurance (one of the two parts of Medicare), with a current tax rate of 2.9 percent. 21. There are a number of other effects, only some of them offsetting, that are ignored in this rough statement.

318

Jeffrey A. Miron and David N. Weil

is the same as a policy in which social security taxes are initially lower, and in which the rest of the government runs a smaller deficit.22 Thus, an observer from 1939 would not be surprised at the size of social security today, although the path by which the current state was reached was not that which had been planned initially.

9.3.2 Changes in the Economic Environment Demographic Changes

One of the factors that has led to recent problems in social security is the changing ratio of elderly recipients to young contributors. That demographic change would be a problem was recognized from the beginning. The actuaries of the CES estimated that the fraction of the population aged 65 and over would rise from 5.4 percent in 1930 to 9.3 percent in 1960 and 12.6 percent in 1990. The figure for 1960 turned out to be too high by 0.1 percent, and that for 1990 too high by 0.4 percent.23 This accuracy in forecasting the percentage of the population that would be elderly conceals two offsetting errors. The CES assumed no change in mortality, although it noted the possibility of a mortality decrease. Indeed, Ransom and Sutch (1986,1988) argue that changes in life expectancy for men over age 60 were insignificant in the six decades prior to the creation of social security, so such an assumption might not have been unwarranted. Compensating for this error, the CES’s assumptions for the birthrate and immigration were far too low. Their estimate for total population in 1990 was 151 million, compared with an actual population of 258 million. The change in the relative size of the elderly and working-age populations has mostly been due to declines in the birthrate. At the same time, changes in longevity have meant that any given benefit structure will produce a larger transfer to each elderly person. Women’s life expectancy at age 60, which had risen by 1 year between 1900 and 1930 (when it was age 76), rose by 3.6 years between 1930 and 1960, and by 2.9 years between 1960 and 1988. The increase for men was similar (Faber 1982; HHS 1992). For a transfer program financed on a pay-as-you-go basis, all that is relevant in determining the tax rate is the ratio of workers to beneficiaries. For a funded system, changes in life expectancy are also important. Thus, the move away from accumulating a trust fund meant that no adjustment had to be made for changes in longevity that were matched by higher-than-expected birthrates. 22. This point was recognized in the early debates over social security financing. Edwin Witte stressed that building up a trust fund, even if the government did so simply by borrowing money from itself, was a matter of “honest bookkeeping.” Similarly, Arthur Altmeyer, in the debates about raising OASI taxes during the war, argued for building up a trust fund as a way of making concrete the promise to pay future benefits. See Parker (1942, chap. 21) and Derthick (1979, chap. 11). 23. CES (1935, table 13, p. 24; CES 1937, 139-41). The ACSS quotes a more extreme estimate by the National Resources Committee that by 1980, 14-16 percent of the population would be aged 65 and over (ACSS [1938] 1977, 15).

319

The Genesis and Evolution of Social Security

Changes in the Nature of Retirement

As discussed above, there is disagreement over how the prevalence of retirement changed in the years before the creation of social security. But in the years since, the change has been dramatic. The labor force participation rate for men aged 65 and over fell from 58 percent in 1930 to 15.6 percent in 1993. Although some of this decrease was due to changes in life expectancy, agespecific participation rates have fallen steeply as well (Costa 1996; Lumsdaine and Wise 1994). At the same time, more people lived to reach retirement. Using cross-sectional life tables, the probability that a 40-year-old man would live to age 65 rose from .61 to .80 between 1930 and 1990. Although labor force participation declined, the economic status of the elderly improved. Private pensions expanded in the period after World War 11. In 1984, 39 percent of people aged 65 and over were collecting pension benefits from some source (Lumsdaine and Wise 1994). The poverty rate among people aged 65 and over was 12.2 percent in 1993, compared with 15.1 percent for the population as a whole. In 1970, the poverty rate among the elderly had been 24.6 percent (U.S. Bureau of the Census 1995, tables 747 and 748). The increased affluence and independence of the elderly was reflected in changes in living arrangements. The probability of a retired man aged 65 or over being a household head rose from less than 65 percent in 1930 to roughly 90 percent in 1990 (Costa 1996). Prior to the depression, retired workers were, to a large extent, physically unable to work. In subsequent decades, the link between retirement and disability was greatly weakened. In 1941, 3 percent of men who began receiving social security at age 65 said that they had retired because they preferred leisure to work. By 1982, the figure had risen to 48 percent (Costa 1996, chap. 7). Costa shows that the age-specific prevalence of degenerative disease among the elderly has fallen significantly and that health has become less important to the labor force decisions of the elderly over the course of the twentieth century. Retirement has become a time of vigor and recreation. Together, these shifts constituted a change in the very nature of retirement. Retirement was no longer an unusual state, comparable to illness, into which some unlucky people would fall. It became an expected part of life. How much of this change is itself due to social security is a matter of debate (see Costa 1996; Lumsdaine and Wise 1994). But the change also had implications on what social security could or should do. The original conception of social security had been as a form of insurance against income loss due to age and infirmity. The designers of the system repeatedly stressed that it was a form of “social insurance.” The very name “social security” emphasized the unpredictability of the needs to be addressed. But the essence of insurance is that not everyone experiences a claim on it. Over time, retirement went from being unusual to being the norm. The area in which the change in the nature of retirement is most closely

320

Jeffrey A. Miron and David N. Weil

reflected is in the earnings test. J. Douglas Brown, who chaired the ACSS, stated the original argument for the earnings test: “The reason for the earnings test is . . . to assure that the protection afforded by the OASDI system is keyed to an adequate protection against hardship so far as the funds of the system permit. Where normal earnings continue, even if the potential beneficiary is otherwise eligible, the assumption is that no benefits are required for adequate protection” (Brown 1977, 12-13). With retirement a normal part of life, the adequacy view of social insurance, under which an earnings test made sense, was no longer appropriate. The change in the environment brought equity considerations to the fore. 9.4

Conclusion

Although social security has grown enormously since it was created, this growth represents, for the most part, the unfolding of the program’s original design. Since OASI paid benefits only to retired workers with a history of contributions, it was inevitable that benefit payments would grow slowly. And although coverage in the program was extended until almost all workers were covered, this was consistent with the plans of the program’s creators. Of the rise in the payroll tax rate from 2 percent in 1937 to 11.2 percent today, part represents an increase that had been scheduled from the plan’s creation, and much of the rest reflects the failure of the program to build up an initially scheduled trust fund. The benefit replacement rate for retiring workers today is almost exactly what had been planned, as is the ratio of benefits to payrolls. Thus, our most important conclusion is that, in a mechanical sense, there has been a surprising degree of continuity in social security since the end of the Great Depression. But this conclusion applies only to the mechanical functioning of the program itself and not to the role that it plays in the economy. Although social security developed along the lines that had originally been laid out during the Great Depression, the world changed around it. Most important, retirement went from being unusual to being commonplace. The framers of social security had stressed that the program was a form of social insurance against the hazard of lost wages in old age. The retirement that they pictured was a hazard precisely because it was unusual at the time: many workers never lived to age 65, and many who did kept working. But in the ensuing six decades, a long, healthy period of retirement has become commonplace. That the program could continue to operate as planned while the world around it changed is in large part attributable to luck. Unforeseen changes in life expectancy, fertility, immigration, women’s labor force participation, retirement ages, inflation, and real wage growth happened to crudely offset each other. While parts of social security, particularly the earnings test, were adjusted to suit changing conditions, the overall structure of the program was not. What had started as an insurance program, which gave money to people in the unlikely events that they reached old age and were unable to work, instead be-

321

The Genesis and Evolution of Social Security

came a transfer program of which most people could expect to be recipients a t some point. Current proposals to reform the program-for example, to replace it with the equivalent of individual retirement accounts-reflect this change in environment (see, e.g., Feldstein 1996). Regarding the lasting impact of the Great Depression, our conclusion is that there was surprisingly little. OASI seems t o have been shaped more by existing precedents and by political considerations than by the specific conditions of the depression. And OAA, which was a response to the immediate crisis of the depression, has faded in importance as an old-age program. O f course, absent the depression OASI might not have been created a t all, but we have not tried to address this issue.

References Advisory Council on Social Security (ACSS). (1938) 1977. Final report of the Advisory Council on Social Security. In Essays on social security, by J. Douglas Brown. Princeton, N.J.: Princeton University, Industrial Relations Section. Ball, Robert M. 1988. The original understanding on social security: Implications for later developments. In Social security: Beyond the rhetoric ofcrisis, ed. Theodore R. Marmor and Jerry L. Mashaw, 17-39. Princeton, N.J.: Princeton University Press. Boskin, M., and D. Puffert. 1988. The financial impact of social security by cohort under alternative financing assumptions. In Issues in contemporary retirement, ed. Rita Ricardo-Campbell and Edward Lazear, 207-42. Stanford, Calif.: Hoover Institution Press. Brown, J. Douglas. 1969. The genesis of social security in America. Princeton, N.J.: Princeton University, Industrial Relations Section. . 1977. Essays on social security. Princeton, N.J.: Princeton University, Industrial Relations Section. Carter, Susan B., and Richard Sutch. 1996. Myth of the industrial scrap heap: A revisionist view of turn-of-the-century American retirement. Journal of Economic Hist o v 56 (1): 5-38. Cates, Jerry R. 1983. Insuring inequality: Administrative leadership in social securio, 1935-54. Ann Arbor: University of Michigan Press. Committee on Economic Security (CES). 1935. Report to the president ofthe Committee on Economic Security. Washington, D.C.: Government Printing Office. . 1937. Social security in America: Thefactual background as summarizedji-om staff reports to the Committee on Economic Security. Washington, D.C.: Social Security Board. Costa, Dora. 1996. The evolution of American retirement: 1880-1990. Cambridge, Mass.: Massachusetts Institute of Technology, Department of Economics. Manuscript. Derthick, Martha. 1979. Policymaking for social security. Washington, D.C.: Brookings Institution. Eliot, Thomas H. 1961. The legal background of the social security act. Speech. Available at http://gopher.ssa.gov/history/eliot2.html. Faber, J. F. 1982. Life tables for the United States, 1900-2050. Washington, D.C.: U S . Social Security Administration, Office of the Actuary.

322

Jeffrey A. Miron and David N. Weil

Feldstein, Martin. 1996. The missing piece in policy analysis: Social security reform. American Economic Review 86 (2): 1-14. Graebner, William. 1980.A history of retirement: The meaning andfunction of an American institution, 1885-1978. New Haven, Conn.: Yale University Press. Gratton, Brian. 1996. The poverty of impoverishment theory: The economic well-being of the elderly, 1890-1950. Journal of Economic History 56 (1): 39-61. Haber, Carole, and Brian Gratton. 1994. Old age and the search for security. Bloomington: Indiana University Press. Kotlikoff, Laurence. 1986. Deficit delusion. Public Interest 84: 53-65. Leff, Mark. 1988. Speculating on social security futures: The perils of payroll tax financing 1939-59. In Social security: Thejrst halfcentury, ed. Gerald D. Nash, Noel H. Pugach, and Richard F. Tomasson. Albuquerque: University of New Mexico Press. Lubove, Roy. 1968. The strugglefor social security. Cambridge, Mass.: Harvard University Press. Lumsdaine, Robin, and David Wise. 1994. Aging and labor force participation: A review of trends and explanations. In Aging in the United States and Japan: Economic trends, ed. Yukio Noguchi and David A. Wise, 7-42. Chicago: University of Chicago Press. Moen, Jon R. 1994. Rural nonfarm households: Leaving the farm and the retirement of older men, 1860-1980. Social Science History 18 (1): 55-75. Myers, Robert J. 1993. Social security, 4th ed. Philadelphia: University of Pennsylvania Press. Parker, J. S . 1942. Social security reserves. Washington, D.C.: American Council of Public Affairs. Ransom, Roger, and Richard Sutch. 1986. The labor of older Americans: Retirement of men on and off the job, 1870-1937. Journal of Economic History 46 (1): 1-30. . 1988. The decline of retirement in the years before social security: U.S. retirement patters, 1870-1940. In Issues in contemporary retirement, ed. Rita RicardoCampbell and Edward P. Lazear, 3-26. Stanford, Calif.: Stanford University Press. Romer, Paul. 1995. Preferences, promises, and the politics of entitlement. In Individual and social responsibility: Child care, education, medical care, and long-term care in America, ed. Victor R. Fuchs. Chicago: University of Chicago Press. Schlesinger, Arthur M., Jr. 1958. The-coming of the New 5eal. Boston: Houghton Mifflin. Steuerle, C. Eugene, and Jon M. Bakija. 1994. Retooling social security for the 21st century: Right and wrong approaches to reform. Washington, D.C.: Urban Institute Press. U S . Bureau of the Census. 1995. Statistical abstract of the United States: 1995, 115th ed. Washington, D.C.: Government Printing Office. U.S. Department of Health and Human Services (HHS). Social Security Administration. 1992, 1994% 1995. Annual statistical supplement to the Social Security Bulletin. Washington, D.C.: Government Printing Office. . 1994b. Fast facts andjgures about social security. Washington, D.C.: Govemment Printing Office. Weaver, Carolyn L. 1987. Support of the elderly before the depression: Individual and collective arrangements. Cat0 Journal 7 (2): 503-25. Willcox, Alanson. 1938. Basic policies under Social Security Act: Argument against system analyzed. The Annalist, 31 August. Zevin, B. D., ed. 1947. Nothing to fear: The selected addresses of Franklin Delano Roosevelt 1932-45. London: Hodder and Stoughton. Zinn, Howard, ed. 1966:New Deal thought. New York: Bobbs-Menill.

10

From Smoot-Hawley to Reciprocal Trade Agreements: Changing the Course of U.S. Trade Policy in the 1930s Douglas A. Irwin

From the Civil War up to the Smoot-Hawley tariff of 1930, Congress retained exclusive authority over U.S. tariffs, which for the most part consisted of a single-column schedule of nonnegotiable, nondiscriminatory import duties. Politicians fought over the height of those duties-the Republicans enacted high protective tariffs (such as Smoot-Hawley) when they were in power, and the Democrats enacted more moderate tariffs when they were in power-but not over the authority of Congress in setting those duties. Following their electoral sweep in 1932, the Democrats undertook an innovation in trade policy making by passing the Reciprocal Trade Agreements Act (RTAA) of 1934. By this legislation, Congress granted the president the authority to reach tariff reduction agreements-agreements that did not require congressional approval-with foreign countries. Republicans opposed this delegation of tariff-making power to the president and for several years threatened to repeal the RTAA. By the late 1940s, however, Republicans generally accepted the status quo of (qualified) executive authority over tariffs, which had fallen from over 50 percent in the early 1930s to about 13 percent. Although Republicans succeeded in halting further trade liberalization through the 1950s, the trade agreements program survived and is the basis for current U.S. commercial policy. Did the Great Depression bring about this fundamental shift in the conduct of U.S. trade policy? Or was it a natural outcome of an underlying movement toward the delegation of trade policy-making powers to the executive that is evident in earlier years? Were the trade agreements that resulted from this delegation of authority responsible for the substantial reduction in the average U.S. Douglas A. Irwin is professor of economics at Dartmouth College and a faculty research fellow of the National Bureau of Economic Research. The author thanks Robert Baldwin and Douglas Nelson for their helpful comments on an earlier draft of this paper.

325

326

Douglas A. Irwin

tariff? Why did the Republicans change their position and not kill the RTAA when they had the chance in 1946? This paper argues that neither political party showed a serious inclination to alter the mechanism of congressional tariff making until the early 1930s. To judge from the previous historical pattern, the Democratic response to the Smoot-Hawley tariff should have been to repeal it and enact an additional unilateral tariff reduction. This option was considered but was rejected. What was different about the early 1930s, and what brought about the delegation of powers to the executive instead, was not so much the Great Depression in the United States per se, this paper argues, but the depression as an international phenomenon. The U.S. depression did foreclose a unilateral tariff reduction as a political option in the short run but did not necessitate delegation. Rather, the economic collapse in Europe and elsewhere led to a dramatic rise in foreign trade barriers and discriminatory measures against U.S. goods. This feature of the international economic environment demanded a response different from the autonomously determined tariffs of previous decades. The Democrats saw reciprocal trade agreements as accomplishing several objectives: they would moderate the tariff code to their liking, make reversal of those lower tariffs difficult should they lose power, and promote economic recovery by reducing foreign tariffs on U.S. exports. Furthermore, it was not the Great Depression but the changed U.S. economic position in the world after World War I1 that blunted Republican opposition to the RTAA and ensured its survival. After the war, export-oriented economic interests-manufacturers and labor in particular-stood to benefit from an open trade regime and supported the RTAA. As a result, Republicans no longer argued for repeal of the RTAA as much as for making trade agreements conditional on congressional approval and inclusive of an “escape clause” provision to protect domestic interests harmed by imports. The nascent postwar consensus about America’s global leadership responsibilities forced Republicans into accepting the General Agreement on Tariffs and Trade (GATT) in 1947 as a necessary component of Europe’s economic recovery and America’s national security. Thus, while the Great Depression as an international phenomenon helped determine the mechanism (reciprocal trade agreements rather than unilateral action) by which the Democrats reduced tariffs, the U.S. global economic position after World War I1 ensured that the Republicans would not reverse the tariff reductions once they achieved power. The RTAA radically changed the process of U.S. trade policy making, but it actually had a relatively minor effect on the height of U.S. tariffs. The United States experienced major swings in the average ad valorem tariff (tariff revenue as a share of dutiable imports) during this period-from 40 percent in 1929 to 59 percent in 1932 to 14 percent in 1948, as illustrated in figure 10.1. These swings were produced less by conscious policy decisions to change tariff rates than by import price fluctuations acting on specific duties, which were levied on about two-thirds of dutiable imports and whose ad valorem equivalent was inversely related to prices. The dramatic escalation of the tariff during 1929-32

327

From Smoot-Hawley to Reciprocal Trade Agreements

10 , , , , , I , , , , I , , , , , , ( , , I ( , , , I , , , , I , , ( , I , , , , I , , , , I , , , , , , , , , 00 05 10 15 20 25 30 35 40 45 50 ! Year Fig. 10.1 Average ad valorem U.S. tariff rate, 1900-1955

Source: U.S. Department of Commerce, Bureau of Economic Analysis, Historical Sfatisticsof the United States, from Colonial Times to 1970 (Washington, D.C., 1975).

was due more to plunging import prices than to higher Smoot-Hawley tariff rates; the subsequent drop in the tariff was overwhelmingly due to higher import prices rather than reciprocal trade agreements with other countries. This paper is organized as follows. Section 10.1 sets out the historical pattern of congressional tariff making in the 40 years prior to the depression, along with the very minor institutional tinkering that took place. Section 10.2 discusses the Smoot-Hawley tariff of 1930 and its relationship to U.S. trade during the Great Depression. Section 10.3 examines why the newly elected Democrats opted for the RTAA in 1934 and how the RTAA was implemented. Section 10.4 focuses on the continuation of the trade agreements program into the postwar period, the waning Republican opposition to the RTAA, and the emergence of a bipartisan consensus in favor of the new status quo. Section 10.5 asks whether the Great Depression constituted a defining moment for U S . trade policy and summarizes the paper's main conclusions.

10.1 Tinkering with Tariffs, 1890-1930 In the decades after the Civil War, the tariff served two primary purposes: it raised revenue for the federal government and protected certain domestic

industries from import competition. The two political parties differed in the

328

Douglas A. Irwin

emphasis they put on these objectives. The Republicans, drawing their support from manufacturers in the Northeast and Midwest, advocated a high protective tariff. They argued that such tariffs preserved the home market for domestic producers and secured high domestic wages for labor by making foreign producers pay for their market access. The Democrats, drawing their support from farmers in the South, believed the tariff should be designed mainly for revenue purposes with moderate protection given to import-competing interests. They womed that too high a tariff might foster monopolies and excessively burden consumers. Progressive Republicans, often in a coalition with the Democrats, agreed that the tariff should be protective but also wanted to reform the tariff system to reduce its inequities and limit its costs to c0nsumers.l The tariff ranked among the most important political issues in the post bellum United States. In the half-century after the Civil War, however, Republicans completely dominated Congress and the executive branch. As a result, U.S. tariff rates remained high. From 1867 to 1912, Democrats controlled both houses of Congress and the presidency for just one brief period (1893-95), during which time they passed the Dingley tariff of 1894, which marginally reduced tariff rates. Less than three years later, the Republicans regained power and pushed tariffs back up again. In 1913 the Democrats regained control of Congress and the presidency and promptly passed the Underwood tariff, the first thoroughgoing downward revision in import duties since the Civil War. Yet this effort was cut short by the outbreak of World War I, when special emergency controls on trade were imposed, and then reversed by the Republicans after the war. Thus, in the brief periods in which they controlled both Congress and the presidency, the Democrats reduced import tariffs (1894,1913). When the Republicans regained control of government, they promptly raised import tariffs (1897, 1922), and they revised or raised them at other times as well (1883, 1890, 1909, 1930).’ Throughout this period, the focal point of U.S. tariff setting was the Congress, which derived its powers from Article I, Section 8, of the Constitution giving it the power to impose and collect import duties and “to regulate Commerce with foreign Nations.” For most of this period, the tariff code was a single-column list of import duties. Congress treated the determination of import duties on hundreds of individual products as an exclusively domestic concern within its own province. For standard collective action reasons, domestic interests seeking high tariffs for protection against import competition dominated the process by which Congress set the duties, while export interests and consumers were underrepresented in the political process. In serving domestic (special) economic interests, Congress’s tariff was set I. As Kenkel put it, “Like the farmers and small-scale manufacturers whom they represented, the middle-westem [Republican progressive] congressmen believed in the principle [of protection] but condemned its application by party leaders who, they charged, responded to the demands of the giant enterprises” (1983, 3). 2. Taussig (1931) remains the classic reference on U.S. tariff history during this period.

329

From Smoot-Hawleyto Reciprocal Trade Agreements

independently of (and generally without regard for) the tariff policies of other countries. Congress was not the best forum in which the foreign policy ramifications of trade policy were given due consideration. The president, whose national constituency and foreign policy responsibilities might be expected to moderate any tariff produced by Congress, was often a minor actor in the tariffsetting exercise. The president could, in principle, negotiate lower U.S. tariffs in conjunction with other countries, but in practice he could not because any tariff treaty would require the approval of two-thirds of the Senate. There were some pressures to change this otherwise smoothly functioning process of congressional tariff making. An ever expanding list of imported goods complicated the task of gathering and processing information on trade matters. Tariff legislation became lengthier and increasingly time con~uming.~ The end result also left many dissatisfied. Aggrieved interests complained that Congress was not getting or acting on the right information. Business groups such as the Chamber of Commerce and the National Association of Manufacturers did not want to end protection but wanted to “take the tariff out of politics” and place it on a more impartial, rational footing, less subject to the whims of special interest politics. Export interests complained that the system was too rigid, not allowing for lower tariffs on intermediate goods or for reciprocal bargaining to reduce foreign barriers against U.S. exports. To hold off tariff-reform-minded progressives who complained that the tariff merely served the interests of big business, Republicans spoke of developing a “scientific” tariff-that is, one based on certain principles and not one that just reflected special interest coalitions. Were these pressures pushing the Congress to delegate some powers or adopt some form of reciprocal trade agreements policy in the decades prior to the Great Depression? During this period the political parties in Congress did struggle with two key issues dealing with the process of tariff making: (1) how to improve or reform the system of formulating the tariff and (2) how to enable the president to achieve reciprocity (equal market access abroad) in the face of foreign discrimination. But a brief examination of each issue reveals that there was no movement toward RTAA-type legislation prior to the early 1930s. 10.1.1 The Tariff Commission Movement Despite the use of independent regulatory commissions in other economic matters, Congress proved reluctant to relinquish its control over the tariff to any other body. Progressive Republicans championed the cause of an independent tariff commission as a move toward a less political tariff, but most Republicans and Democrats questioned the desirability of such a move. Congress resisted not just the delegation of tariff-making powers to a separate agency 3. Schattschneider (1935, 23) shows that pre-Civil War tariff legislation ran fewer than 20 pages, reached about 100 pages by the turn of the century, and amounted to 200 pages with SmootHawley. Tariff bills typically took 6 to 12 months to get through Congress around the turn of the century, but it took over 18 months in the case of Fordney-McCumber and Smoot-Hawley.

330

Douglas A. Irwin

but even any reliance on such an agency to provide impartial advice about the tariff or simply unbiased information about conditions in trade and industry. This opposition was based on the following consideration: If an independent tariff agency was to do more than just collect data and was to actually give advice on the proper height of tariffs, it would need a criterion by which to determine whether a tariff was too high or too low. The two political parties differed fundamentally as to what the guiding principle in setting tariffs should be. Each party feared that the agency could become stacked with members from the other party or have as its mandate an objectionable criterion for setting tariffs. Furthermore, why bother setting up such an agency when Congress could not commit itself to adhere to a commission’s advice? Nevertheless, some experimentation did take place. In the early 188Os, a Republican Congress created a temporary tariff commission to gather business views on the tariff. Although its report was essentially ignored, the commission made a suggestion that Republicans later embraced: the tariff should be designed to equalize the costs of production between domestic and foreign producers. Advocates of this “scientific” tariff reasoned as follows: because wages in the United States exceeded those abroad, U.S. production costs were higher than foreign costs, and therefore a tariff to equalize these costs would permit domestic and foreign producers to compete on an equal footing, thereby keeping U.S. wages high. How such a calculation could be done for the thousands of items in the tariff code and the dozens of countries from which products came was never worked out. Democrats heatedly denied that this “scientific” principle of tariff design had any merit. In 1909 President Taft set up a tariff board (an executive advisory group, not a creature of Congress) to provide advice on tariff reform. Democrats vociferously opposed creating any permanent tariff bureaucracy on the grounds that it would entrench the system around a bad principle-that of setting tariffs to equalize differences in the costs of production of domestic and foreign producers. Congressional Republicans were also suspicious of presidential meddling with the tariff. The board went defunct in 1912. Content with the tariff reductions enacted in 1913, President Wilson and many congressional Democrats continued to oppose the creation of a tariff commission. But Wilson was eventually persuaded by some cabinet members (and the advice of Harvard economics professor Frank Taussig, who was appointed the first chairman of the Tariff Commission) that an independent, nonpartisan, fact-finding agency could provide useful information on trade conditions at home and abroad, particularly given the wartime disruptions to commerce. Rep. Oscar Underwood (D-Ala.) saw such an agency as a threat to the low tariffs he had ushered through Congress in 1913, but his resolution to stop the proposal was defeated by a vote of 55-5 (all opposing votes were cast by Democrats). Title VII of the Revenue Act of 1916 established the U.S. Tariff Commission (USTC). The USTC was created only to provide objective information to the

331

From Smoot-Hawley to Reciprocal Trade Agreements

Congress about trade matters, and no delegation of congressional authority was involved. Composed of six members (no more than half from a single political party), the USTC would not be involved in recommending particular tariff rates, and hence, there was no provision for how the tariff should be set. During its first five years in operation (1917-22), the USTC produced several reports on trade policy and made recommendations on mundane administrative details (such as improving the tariff classification schedule). Latent Democratic fears about the potential misuse of the USTC were confirmed when the Republicans got their next shot at writing tariff legislation. A long-standing concern about the tariff code was that it was too rigid in the face of changing trade conditions. Section 315 of the Fordney-McCumber tariff of 1922 introduced the “flexible” tariff provision, which authorized the president to increase or decrease a tariff rate by as much as 50 percent whenever he determined, after investigation by the USTC, that such a change was necessary to equalize differences in the cost of production between the United States and foreign countries. Although this provision was hardly ever employed over the course of the 1920s, when it was used it resulted more frequently in higher than in lower import d ~ t i e sDemocrats .~ were dismayed by the introduction of the costs-of-production principle as a bureaucratic objective and by the operation of the flexible tariff provision.s To conclude briefly, up to the Great Depression the Congress never seriously considered delegating its tariff-making powers to another agency. 10.1.2 Tariff Bargaining by the Executive As a practical matter, the congressionally set tariff proved to be nonnegotiable. Under his foreign policy powers, the president could always negotiate a tariff reduction treaty with another country. However, the measure would not just require the approval of two-thirds of the Senate but implicitly require House approval as well because it would involve a revenue matter. Knowing that the executive could not commit to the implementation of any signed agreement, foreign countries were reluctant to negotiate with the United States. Just three major trade agreements-with Canada (1855-66), Hawaii (1876-1900), and Cuba (1903-34)-were enacted during this period. The Tariff Act of 1897 was the first to authorize the president (within two years) to reduce certain import duties (by no more than 20 percent, in agreements lasting no more than five years) in conjunction with countries giving equivalent tariff concessions. Any agreement would still require the ap4. During 1922-29, the USTC conducted some 80 investigations and issued some 40 reports. The president raised import duties in 33 of 38 cases. The duty reductions were on such commodities as phenol, live bobwhite quails, cresylic acid, and paintbrush handles. See Kelley (1963, 22). 5. In drafting the Smoot-Hawley tariff of 1930, Democrats and progressive Republicans formed a coalition to wrest control of the flexible tariff from the executive branch and return it to the Congress. President Hoover threatened to veto any tariff bill with that provision, and in the end the coalition was defeated.

332

Douglas A. Irwin

proval of both the House and Senate, but by explicitly inviting such agreements Congress implied that it would approve them, even promising to consider any agreement without amendment. As the qualifications suggest, however, the invitation for the president to undertake negotiations was rather tepid and carefully circumscribed. Congress’s endorsement of reciprocity ultimately proved empty because treaty opponents ensured that not one of the eleven negotiated agreements even came up for a vote. Furthermore, the tariff act in which this provision appeared was apparently set up as a classic bargaining duty: tariff rates were set 20 percent higher than otherwise to take into account their later reduction under presidential agreements, according to the USTC (1919, 202-3). In the end, the United States had a high bargaining tariff without having approved any of the bargains. The Tariff Act of 1913 also had a provision authorizing the president to undertake tariff negotiations, but no agreements were negotiated. Congress more frequently embraced a different kind of reciprocity, one to eliminate foreign discrimination against U.S. goods by imposing penalties rather than offering concessions. The McKinley tariff of 1890, for example, authorized the president to impose penalty duties on the importation of sugar, molasses, coffee, tea, and hides from foreign countries that had “unequal and unreasonable” tariffs on U.S. goods. Reciprocity here was not the mutual reduction of tariffs but the elimination of foreign discrimination. It was to be achieved not with the carrot of bargaining but with the stick of retaliation. In 1909, Congress shifted gears and passed a two-column, “maximumminimum” tariff schedule. The maximum schedule was designated as the general tariff, but if the president determined that another country’s policies did not “unduly” discriminate against U.S. products, he could apply the minimum schedule to that country’s goods. In this instance, the president rendered the maximum schedule moot by determining that no foreign country “unduly” discriminated against the United States. All of these measures, however, amounted to half-hearted tinkering and proved to be short lived. The penalty duties in the McKinley tariff of 1890 were abolished by the Democratic tariff of 1894, reinstituted by the Republicans in 1897, abolished by the Democrats once again in 1913, and not reintroduced by the Republicans in 1922. The 1897 invitation for the president to bargain was absent in the 1909 law, reappeared again in 1913, but was left out of both the 1922 and 1930 tariff acts. Before 1922 “the United States tariff, although negotiable in principle, had not been very negotiable in fact,” writes Kelley (1963, 27): “After 1922, even the principle of negotiability was discarded.” A Republican scientific tariff based on the equalization of costs of production could not be reconciled with a tariff that could be arbitrarily bargained away. Congress’s tight reins on tariff making also constrained the president from exercising greater international economic leadership in the interwar period. Although encouraging freer, nondiscriminatory trade policies was one of President Wilson’s Fourteen Points promulgated after World War I, he was unable

333

From Smoot-Hawley to Reciprocal Trade Agreements

to follow through with any substantive effort in this direction. At international economic gatherings through the 192Os, particularly the 1927 World Economic Conference, the State Department could offer platitudes but not concrete proposals. Two unheralded but key changes took place during this period that would have important implications for the future course of U.S. trade policy. First, the income tax, which the Democrats had consistently linked to tariff reform, dramatically reduced the dependence of the federal government on revenues from import duties after 1916. Tariffs generated over 90 percent of federal revenue prior to the Civil War, about 50 percent from 1870-1910, but only about 10 percent of federal revenue in the 1920s. The tariff was now free to be set with objectives other than revenue in mind. Second, with almost no fanfare the United States announced in 1923 that it would adopt an unconditional most-favored-nation (MFN) policy. Up to that point, the United States had adhered to a conditional MFN policy, in which U.S. tariff concessions to one country would not be extended to others unless a reciprocal reduction was offered. Under an unconditional MFN policy, any negotiated U.S. tariff reduction would be automatically applied to all countries that had an MFN treaty with the United States. The 1923 decision formalized the U.S. commitment to nondiscrimination as a ruling principle in international commerce. At the time this was a minor change in the administration of trade policy. “Because the United States conditional MFN policy resulted in relatively little discrimination,” Kelley notes, “the adoption of an unconditional MFN policy in 1923 would not be of such major significance had the United States tariff continued to be virtually non-negotiable” (1963, 35). Because the Republicans did not anticipate any major tariff negotiations they willingly adopted a policy that became the cornerstone of such negotiations. The following should now be clear: right up to the Smoot-Hawley tariff, Congress was reluctant to delegate any of its tariff-making powers either to an independent agency in order to obtain better information or to the executive in order to reach reciprocal tariff reduction agreements. 10.2 Smoot-Hawley and the Great Depression The Smoot-Hawley tariff of 1930 has been popularly portrayed as an aberration, an atrocious piece of tariff legislation whose extremity reflects special interest logrolling run amok. Taking the previous half-century into one’s perspective, however, suggests that Smoot-Hawley was not that much out of the ordinary. Indeed, Smoot-Hawley fits in well with the pre-World War I pattern of Republican control of the tariff described above. Once the Republicans regained control of Congress and the presidency in the 1920 election, they promptly raised tariffs. Following the Emergency Tariff Act of 1921, the Fordney-McCumber tariff of 1922 formalized and extended the sharp increase in import duties: Republicans aimed not just to reverse the

334

Douglas A. Irwin

Democratic tariff reduction of 1913 but to raise rates further to insulate domestic industries from postwar competition from Europe. As firms lined up with complaints about import competition and requests for higher tariffs, Rep. Joseph Fordney (R-Mich.), chairman of the House Ways and Means Committee, reportedly stated that it was his duty to “give the boys what they wanted” (quoted in Kenkel 1983, 152). Partly as a result, the average tariff (shown in fig. 10.1) returned to where it had been just prior to the war but still below where it had been around the turn of the century. In view of the increasing economic difficulties that afflicted agriculture during the 1920s, it was not unnatural for the Republicans to call for another upward revision of the tariff during the election campaign of 1928. (Recall that several times previously the Republicans had revised their own tariff.) Having secured the presidency once again with even greater congressional majorities, Republicans began work on the Smoot-Hawley tariff bill in January 1929. When it achieved final passage 18 months later, the vote was predictably partisan: in the House, 92 percent of Republicans favored the measure, and 91 percent of Democrats opposed it; in the Senate, 78 percent of Republicans favored the bill, and 86 percent of Democrats opposed it. The Smoot-Hawley tariff was not particularly unusual in the degree to which it raised import duties. According to the USTC, had Smoot-Hawley tariff rates been applied to actual imports observed in 1928, the tariff would have pushed the average ad valorem equivalent rate of duty from 34.6 percent (under the Fordney-McCumber tariff) to 42.5 percent, an increase of about 23 percent. This was not as significant a tariff increase as Fordney-McCumber, which the USTC suggested (in a similar fixed import weight calculation) pushed the average tariff rate up 64 percent from the Democratic Underwood tariff of 1913 (see Irwin 1998). Smoot-Hawley helped boost the average tariff to where it had been around the turn of the century. Yet Smoot-Hawley was unique on several dimensions. The legislation took much longer (from January 1929 to June 1930) to go from House Ways and Means Committee hearings to final congressional passage than most previous tariff legislation. At almost 200 pages the bill was significantly longer, more complex, and more controversial than its predecessors. President Hoover’s promise of “limited tariff revision” to help farmers was compromised by the bill, which overhauled the entire tariff and included much higher duties for manufacturers as well. Public opposition to the bill was much more vocal than to previous tariff acts either because special interest logrolling was more blatant or because there was greater sensitivity to such logrolling.6Taussig (1931, 499) speaks of Hoover’s being “besieged” by requests to veto the bill as farmers objected, newspaper editorialists moaned, foreign governments protested, 6. Schattschneider (1935) remains the classic reference on the politics of the Smoot-Hawley tariff. See also Irwin and Kroszner (1996) for an analysis of logrolling in Senate roll call votes on individual tariffs, and Callahan, McDonald, and O’Brien (1994) for an analysis of factors determining the final congressional votes. Eichengreen (1989) provides an excellent general discussion of Smoot-Hawley.

335

From Smoot-Hawley to Reciprocal Trade Agreements

and 1,028 economists signed a petition urging the president not to sign the legislation. The Smoot-Hawley tariff was not a response to the Great Depression: preparation for tariff revision began in late 1928 in reaction to the severe economic distress faced by farmers, well before the stock market crash or the slide in aggregate output and employment. Instead, Smoot-Hawley became infamous as a result of the economic catastrophes that occurred in its wake. SmootHawley was later blamed for the 40 percent plunge in the volume of U.S. trade in the two years after its imposition, for poisoning international trade relations by triggering a wave of foreign tariffs that put world commerce on a downward spiral, and even for turning a modest recession into the Great Depression. The Smoot-Hawley tariff, however, was almost surely not responsible for causing the Great Depression. There are no strong theoretical or empirical grounds for maintaining that higher average tariffs led to business cycle downturns; the larger Fordney-McCumber tariff increase, for example, was followed by an economic recovery.’ Pushing up the average tariff from 40 percent to 47 percent, as in the case of Smoot-Hawley, we shall see, resulted in just a 5-6 percent increase in the relative price of imports at a time when imports were only 4 percent of GNP. It is not even clear whether the tariff exacerbated or ameliorated the depression. Two studies of the macroeconomic effects of Smoot-Hawley differ as to the direction of the effect, although they arrive at comparable magnitudes-which are small relative to the depression itself. Crucini and Kahn (1996) argue that tariff-induced distortions to capital accumulation and foreign retaliation could have brought about a 2 percent decline in U.S. GNP, while Eichengreen (1989) argues that the Keynesian-type stimulus from Smoot-Hawley dominated any foreign retaliation and could have increased GNP by about 2 percent. If Smoot-Hawley did not play a major role in promoting or inhibiting the Great Depression, it bears part of the responsibility for the collapse of trade in the early 1930s. Even here, however, Irwin (1998) finds that declining income rather than the higher tariff accounts for most of the observed 40 percent fall in import volume between 1930 and 1932. The volume of imports fell 15 percent, for example, in the year prior to the imposition of the Smoot-Hawley tariff, a period when real GNP fell 7 percent. But the depression exacerbated the impact of the tariff through an unusual mechanism: the ad valorem equivalent of specific duties, which were levied on roughly two-thirds of dutiable imports, rose during periods of deflation. The 50 percent fall in import prices between 1929 and 1932, therefore, dramatically raised the real effective tariff.8 In fact, the deflation-induced increase in the tariff exceeded that of the Smoot7. Dornbusch and Fischer write in the context of Smoot-Hawley that “from either a Keynesian or a monetarist perspective, the tariff by itself would have been an expansionary impulse in the absence of retaliation. In the Keynesian view, the reduction in imports diverts demand to domestic goods; in the monetarist view the gold inflow increases the domestic money stock if not sterilized” (1986,468-69). 8. On this point see Crucini (1994) and Irwin (1996).

336

Douglas A. Irwin

Hawley legislation: Congress pushed up duties about 20 percent, from 40 percent to 47 percent average ad valorem during 1929-30, while deflation increased the duties an additional 30 percent to a peak of almost 60 percent in 1932, as shown in figure 10.1. In all, the effective tariff went up nearly 50 percent in the early 1930s, sufficient to raise the relative price of imports by about 15 percent and reduce import volume by about 12-17 percent (ceteris paribus). If the economic contraction was primarily responsible for the precipitous drop in imports, what accounts for the roughly equal drop in export volume between 1930 and 1932? (Fig. 10.2 shows that both exports and imports-as percentages of GNP-fell sharply in the early 1930s, with little change in the balance of trade.) The symmetric decline in U.S. exports could have been brought about by declining foreign demand due to (1) the inability of foreign countries to earn dollars from exports to the United States, ( 2 ) declining foreign incomes due to the Great Depression abroad, or (3) higher foreign tariffs and other trade restrictions. Smoot-Hawley has been blamed for bringing on these higher foreign trade barriers and for exacerbating world trade tensions. Three explanationshave been proposed for how U.S. policy relates to higher

...................................

25

30

I

35

40

45

~~~

-E X ~ O ~ -----~ S

Fig. 10.2 lkade as a percentage of GNP, 1921-55

Imports

I

50

C

5

Source; U.S. Department of Commerce, Bureau of Economic Analysis, Historical Statistics of the United States, from Colonial Times to 1970 (Washington, D.C., 1975).

337

From Smoot-Hawleyto Reciprocal Trade Agreements

tariffs abroad: Smoot-Hawley may have (i) spawned direct retaliatory measures against the United States, (ii) signaled a breakdown in policy discipline and triggered tariff increases as other countries followed the U.S. example, or (iii) had no impact on foreign tariffs, which increased for the same domestic political economy reasons that U.S. tariffs were raised. To some degree, all three factors were at work. As Jones (1934) has documented, Canada, Spain, and Switzerland provide the clearest examples of countries that retaliated with specific, discriminatory tariffs against U.S. product^.^ The League of Nations stressed the second factor, arguing that Smoot-Hawley “was the signal for an outburst of tariff-making activity in other countries, partly at least by way of reprisals” (1933, 193). Finally, Eichengreen (1989) argues that purely domestic considerations would have led to higher tariffs in Britain and elsewhere, even if Congress had not passed Smoot-Hawley. A somewhat impressionistic reading of the evidence (consistent with Eichengreen’s 1989 conclusion) suggests that factors ii and iii were the most important, with factor iii probably the dominant one. At any rate, the international economic climate in 1932 was dramatically different from what it had been in 1929. World trade volume fell 26 percent and world industrial production plummeted 32 percent between 1929 and 1932. Widespread protectionism-in the form of tariffs, import quotas, foreign exchange restrictions, and the like-materialized ovemight.’O What was the response of the Hoover administration to these developments? Through 1931 and 1932, the response on the trade policy front was inaction: no solutions were offered, and no major change in course was considered.

10.3 The Democratic Response: The Reciprocal Trade Agreements Act of 1934 The election of 1930 handed the Democrats control of the House and, with the aid of progressive Republicans, a working majority in the Senate on trade issues. This provided the Democrats with their first chance since the Wilson administration to change the course of U.S. trade policy. Democrats and progressive Republicans joined forces behind the Collier bill, introduced in January 1932. Disillusioned by the way the flexible tariff provision had been employed simply to raise tariffs during the 1920s, Democrats and progressive Republicans sought to deprive the president of the ability to make adjustments in tariff rates and to give Congress that authority instead. The bill also authorized the president to undertake negotiations with other countries to reduce tariffs. While this is evidence of congressional sentiment in favor of tariff bargaining, any agreement would still require Senate approval. 9. See the detailed study of Canada’s reaction to Smoot-Hawley by McDonald, O’Brien, and Callahan (1996). 10. The figures are from League of Nations (1934, 45). On the spread of protectionism across Europe, see League of Nations (1942) and Liepmann (1938).

338

Douglas A. Irwin

The Collier bill passed the House in a partisan vote, 100 percent of Democrats in favor, and 94 percent of Republicans against. The Senate later approved the measure by a similarly partisan vote, whereupon it was promptly vetoed by President Hoover in May 1932. In his veto message, Hoover stated that there has “never been a time in the history of the United States when tariff protection was more essential to the welfare of the American people than at present” (1977,205-7). The tariff was “solely a domestic question” and economic conditions made its maintenance “imperative.” An international conference was inappropriatebecause it might lead to the “abandonment of essential American policies.” Hoover also objected to the elimination of presidential authority over the flexible tariff. Democrats got a better opportunity to put their mark on tariff policy when they were swept into power in the election of 1932. The Democratic platform and candidate Franklin Roosevelt had criticized the Smoot-Hawley tariff during the election campaign and proposed reciprocal trade agreements with other countries. In view of the high unemployment rate, however, Roosevelt also promised not to strip away protection for American industry. Despite Roosevelt’s equivocations over tariff policy, his secretary of state, Cordell Hull, was staunchly in favor of lower tariffs through international negotiations and was the main force behind the RTAA. 10.3.1 Formulating the RTAA Had they come to power in ordinary circumstances, the Democrats might simply have enacted a unilateral tariff reduction as they had done in 1913. Some Democrats proposed just this. In August 1931, for example, Sen. Kenneth McKellar (D-Tenn.) suggested repealing the Smoot-Hawley tariff and enacting an immediate, across-the-board tariff cut of 25 percent. A draft speech for Roosevelt prepared by Cordell Hull in September 1932 contemplated slashing tariffs unilaterally by 10 percent and then negotiating further reductions (see Tasca 1938, 14; Haggard 1988, 106). But the economic situation in 1933 differed from that in 1913 on two dimensions and prevented this course of action. First, the U.S. economy was in the depths of the depression. Most Democrats wanted lower tariffs, but in the midst of high unemployment and stagnant output they could not muster the political support to seriously consider a unilateral tariff cut.” Second, foreign tariffs escalated sharply during 1930-32 and had been supplemented with quantitative restrictions and exchange controls. Table 10.1 shows how tariffs in the major U.S. export markets increased during this period. While SmootHawley might have spawned some of those barriers, a repeal of Smoot-Hawley was not going to eliminate them. Thus, the domestic and foreign economic situations together undercut the case for unilateral action. Many feared that a repeal of Smoot-Hawley would 11. Hull later wrote that “it would have been folly to go to Congress and ask that the SmootHawley Act be repealed or its rates reduced by Congress” (1948, 1:358).

339

From Smoot-Hawley to Reciprocal Trade Agreements

Table 10.1

Country Canada United Kingdom Germany Japan France Argentina Italy Netherlands Australia

Average Tariff of Major U.S. Trade Partners, Selected Years Share of U S . Exports in 1928

1913

1928

1932

1949

18.2 16.5 9.1 5.6 4.7 3.5 3.2 2.8 2.7

26.1 4.3 6.3 9.3 9.2 17.6 7.4 0.4 17.9

23.3 9.9 7.9 5.5 8.7 18.8 6.7 2.1 22.4

27.4 23.1 23.8 5.4 17.5 28.8 23.5 4.7 41.2

15.7 35.7 7.3 3.2 10.9 23.6 7.9 4.8 18.6

RTAA Country by 1939

J J J

J

Sources: Figures are not comparable across countries. For Canada, figures are average tariff applied to U.S. imports (tariff revenue divided by dutiable imports from the United States) from Government of Canada, Canada Year Book (Ottawa, various issues). For other countries, average tariff is defined as tariff revenue divided by total imports from Mitchell (1983, 1992, 1995). US. export data from U.S. Department of Commerce, Foreign Trade and Navigation of the United States, 1928 (Washington, D.C., 1929).

be contractionary unless opportunities were created for increasing U.S. exports. Entrenched foreign trade barriers, it was believed, could be reduced only through negotiation. Some business groups also began to coalesce behind the principle of reciprocal tariff reductions. As a result, there was virtually no sentiment in Congress for unilateral tariff reduction and much greater support for reciprocity. In April 1933, President Roosevelt announced his intention to request negotiating authority from the Congress to undertake tariff reduction agreements with other countries. Secretary of State Hull departed for the London Monetary and Economic Conference in June with the expectation that such negotiating authority would be forthcoming and that he could make substantive tariff reduction proposals. In London, Hull quickly discovered two things. First, the administration had decided to postpone its request for negotiating authority to concentrate on passage of the National Industrial Recovery Act (NIRA) and the Agricultural Adjustment Act (AAA) through Congress. Roosevelt’s political advisors feared that adding ambitious trade legislation to the Hundred Days agenda might overload Congress and jeopardize the timely passage of the domestic components of the New Deal. As a result, though delegates in London proposed him as chairman of a committee on trade barriers, Hull “promptly declined because the American delegation had been stripped of any real authority to function in this field” (1948, 1:258). Second, there was, at any rate, no support for a multilateral tariff reduction conference among the London participants. As Hull put it: In earlier years I had been in favor of any action or agreement that would lower tariff barriers, whether the agreement was multilateral . . . regional . . . [or] bilateral. . . . But during and after the London Conference it was mani-

340

Douglas A. Irwin

fest that public support in no country, especially our own, would at that time support a worth-while multilateral undertaking. My associates and I therefore agreed that we should try to secure the enactment of the next best method of reducing trade barriers, that is, by bilateral trade agreements which embraced the most-favored-nation policy in its unconditional formmeaning a policy of nondiscrimination and equality of treatment. (1948, 1:356) The administration finally put forth its proposal in early March 1934. The president requested negotiating authority to reduce U.S. tariffs in trade agreements that would not require congressional approval. Roosevelt stated “that a full and permanent domestic recovery depends in part upon a revived and strengthened international trade and that American exports cannot be permanently increased without a corresponding increase in imports” (quoted in Tasca 1938, which reproduces the text of Roosevelt’s message). The contrast between the House Ways and Means consideration of the RTAA (the final provisions of which will be discussed shortly) and of SmootHawley could not be starker: in 1929 the committee heard 1,131 witness (none from the executive branch) over 43 days, generating 11,200 pages of published testimony (plus index) in 18 volumes; in 1934 the committee heard from 17 witnesses (7 from the administration) over 5 days, in testimony amounting to 479 pages in 1 volume. While the 173 pages of Smoot-Hawley legislation took 18 months to work its way through Congress, the RTAA’s 3 pages took just 4 months (Schnietz 1994; Congressional Record 1930, 72, pt. 10:10761). Like Smoot-Hawley, however, passage of the RTAA was almost strictly partisan. The House approved the bill on 20 March with 96 percent of the Democrats voting in favor, 98 percent of Republicans voting against. The Senate approved the measure in early June with 93 percent of Democrats in favor, 85 percent of Republicans opposed. Roosevelt signed it on 12 June. Cordell Hull later recalled that “in both House and Senate we were aided by the severe reaction of public opinion against the Smoot-Hawley Act” (1948, 1:357). The notion that the passage of the RTAA constituted a repudiation of Smoot-Hawley by Congress persists to this day. But the passage of the RTAA did not reflect a change in the underlying tariff preferences of individual politicians, just a change in the partisan composition of Congress, namely, the Democratic majority. By examining the votes of all members of Congress who voted on both Smoot-Hawley and the RTAA, Schnietz (1994) clearly demonstrates that Congress did not “learn” that the Smoot-Hawley tariff was bad. The learning hypothesis implies that members who voted for Smoot-Hawley later regretted its harmful consequences and voted for the RTAA. But only 6 of the 178 House members who voted on both measures demonstrate this type of learning, and all were Democrats. No House Republican who voted for Smoot-Hawley also voted in favor of the RTAA.’* Whether the electorate “learned” to prefer the tariff views of the Democrats is doubtful because 12. Of the 47 senators who voted on both bills, only 2 Democrats and 1 Republican ‘‘learned.’’

341

From Smoot-Hawley to Reciprocal Trade Agreements

Smoot-Hawley played an insignificant role in the Democrats’ electoral success in 1930 and 1932. The RTAA, which was technically an amendment to Smoot-Hawley, had the following provisions: The president was authorized to enter into tariff agreements with foreign countries. The president could proclaim an increase or a decrease in import duties by no more than 50 percent but could not transfer any article between the dutiable and free lists. The proclaimed duties would apply to imports from all countries on an unconditional MFN basis. The president’s authority to enter into foreign trade agreements would expire in three years. Any agreement could be terminated after three years with six months’ notice; otherwise, it would survive indefinitel~.‘~ The most important change from previous legislation was that Congress waived its approval of the agreements. Congress also explicitly endorsed the unconditional MFN clause, under which tariff reductions negotiated with one country would be automatically extended to others. On the House floor a key amendment was offered to the president’s proposal that served to ensure ongoing congressional influence over trade policy. While the president had proposed (with the concurrence of the Ways and Means Committee) no time limit to his negotiating authority, the House instead chose to limit that authority to three years, after which it would automatically terminate without renewal by Congress. This feature of the RTAA strengthened Congress’s hand because the threat of nonrenewal of negotiating authority would keep the executive branch politically sensitive to the legislature’s concerns. Had this provision not been added, a two-thirds majority of both houses of Congress would have been required to override a presidential veto of a bill stripping away his negotiating authority. Unlike Smoot-Hawley, Congress’s consideration of the RTAA attracted virtually no participation by interest group^.'^ Haggard reasons that “in contrast to 1930 . . . when interest groups were the main protagonists and specific tariff rates the issue, the most important issues at stake in 1934 were institutional, centering on the transfer of authority from Congress to the executive” (1988, 112). The RTAA was simply enabling legislation, and no one knew how the authority would be used, how successful the negotiations would be, or how extensive the agreements might be. When the RTAA was passed, Congress could not anticipate how important the legislation would become or even whether it would be sustained by future Congresses. In view of the many short13. The complete text of the RTAA is in Tasca (1938, 306-8). 14. For this reason, Pastor writes that it is “not surprising that there are few interest group political analyses of the 1934 Trade Act” (1980, 91).

342

Douglas A. Irwin

lived trade policy experiments of the past three decades, it was not obvious that the RTAA would necessarily bring a lasting, durable change in U.S. trade policy making. Perhaps this accounts for the minimal participation of interest groups, even among export associations, in the RTAA’s passage. Bailey, Goldstein, and Weingast (1997) and Schnietz (1994) interpret the RTAA as being an institutional mechanism cleverly designed by the Democrats to lock in their preferred tariff le~e1.I~ By making reciprocal tariff reductions via the RTAA rather than a treaty, only half of Congress (rather than two-thirds) was needed to enact lower tariffs. If the RTAA succeeded in reducing tariffs, it also would constrain future (Republican) politicians from easily reversing the policy since it was made in the context of a foreign trade agreement. With hindsight this interpretation rings true, but the RTAA’s success was not guaranteed. The RTAA could easily have been reversed by the Republicans had they been returned to power, and until the early 1940s they explicitly vowed in their party platform to halt and perhaps reverse any tariff changes brought about by the RTAA. (Section 10.4 will discuss the emergence of the postwar consensus in favor of the RTAA.) 10.3.2 Implementing the RTAA During the first year of the RTAA, Secretary of State Hull’s most important battles were conducted within the administration. According to Hull, “The greatest threat to the trade agreements program came not from foreign countries, not from the Republicans, not from certain manufacturers or growers, but from within the Roosevelt administration itself, in the person of George N. Peck, former chief of the A M ’ (1948, 1:370), who in 1934 became the president’s foreign trade adviser (operating outside the State Department). Peck was not alone within the administration in viewing lower tariffs as conflicting with New Deal programs aiming to promote domestic economic recovery by raising prices. In fact, the NIRA and AAA gave the president the authority to block imports that interfered with this objective.16 The State Department consolidated its control over trade policy after Peck resigned in November 1935. That State, rather than other agencies such as the Commerce Department, was the locus of trade negotiations proved to be important. Led by a staunch supporter of freer trade, Hull’s State Department was more insulated from domestic economic interests than other agencies and was able to focus on the diplomatic and economic benefits of trade agreement^.'^ 15. See also Nelson (1989) for a discussion of the domestic political sources of the changes in U.S. trade policy during this period. 16. Peck was also a sharp critic of unconditional MFN, which he called “unilateral economic disarmament” for fear it would lead to free-riding by other countries. 17. Haggard notes that “giving the State Department a central role in trade policy introduced broader international economic and political considerations onto the policy agenda while providing a strong institutional base for free-traders” (1988, 93).

343

From Smoot-Hawleyto Reciprocal Trade Agreements

An interagency committee (dominated by the State Department but also including the USTC and the Departments of Agriculture, Commerce, and Treasury) oversaw the trade agreements program. Once the committee announced its intent to negotiate with a particular country, time was set aside to receive public comment, usually from exporters wishing for certain foreign tariffs to be reduced or domestic producers worried about U.S. tariff cuts on certain commodities. In the negotiations the United States would offer not across-theboard horizontal tariff reductions but specific tariff cuts on a product-byproduct basis. This selective approach was to be pursued, according to the RTAA, “in accordance with the characteristics and needs of the various branches of American production.” To mitigate the problem of “free-riding” by other countries, tariffs would be reduced only on commodities in which the negotiating country was a “principal supplier.” The United States had mixed results in bringing other countries to the bargaining table during the 1930s. Table 10.1 lists the largest U.S. export markets prior to the depression. By 1936 agreements had been reached with only three of those countries: Canada (America’s largest trading partner), France, and the Netherlands. Germany had requested negotiations, but the United States refused because of Germany’s discriminatory trade practices (such as barter arrangements with southeastern Europe). Japan, Argentina, and Australia expressed no interest in negotiating, and talks with Italy broke down. The United States also approached Britain, the second largest foreign market for U.S. goods, in search of a trade agreement. Britain was unenthusiastic; it had just established a higher tariff with colonial preferences, and the United States took just 6 percent of British exports. As war approached in Europe, however, Britain saw diplomatic advantages to signing an agreement. The result, in Rooth’s words, was “a limited and unspectacular treaty, produced by difficult and protracted negotiations” (1992,303). Furthermore, it was in effect for less than a year before Britain entered World War I1 in late 1939. By 1940, the United States had signed agreements with 21 nations that accounted for over 60 percent of U.S. trade. (Table 10.2 lists the countries with which the United States signed trade agreements.) How extensively did these Table 10.2

Reciprocal Trade Agreements, 193&39

Year When Treaty Became Effective 1934 1935 1936 1937 1938 1939

Country Cuba Belgium, Haiti, Sweden Brazil, Colombia, Canada, Finland, France, Guatemala, Honduras, Netherlands, Nicaragua, Switzerland Costa Rica, El Salvador Czechoslovakia, Ecuador United Kingdom

344

Douglas A. Irwin

trade agreements reduce the average U.S. tariff? As reported in Tasca (1938, 188 ff.), the USTC calculated the duties that would have been collected in 1934 had the tariff resulting from the first 13 country agreements implemented by 1936 been in effect. This fixed-weight measure indicates a decline in the average ad valorem tariff from 46.7 percent to 40.7 percent, a 13 percent drop, not much more than half of the Smoot-Hawley increase. The extent of the tariff reductions varied substantially across commodities. By far the largest cuts were concentrated in the “spirits, wines, and other beverages” category of the tariff schedule, with almost no reductions in textile tariffs. Another calculation over a slightly longer time period yields a tariff reduction on the same order of magnitude. As measured by the ratio of tariff revenue to dutiable imports, the average ad valorem tariff declined from 46.7 percent in 1934 to 37.3 percent in 1939. This amounts to a 20 percent reduction in the average tariff, pushing it just below its pre-Smoot-Hawley level of 40.1 percent in 1929. Only part of this reduction is due to the RTAA. Import prices rose 11 percent between 1934 and 1939 and reduced the ad valorem equivalent of the specific duties that had been nominally fixed in the Smoot-Hawleytariff. Using Irwin’s (1996) estimate of the elasticity of the average tariff with respect to import prices of -0.64, higher import prices reduced the tariff by 7.1 percent (or 3.3 percentage points). This calculation implies that the RTAA reduced the tariff by 12.8 percent, quite close to the USTC’s estimate of 13 percent. As a result, negotiated tariff reductions (as opposed to higher import prices) accounted for two-thirds of the overall tariff cut during 1934-39. What was the impact of this tariff cut on U.S. imports, the pattern of U.S. trade, and the economic recovery from the depression? If lower tariffs were fully passed through to import prices, the tariff reduction attributable to the RTAA would have reduced the relative price of imports by 4 percent, sufficient to increase the volume of imports by about 3.3 percent (given Irwin’s 1998 estimate of the price elasticity of import demand of -0.8). The actual volume of imports rose 25 percent during 1934-39, so the RTAA apparently made just a modest contribution to the recovery of trade after 1933. There is some evidence in Tasca (1938, 265 ff.) and elsewhere that the pattern of U.S. trade shifted toward countries with whom trade agreements had been signed. There is no evidence that trade was a particularly strong component of the economic recovery. The contribution of net exports to real economic growth after 1933, for example, was often negative.

10.4 Perpetuating the RTAA: Political Consensus and the Postwar Order Republicans voted overwhelmingly against the RTAA in 1934. Fearing the unraveling of protective tariffs and the loss of congressional control over trade policy, Republicans raised economic and constitutionalquestions about the act. They complained that any move to reduce tariffs would jeopardize the ongoing recovery from the depression. They also charged that the RTAA involved an

345

From Smoot-Hawley to Reciprocal Trade Agreements

unconstitutionaldelegation of legislative taxation powers to the executive. Sen. Arthur Vandenberg (R-Mich.) called the RTAA “fascist” for giving the president the ability to tax imports. The Republican platform of 1936 vowed to “repeal the present reciprocal trade agreement law,” deeming it “destructive” for “flooding our markets with foreign commodities” and “dangerous” for entailing secret negotiations without legislative approval. Republicans voted overwhelmingly against renewing the RTAA in 1937 (New York Rmes, 3 January 1934, 8; Isaacs 1948,258-59). By 1940, however, the Republican position had softened. That year’s platform, which endorsed “genuine” reciprocity agreements, did not call for the repeal of the RTAA. Rather, it condemned the RTAA for operating “without adequate hearings, with undue haste, without proper consideration of our domestic producers, and without congressionalapproval” (quoted in Isaacs 1948, 267). Still, Republicans voted overwhelmingly against the 1940 renewal, with 96 percent of House and 100 percent of Senate Republicans against it. In a show of wartime unity, a majority of Republicans supported a two-year renewal of the RTAA in 1943, but World War I1 had effectively brought the trade agreements program to a standstill. During 1940-43, new agreements with Argentina, Mexico, Iran, Peru, and Uruguay (and supplementary agreements with Canada and Cuba) were signed, and during 1943-45 a single accord (with Iceland) was reached. These agreements did not entail significant tariff reductions, and none had a major impact on trade. The biggest question facing the RTAA in 1945 was its postwar survival. The president’s negotiating authority, necessary to complete the ambitious postwar plans for international commercial policies, expired in mid- 1945. In addition, the 50 percent maximum reduction in tariffs specified in the original 1934 act had been made on over 40 percent of dutiable imports (USTC 1948, pt. 2: 14). President Roosevelt sought the authority to reduce tariffs another 50 percent over the next three years. Republicans again tried to restrict the president’s powers. Pastor notes that the House leadership helped defeat “twelve different amendments which would have given Congress veto power, reduced or eliminated the new authority, or help several special interests” (1980, 95). In the House, the final vote was again largely along party lines: 95 percent of Democrats favored renewal, while 8 1 percent of Republicans were opposed. In the Senate, the 88 percent of Democrats in favor were joined by a surprising 43 percent of Republicans. Irwin and Kroszner (1997) find that this Republican support came from senators in states with an above average export orientation of producers. A bipartisan bloc in favor of the RTAA was beginning to emerge. Unlike plans for the international monetary system, postwar international trade arrangements materialized slowly. The United States aimed to convert 18. The Republican platform of 1944 announced that the party would cooperate in “removal of unnecessary and destructive barriers to international trade,” but that it wanted to “maintain [a] fair protective tariff. . . so that the standard of living of our people shall not be impaired” and that the tariff should be modified “only by reciprocal bilateral trade agreements approved by Congress” (quoted in Isaacs 1948,274-75).

346

Douglas A. Irwin

the piecemeal, bilateral RTAA approach into a broader, multilateral process. In November 1945, the State Department circulated the document “Proposals for Expansion of World Trade and Employment,” which envisioned an International Trade Organization (ITO) having wide-ranging authority over global commerce, including trade policy, cartels, commodity agreements, employment, economic development, and international investment. Negotiations to finalize the IT0 charter took another two and a half years, and while the United Nations eventually approved the charter in 1948, it proved stillborn due to lack of U.S. support. (In 1950, the Truman administration finally decided that the IT0 would not be submitted for congressional approval.) The commercial policy articles of the IT0 charter, however, were negotiated and finalized in Geneva between April and October 1947 and resulted in the GATT. The GATT set out the principles governing the trade policy of signatory countries, including first and foremost unconditional MFN treatment for all GATT signatories. The agreement also embodied tariff reductions negotiated by the 23 participating countries. The tariff negotiations were conducted on a bilateral, product-by-product basis (using the principal supplier rule). The lower negotiated rates were then generalized to other participants via the MFN clause and considered bound at those rates. In the case of the United States, these tariff reductions became effective on 1 January 1948 and did not require congressional approval. How extensive were the tariff reductions introduced by the first GATT negotiating round at Geneva? The USTC (1948, 19) calculated that had the Geneva duties been in effect in 1947, the average ad valorem rate on dutiable imports would have been 15.3, instead of the actual 19.4, amounting to a 21.1 percent tariff reduction. But higher import prices brought an additional reduction by lowering the ad valorem equivalent of specific duties. Import prices rose 10.5 percent between 1947 and 1948, which would have reduced the tariff by about 7.0 percent, according to Irwin’s (1996) calculations. So in this one pivotal year, fully one-third of the total U.S. tariff reduction-from 19.4 percent in 1947 to 13.9 percent in 1948-was due to higher import prices. Remarkably, the immediate postwar trade liberalization of the United States effectively ended with the Geneva agreement in 1947. The Geneva negotiations were followed up by negotiating rounds at Annecy (1949), Torquay (1950-51), and Geneva again (1955-56), which resulted in additions to the GATT club but negligible tariff reductions (about 2.5 percent in each round, on average). The Dillon Round (1961-62) also accomplished little. Not until the Kennedy Round, initiated in 1962 (but not completed until 1967) owing to U S . concerns about the impact of the European Economic Community on its exports, was there another significant reduction in U.S. tariffs through multilateral negotiations.I9 19. According to Koch, “An important factor [behind the passivity during the late 1940s and through the 1950~1was the growing protectionism in the United States. . . . There was a feeling that the United States had given away concessions without any real corresponding benefit, as the European countries were slow in eliminating their discrimination against dollar goods” (1969, 82,

347

From Smoot-Hawley to Reciprocal Trade Agreements

But the U.S. tariff wall had been breached: the average ad valorem tariff, which stood at 59 percent in 1932, was just 12 percent by 1954. The average tariff had been reduced by 80 percent. Yet only about 29 percent of this reduction can be attributed to lower tariff rates arising from negotiated trade agreements. The silent and gradual erosion of specific duties (last set in nominal terms by the Smoot-Hawley duties in 1930) by inflation was the overwhelming source of the reduction in tariffs. This finding does not imply that the RTAA and subsequent GATT agreements were irrelevant. Although U.S. tariffs would have declined significantly in the absence of any trade agreements, the RTAA-GATT mechanism raised the cost of possible congressional action to unravel these de facto reductions. The RTAA mechanism formally bound the tariffs at lower levels against the temptation to raise them. But what really ensured the sanctity of the lower tariffs was that Republicans, even after having won control of Congress in the 1946 election, now accepted the trade agreements program and allowed the Geneva negotiations that created the GATT to continue. Early in 1947, Rep. Thomas Jenkins (R-Ohio) of the Ways and Means Committee introduced a resolution to halt the Geneva negotiations until the USTC had studied the impact of lower tariffs on U.S. industry. Two senior Republican senators (including Arthur Vandenberg, who had earlier called the RTAA fascist) reached an agreement with the State Department in which the Jenkins resolution would be killed in exchange for an “escape clause” that would establish procedures allowing domestic industries to obtain relief from import competition. President Truman issued an executive order stating that in all future trade agreements, the United States could withdraw or modify concessions ‘‘if, as a result of unforeseen developments and of the concession granted by the United States on any article in the trade agreement, such article is being imported in such increased quantities and under such conditions as to cause, or threaten, serious injury to domestic producers of like or similar articles” (quoted in Leddy and Norwood 1963, 127).20This language was included as Article 19 of the GATT. Republicans accommodated, to a limited extent, special interests adversely affected by lower tariffs but also implicitly accepted the past and future tariff reductions as a fait accompli. Thus, in their one big opportunity to halt the continuation of the trade agreements program into the postwar period, the Republicans balked. Yet, although the Geneva negotiations were allowed to proceed unchallenged, the Republicans did bring a screeching halt to further tariff reductions and additional trade agreements. The Republican Congress extended presidential negotiating authority in 1948 for just one year. Included in the renewal was the 84) and had failed to dismantle their colonial preferences as well. Also see Irwin (1995) for a discussion of the early GATT rounds. 20. The escape clause was used sparingly between 1947 and 1951. Twenty-one applications for import relief were received, resulting in just two tariff increases (on women’s fur felt hats and hat bodies and on hatters’ fur) because most applications were dismissed after preliminary investigations (Leddy and Norwood 1963, 128).

348

Douglas A. Irwin

“peril point” provision, which required the USTC to calculate the protection necessary to prevent serious injury to domestic producers. (Although the president was not required to act on this information, if tariffs were cut below this level he would have to provide Congress with an explanation.) In 1949, a new Democratic Congress repealed the previous extension with its peril point provision and enacted a new (retroactive) three-year extension of negotiating authority. Like the prewar pattern, these congressional votes were largely partisan, although an increasing number of Republicans were supporting the Democrats. By the early 1950s, the partisan division over trade policy had largely evaporated.*I A Democratic Congress in 1951 extended RTAA authority for two years but accepted the Republican idea of peril points and required the USTC to investigate injury complaints caused by concessions in trade agreements.22 A Republican Congress in 1953 extended negotiating authority, but just for a single year in exchange for President Eisenhower’s pledge not to pursue further trade agreements. These votes were largely bipartisan: both parties endorsed the principle of executive leadership on trade matters by continuing to extend presidential negotiating authority, although now always with safeguards for domestic import-competing producers. Neither party seriously considered reversing the existing tariff reductions, but they were also content to bring a pause to further trade liberalizati~n.~~ Why did the 80 percent tariff reduction over two decades, which would have been unthinkable in the mid-l930s, survive so easily in the postwar period? Several reasons can be given. Public opinion was not hostile to the new approach to trade policy. In May 1945, a survey by the American Institute of Public Opinion found that when informed voters were asked whether the trade agreements program should be continued, 75 percent answered yes, 7 percent answered no, and 18 percent were uncertain. Asked whether it was good to reduce U.S. tariffs under the trade agreements program, 57 percent said yes, 20 percent said no, and 23 percent were uncertain (see Bauer, Pool, and Dexter 1963, 81 ff.). Despite the usual caveats that come to mind when considering public opinion polling, virtually all polls during the late 1940s and early 1950s show a clear majority in favor of lower tariffs in the context of reciprocal trade agreements. Such favorable public sentiment reflected the fact that powerful economic 21. Watson (1956) analyzes congressional voting on trade bills during this period and finds that regional patterns replaced partisan ones. 22. Meeting with congressional leaders shortly after his election, however, President Eisenhower found that some Republicans “even hoped we could restore the Smoot-Hawley Tariff Act, a move which I knew would he ruinous” (Eisenhower 1963, 195). 23. Congress was also suspicious of the GATT. In every renewal of negotiating authority during the 1950s (1951, 1953, 1954, 1955, 1958), Congress included the following disclaimer: “The enactment of this Act shall not be construed to determine or indicate the approval or disapproval by the Congress of the Executive Agreement known as the General Agreement on Tariffs and Trade” (quoted in Kelley 1963, 107).

349

From Smoot-Hawley to Reciprocal Trade Agreements

interests-both business and labor groups-had swung behind the RTAA. This was not altogether evident from looking at the pro-protection share of oral and written testimony at congressional hearings, which remained roughly constant at 80 percent through late 1940s through 1950s, according to Verdier (1994,210).But the many narrow industry interests against lower tariffs (such as milk, mushroom, and wool producers) belied the widespread support for lower tariffs among much larger organizations. In 1945, major labor organizations, such as the Congress of Industrial Organizations (representing over 6 million workers), the United Automobile and Aircraft Workers, the Textile Workers’ Union of America, and the Amalgamated Clothing Workers of America, supported extension of the RTAA. Business groups did as well, including the Chamber of Commerce, the American Farm Bureau Federation, and the American Bankers Association. With European and Asian economies devastated by war, these groups were poised to benefit from greater exports and thus supported an open world trading system. And although European and Japanese industries began to recover soon after the end of the war, trade liberalization entailed very small economic costs in terms of labor displacement due to import competition. Imports as a percentage of GNP in the United States rose from 2.3 percent in 1946 to 3.1 percent in 1950 and remained at this level through the 1950s. During the high-tariff period of the 1920s, by contrast, imports had been at 4.5 percent of GNP. Meanwhile, exports accounted for 4.5 percent of GNP throughout 1950s. On mercantilist grounds, politicians found freer trade to be a “winning” policy. In addition, as the cold war intensified, support for freer trade with Western Europe and other allies dovetailed with the U S . national security agenda. Lower tariffs and open markets, like military aid and the Marshall Plan, were linked to American national security and the containment of communism. The change of heart of Arthur Vandenburg epitomized the Republican shift: once the high-tariff isolationist of the 1930s who called the RTAA fascist, he became the anticommunist internationalist of the 1940s who accepted lower tariffs as part of the fight for democracy and economic recovery in Along with changes in underlying economic interests, foreign policy considerations muted postwar Republican opposition to the RTAA. It would be completely misleading to conclude that Congress, through its repeated delegation of trade powers to the president, abdicated any role in the formation of trade policy. Although Congress never again wrote the tariff code that set rates of import duty, it influenced the agenda by keeping the president on a short negotiating authority leash and enacting various forms of procedural escape clauses that were expanded over the postwar period. 24. Verdier notes that “economic isolationism was discredited and fell into such disarray that the Republican-prudently, it turned out-decided to eschew a frontal attack on the RTA during the renewal of 1948 and confine their attacks instead to procedural matters (peril points and the independent Tariff Commission)” (1994, 204); see also Bailey 1997.

350

Douglas A. Irwin

10.5 Conclusions Did the Great Depression constitute a defining moment for U.S. trade policy? In view of the sharp and lasting policy changes brought about during the 1930s, the answer appears to be yes. In 1930, the Congress was firmly in control of the tariff; after 1934, the executive largely dominated tariff making. In the early 1930s, the average tariff was over 50 percent; by 1939, it was below 40 percent, and by 1949, it was below 15 percent. While the Great Depression did play a role in bringing about these changes, a closer study of U.S. trade policy during this period qualifies this conclusion. The international nature of the Great Depression gave rise to higher foreign trade barriers against U.S. exports, which in turn prompted the executive delegation feature of the RTAA. Had the depression been confined to the United States and foreign tariffs remained low, there may have been no need for an RTAA. Once the political prerequisite of a domestic economic recovery had been established, the Democrats could have repealed Smoot-Hawley and enacted additional unilateral tariff cuts (much as they had done after the 1892 and 1912 elections). They did not do so after 1932 in part because the higher foreign barriers gave the United States a greater incentive to abandon its independent tariff policy and engage in bargaining.25 Although the RTAA was born during the depression, it was not institutionalized until the late 1940s. During World War I1 the Republican position evolved from outright opposition to the RTAA to making agreements conditional on congressional approval and inclusive of safeguards to protect importcompeting domestic interests. It was not the Great Depression per se but the new economic and political position of the United States in the world resulting from the war that made a return to Smoot-Hawley virtually unthinkable. This facilitated the emergence of a bipartisan consensus in support of the executive trade agreements framework. Finally, the trade agreements negotiated under the auspices of the RTAA only modestly reduced the level of U.S. tariffs. Three-quarters of the postSmoot-Hawley tariff reduction was accomplished through higher import prices that eroded specific duties last set in 1930. This reduction would have occurred even in the absence of the RTAA as long as the specific duties in the SmootHawley tariff remained unchanged.

25. Lake (1998) discusses the international constraints on U.S. tariff policy during this period, and Hillman and Moser (1996) develop a model in which greater political support for trade liberalization can arise from reciprocal negotiations than from unilateral actions.

351

From Smoot-Hawley to Reciprocal Trade Agreements

References Bailey, Michael. 1997. Security and trade: Issue linkages and the political foundations of U.S. trade policy. Washington, D.C.: Georgetown University. Unpublished paper. Bailey, Michael, Judith Goldstein, and Barry R. Weingast. 1997. The institutional roots of American trade policy: Rules, coalitions and international trade. World Politics 40 (April): 309-38. Bauer, Raymond A., Ithiel de Sola Pool, and Lewis Anthony Dexter. 1963. American business and public policy: The politics of foreign trade, 2d ed. Chicago: AidineAtherton. Callahan, Colleen, Judith McDonald, and Anthony O’Brien. 1994. Who voted for Smoot-Hawley? Journal of Economic History 54 (September): 683-90. Congressional Record. 1930. 71st Cong., 2d sess. Vol. 72, pt. 10. Crucini, Mario J. 1994. Sources of variation in real tariff rates: The United States, 1900-1940. American Economic Review 84 (June): 732-43. Crucini, Mario J., and James Kahn. 1996. Tariffs and aggregate economic activity: Lessons from the Great Depression. Journal of Monetary Economics 38 (December): 427-67. Dornbusch, Rudiger, and Stanley Fischer. 1986. The open economy: Implications for monetary and fiscal policy. In The American business cycle: Continuity and change, ed. Robert J. Gordon. Chicago: University of Chicago Press. Eichengreen, Barry. 1989. The political economy of the Smoot-Hawley tariff. In Research in economic history, ed. Roger Ransom, vol. 12. Greenwich, Conn.: JAI. Eisenhower, Dwight D. 1963. The White House years: Mandate for change, 1953-1956. Garden City, N.J.: Doubleday. Haggard, Stephan. 1988. The institutional foundations of hegemony: Explaining the Reciprocal Trade Agreements Act of 1934. International Organization 42 (Winter): 91-1 19. Hillman, Arye, and Peter Moser. 1996. Trade liberalization as politically optimal exchange of market access. In The new transatlantic economy, ed. Matthew Canzoneri, Wilfred Ethier, and Vittorio Grilli. New York: Cambridge University Press. Hoover, Herbert. 1977. Public papers of the president of the United States: Herbert Hoovel; 1932-33. Washington, D.C.: Government Printing Office. Hull, Cordell. 1948. Memoirs, 2 vols. New York: Macmillan. Irwin, Douglas A. 1995. The GATT’s contribution to economic recovery in postwar Western Europe. In Europe’s postwar recovery, ed. Barry Eichengreen. New York: Cambridge University Press. . 1996. Changes in U.S. tariffs: Prices or policies? NBER Working Paper no. 5665. Cambridge, Mass.: National Bureau of Economic Research, July. . 1998. The Smoot-Hawley Tariff A quantitative assessment. Review of Economics and Statistics. Forthcoming. Irwin, Douglas A., and Randall S. Kroszner. 1996. Logrolling and economic interests in the passage of the Smoot-Hawley tariff. Carnegie Rochester Series on Public Policy 45 (December): 173-200. . 1997. Interests, institutions and ideology in the Republican conversion to trade liberalization, 1934-1945. NBER Working Paper no. 61 12. Cambridge, Mass.: National Bureau of Economic Research, July. Isaacs, Arthur. 1948. International trade: Tariff and commercial policies. Chicago: Irwin. Jones, Joseph M. 1934. Tariff retaliation: Repercussions of the Hawley-Smoot bill. Philadelphia: University of Pennsylvania Press. Kelley, William B., Jr. 1963. Antecedents of present commercial policy, 1922-1934. In

352

Douglas A. Irwin

Studies in United States commercial policy, ed. William B. Kelley, Jr. Chapel Hill: University of North Carolina Press. Kenkel, Joseph F. 1983. Progressives and protection: The search for a tariff policy, 1866-1936. Lanham, Md.: University Press of America. Koch, Karin. 1969. International trade policy and the GATT: 1947-1967. Stockholm: Almquist and Wiksell. Lake, David A. 1988. Powel; protection, andfree trade: International sources of U.S. commercial strategy, 1887-1 939. Ithaca, N.Y.: Cornell University Press. League of Nations. Various years. World economic survey. Geneva: League of Nations. . 1942. Commercial policy in the interwar period: International proposals and national policies. Geneva: League of Nations. Leddy, John M., and Janet L. Nonvood. 1963. The escape clause and peril points under the trade agreements program. In Studies in United States commercial policy, ed. William B. Kelley, Jr. Chapel Hill: University of North Carolina Press. Liepmann, H. 1938. Tariff levels and the economic unity of Europe. London: Allen and Unwin. McDonald, Judith A., Anthony Patrick O’Brien, and Colleen M. Callahan. 1996. Tariff wars: Canada’s reaction to the Smoot-Hawley tariff. Bethlehem, Pa.: Lehigh University. Unpublished paper. Mitchell, Brian R. 1983. International historical statistics: The Americas and Australasia. Detroit, Mich.: Gale Research. . 1992. International historical statistics: Europe, 1750-1988, 3d ed. New York: Stockton. . 1995. International historical statistics: Africa, Asia, and Oceania, 17501988, 2d ed. New York: Stockton. Nelson, Douglas. 1989. Domestic political preconditions of U.S. trade policy: Liberal structure and protectionist dynamics. Journal of Public Policy 9 (January-March): 83-108. New York Times. 1934. 3 January. Pastor, Robert A. 1980. Congress and the politics of U S . foreign economic policy, 1929-1976. Berkeley and Los Angeles: University of California Press. Rooth, Tim. 1992. British protectionism and the international economy: Overseas commercialpolicy in the 1930s. New York: Cambridge University Press. Schattschneider, E. E. 1935. Politics, pressure, and the tar@ New York: Prentice-Hall. Schnietz, Karen E. 1994. The 1934 Reciprocal Trade Agreements Act: Partisan institutional protection of liberal trade policy. Houston, Tex.: Rice University. Unpublished paper. Tasca, Henry J. 1938. The reciprocal trade policy of the United States. Philadelphia: University of Pennsylvania Press. Taussig, Frank. 1931. A tariff history of the United States. New York: Putnam. U.S. Tariff Commission (USTC). 1919. Reciprocity and commercial treaties. Washington, D.C.: Government Printing Office. . 1948. Operation of the trade agreements program: June 1934 to April 1948. Washington, D.C.: Government Printing Office. Verdier, Daniel. 1994. Democracy and international trade: Britain, France, and the United States, 1860-1990. Princeton, N.J.: Princeton University Press. Watson, Richard A. 1956. The tariff revolution: A study of shifting party attitudes. Journal of Politics 18 (November): 678-701.

11

The Great Depression as a Watershed: International Capital Mobility over the Long Run Maurice Obstfeld and Alan M. Taylor

The era of the classical gold standard, circa 1870-1914, is rightly regarded as a high-water mark in the free movement of capital, labor, and commodities among nations. After World War I, the attempt to rebuild a world economy along pre- 1914 lines was swallowed up in the Great Depression and in the new world war that the depression bred. Only in the 1990s has the world economy achieved a degree of economic integration that, in the assessment of Sachs and Warner (1995), rivals the coherence already attained a century earlier. This development broadly fulfills the hopes for the world economy that United States policymakers held at the close of World War 11, albeit within an institutional and policy context far different from the one they designed or even envisioned. Why did the network of world trade suddenly collapse in the depression, and how did the collapse itself influence the subsequent process of regeneration? This chapter is primarily concerned with one aspect of international commerce, capital movements, although the forces restraining international capital movements are not fully understandable without an appreciation of the natures and purposes of related restraints on other kinds of trade. In the present chapter we document empirically the ebb and flow of international capital mobility over more than a hundred years and propose a simple Maurice Obstfeld is professor of economics at the University of California, Berkeley, and a research associate of the National Bureau of Economic Research. Alan M. Taylor is assistant professor of economics at Northwestern University, a faculty research fellow of the National Bureau of Economic Research, and a 1997-98 National Fellow at the Hoover Institution. The authors are grateful to Ham Connolly, Matthew Jones, and Ryan Edwards for superb assistance and to Lance Davis, Rudiger Dornbusch, and Pierre Sicsic for helpful discussions. Special debts are owed to Bany Eichengreen for numerous suggestions and references and to Michael Bordo for extensive comments on several drafts. Financial support was provided by the National Science Foundation and by the Center for German and European Studies at the University of California, Berkeley.

353

354

Maurice Obstfeld and Alan M. Taylor

framework for interpreting the forces that gave rise both to the disintegrative trend initiated by the depression and to the boom in global capital mobility of recent years. After first reviewing briefly the economic functions of international capital movements, we review some empirical evidence. Our quantitative indicators-and a wider literature-are in broad agreement that international capital mobility was considerable during the days of the classichi gold standard and under the reconstituted gold standard through around 1930.They likewise agree that mobility contracted sharply as a result of the depression and World War I1 and then slowly reemerged, albeit with significant reversals, starting in the late 1950s. With the quantitative history laid out for interpretation, we next chronicle the major vicissitudes of the international capital market starting with World War I. Financial controls deployed in the course of all-out war began to be relaxed later during an interregnum of floating exchange rates that was widely viewed as a prelude to the “normalcy” of a restored gold standard. Nearly universal (again) by the end of the 1920s, the gold standard was shattered by the Great Depression. Many countries utilized controls of some sort in their attempts to maintain gold parities, and international finance became fragmented as even free exchange countries struck bilateral deals with Germany and other exchange controllers. Controls spread and intensified during World War 11. Bilateral trade arrangements proliferated after the war in a scramble to husband scarce hard-currency-mostly U.S. dollar-resources. The gradual restoration of convertibility in Europe (and elsewhere), fostered in part by American Marshall aid and the European Payments Union, promoted growing world trade and, with it, a growing risk of pressures on the pegged exchange rates mandated by the postwar Bretton Woods system. The breakdown of fixed dollar rates in the early 1970s led, in turn, to extensive liberalization of capital movements by the United States and Germany starting in 1974, by Japan starting in 1979,and by Britain in the same year. Most of Europe, and much of the developing world, followed suit starting around 1990. Evidently, individual country experiences have differed, as have the motivations for external liberalization and the institutional setting in which it has taken place. These differences notwithstanding, financial openness has now reached a depth, universality, and resiliency comparable to that of the classical gold standard era. But that development was consummated only in the 1990s. Secular movements in the scope for international lending and borrowing may be understood, we shall argue, in terms of a fundamental macroeconomic policy trilemma that all national policymakers face: the chosen macroeconomic policy regime can include at most two elements of the “inconsistent trinity” of (i) full freedom of cross-border capital movements, (ii) a fixed exchange rate, and (iii) an independent monetary policy oriented toward domestic objectives. If capital movements are prohibited (element i is ruled out), a country on a fixed exchange rate can break ranks with foreign interest rates

355

International Capital Mobility over the Long Run

and thereby run an independent monetary policy. Similarly, a floating exchange rate (element ii is ruled out) reconciles freedom of international capital movements with monetary policy effectiveness (at least when some nominal domestic prices are sticky). But monetary policy is powerless to achieve domestic goals when the exchange rate is fixed and capital movements are free (element iii is ruled out), since intervention in support of the exchange parity then entails capital flows that exactly offset any monetary policy action threatening to alter domestic interest rates. Recognition of the policy trilemma leads to a central proposition of this paper. Capital mobility has prevailed and expanded under circumstances of widespread political support either for an exchange-rate-subordinated monetary policy regime (e.g., the gold standard) or for a monetary regime geared mainly toward domestic objectives at the expense of exchange rate stability (e.g., the recent float). The middle ground in which countries attempt simultaneously to hit exchange rate targets and domestic policy goals has, almost as a logical consequence, entailed exchange controls or other harsh constraints on international transactions.' The Great Depression stands as a watershed in that it was caused by an illadvised subordination of monetary policy to an exchange rate constraint (the gold standard), which led to a chaotic time of troubles in which countries experimented, typically noncooperatively, with alternative modes of addressing the fundamental policy trilemma. Interwar experience, in turn, discredited the gold standard and led to a new and fairly universal policy consensus, one that shaped the more cooperative postwar international economic order fashioned by Harry Dexter White and John Maynard Keynes but also implanted within that order the seeds of its own eventual destruction a quarter-century later. The global financial nexus that since evolved is based on a solution to the basic open economy trilemma quite different than that envisioned by Keynes or White-one that allows considerable freedom for capital movements, gives the major currency areas freedom to pursue internal goals, but largely leaves their mutual exchange rates as the equilibrating residual.

I. Our interpretation is consistent with the view in the political science literature that purposeful government control is the key factor determining the degree of international financial integration. See, e.g., Helleiner (1994) and Kapstein (1994), and the references they list. Also relevant to our analysis is the paper by Epstein and Schor (1992), who link the existence of controls to the balance of power between labor-oriented interests favoring Keynesian macroeconomic policies and financial-oriented interests favoring inflation containment. We stop short in this paper of a formal econometric analysis of the determinants of capital controls. Alesina, Grilli, and Milesi-Ferretti (1994) and Grilli and Milesi-Ferretti (1995) carry out panel studies of the incidence of capital controls (for 20 industrial countries over the years 1950-89 and for 61 industrial and developing countries over the years 1966-89). Consistent with our interpretation, they find that more flexible exchange rate regimes and greater central bank independence lower the probability of capital controls. In chap. 12 in this volume, Bordo and Eichengreen discuss these results in greater detail and reconsider the effects of postwar capital controls on macroeconomic outcomes.

356

Maurice Obstfeld and Alan M. Taylor

11.1 Capital Mobility: Implications and Historical Evidence This section discusses the ramifications of capital mobility in the international economy and reviews the current evidence we have as to the rise and fall of capital mobility over the long run. In subsection 11.1.1 we discuss the functional importance of capital mobility in historical perspective, touching on issues of economic efficiency and the long-run convergence process. We then document some basic evidence on the evolution of the global capital market, namely, the extent of capital flows as measured by countries’ balances on current account, equal to net foreign investment. This is followed by an examination of the evolution of nominal interest differentials as a guide to the extent of market integration. After that, international differences in real domestic interest rates, taken as measures of the required return to capital, are looked at. Along the way we draw on related literature devoted to more sophisticated econometric tests of market integration, but all evidence points to the latter part of the interwar period, the period of the Great Depression, as the era of weakest financial integration: capital flows were small, countries behaved like closed economies in the capital market, and real and nominal price (interest rate) differentials expanded. 11.1.1 Functions of Capital Mobility We think it important to highlight the functions of capital mobility both for efficiency and for policy. The efficiency implications are clear enough: When capital is immobile, it may not be free to seek out the highest return within the global economy. Nor can countries smooth temporary fluctuations in consumption through international borrowing and lending or diversify risks through the exchange of assets with uncertain and imperfectly correlated payoffs. From the 1930s investors had good reason to be cautious of committing funds to foreign countries that had (or might enact) forms of capital control, since returns would be compromised by controls intended to tax or prevent the repatriation of profits. Cumulating over many years or decades, such disincentives could tend to produce a global misallocation of capital, with an inefficient and excessive amount of capital remaining in capital-abundant (rich) countries and too little flowing to capital-scarce (poor) counties (Lucas 1990). Domestic savings could be biased toward use in domestic investment activity (Feldstein and Horioka 1980). These tendencies do indeed seem to be characteristic of much of the postwar period, as the evidence below suggests. The outcome was not discouraged by the tolerant and benign view of capital controls that prevailed for several decades in the aftermath of the Great Depression. Such a historical process would naturally have implications for the crosscountry patterns of economic growth and development: To the extent that an excess of capital remained in capital-rich countries the process of convergence could have been retarded. Beyond just the evolution of productivity levels, such capital misallocation would have distributional implications. Inefficiently allo-

357

International Capital Mobility over the Long Run

cated capital would be earning low returns in capital-abundant locales, and capital-scarce areas would be characterized by inefficiently low wages. Furthermore, the process of capital market integration would reinforce factor price equalization via trade and integration in goods markets (absent large-scale labor migration in the late twentieth century), a historical process deservedly attracting new attention (Wood 1994; Williamson 1996). Another potential function of an open capital market under conditions of high capital mobility is as a disciplining device on policymakers. Unsound policies may be limited by the threat of massive capital outflow. This form of commitment mechanism might be seen as supplementing, or possibly even dominating, any disciplinary power inherent in a commitment to an exchange rate peg. The full set of gold standard “rules of the game” has been characterized as a form of commitment mechanism or reputational device because of the explicit constraints on monetary and fiscal policy (Bordo and Kydland 1995; Bordo and Rockoff 1996). These advantages of international capital mobility have all motivated its growth after periods of full or partial capital market collapse. As in the case of conventional commodity trade, however, the process of market reintegration itself has required and reflected a political context enabling countries to overcome the coordination problems that often can obstruct the gains from trade. 11.1.2 Trends in Current Accounts Following Eichengreen (1992b) and Taylor (1996b), we turn to some basic measures of the extent of capital flows throughout the past century. A sense of the changing patterns of international financial flows can be gleaned by examining their trends and cycles. We focus on the size of the current account balance CA as a fraction of national income !l Thus, (CA/Y),becomes the variable of interest for country i in period t. By dint of the national income identity, the current account equals the difference between national saving S and domestic investment I , CA = S - 1.

and thus corresponds to net foreign investment or, alternatively, to the level of net capital outflow. Thus, the size of the current account measures the extent to which the economy is borrowing abroad to finance its investment (CA < 0) or is lending abroad to facilitate foreign investment in excess of foreign saving (CA > O).2 2. Formally, we make definitions as follows. GNP is given by Y = Q + rB = C + I + G + NX + rB, where Q is GDP, C is private consumption, I is investment, G is public consumption, NX is net exports, and r is the return on net foreign assets B. The current account CA satisfies CA = NX + rB = ( Y - C - G ) - I = S - I, where S = Y - C - G is national saving. Finally, note that the dynamic structure of the current account and the external credit position is given by the equality of the current account surplus (CA) and the capital account deficit (-AX), so that B,+,- B, e -KA, = CA,.

358

Maurice Obstfeld and Alan M. Taylor

Table 11.1 and figure 11.1 present the basic trends in foreign capital To measure the extent of capital flows we use the mean absolute value pICm,, in each cross section at time t. Quinquennially averaged data are used in the figure, and in the table we present data for selected periods. By this measure, the average size of capital flows in our sample was often as high as 4-5 percent of national income before World War I.4 At its first peak it reached 5.1 percent in the overseas investment boom of the late 1880s. This dropped back to around 3 percent in the depression of the 1890s. The figure approached 4 percent again during the years 1910-14, and wartime lending pushed the figure over 5 percent in the years 1915-19. Flows diminished in size in the 1920s, however, and international capital flows were less than 1.5 percent of national income in the late 1930s.Again, wartime loans raised the figure in the 1940s, but in the 1950s and 1960s, the size of international capital flows in this sample reached an alltime low, around 1 percent of national income. Only in the late 1970s and 1980s have flows increased, though not to levels above those of a century ago. These basic descriptive data do illustrate the record of capital flows and offer prima facie evidence that the globalization of the capital market has been subject to major dislocations, most notably over the interwar period, with a dramatic contraction of flows seen during the depression of the 1930s. Moreover, this low level in the volume of flows persisted long into the postwar era. More sophisticated analysis of the quantity (flow) data is, of course, possible. The current account identity may be examined through a study of the relationship between domestic saving and investment, following Feldstein and Horioka (1980). Applying such techniques to 150 years of panel data as described above strengthens the basic impression that capital mobility reached its low point in the 1930s (Eichengreen 1992b; Obstfeld 1995; Taylor 1996b). 11.1.3 Evidence on Nominal Interest Rates Perhaps the most unambiguous indicator of capital mobility is the relationship between interest rates on identical assets located in different financial centers (see the discussion in Obstfeld 1995). Thus, for example, interest rates on Euromark deposits in London in recent years have been quite close to those on comparable deutsche mark deposits in Bonn. The great advantage of comparing onshore and offshore interest rates such as these is that relative rates of return are not affected by pure currency risk.s 3. The open circles in fig. 11.1 and in fig. 11.3 denote gaps in data coverage due to the two world wars. The positions of the circles are determined by the incomplete sample of countries for which data are available. 4. In that era the main lender was Britain, whose current account surplus was often in the range of 5-10 percent of GDP (Edelstein 1982). 5. Eichengreen (1991) presents similar data for the interwar period, as does Marston (1995) for the postwar period. Under a fixed-rate regime such as the gold standard, another arbitrage-like test of financial market integration asks whether nominal interest differentials in different currencies are consistent with the maximal allowable exchange rate fluctuation band (Goschen 1861; Weill 1903; Morgenstern 1959; Officer 1996). Such a test relies on the maintained hypothesis that the

Table 11.1 Period

Extent of Capital Flows since 1870 (mean absolute value of current account as a percentage of GDP) ARG

AUS

CAN

DNK

FRA

DEU

ITA

JPN

NOR

SWE

GBR

USA

All

18.7 6.2 2.7 4.9 3.7 1.6 4.8 3.1 1.0 1.9 2.0

8.2 4.1 3.4 4.2 5.9 I .7 3.5 3.4 2.3 3.6 4.5

7.0 7.0 3.6 2.5 2.7 2.6 3.3 2.3 1.2 I .7 4.0

I .9 2.9 5.1 1.2 0.7 0.8 2.3 1.4 1.9 3.2 1.8

2.4 1.3

1.7 1.5 2.4 2.0 0.6 2.0 1.O 2.1 2.7

1.2 1.8 11.6 4.2 1.5 0.7 3.4 1.4 2.1 1.3 1.6

0.6 2.4 6.8 2.1 0.6 1.0 1.0 1.3 1.0 1.8 2.1

1.6 4.2 3.8 4.9 2.0 1.1 4.9 3.1 2.4 5.2 2.9

3.2 2.3 6.5 2.0 1.8 I .5 2.0 1.1 0.7 1.5 2.0

4.6 4.6 3.1 2.7 1.9 1.1 7.2 1.2 0.8 1.5 2.6

0.7 1.o 4.1 1.7 0.7 0.4 1.1 0.6 0.5 1.4 1.2

3.7 3.3 (5.1) 3.1 2.1 1.2 (3.2) 1.9 1.3 2.2 2.3

~~

1870-89 1890-1 9 13 1914-18 1919-26 1927-3 1 1932-39 1940-46 1947-59 1960-73 1974-89 1989-96

-

2.8 1.4 1.o -

1.5 0.6 0.8 0.7

Source: See Taylor (1996b). Some estimates revised. Notes; Annual data. Parentheses denote average with some countries missing.

360

Maurice Obstfeld and Alan M. Taylor

0

i 0

0 4 . 1870

I

1880

i

~

1890

1

I

1900

~

!

1910

1920

~

1930

I

1940

0

I

!,

1950

I

I

1960

~i

i

1970

,

lI

~

:

1980

1990

Fig. 11.1 Extent of capital flows since 1870 (mean absolute value of current account, 12 countries, quinquennia, percentage of GDP) Source: See table 11.1. Nore: Open circles denote wartime samples.

For much of the period we study in this paper, a direct onshore-offshore comparison is impossible. However, the existence of forward exchange instruments allows us to construct roughly equivalent measures of the return to currency-risk-free international arbitrage operations. Forward exchange trading-in which two parties contract to exchange currencies at a predetermined rate on an agreed date-is one way to conduct international interest rate arbitrage free of currency risk. Using monthly data on forward exchange rates, spot rates, and nominal interest rates for 1921-96, we assess the degree of international financial market integration by calculating the return to covered interest arbitrage between financial centers. For example, a London resident could earn the gross sterling interest rate 1 +:i on a London loan of one pound sterling. Alternatively, he could invest the same currency unit in New York, simultaneously covering his exchange risk by selling dollars forward. He would do this in three steps: Buy e, dollars in the spot exchange market (where e, is the spot price of sterling in dollar terms); next, invest the proceeds for a total of e,( 1 + i,) (where i, is the nominal dollar interest rate); and, finally, sell that sum of dollars forward for e,( 1 + i,)& in sterling (where A, the forward exchange rate, is the price of forward sterling in terms of forward dollars). The net gain from borrowing in London and investing in New York,

exchange rate band is credible (though not on uncovered interest parity) and more recently has been interpreted as a test of exchange rate credibility (Svensson 1991; Giovannini 1993; Marston 1995, chap. 5 ; Hallwood, MacDonald, and Marsh 1996).

361

International Capital Mobility over the Long Run

is zero when capital mobility is perfect and the interest rates and forward rate are free of default risk. We can pursue a similar arbitrage calculation before 1920, but with minor modifications to correspond to historical practice and the prevailing financial instruments and institutions of that earlier time. Forward exchange markets functioned before World War I (Einzig 1937, chap. 7), but they were comparatively thin before the period of floating exchange rates that followed the war. For the period 1870-1920, we consider a different type of sterling interest rate arbitrage operation between London and New York, the dominant market of its kind, going through the New York market for 60-day sterling bills (for a discussion of this market, see Perkins 1978). Sixty-day sterling bills were promises to pay the bearer one pound sterling, usually in London, after 60 days. Thus, the dollar price of a sterling bill is the New York price of forward sterling in terms of current dollars. Rather than lending a pound in London at the gross interest rate prevailing there, one could instead buy e, sight dollars and use these to purchase e,lb,pounds payable in 60 days (where b, is the New York dollar price of a sterling bill). The net gain from borrowing in London to buy sterling bills in New York would be

which again is zero under perfect integration of the New York and London financial markets.6 Table 11.2 and figure 1I .2 present some evidence on covered interest arbitrage on the dollar-sterling exchange since 1870, showing the above differentials for the years from 1870 to the present. Differential returns are calculated as annual rates of accrual. Some concerns about the data warrant mention.' First, as described above, the two measures of market integration that we calcu6. This arbitrage argument underlies the calculation in the textbook by Spalding (1915, chaps. 5-6), e.g., although he assumes perfect international financial integration. Notice that, unlike in the case of covered interest arhitrage, differential default risk between the New York and London markets is not important here. An implicit forward exchange rate based on the ''long'' exchange rate b, is given byf:"'p1"" = b,( 1 + i,). 7. The data were collected from various sources. Exchange rates: Before 1921, spot and 60-day sterling bill exchange rates (in U.S. dollars) are from the Financial Review or Commercial and Financial Chronicle; 1921-36 spot and 90-day forward rates are from Einzig (1937); 1937November 1946 spot and 90-day forward rates are from the Economist; December 1946-May 1947 spot and 30-day forward rates are from the Wall Street Journal; June 1947-1965 spot and 90-day forward rates are from the New York Times; and thereafter, spot and 90-day forward rates are from OECD Historical Statistics and (starting in January 1976) from Reuters (as reported by Datastream). The Einzig foreign exchange data are monthly averages, whereas all other exchange rates are taken at or near the end of the month. Merest rates: For 1870-1920, three-month rates on London bank bills are from Capie and Webber (1985). data taken at or near end of month; 1921-36, month average data on London and New York three-month market discount rates are from Einzig (1937). U.K. interest rate data for 1937-April 1971 are three-month London bank bill rates from Capie and Webber (1985); May 1971-April 1991, three-month London bank bill rates are from Datastream; and May 1991-April 1996, five-month London bank bill rates are from

362

Maurice Obstfeld and Alan M. Taylor

Table 11.2

Nominal Interest Parity since 1870 (U.S.-U.K. covered domestic interest differentials,percent per annum) Period

Mean

S.D.

1870-89 1890-19 13 1914-18 1919-24 1925-30 1931-39 1940-46 1947-56 1957-67 1968-79 1980-89 199W96

1.02 0.60 0.05 1.05 -0.58 0.12 -0.44 1.29 0.02 0.10 -0.64 0.25

0.66 0.39 0.70 2.06 0.95 2.05 0.31 3.50 0.54 2.01 1.28 0.47

Sources: See text. Notes: Monthly data; various maturities. Sixty-day bills before 1920; forward exchange after 1920.

late refer to different arbitrage possibilities before and after 1920, and thus comparability across this break cannot be assured. Second, the forward and sterling bill transactions appear at different maturities in our data set: through 1920 we deal with two-month rates, afterward with three-month rates. Third, most data are observed at or near the end of month, but all data for the years from 1921 to 1936 are averages of weekly numbers. Averaging has the effect of dampening measured volatility for part of the interwar period. Fourth, data from World War I1 reflect rigidly administered prices and have no capital mobility implications. Fifth, the data used are not closely aligned for time of day (and even differ as to day in some cases), so that some of the monthly deviations from nominal interest parity that we calculate may be spurious. Sixth and finally, this exercise is being performed here only for a single country pair, the United States and United Kingdom. In future work we hope to compile similar data for more countries, including France and Germany, to permit an evaluation of covered interest arbitrage between other markets. Despite these many caveats, the figures are revealing and supportive of the conventional wisdom. Differentials are relatively small and steady under the pre-1914 gold standard but start to open up during World War I. They stay Datastream (all at or near end of month). For 1937-April 1940, rates on banker’s acceptances in New York are from the Ecunomist, at or near end of month; May 1940-May 1947, the same rates are monthly averages as reported by the Federal Reserve; June 1947-1965,30- to 90-day banker’s acceptance interest rates in New York come from the New York Times, observed at or near month’s end; January-April 1966, rates are month averages of 90-day banker’s acceptance rates reported by the Federal Reserve; and starting in May 1966, the Federal Reserve reports month-end data, which we have used in the calculations.

363

International Capital Mobility over the Long Run

-4 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 ~

1-t.

mean +s.d

J

Fig. 11.2 Nominal interest parity since 1870 (U.S.4J.K. covered domestic interest differentials, annual, percent per annum) Source: See table 11.2.

quite large in the early 1 9 2 0 Differentials ~~ diminish briefly in the late 1920s but widen sharply in the early 1930s. There are some big arbitrage gaps in the late 1940s through the rnid-l950s, but these shrink starting in the late 1950s and early 1960s, only to open up again in the late 1960s as sterling is devalued and as the Bretton Woods system unravels in the early 1970s. Interest differentials have become small once again only in the most recent years of floating exchange rates. Thus, the Great Depression, perhaps as part of a much broader interwar phase of disintegration, stands out as an event that transformed the world capital market and left interest arbitrage differentials higher and more volatile than ever before. Disintegration lasted for several decades, and large nominal return differentials persisted into the 1980s.

1 1.1.4 Evidence on Real Interest Rates A basic standard for market integration remains the law of one price. This is usually interpreted for capital markets as implying some test for the equalization of real costs of capital, typically real interest rates. It is well known, 8. The rather stable premium on New York loans before World War I (which appears to fall in the early twentieth century) probably reflects a less liquid market. (The London reference rate is a high-quality bank bill rate.) For a comparative discussion of the New York and London capital markets before 1914, see Davis and Cull (1994, chap. 4).

364

Maurice Obstfeld and Alan M. Taylor

however, that real interest rate convergence is a very strong criterion for market integration, resting as it does not only on perfect capital mobility but also on two supplementary parity relationships, either of which may fail to hold and which are not directly relevant to capital mobility: uncovered interest parity (UIP) and purchasing power parity (PPP).’ Clearly, risk premia can modify UIP even in a world of free trade and frictionless asset markets, so UIP cannot be a relevant precondition for free capital mobility. As is well known, however, it is hard to devise reasonable models in which currency risk premia are large (Lewis 1996). As for PPP, it may fail even over the long run because nontraded goods enter consumption price indexes. If capital is mobile and technologies ultimately converge internationally, however, there will be a tendency for countries’ relative prices of nontradables in terms of tradables to be equalized as time passes. The mechanism bringing about this equality is akin to that underlying the factor price equalization proposition in trade theory (see Obstfeld and Rogoff 1996, chap. 4).It can work even without capital mobility, of course, but is likely to be speeded by the technology transfer that international capital mobility may facilitate.’O Thus, even though PPP is sometimes asserted to be a proposition about goods market integration, capital mobility can indeed be relevant to the issue, hence to the international equality of expected real interest rates. Studies of long-run patterns of PPP may be very suggestive of the likely periods in which real interest parity holds or fails. Taylor (1996a) works with a 20-country panel data set and shows that the interwar period, and especially the 1930s, represents an era of marked deviation from PPP in the cross section. In contrast, the recent float shows evidence of a return to a level of conformity with PPP conditions not 9. Uncovered (or “open”) interest parity requires that the expected rate of relative currency depreciation over the relevant horizon equal the corresponding nominal interest rate difference between the two currencies: E,et+,- et = i,

-

Q

Here e, is now the log spot exchange rate (defined as the domestic currency price of foreign currency) and i, and :i are the home and foreign currency interest rates. Purchasing power parity (strictly, the strong relative version of PPP) implies that expected exchange rate changes equal expected inflation differentials:

EF,,, - e, = E ? T + , - Err:, ’ where T,+, = p,,, - p, and a : , = p;, - p: are the inflation rates in the two currencies based on log price levels p, and p:. UIP and PPP together imply the equalization of ex ante real interest rates:

7 = i, - E,T,+,= : i

-

E,T~;, = 7*

(Since UIP compares assets with the same country/political risk characteristics, which may be freely traded against each other, capital mobility comes in by ensuring that interest rates on a given currency are the same the world over, as discussed earlier.) 10. Transport costs or regulations impeding the international shipment of rrudubles will, however, weaken any tendency for countries’ price levels to converge.

365

International Capital Mobility over the Long Run

seen since the classical gold standard of the late nineteenth century These findings beg the question whether real interest parities show similar patterns of historical evolution, as one might expect. The real interest parity relationship between prices in two physically separate, but economically integrated, markets could be tested in a number of ways. Prices may be equalized save for some transaction cost gap; or they may equalize in the long run but exhibit short-run deviations. Such concerns make the test of strict and permanent equality only the most extreme or strict criterion for integration. If integration is not viewed as a zero-one variable, the degree of integration becomes a valid object of research. Exploring both the equilibrating forces of adjustment (promoting convergence in prices) and the nature of disequilibrating shocks to the system (promoting divergence), we may ask: How did the system perform in terms of the overall deviations from real interest rate equality that were actually recorded across time? The dispersion of the real interest rate is of interest because it does indeed summarize two facets of the capital market: the size of market-specific shocks (impulses) and the capacity of the system to smooth out these shocks across space (propagation). The two together describe the stability and unity of the global capital market. Naturally, a system with uniform price shocks across regions will never be put to the test in terms of its adjustment capacity; and a system with good adjustment dynamics could cope with pronounced local disturbances and still dissipate the shocks so as to equalize rates of return quickly. The patterns of real interest rate dispersion offer preliminary evidence as to the working efficiency and stability of the world capital market, and the basic record for our sahple of countries is indicated by table 11.3 and figure 11.3. Shown there is the standard deviation of real ex post interest rates for a sample of 10 countries from 1880 to 1989. Quinquennially averaged data are used in the figure, and in the table we present data for selected periods." The dispersion measure (the standard deviation) of real rates shows a definite pattern. There was slight convergence in real interest rates after 1880 and before 1914, though nothing as dramatic as the convergence seen prior to 1870 (cf. Lothian 1995). But after 1914, the dispersion of real rates rose sharply. It fell slightly from the late 1920s to the early 1930s but then increased again. Dispersion hit a peak in the late 1940s, and then convergence in real interest rates was again seen during the early phase of Bretton Woods, such that, after 1960, dispersion levels had returned to their pre- I914 levels. Dispersion has been flat almost ever since, with some divergence apparent upon the collapse of the Bretton Woods system in the early 1970s. The dispersion data thus confirm again the textbook characterization of the 11. The method echoes the recent study of Lothian (1995), and we find similar patterns here. Eichengreen (1991) compares and analyzes real interest differentials over subperiods of the interwar period.

366

Maurice Obstfeld and Alan M. Taylor Real Interest Parity since 1870 (dispersion of real interest rate, percent per annum)

Table 11.3

AUS

BEL

CAN

FRA

DEU

ITA

NLD

SWE

GBR

All Countries S.D.

4.5 2.8 7.9 3.5 2.9 3.3 3.3 4.4 1.3 1.9 3.3

3.8 4.1

4.3 2.1 1.3 2.2 1.1 1.2 3.2 1.8 1.0 1.2 2.0

4.8 4.2

1.6 2.4 0.7 6.1 2.9 2.6

1.8 1.8 11.0 9.4 4.8 4.7 38.5 4.4 2.0 2.2 2.6

3.3 2.8 5.8 4.2 1.7 3.8 5.5 4.2 1.7 2.3 1.8

3.2 2.6 8.3 5.8 1.8 3.4 5.2 3.0 1.4 1.8 1.8

3.3 1.8 6.2 4.6 1.3 3.0 3.9 2.1 1.3 3.2 1.5

4.2 3.4 (9.1) 20.2 6.0 6.5 (22.3) 6.0 1.6 3.7 1.7

Average Absolute Differential Relative to USA Period

1870-89 1890-1913 1914-1 8 1919-26 1927-3 I 1932-39 1940-46 1947-59 1960-73 1974-89 1989-96

-

34.1 7.6 3.3 10.7 3.3 1.3 2.4 2.4

-

13.5 6.6 10.7 -

4.5 0.9 1.5 2.3

-

4.5 1.8 2.4 1.3

Source: Unpublished data from Michael Bordo, with some series extended. Notes: Annual data. Parentheses denote average with some countries missing. 0 0

25 301

"I 15

1870

1880

1890

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

Fig. 11.3 Real interest parity since 1870 (dispersionof real interest rate, 10 countries, quinquennia, percent per annum) Source: See table 11.3. Nore: Open circles denote wartime samples.

evolution of international capital market integration. Integration was relatively high in the late nineteenth and late twentieth centuries-though arguably no better now than under the gold standard-and it was subject to a massive dislocation in the interwar period. Thus, together with the evidence on the extent of capital flows, this subsection again offers support for the view that the depression marked a low point in the modern history of international capital mobility.

367

International Capital Mobility over the Long Run

11.2 World War I and the Interwar Period World War I demonstrated the capacity of governments to radically alter exchange rates and price levels, often with the assistance of explicit controls beyond and above normal central bank operations. These newly found powers were not quickly forgotten. In the early 1920s, they were used to ease the deflationary adjustments of economies seeking to repeg to gold and were abused in the monetary mayhem surrounding the hyperinflations in several European states. When the fleeting interwar gold exchange standard took form after 1925 the older laissez-faire approach to the exchanges was briefly reestablished as many countries eliminated or relaxed exchange controls, but the international financial crisis of 1931 dealt a final blow to the old orthodoxy. Sterling’s departure from its gold peg in September 1931 heralded the demise of the gold-based system as well as the return of exchange controls, “in many ways to an even greater extent than during and after the war” (Einzig 1934, 1-2). Out of the resulting economic and political turmoil would emerge the new consensus on international macroeconomic coexistence embodied in the Bretton Woods agreement of 1944. The effectiveness of exchange controls varied greatly. Naive policies contained loopholes through which regulations on capital flows might be evaded. Certain controls proved hard or even impossible to implement, but others, when sufficiently refined by the increasingly cunning authorities, served their purpose. A measure of the impact of such policies was the common appearance of the “black bourse” in some of the most tightly controlled economies. Free market rates often diverged widely from official rates. This added further uncertainty to foreign exchange markets already subject to frequent, often violent, fundamental movements after floating rates appeared in the wake of gold standard suspensions during and after World War I. With the worlds nominal anchor removed, massive exchange risks reentered the calculation of every foreign investor. Controls, if they threatened to compromise the secure and full repatriation of profits or principal, heightened risk further and could prompt capital flight or the collapse of lending. Speculative activity in the forward market, and the emerging threat to central banks and treasuries posed by increasing volumes of highly liquid, “hot” money, prompted even greater caution in the bureaucratic supervision of foreign exchange transactions. Exchange controls thus compounded a deteriorating framework for international capital flows. 11.2.1 World War I and the Return to Gold Direct controls over private exchange transactions were rarely employed under the gold standard before 1914. Central banks occasionally used “normal” measures to support exchange rates, broadly defined to include moral suasion over banks, direct interventions to alter gold export and import points, and

368

Maurice Obstfeld and Alan M. Taylor

other formally noncoercive devices. But if a central bank could no longer defend the exchange rate through such measures, as often occurred in Latin America, the rate was generally set free to float with no control employed. Within Europe, the credibility of exchange parities was bolstered by Britain’s hegemony within the world financial system and its espousal of free trade, as well as by central bank cooperation and the overriding and largely unquestioned commitment of central banks to the goal of gold convertibility at an unchanging par (Kindleberger 1986; Eichengreen 1992a). Credibility ensured that capital movements were usually stabilizing. The high degree of international capital mobility was promoted by the gold standard regime; and by reducing actual gold movements, capital mobility in its turn helped the system to function smoothly. The Great War destroyed this equilibrium, and the classical gold standard too. Initially, countries kept up the appearance of the gold standard, maintaining official gold coinage, pegging official exchange rates, and, on paper, permitting the movement of gold; but obstacles and regulations, as well as heightened susceptibility to patriotic appeals, prevented normal functioning according to the rules of the game (Eichengreen 1992a, 67). The belligerent countries were the first to enforce controls. Wartime needs drove their trade balances into deficit, and monetization of fiscal deficits drove inflation, though to widely differing degrees in the several countries.Although exchange control became an “obvious necessity” in these circumstances, countries did not produce a full-blown, cut-and-dried system of controls at the outbreak of war. A gradual implementation of ever stricter controls ensued, although trading with the enemy was quickly terminated. In 1914 and early 1915, belief that the war would be short and swift kept the exchanges fairly stable. It was not until later in 1915 that general foreign exchange transactions came under restriction as the exchanges became more volatile (Brown 1940,59-63; Einzig 1934,22-23; League of Nations 1938,9). Allied experiences varied considerably. The British began in 1915 by pegging the dollar rate of sterling, with the British Treasury acting via J. P. Morgan in New York to support sterling at $4.7640 using gold and dollar reserves. In the early war years the country often came near to exhausting its reserves, as recounted by Keynes (1978, 10-12). After 1917, it was the U.S. Treasury that supplied the required funds, and the peg continued. France employed similar methods to defend a franc-sterling peg, albeit with both “passive” and “active” intervention by the Bank of France. In the later stages of the war exchange controls grew much stricter than in Britain. Whereas appeals to patriotism and other types of moral suasion had sufficed to discourage outbound capital transfers for a while, in the end tougher measures were needed. France’s more severe inflation problems undermined the credibility of the peg, and capital outflows were harder to tame. Italy likewise pursued a policy of pegging against sterling. Like these peggers, allied powers that did not peg their currencies

369

International Capital Mobility over the Long Run

nonetheless resorted to exchange controls toward the end of the war. Even the United States applied direct controls after entering the war in 1917, despite its strong trade balance, as a result of the dollar’s appreciation against several neutral currencies (Einzig 1934,28-29; Eichengreen 1992a, 73). In Germany, the mark was never pegged to another currency during the war, and the Reichsbank spent a mere 450 million marks on intervention to defend the mark in times when depreciation threatened to become a burden-a tiny fraction of the inter-Allied resources devoted to currency support. However, Germany’s trade was effectively blocked by the Allies, so its adverse net export balance was rather small. Exchange restrictions did not come into force until 1916 and were only mild until they were dramatically reinforced in 1917 (Einzig 1934,29-30). Still, Germany employed strong forms of compulsion to mobilize residents’ foreign securities (Eichengreen 1992a, 83). The Armistice gave hope that wartime exchange controls would soon be removed and the prewar state of affairs would soon prevail. The removal of controls was one of the few things the Brussels Conference of the League of Nations (1920) could agree on (Eichengreen 1992a, 154-55). In the United States exchange control was dismantled, and in Britain controls had largely ended by the time sterling rejoined the gold standard in 1925. The dollar peg ended in March 1919, and sterling was cut loose to take care of itself. The authorities refrained from direct control measures; however, occasional weak embargoes on British foreign loans were enforced starting in 1924 to bolster the currency as it inched back toward gold parity.I2By floating in 1919, Britain was able to open its capital market relatively quickly after the war’s end. Elsewhere, however, stability proved elusive, and exchange controls had to be maintained or reinforced after the war as many countries descended into economic chaos (Einzig 1934, chaps. 4-5). When inter-Allied support ended, a rapid flight from the franc ensued with rampant bear speculation, and the Bank of France remained neutral and impassive, preferring to husband its gold stock rather than intervene in a probable losing cause. Harsh exchange controls were promulgated, but it was not until after the franc stabilized in 1926, and vast sums were repatriated, that all could see how ineffective the controls had been. Moreover, other factors impinged on capital flows, notably the fierce controversy over the capital levy. With a broader franchise, political groups representing labor now tried to force capital to shoulder a larger part of the fiscal burden. Deadlock persisted as governments came and fell. In 1925, the capital levy was nearly adopted by the government, a 10 percent tax on all wealth over 10 years, and although the govern12. The British government, for revenue reasons, also levied a stamp tax on foreign bearer bonds. See Moggridge (1971), who concludes, however, that British government suasion over foreign lending was largely ineffective in keeping capital at home and that the stamp tax could be evaded. His conclusion receives support from fig. 11.2, which shows that from 1924 through 1930, sterling interest rates in London frequently exceeded the covered sterling return on New York loans.

370

Maurice Obstfeld and Alan M. Taylor

ment fell on a no-confidence vote by the Senate, the capital levy idea was only killed for good by the Poincark government’s fiscal stabilization package in late 1926. In the interim, the lingering possibility of a wealth tax sent capital fleeing abroad (Einzig 1934, chaps. 4-5; Eichengreen 1992a, 172-79). As inflation seized Germany’s economy between 1919 and 1923, even tighter exchange restrictions were deployed to halt the slump in the mark. Exporters and importers had all exchange requests subjected to government approval, and indirect controls were used to restrict imports. Even so, capital flight from Germany went on unabated, the black bourse in Cologne was very active, and the western border became a major point of leakage. Exchange controls in the successor states of the Austro-Hungarian empire took even more esoteric In Italy, the postwar Fascist regime enjoyed greater success in controlling the exchanges, however, largely as a result of its extraordinary powers of enforcement. In most countries, restrictions eventually were relaxed following stabilization. Still, there were delays; for example, although the mark stabilized in 1923, the last restriction was not removed until 1926. And very often stabilization had only been achieved by dint of exchange controls in the interim. Nonetheless, by 1927, most of the world’s market economies had returned to “normalcy” in the form of pegged exchange rates and some form of gold standard.

11.2.2 Interwar Exchange Control In sharp contrast with the laissez-faire approach normal during the classical gold standard era prior to 1914, the interwar period saw a marked increase in the adoption of policies to control not only international capital flows but foreign exchange transactions in general (Einzig 1934; Gordon 1941). Controls over foreign currency transactions took several forms. In assessing how controls affect capital mobility, we are primarily concerned with measures that would have been viewed as “abnormal” under the gold standard-steps taken to defend or change the course of the exchange, and covering direct measures such as loan embargoes and foreign exchange rationing and “indirect” measures to influence the foreign trade or foreign loan markets. Such measures were attempted fitfully in the 1910s and 1920s, but their reappearance in “extreme forms” dated from the crisis of 1931 (Bratter 1939, 274).j4 Such interventions served a variety of purposes of concern to the policymakers: to counteract the transfer of liquid balances, the flight of national capital, the possibility of speculation, fluctuations in the trade balance, or exchange 13. It was during a brief period in 1919 that Joseph A. Schumpeter served as Austrian finance minister. He favored a capital levy. 14. In many ways the direct and indirect measures are interchangeable as policy instruments. Differential exchange rates according to type of good may yield the same relative price structure as a tariff schedule. Barter arrangements in trade resemble the outcome of a strict bilateral exchange-clearingarrangement.Thus, although direct measures impinge directly on foreign capital movements, so too do the indirect measures.

371

International Capital Mobility over the Long Run

controls of other countries. Controls could be called upon to offset mere dayto-day fluctuations or problems, to stem persistent speculation or capital flight, to smooth predictable seasonal and other normal tendencies, or to attempt to reverse fundamental trends. In many cases such attempts to distort capital account transactions were also complemented by commercial policies (tariffs and quotas) aimed at inhibiting the volume of current account transactions. In fact, given the balance-of-payments identity, policymakers viewed certain exchange control and tariff policies as pure substitute^.'^ From a macroeconomic viewpoint, controls enabled a government to maintain (at least nominally) a pegged exchange rate, while simultaneously using interest rate policies and other policies based on divergences between internal and external prices to attain domestic economic objectives. Exchange controls became “among the best-hated” forms of government interference in free markets in the eyes of observers and market participants (Einzig 1934, 106). Controls were criticized for causing exchanges to diverge from their fundamental levels (though identifying fundamental levels proved elusive in the interwar chaos) and for their damaging effect on international trade and finance (though the effect of exchange controls here could not be easily differentiated from the corrosive effects of tariffs, quotas, and other commercial policy choices). Even if not implicated on these charges, exchange controls were subject to even more stinging criticism, facing ridicule for being “utterly inefficient and impossible to enforce” (Einzig 1934, 107). The nettlesome interferences with the exchange were thus in vain, critics charged, on account of weak policing and enforcement and the numerous loopholes that savvy exchange dealers could easily exploit to circumvent the intent of the restriction. If the speculators proved strong enough for the task, the authorities faced certain defeat, and the incentive to exploit loopholes only loomed larger as the exchanges moved further from fundamentals, inviting arbitrage. Such was undoubtedly a major weakness of the early and rudimentary controls seen in the 1920s, as in the French and Belgian cases. Embargoes on loan issues might fail if investors were willing to purchase issues in a third country or if short-term trade credits could be disguised and employed to finance longer term capital flows. Partial controls could be futile, as transactions might be easily disguised in false categories, necessitating full-blown supervision of every transaction (Nurkse 1944, 165). Evasion could never be totally eliminated, but authorities learned the lessons of failed controls and became more ruthless in imposing and enforcing trading restrictions as the 1930s wore on. All countries had access to a variety of measures, beginning with unofficial discouragement of capital export through 15. Bratter again: “In effect control of the volume of foreign exchange transactions with foreign countries amounts to determination of the value or volume of goods and services exchanged with foreign countries. Exchange control accomplishes the purpose of a protective tariff or an import embargo. And it has the further ‘advantage’ that it often operates secretly as to the details” (1939, 274).

372

Maurice Obstfeld and Alan M. Taylor

moral suasion, official prohibition with the cooperation of the banks, or direct prohibition without the help of the banks. More desperate measures included even more restrictive allocations of exchange for loans and import, the compulsory surrender of export proceeds, and, finally, the complete suspension of free dealings, that is, a crackdown on the black bourse.16 By the 1930s the criticism that the controls were ineffective could be said to have lost much of its force (Einzig 1934, 112). Capital controls were now binding on the global capital market to an unprecedented extent. Although devised primarily as a response to short-run problems with capital flight, even the prospect of modest barriers to outward flows undermined the efficient allocation of global capital. As Ellis succinctly summarized, capital controls “may interfere with the tendency of capital to bring its marginal employments to equality and thus maximize yields . . . in preventing capital repayments, exchange control effectively discourages the investment of new foreign capital. Since the ‘natural‘ direction of capital flow was toward the debtor (now exchange control) countries, this is probably the more serious consequence” (1941, 22). In addition, distributional conflicts over the tax burden raged on between labor and capital, and the possibility of a capital levy in some countries “hung over investors like a fiscal sword of Damocles, discouraging saving, provoking capital flight, and heightening the fiscal crisis” (Eichengreen 1992a, 107). Over the course of the 1930s, external debt default became widespread in Europe and Latin America (Cardoso and Dornbusch 1989, 1394). All of these factors contributed to the dramatic fall in international capital mobility documented in the previous section. 11.2.3 Controls as a Reaction to the Great Depression, 1931-39 The global Great Depression and the financial instability accompanying it were directly responsible for the sharp turn toward exchange control in much of the world. Stability on the exchanges came to an abrupt end in 1931, though trouble had been brewing longer in many countries, especially at the periphery (Einzig 1934, chap. 6; League of Nations 1938, 10-11; Ellis 1941, 7; Yeager 1976, chap. 17). Currency crises in 1931 led to flights from the Austrian schilling, the Hungarian pengo, and the German mark following the Creditanstalt collapse. It appeared that exchange control might be the only policy alternative since when “flight psychology” prevailed “no increase in the discount rate may be sufficient to deter it. Indeed an increase in the discount rate, by shaking confidence further, is apt to produce the opposite effect”; yet, confoundingly, “the introduction of control itself. . . tended to upset confidence further, increasing the urge to export capital” making the exercise “self-aggravating to some extent” (Nurkse 1944, 162-63). Policymakers groped for a solution. In July 1931 a flight from sterling began, leading to gold standard suspen16. Kindleberger (1984,392) notes the very desperate measures favored in fascist Germany and Italy, where punishments continuously increased in seventy until they included the death penalty in both countries.

373

International Capital Mobility over the Long Run

sion in September. Facing high unemployment, the British government had no stomach for an aggressive defense of the pound through budgetary retrenchment, which would have required scaling back the dole. Nor did the Bank of England carry out an aggressive interest rate defense. Instead, bank rate was raised shortly before the announcement of the gold standard‘s suspension “as a measure of reassurance against inflation” (Sayers 1976,412). Soon the dollar and other currencies were exposed to runs, and flight from the yen drove Japan off gold before year’s end. Not all currencies fell from gold immediately, but the fear grew. In such circumstances, exchange controls inevitably returned to prominence: governments fought off depreciation and convertibility crises with intervention, exchange restrictions, and other forms of exchange control (Nurkse 1944, chap. 7). Simple intervention usually proved ineffective in the face of continued gold drain, as with Germany, Austria, and Hungary in the summer of 1931. Coffeehouse transactions on the black market soon undermined rationing through the banks. German restrictions were severe, foretelling the blocked balances and other obstructions to come. In July 1931 a partial transfer moratorium was announced, suspending principal payments and later extended in a full standstill agreement with Germany’s creditors. Only thus was a collapse of the mark prevented. Both Austria and Germany’s banking systems stood on the verge of collapse, and choosing to sacrifice gold convertibility for bank stability, the governments adopted exchange control. In fall 1931 Britain promulgated several mild exchange restrictions following suspension and lasting for six months, primarily to prevent capital flight. In general during the 1930s, Britain employed relatively limited controls, ranging from persuasion, to an embargo on large foreign bond issues, to official restrictions applied by banks that remained loyal to policy goals. But these measures were far from comprehen~ive.’~ The United States, under the Hoover administration, continued to maintain dollar convertibility into gold at $20.67 per ounce. As an accompaniment to President Roosevelt’s suspension of gold convertibility in 1933, however, the United States began to deploy informal pressures similar to those used in Britain, though occasionally enforcing official supervision of banks when an assumption of loyalty could not be taken for granted. In Japan, back on gold only since 1930 and suffering the fiscal strains of the Manchurian campaign, a gold embargo was applied to stem severe losses in 1932; depreciation heralded the end of convertibility and the application of more restrictions on foreign exchange to prevent capital flight. France also generally avoided direct measures, relying on tariffs, quotas, and other commercial policies to keep the trade balance favorable and gold stocks plentiful-but the 17. Stewart noted: “There is, first of all, complete freedom of transferring pounds sterling into foreign currencies and, secondly, there are ample facilities for the purchase in London of foreign securities. The inconsistency of keeping these channels open while maintaining a strict embargo on new foreign issues has been severely criticised” (1938, 57).

374

Maurice Obstfeld and Alan M. Taylor

gold bloc could not hold on forever (Yeager 1976, chap. 18). Italy’s government made very effective use of unofficial restrictions by dint of the powerful command of the banking system at central government level, and capital flight in 1935 forced Belgium into very stringent controls (Einzig 1934, chap. 6; Eichengreen 1992a, chap. 9).18 The tendency toward the forcible confinement of foreign exchange within borders was perhaps most famously institutionalized in the widespread adoption of the system of “blocked balances” in Central European and Latin American countries, and notably in Germany under the Gold Discount Bank (Einzig 1934, chap. 12; Ellis 1941, 13-17). Simply put, “blocked currencies” enshrined the idea that debtors could make debt payments not with foreign exchange but with domestic currency placed in special, earmarked accounts, funds that the creditor might only then use in limited ways, for renewed direct investment in the debtor country or to buy more of the debtor’s exports. Thus, the blocked account became a new payoff option unilaterally imposed by debtors and effectively defaulting on the terms of their original loan contract^.^^ Moreover, payment into a blocked account was often illusory as a financial transaction, entailing no shift in the structure of international indebtedness, affording no liquidity to the creditor, and usually enforcing no loss of liquidity on debtor banks that often maintained currency issues backed by blocked accounts.20Owing to this vehicle for credit creation, blocked accounts were easily manipulated “for disguising the insolvency of the debtors, and especially of one particular debtor-the Government of the debtor country” (Einzig 1934, 126-27). Thus, an insolvent government might pay off debts into its blocked account then relend to itself out of the same funds. Inevitably, claims on such blocked accounts soon began trading on the secondary market at a heavy discount. An international market soon developed in the 1930s for four types of German marks, six types of Hungarian pengoes, and many other blocked currencies. Market rates diverged dramatically from the official par rates of the exchange-controlled domestic currency.*’ Germany, Austria, and Hungary all developed complex systems of blocked currencies and bilateral clearings (Ellis 1941; League of Nations 1938, 16; Yeager 1976, 368-71). Many other countries in Central and Southern Europe followed suit, causing the return of virtual barter conditions in many goods markets and stifling foreign investment. However, the German case remains 18. Italy’s controls were “so stringent as to render her gold bloc status meaningless” (Eichengreen 1992a, 357). The Belgian controls admitted loopholes and were rendered immaterial within weeks as speculators, anticipating a devaluation, provoked one (Eichengreen 1992a, 362-63). 19. This payment was an option typically more injurious to the creditor than even a temporary moratorium-a suspension that might only for a time prevent the discharge of debts but that did not inflict any change in the final terms of settlement. 20. Put another way, the banks treated the accounts as reserves, rather than as earmarked funds not strictly available. 21. Discounts were low for countries whose exports were in demand but very high for currencies whose only use was for very unattractive direct foreign investment in the debtor country.

375

International Capital Mobility over the Long Run

the example par excellence of this type of exchange control-and by the late 1930s it had persisted beyond the point where it was economically defensible, seeming more a tool of national and international political power via favorable allocations of trading activity to domestic agents and foreign trading partners. Thus, after the immediate 1931 crisis, the reasons for keeping exchange control mutated, and the control “introduced in the first instance mainly to prevent capital exports soon shifted its emphasis to the control of commodity imports” (Nurkse 1944, 166). Political as well as economic concerns surfaced, with the free market or black bourse as the only recourse for all but a few restricted transactions. “De fucto and sub rosa devaluation transformed the official rate of exchange to a mere face-saving fiction” (Ellis 1941,293). In contrast to Germany, Austria long before the Anschluss was already relaxing controls (Ellis 1941, chap. 2). Comparable measures to relax exchange controls and bilateral constraints were to be seen in Romania, Yugoslavia, Hungary, Czechoslovakia, Bulgaria, and elsewhere in the exchange control bloc (League of Nations 1938,40-45). Bilateral exchange clearing was beginning to be seen as a welfare-reducing, trade-diverting choice justified by “ulterior ends”; one such end was protection, which “appeared as a by-product of attempting to defend the currency, but it proved to be so welcome a by-product as certainly to become an end itself” (Ellis 1941, 297). By obstructing trade along lines of comparativeadvantage, clearings frequently depressed domestic exports, only exacerbatingthe shortage of foreign exchange that exchange controls were supposed to alleviate. More and more countries turned away from trading under such constraints with countries in the clearing bloc (League of Nations 1938, 24-37; Nurkse 1944, 177-83).22 In Latin America, countries both depreciated their currencies and joined the movement toward exchange control as the depression deepened and after sterling left gold (Bratter 1939; Nurkse 1944, 162). Most also defaulted on their foreign debts, an event that had a profound negative impact on subsequent capital inflows to the region, as many defaults were not settled until the 1940s and even the 1950s (Diaz Alejandro 1983, 27). Controls were initially a response to balance-of-payments crises resulting from a collapse of primary product prices and quanta in export markets, the stickiness of import demands, and large fixed nominal debt obligations. However, controls were generally less rigid than in Europe, with a liberal attitude taken to foreign exchange transactions outside normal channels-thus, some capital account transactions were permitted and black markets were tolerated, while in Europe such flows were strictly controlled; and Latin American countries were generally less inclined to adopt bilateral clearing arrangements save under duress (League of Nations 1938, 17; Nurkse 1944, 170). The key instrument was the rationing of 22. Absurd examples of trade diversion included the import of raw materials in a bilateral clearing deal and subsequent reexport at a large loss to a free currency country, undercutting the original producer, simply as a means for the reexporter to obtain foreign exchange (League of Nations 1938, 35).

376

Maurice Obstfeld and Alan M. Taylor

exchange for different uses according to government priorities, implying multiple exchange rates (League of Nations 1938, 15). High priorities were usually debt service (unless in default) and essential imports.23The more “reactive” countries soon adopted controls: Argentina, Bolivia, Brazil, Chile, and Colombia (in 1931); followed by Costa Rica, Nicaragua, Paraguay, and Uruguay (1932); Ecuador (1933); Honduras (1934); and Venezuela ( 1936).24Argentina, Brazil, Chile, and Colombia were locked in clearing agreements with Germany, and these plus Costa Rica and Uruguay controlled trade along bilateral lines via exchange controls or clearing agreements. Such arrangements had marked consequences for regional trade, and a good deal of Latin American trade was canalized bilaterally not by choice but by the actions of European trading partners and to the detriment of rival markets. In many countries, trading with Nazi Germany under special internationally nontransferable commodity-specific marks was seen to have dramatically altered the composition of trade.*’ Although the Pan American Union called for moves to abolish controls in 1936, the 1937 recession again exposed the underlying weakness of the periphery’s balance-of-payments position, and no concrete action was taken to lift controls before the war (Bratter 1939, 286). Australia was also typical of peripheral primary producers caught in a balance-of-payments crisis and facing unsustainable capital outflows as early as 1929. The terms of trade had dived from a peak in 1924, reflecting oversupply in the wool market. Mild controls appeared first in the form of foreign exchange rationing, and soon the currency slipped outside the gold points. It was 8 percent off par by April 1930. A steady devaluation ensued, and a black market appeared, compromising the policy of rationing via the banks. Eventually the system broke down, and the currency was devalued to 30 percent be23. The discovery that such policies could radically alter the shape of foreign trade and the level of domestic economic activity eventually allowed new and broader purposes of economic control to motivate the use of exchange control, beyond the presumably temporary intent to manage transitory payments crises. It was partly thus that “reactive” policies of the 1930s paved the way for a transition to import substitution strategies in the 1940s and 1950s. See Fishlow (1971) and Diaz Alejandro (1984). 24. Of these, only Venezuela permitted a completely “free” parallel market; other countries intervened to greater or lesser extents. More “passive” countries such as Cuba, the Dominican Republic, Ecuador, El Salvador, Guatemala, Haiti, Mexico, Panama, and Peru did not institute any controls in the 1930s (Bratter 1939, 280-81). The methods of exchange control varied. E.g., in Argentina, the government still favored allocation of foreign exchange to balance bilateral trade, much to U.S. consternation, and largely as a result of the 1933 Roca-Runciman Treaty with Britain-itself a deal struck to offset British imperial trade preferences established in the Ottawa Treaty (Bratter 1939,279-81; Salera 1941). A much stricter regime of controls held sway in Uruguay-four varieties of exchange rate were subject to manipulation, bilateral clearing arrangements were even more constraining, and attempts to favor particular products and trading partners more pervasive (Bratter 1939,281-82). 25. Between 1929 and 1937, the British shares of imports fell in Brazil (19.2 percent to 12.1 percent), Chile (17.7 percent to 10.9 percent), and Peru (15.0 percent to 10.3 percent), while German import shares to all three rose (Brazil, 12.7 percent to 23.9 percent; Chile, 15.5 percent to 26.1 percent; and Peru, 10.0 percent to 19.7 percent; all figures from Bratter 1939, 284).

377

International Capital Mobility over the Long Run

low par at the start of 1931. Canada followed similar steps to limit gold export and convertibility, inevitably leading to devaluation of its currency (Eichengreen 1992a, 232-36,240). 11.2.4 Policy Outcomes Once the initial crisis of 1931 passed, policymakers faced a choice: on the one hand they could treat exchange controls as a temporary expedient for the crisis and thereafter work toward free exchanges, sacrificing policy autonomy; or else they could retain and enhance the security of their controls and so expand the range of policy options. Thus, by the mid-l930s, countries could be classified as “free currency” countries (whether on or off the gold standard) or “exchange control” countries (League of Nations 1938; Eichengreen 1992a, 339). (Of course, even countries in the former group could employ informal measures of capital account control, such as moral suasion, in the interest of exchange rate or balance-of-payments targets.) Table 11.4 illustrates some of the contrasts between the exchange rate experiences of the two groups. The free currency group included much of Scandinavia and Western Europe. Among these, Belgium, the Netherlands, Switzerland, and France (along with the United States until 1933) were in the “gold bloc” and avoided devaluation only through strong indirect measures (e.g., tariffs and quotas) in the early 1930s. By contrast, exchange control economies included Germany, Austria, Hungary, and neighboring countries to the east, plus Turkey, Italy, and the Baltic states-a largely Central and southeastern European grouping. In the latter group, generally severe exchange controls allowed governments the freedom to maintain parities (or tolerate only relatively mild devaluations) without fear of speculative attacks, as capital flight was severely contained.z6 Recent academic writing has emphasized the role of the international gold standard in propagating the Great Depression, showing systematically how countries that maintained gold parities and continued approximately to follow other gold standard rules of the game through the mid-1930s suffered much sharper output declines and deflati~n.~’ Countries willing to devalue could 26. Equivalently, Eichengreen (1992a. 258) uses a three-category classification consisting of “exchange controlled,” ‘‘sterling area,” and “gold bloc”-in practice, those not on controls or pegged to gold chose a sterling peg, with a few exceptions (e.g., Canada, which pegged to a sterling-dollar basket). After 1935 the gold bloc collapsed. France, Switzerland, and the Netherlands departed from their earlier policies, but without control. Some exchange control countries did choose to devalue-e.g., Italy, Czechoslovakia, and Greece-and some no longer adhered to official rates of exchange. The conclusion of the Tripartite Agreement between Britain, France, and the United States lent a modicum of stability and a veneer of cooperation to international monetary arrangements, and conditions improved until the recession of 1937 (League of Nations 1938; Eichengreen 1992a). 27. See, e.g., Choudhri and Kochin (1980), Diaz Alejandro (1983), Eichengreen and Sachs (1985), Hamilton (1988). Temin (1989), Campa (1990), Eichengreen (1992a), Bernanke (1995), and Bernanke and Carey (1996). These writers have followed on a nonformal tradition that quite clearly appreciated the basic monetary forces at work in propagating and prolonging the depression

378

Maurice Obstfeld and Alan M. Taylor

Table 11.4

Currency Depreciation in the 1930s (percentage depreciation relative to official gold panty) Country Exchange control countries Bulgaria” Germany” Hungaryb Romania” Latvia“ Turkey” Italy” Czechoslovakiaa Austria’ Yugoslaviab Estoniah Denmarkb Uruguayb Argentinah Free currency countries (gold bloc) France” Netherlands” Switzerlanda Belgium” Poland“ Free currency countries (devaluers) Irelandb South Africab United States” Canadab United Kingdom” Swedena Norway” Finland“ New Zealandb Australiab

1932

1935

0.0 0.0 0.0 0.0 0.0 1.7 1.5 0.0 22.0 6.8 0.0 29.7 54.5 39.4

0.0 0.0 0.0 0.0 1.7 1.9 6.3 16.2 22.0 23.0 39.9 51.5 53.9 54.3

0.0 0.0 0.0 0.0 0.0

0.0 0.0 0.0 3.2 0.0

28.0 2.1 0.0 11.9 25.2 25.9 26.9 36.4 34.2 42.5

40.2 40.8 40.8 40.9 41.9 45.6 47.0 50.4 52.3 52.6

Source: League of Nations (1938,50-5 1). “Annual average. bMonthly average for March 1932 or 1935.

but that lacked a rigorous analytical and statistical framework for representing their global scope. Thus, Edward M.Bernstein, Harry Dexter White’s deputy at the U.S. Treasury during the Bretton Woods negotiations and the first research director of the International Monetary Fund, recalled in 1984 that “we [at the Treasury] held that the Great Depression was caused by the interaction of the wartime inflation and the traditional gold standard.. . . The Great Depression did not end until every country had abandoned the gold parity of its currency” (Black 1991,98). See also Haberler’s (1976) evaluation. Eichengreen (1992a) cites Ralph Hawtrey and Lionel Robbins as early precursors. Unfortunately, the insular focus of much American macroeconomic thinking for at least 35 years after World War I1 tended to blind many U.S. scholars to the powerful international monetary transmission mechanism at work during the depression.

379

International Capital Mobility over the Long Run

lower the relative price of national output and expand their money supplies, boosting effective demand and employment while retaining a relatively open capital market. Exchange control countries addressed the macropolicy trilemma by eliminating capital movements. While officially maintaining 1931 gold parities, these countries effectively devalued their currencies through a maze of restrictions on foreign exchange acquisition. Elimination of dependence on international capital markets in some cases increased the scope for domestic fiscal expansion, as in Germany. But countries in the gold bloc, despite resort to conventional trade policies, felt the full force of the policy trilemma, maintaining initial gold parities and free foreign exchange markets only at the cost of a deep and protracted domestic slump. Econometric evidence points to independent roles for controls and exchange depreciation in mitigating the effects of the depression. The equation below is a 26-country cross-sectional regression of the 1929-35 cumulative rate of growth of industrial production, AIP, on a constant and two dummy variables. FIXED takes the value one for countries that held their official exchange rates fixed at 1929 levels longer than the United States (which severed the dollar’s link to gold in April 1933) and takes the value zero for others. CONTROLS equals one for countries classified by the League of Nations in the mid-1930s as exchange control countries and is zero for free exchange countries. (The exchange control group comprises a wide variety of control strategies, some much more stringent than others, and omits countries that applied controls only fleetingly.) The result of estimation (with standard errors in parentheses) is

AIP = 0.028 - 0.261 FIXED (0.060) (0.080)

+ 0.213 CONTROLS,

R2 = 0.41.

(0.079)

As is now well known, countries that retained fixed exchange rates suffered harsher real contractions. According to the preceding equation, they experienced (on average) over the years 1929-35 a 26 percent output decline avoided by countries that devalued. However, controls (which usually implied de facto devaluation) had a significant mitigating effect on the extent of output decline due to fixed exchange rates.28 The output effects of controls are mirrored by the behavior of the price level, 28. The exchange control countries in the sample are Argentina, Austria, Czechoslovakia, Denmark, Estonia, Germany, Greece, Hungary, Italy, Japan, Latvia, and Romania. The free exchange countries are Australia, Belgium, Canada, Finland, France, Netherlands, New Zealand, Norway, Poland (which imposed controls only in 1936). Spain, Sweden, Switzerland,the United Kingdom, and the United States. (Some countries, such as Argentina, Austria, and Denmark, both devalued early and imposed controls.) Wholesale price index data come from League of Nations, World Production and Prices, I937/38. Industrial production data come from the same source, except for Argentina, Australia, and Switzerland, the data for which are used in Bernanke and Carey (1996) and were generously provided by Ben Bernanke.

380

Maurice Obstfeld and Alan M. Taylor

as shown in the following regression, in which the dependent variable is the cumulative 1929-35 rate of wholesale price index (WPI) inflation: AWPI = - 0.157 - 0.227 FIXED

(0.026) (0.035)

+ 0.082 CONTROLS,

R2 = 0.66.

(0.034)

Here we see again the familiar deflationary effect of fixed exchange rates, but also a significant countereffect of controls on the price level. Though statistically significant, that effect is small because the “devaluation” implied by controls did not generally lead to significant monetary expansion relative to the world average. Fundamentally, these diverse experiences underscored the unattainable nature of the economic “trinity”: of three desirable policy goals, exchange rate stability, full employment, and free capital mobility, only two of three are mutually compatible. The free currency devaluers discarded exchange stability and gained the freedom to pursue expansionary fiscal and monetary policies. The exchange control countries sought the same freedom by inhibiting capital mobility and, further, manipulated the levers of thoroughgoing exchange control and discriminatory trading in pursuit of domestic goals. Notably, neither group considered a full return to gold standard orthodoxy,requiring the neglect of the full-employment goal and commitment to the other two goals, exchange parity and free exchanges-a testament to the transformation in the political economy of macroeconomic management, the power of new interest groups and enfranchised voters, and the resulting unwillingness of governments to tolerate deflation and labor unrest in a distributional fight under conditions of downward wage inflexibility (Eichengreen 1992a). Much of the motivation for maintaining pegged exchange rates, both in gold bloc and exchange control countries, was the fear of hyperinflation and the attendant social conflict, as witnessed all too recently in Central and Eastern Europe. That fear was present, though not dominant, even in countries that chose open devaluation (see Eichengreen 1992a, 292; Nurkse 1944, 166; Sayers 1976, 412). Ironically, exchange control, itself so inimical to the liberal principles of orthodox finance, nonetheless facilitated the persistence of orthodox monetary policies in those countries least willing, given recent inflationary experience, to sacrifice the nominal anchor of their official gold parity. Even in peripheral Latin America, “memories of wild inflation under inconvertible paper during the late nineteenth century, memories still fresh during 1929-3 1, The coefficient on CONTROLS in the last regression implies that, on average in the sample, imposing exchange controls nearly offset the negative output effect of not devaluing. This result appears at odds with the conclusion in Eichengreen (1992a. 350, table 12.1, col. [4]) that exchange control countries did better than gold bloc countries but much worse than devaluers. However, the exchange control group underlying the last regression is larger than Eichengreen’s, including, in addition to his observations, Argentina, Japan, Romania, Greece, Latvia, Estonia, and Denmark.

381

International Capital Mobility over the Long Run

hampered and slowed down the adoption of more self-assured and expansionist policies” (Diaz Alejandro 1983, 18). The exchange control countries, burdened by foreign debts and precarious reserve levels when the 1931 crisis hit, could maintain their exchange parities in no other way but through controls. Once in place, however, controls were in most cases difficult to contain and were found to have other uses (albeit at foreigners’ expense). The gold bloc countries, in contrast, had the financial resources to cling to gold parities without radical controls but as a result were defenseless against the deflationary forces of the depression. As much as anything, the experience of these countries discredited the last adherents of gold standard orthodoxy and opened the way for the new and interventionist Keynesian approach to international monetary relations that would prevail after the Second World War.

11.3 World War 11, Its Aftermath, and the Resurgence of Capital Mobility Private international capital mobility reached a nadir during and after World War 11, with much of the world left in the grip of bilateral payments arrangements. The postwar international economic order negotiated at Bretton Woods in 1944, and inaugurated with the declaration of currency par values in 1946, mandated convertibility for current account but not necessarily for capital account transactions. Even current account convertibility proved hard to attain, however, in the circumstances following the war. Only at the end of 1958 was external (Le., nonresident) convertibility on current account restored for the main European c ~ r r e n c i e sThe . ~ ~ following decade was characterized by increasing capital mobility, but also by speculative tensions that prompted industrial countries to intensify capital controls in an attempt to shore up the system of fixed exchange rates. These measures proved insufficient, and the modern era of floating dollar exchange rates finally dawned in 1973. Since then, the international flow of capital has expanded dramatically. 11.3.1 Wartime Intensification of Exchange Control The onset of renewed war in 1939 led to an intensification of exchange control. In a memorandum written for British Treasury officials in September 1939, Keynes recalled of the emergency measures taken during World War I, “Complete control was so much against the spirit of the age, that I doubt it 29. A currency is externally convertible if foreigners who hold it (but not necessarily residents of the issuing country) may exchange it freely for other currencies or for domestic goods and assets. The currency is externally convertible for current transactions if foreigners who have acquired it through exports or receipts of asset income can convert it into other currencies or domestic goods. (In contrast, a currency is internally convertible when domestic residents may freely exchange it for other currencies.) For a discussion of various notions of convertibility, see McKinnon (1979,3-7).

382

Maurice Obstfeld and Alan M. Taylor

ever occurred to any of us that it was possible” (Keynes 1978, 10). Countries drew heavily on their interwar experiences with controls to mobilize their foreign exchange resources for all-out conflict. By March 1940, dealings in nearly all the world‘s major currencies, the two important exceptions being the U.S. dollar and the Swiss franc, were subject to some form of exchange control (Mikesell 1954, 15).An additional advantage of restricting capital outflows in wartime was that governments might thereby borrow at artificially low rates of interest. Britain introduced controls in August and September 1939, initially regulating residents’ purchases of foreign currencies but neither blocking sterling balances held by nonenemy aliens nor preventing all sterling transactions between residents and nonresidents. An offshore market in “free” sterling consequently developed. As Keynes forcefully pointed out, nonresidents could buy British exports with free sterling, thus depriving the country of badly needed hard currency-basically, U.S. dollars or Swiss francs (1978, 158-71). This loophole and others were closed early in June 1940 (Mikesell 1954, 16), the same month Keynes took up a formal advisory position at the Treasury.30The interwar sterling bloc, previously a loose association of countries pegging to the pound, narrowed its membership and transformed itself into the Sterling Area, within which similar external exchange controls were enforced but internal currency transactions, including capital movements, were free.31 11.3.2 Capital Mobility in the Bretton Woods System Well before the end of the war, officials in Allied treasuries were turning their minds toward designing a postwar international economic order. In 1941 and 1942, respectively, John Maynard Keynes in Britain and Harry Dexter White in the United States circulated different draft plans for postwar institutions designed to aid in the maintenance of exchange stability, macroeconomic stability, and orderly, generally nondiscriminatory trading relations among nations. White’s plan would, in 1944, become the basis for the Bretton Woods agreement that led to establishment of the International Monetary Fund (IMF), the World Bank, and the General Agreement on Tariffs and Trade (GATT). Both plans are instructive, however, for the light they throw on official and academic attitudes toward the role of capital movements.32In essence, the plans reflected a broad policy consensus, growing out of the experience of the depression, that the global economy would not necessarily be smoothly self30. In August, Keynes was placed on the Exchange Control Conference. 31. The Sterling Area’s holdings of hard currencies were centralized at the Bank of England, which also supplied these resources when needed by area members. Both internally and outside of the Sterling Area, sterling was inconvertible into hard currencies or gold. 32. Various drafts of the Keynes and White plans are reproduced in Horsefield (1969). The French and the Canadians also advanced proposals (the latter was known colloquially as the “offWhite” plan).

383

International Capital Mobility over the Long Run

regulating were wartime controls to be dismantled, so that exchange rates and international capital movements would both have to be closely ~ o n t r o l l e d . ~ ~ Keynes’s plan stepped back from the extreme economic nationalism he had flirted with in his famous 1933 article on “National Self-Sufficiency” (Keynes 1982, 233-46).34But the plan’s basic premise was that heavy government management of macroeconomic policies and exchange rates should be deployed in the defense of internal macroeconomic stability and that such a resolution of the policy trilemma presupposed extensive restrictions over not only capital movements but foreign exchange transactions in general. Keynes’s plan proposed an International Clearing Union (ICU) that would facilitate multilateral trade among members and extend credit (within limits) to cover current account deficits. To these ends, countries with external surpluses would accumulate claims on the ICU, and countries with deficits, liabilities. Such credits of “bancor,” as the new international currency was called, fixed in gold value and in terms of national currencies, would be used by countries to settle international accounts, much as gold had been in an earlier era. The instability associated with fluctuating interwar exchange rates remained a powerful influence over attitudes toward postwar monetary relations. In Keynes’s view, floating rates were to be rejected both for their disruptive effects and as a reversion to discredited laissez-faire economics. Exchange values under the ICU were not to be “unalterably” fixed, however; far from it. Instead, Keynes’s conception, as expressed several years later in defending the proposed IMF in the House of Lords, was that “we are determined that, in future, the external value of sterling shall conform to its internal value as set by our own domestic policies. . . . Instead of maintaining the principle that the internal value of a national currency should conform to a prescribed de jure external value, [the Bretton Woods plan] provides that its external value should be altered if necessary to conform to whatever de fact0 internal value results from domestic policies, which themselves shall be immune from criticism by the Fund” (Keynes 1980b, 16-18). In other words, exchange realignments rather than domestic deflation, as under the gold standard, were the preferred tool for rectifying payments deficits and unemployment in Keynes’s system. Domestic policies would be geared toward high employment, with short-term international imbalances being met by overdrafts on the ICU. Keynes’s view on exchange rate adjustment represented a sea change in the attitudes that had prevailed in the gold standard era. As Haberler puts it 33. There were, of course, numerous dissenters from various aspects of this consensus, e.g., Friedman (1953), who argued for floating exchanges and freedom of short-term capital movements, and Viner (1943b), who espoused fixed rates but believed they might be consistent with a liberal capital transfer regime. Some still argued for the gold standard (New York Times, 30 March 1943). 34. Harrod (1951,525-26) ascribes the shift to Keynes’s perception in 1941 that in a new postwar order, Keynesian economics might be applied on a global scale, rather than the national scale he envisaged in the 1930s.

384

Maurice Obstfeld and Alan M. Taylor

in an insightful discussion of the Great Depression’s causes and legacy: “The sanctity of fixed exchange rates was a casualty of the Great Depression. It is true that there had been many exchange-rate changes in the nineteenth century and earlier. But the devaluation of the leading currencies of the world . . .made the operation ‘salonfuhig,’ that is, fit for gentlemen” (1976, 17). Keynes appreciated clearly that by resolving the policy trilemma in favor of internal employment goals and exchange rate management, he was ruling out open capital markets. Indeed, Keynes’s plan embraced exchange control wholeheartedly and explicitly called for curbs on capital movements, with some provision for international long-term capital movements added in as an afterth~ught.~~ The attitude toward private capital movement was set out explicitly in all drafts of Keynes’s plan, for example, the It is widely held that control of capital movements, both inward and outward, should be a permanent feature of the post-war system-at least so far as we &e concerned. If control is to be effective, it probably involves the machinery of exchange control for all transactions, even though a general ‘open license is given to all remittances in respect of current trade. But such control will be more difficult to work, especially in the absence of postal censorship, by unilateral action than if movements of capital can be controlled at both ends. It would therefore be of great advantage if the United States and all other members of the Currency Union would adopt machinery similar to that which we have now gone a long way towards perfecting in this country; though this cannot be regarded as essential to the proposed Union. (Horsefield 1969, 13) White’s alternative plan placed less emphasis on periodic exchange rate adjustment than did Keynes’s and viewed capital movements in a somewhat more favorable light. Dam (1982, 83) quotes a passage from the April 1942 version of the White plan (Horsefield 1969, 47) to support the assertion that White took a creditor’s view of the postwar order, favoring reduced capital controls in contrast with “Keynes’s enthusiasm for capital controls.” In fact, White was referring to generalized exchange controls on the model of interwar Germany in the quoted passage, not specifically to capital controls, and later in the plan advocated a prohibition of IMF resources for funding “illegitimate” capital flows (Horsefield 1969, 49-50). Such a provision would have been necessary in any event to assuage congressional fears that the United States would end up funding unlimited foreign imbalances. White’s plan also called for international cooperation to limit capital flows inspired by “speculation” or tax evasion: “It would be an important step in the direction of world stability if a 35. In Keynes’s conception, central banks would be monopoly dealers in foreign exchange within each country; they in turn would sell foreign exchange to the ICU for bancor credits or settle directly with foreign central banks. As monopoly dealers, the central banks were ideally placed to scrutinize all foreign exchange transactions and deny foreign exchange for purposes of capital transfer (Keynes 1980a, 216). 36. The draft is dated 11 February 1942.

385

International Capital Mobility over the Long Run

member government could obtain the full cooperation of other member governments in the control of capital flows. . . . The assumption that capital serves a country best by flowing to countries which offer most attractive terms is valid only under circumstances that are not always present. . . . A good case could be made for the thesis that a government should have the power to control the influx and efflux of capital, just as it has the authority to control the inflow and outflow of goods and of gold’ (Horsefield 1969,66-67). In reality, Keynes and White were not far apart on the principle that capital flows might need to be regulated, although Keynes’s methods for accomplishing this task were more realistic and therefore much more dirigiste. This shared tolerant attitude toward capital account prohibitions was fully reflected in the eventual Articles of Agreement establishing the IME A major goal of the IMF system was nondiscriminatory multilateral convertibility on current account (as set out in Article VIII). But Article VI(3) stated that “members may exercise such controls as are necessary to regulate international capital movements.” Article VI( 1) prohibited members from using IMF resources “to meet a large or sustained outflow of capital” and even empowered the Fund to request imposition of capital controls in such cases (Horsefield 1969, 193-94). Keynes’s plan had also included the latter feat~re.~’ The United States’ agreement to such provisions in 1944 and 1945 may seem contrary to its natural interests as the premier creditor and financial power of the postwar period. As noted above, however, Congress was concerned about the extent of America’s financial commitment. The allowed restrictions, anyway, seemed unlikely ever to be needed by the United States, would likely apply to other countries’ outflows rather than inflows, and could only ensure New York‘s position as the worlds leading financial center. Business interests in the United States were in any case more concerned with current account convertibility and expanded export opportunities than with capital flows. Moreover, New Deal Washington viewed the financial world with considerable distrust. This distrust was inherent in the Democratic Party’s Jacksonian tradition. But it was greatly heightened by the perceived role of banks and security markets in bringing on the Great Depression. Disillusion with banks and financial markets prevailed in many countries, in fact, and led to a general reduction during the 1930s of central bank independence in favor of treasury dominance. (See Dam 1982, 53. On the United States, see Calomiris and Wheelock, chap. 1 in this volume.) This view provoked stricter regulations on financial markets in the 1930s. It is also reflected in Treasury Secretary Henry Morgenthau’s pronouncement at Bretton Woods that the new institutions would “drive . . . the usurious money lenders from the temple of international 37. In 1956 the IMFs executive directors interpreted Article VI as allowing countries (subject to some mild restrictions) to impose capital controls “for any reason” and “without approval of the Fund” (Horsefield 1969,246).

386

Maurice Obstfeld and Alan M. Taylor

finance” (Gardner 1980, xix), and in his successor, Fred Vinson’s, unilateral shift of IMF and World Bank headquarters from New York to Washington on the purported grounds that “the institutions would be fatally prejudiced in American opinion if they were placed in New York, since they would then come under the taint of ‘international finance”’ (Keynes 1980a, 211). White himself undoubtedly shared these views, arguing in his plan that capital controls “would constitute another restriction on the property rights of the 5 or 10 percent of persons in foreign countries who have enough wealth or income to keep or invest some of it abroad, but a restriction that presumably would be exercised in the interests of the people-at least so far as the government is competent to judge that interest” (Horsefield 1969, 67).38 11.3.3 Postwar Inconvertibility and the European Payments Union At the height of the world war in 1943, the governments-in-exile of the Belgium-Luxembourg union and the Netherlands entered into a bilateral financial agreement that was the first of about 200 that would be in effect in Europe by 1947 and nearly 400 that would be in effect worldwide shortly thereafter (Yeager 1976,407). Under the agreement, the two countries promised to peg their mutual exchange rate by standing ready to purchase the other’s currency. This type of agreement aimed at conserving reserves of hard currency and gold through mutual credit arrangements but in practice entailed controls over resident transactions so as to prevent the buildup of unbalanced positions in partner currencies. A corollary of hard-currency scarcity was a continuation of currency inconvertibility and of wartime prohibitions on private capital movements, which might quickly strip a government of reserves. Domestic financial controls further limited international intermediation and, along with the economic and political instabilities implied by reconstruction, blocked the channels through which potential capital-receiving countries might have borrowed privately abroad. Private international capital movements had essentially dried up. Currency inconvertibility seriously compromised even the gains from current international trade. If country A had a trade surplus with country B, it could not use its surplus accumulation of B’s currency to finance a deficit with country C , as would have been possible under general external currency convertibility. Somehow, country A’s payments would have to be balanced vis-avis both B and C individually, not simply vis-a-vis the rest of the world taken in totality. Bilateral trading agreements may have been superior to blanket, indiscriminate limitations on foreign transactions in allowing for mutual credits (the trade creation aspect), but they had the drawback of shunting demand from the cheapest source of supply worldwide toward countries with extensive demands for domestic products (trade diversion). A system of multilateral clearing held out the potential of easing such constraints and promoting more 38. The evolution of White’s political views is discussed by Rees (1973).

387

International Capital Mobility over the Long Run

efficient global resource allocation, if only a true multilateral payments system could be attained. Unfortunately, universal convertibility, even limited external convertibility, was difficult for individual countries to attain in the immediate postwar circumstances. Restoring convertibility required countries to solve a serious coordination problem. In a setting of general inconvertibility, a single country allowing foreigners to freely convert its currency would face an uncomfortable net drain of foreign exchange reserves: foreign exporters would convert the bulk of their domestic currency earnings into central bank foreign reserves, whereas most of the foreign currency earned by domestic exporters would be unusable. The latter could insist on being paid in their own currency, but this would seriously injure sales, as the home country would likely have its best potential export markets in countries from which it did not itself import much. Furthermore, foreign countries might deliberately restrict imports from the convertible currency country so as to maximize their hard-currency inflow at its expense.39Of course, one solution to the problem would have been for monetary authorities simply to refrain from trading domestic currency for foreign currencies, that is, to let the exchange rate of domestic currency float, as several countries did after World War I. Indeed, Friedman’s celebrated polemic “The Case for Flexible Exchange Rates,” drafted in 1950, explicitly promoted floating rates as a strategy for moving immediately to general currency convertibility (Friedman 1953, 158). This step governments were reluctant to take, out of fear of the currency instability and hyperinflation associated with interwar floating exchange rates. Article VIII of the IMF Articles of Agreement, as noted earlier, called for convertibility on current account and unrestricted freedom of current international payments. Article XIV, however, allowed countries to maintain restrictions contravening Article VIII during a transitional period, and even to introduce new restrictions. Only five years after the start of IMF operations was any member not yet in compliance with Article VIII required to begin annual consultations on its progress with the Fund. At the time the articles were drafted, a five-year breathing space was regarded as allowing a reasonable period for the general return to (current account) convertibility. Nothing of the sort happened. Instead, controls generally proliferated. By 1953, more countries were engaging in multiple currency practices than in 1946, leading Mikesell to the exasperated remark that “the system of fixed exchange parities combined with a complex of neo-Schachtian devices has provided far less exchange stability in the postwar period than did the fluctuating free exchange rates of the 1930s” (1954,25-27). By 1957, only eight countries apart from the United States and Canada-Mexico, Cuba, the Dominican Republic, Guatemala, El Salvador, Honduras, Haiti, and Panama39. See Yeager (1976, 409-10) for further discussion of this “contagion of bilateralism.” See also Triffin (1957, 88-93). The basic mechanisms at work were emphasized by Viner (1943a).

388

Maurice Obstfeld and Alan M. Taylor

had formally accepted the Article VIII obligations (Triffin 1957, 115). The proliferation of controls reflected the same forces preventing unilateral movements toward convertibility by dollar-hungry countries. A classic example is Switzerland,which, to protect foreign exchange reserves, made its franc inconvertible for Europeans while leaving it convertible for residents of the dollar area (see Kaplan and Schleiminger 1989,57; Yeager 1976,409). Some countries responded to the situation by adopting floating exchange rates, IMF norms notwithstanding. Canada dismantled its exchange controls under cover of a floating rate; Mexico, Peru, and Chile likewise floated their currencies; and Churchill’s government in Britain seriously debated a scheme for freeing the pound in 1952 (Cairncross 1985, chap. 9; Kaplan and Schleiminger 1989, chap. 10). Triffin argued that the IMF might have been able to move the world more quickly to convertibility if its structure had facilitated multilateral clearing, for example, through Keynes’s conception of a synthetic international currency. Instead, the Ih4Fblueprint “dealt with the setting up and revision of par values, the elimination of exchange controls, and the Fund’s lending operations as if these problems could be handled with each country individually against a background of general convertibility and stability in world trade and currency arrangements” (Triffin 1957, 137). The coordination problem involved in moving to the latter type of equilibrium from the one left by the war was not addressed. The hazards of a unilateral return to convertibility by war-tom countries are well illustrated by Britain’s abortive attempt to restore multilateral current account convertibility for sterling in July 1947-an experiment that had to be abandoned after only five weeks. In September 1945 a British delegation led by Keynes arrived in Washington to negotiate a loan of dollar reserves. The United States insisted (among other conditions) that sterling’s convertibility on current account be restored no later than one year after the funds (totaling $3.75 billion) became available. Congress and American business interests strongly supported the convertibility provision (as well as an associated trade nondiscrimination clause; see Gardner 1980, 197-98). In particular, these groups felt that the LMF articles’ timetable for restoring convertibility was lax. Immediate convertibility of so widely held a currency as sterling, it was believed, would hasten worldwide freedom of current payments, at the same time easing discriminatory trade practices intended to maximize bilateral trade surpluses with the United States. Britain put aside its misgivings and agreed to these terms: 15 July 1947 emerged as sterling’s convertibility date after congressional approval of the loan midway through 1946.4OBritain’s current account deficit increased sharply after the harsh winter of 1946-47. By the end of June more than half the U.S. loan had been used up (Cairncross 1985, 132). Despite continuing gold and 40. Canada added $1.25 billion to the loan.

389

International Capital Mobility over the Long Run

dollar outflows, however, Britain fulfilled its commitment and declared convertibility on 15 July, hoping desperately that convertibility would raise global confidence in sterling. Instead, reserve outflows accelerated. With only $400 million of the American loan remaining, Britain suspended convertibility on 20 August. The sharp acceleration in dollar losses in July seems largely to have been the result of capital outflows. It was feared that convertibility would be fleeting and that sterling might be devalued once the experiment failed. Now was the time to get dollars, and at a relatively cheap sterling price. (Convertibility did turn out to be very temporary, but the feared devaluation did not come until 1949.) The result was a classic speculative attack. How was an attack on sterling carried out in a world of seemingly pervasive capital controls? Some countries converted preexisting sterling balances into dollars, representing them as current earnings. On the basis of revised balance-of-payments data, Cairncross (1985, 157) argues that this channel was not very important, notwithstanding a long tradition placing much of the blame for the debacle on such conversions (e.g., Gardner 1980, 317-18). More significant were “leads and lags” in trade credits-the practice of accelerating sterling payments and delaying foreign currency receipts in the expectation of a sterling depreciation (Einzig 1968). Finally, some reserves leaked out through capital transfers to other Sterling Area countries. Sterling Area members such as South Africa and Australia borrowed large sums of sterling in London and rapidly used the proceeds for imports from the dollar area.41 The crisis carried two distinct lessons. First, in the circumstances of the immediate postwar years, a single country like Britain with a structural current account deficit due to wartime changes could not unilaterally return to convertibility. Any such return would need to be coordinated among many nations. As Gardner puts it: “The fact is that the negotiators [of the Anglo-American loan agreement] did not fully understand the economics of convertibility. They did not appreciate the difficulty in which Britain might find itself in the event that it went on accumulating inconvertible currencies while other countries, deliberately restricting imports from the United Kingdom, presented large sterling surpluses for conversion. Given this hazard of making one currency convertible in a generally inconvertible world, the use of a rigid time-table was certainly injudicious” (1980, 218). A second lesson of the crisis, one less appreciated at the time, was that damaging speculative crises could occur even under exchange control. Capital controls were porous, certainly porous enough to devastate the slim liquidity bases upon which most countries were operating in the late 1940s. The channels of capital flight revealed in the U.K. convertibility crisis, especially leads and lags, would remain widely operative through the end of the Bretton Woods system, coming strongly into play whenever the prospect of devaluation of41. See Wyplosz (1986) for a formal analysis of speculative attacks under capital controls.

390

Maurice Obstfeld and Alan M. Taylor

fered a large speculative gain over a short period. Indeed, the scope for such phenomena only increased as trade expanded in the 1960s. Even before Britain suspended sterling’s short-lived convertibility, the United States proposed the Marshall Plan with its accompanying call for economic cooperation within Europe. The plan reflected a change in the U.S. “universalist” approach to postwar economic problems, motivated by the perception of a dire political threat to a uniquely important region. American policymakers had hoped that Marshall aid would promote intra-European trade, strengthening Europe’s economies and creating a shared interest in political stability. However, the absence of any multilateral clearing system for intra-European payments frustrated this hope. Under prodding from the U.S. European Cooperation Administration, which had been set up to administer the Marshall funds, the nations of Western Europe in September 1950 created the European Payments Union (EPU). The EPU was a major success in facilitating trade liberalization within Europe (and extending it to territories on other continents that belonged to some European country’s currency area). The EPU worked by every month consolidating each member’s bilateral payments deficits into a net debt to the union, extending some credit but eventually requiring settlement in dollars and gold. This arrangement allowed European country A to use its surplus with European country B to finance its deficit with European country C, despite the inconvertibility of B’s currency. (The IMF, in contrast, could perform no comparable clearinghouse function.)42 The initial success of the EPU allowed some privatization of foreign exchange transactions, which had been concentrated in the hands of central banks. This liberalization allowed private banks to take over some of the EPU’s clearing functions. Over the course of the 1950s several EPU members, notably the United Kingdom and Germany, liberalized foreign exchange transactions further, Germany going much the furthest in allowing residents to retain foreign exchange earnings and to hold foreign assets. (In the United Kingdom, residents could deal among themselves in a managed pool of foreign “investment currency” but otherwise were barred from acquiring foreign assets, while nonresidents until 1967 had to trade sterling securities in a separate market for “security sterling.”) During 1957-58 Europe’s hard-currency reserves rose sharply, the counterpart of a huge U.S. payments deficit. On 27 December 1958, the EPU was terminated by mutual consent, with most members, including France, Germany, Italy, and the United Kingdom, declaring their currencies externally convertible on current account. (The former EPU countries formally accepted their Article VIII convertibility obligations in February 1961. Japan followed in April 1964.) Germany also moved to full convertibility on capital account, so that, as of January 1959 the Bundesbank could declare that “only the payment of interest on foreigners’ balances, the sale of domestic money42. For discussions of the EPU, see Triffin (1957), Kaplan and Schleiminger (19891, and Eichengreen (1993).

391

International Capital Mobility over the Long Run

market paper to foreigners and the taking of foreign loans running less than five years remain forbidden, the object being to check the inflow of ‘hot money’ into the Federal Republic” (Deutsche Bundesbank 1959, 52). Germany’s motives for such wide-ranging liberalization were two. One was the free market ideology characteristic of Economics Minister Ludwig Erhard’s policies. Equally important, however, was the pressure on Germany’s internal liquidity and prices due to the chronic balance-of-payments surplus that had developed after the early 1950s. By liberalizing capital outflows, the authorities hoped to reduce the payments surplus, whereas the remaining capital controls listed in the preceding quotation were intended to discourage capital inflows and provide scope for sterilization operation^.^^ In 1957 Belgium, France, Germany, Italy, Luxembourg, and the Netherlands signed the Treaty of Rome creating the European Economic Community (EEC). The treaty called on signatories to undertake the progressive abolition “between themselves of all restrictions on the movement of capital belonging to persons resident in Member States” (Article 67( l), quoted in Bakker 1996, 279). This provision was viewed as fundamental to the long-term goal of a single European market. In its first directive of May 1960, the EEC Council required member countries to free short to medium-term trade credits, direct investments, and cross-border trades of listed shares. Germany had pushed for full liberalization of capital movements in the negotiations leading to the directive, including movements between EEC members and nonmember states (Bakker 1996,81). In May 1959, seeing a welcome fall in its official reserves and assuming that the policy of encouraging capital exports was working, Germany unilaterally abolished its remaining restrictions on capital import (Yeager 1976, 496). Despite policymikers’ optimism, however, Germany was very shortly to experience the type of policy conflict that ultimately brought the Bretton Woods system down amid escalating capital controls. 11.3.4 The Collapse of Bretton Woods

Only the month after the EEC Council’s directive on liberalization of capital movements, Germany reimposed some of the controls it had abolished a year earlier, hoping to discourage renewed capital inflows. Attempting to restrain a domestic boom through higher interest rates, the Bundesbank found itself frustrated by the large volume of reserve purchases it was obliged to carry out 43. On Germany’s attempts to counteract the inflationary potential of its balance-of-payments surpluses, see Boarman (1964) and Emminger (1977). Germany’s relatively fast productivity growth in the 1950s and 1960s mandated a secular real appreciation of the deutsche mark against the dollar, that is, a rise in Germany’s price level measured in dollars against that of the United States. Given a fixed nominal exchange rate, however, this equilibrating real currency appreciation could occur only through higher inflation in Germany than in America-something German policymakers were largely unwilling to accept. This tension made revaluation inevitable once German capital markets were fully open. In contrast, Japan did accept a higher inflation rate than that of the United States (Obstfeld 1993).

392

Maurice Obstfeld and Alan M. Taylor

to maintain the deutsche mark’s exchange parity. In March 1961, Germany, soon followed by the Netherlands, reluctantly revalued its currency by 5 percent against the dollar. These events heralded a new era in which speculative capital flows continually bedeviled policymakers in Europe and elsewhere. Concerted EEC progress on further capital account liberalization bogged down, and fear of speculation made any open discussion of exchange parity changes impossible. Italy suffered a balance-of-payments crisis in 1964, which it beat back with the help of loans from the United States and the IMF. Britain entered a prolonged period of crisis in the same year, giving in finally and devaluing sterling in November 1967. Nonetheless, individual European countries did take some liberalizing steps in the early and mid-1960s. Italy allowed its residents more freedom to invest abroad. France, enjoying a strong balance of payments during the mid-l960s, unilaterally eased its controls in 1967, motivated in part by a desire to promote the role of Paris as a global financial center (Bakker 1996, 101). However, the May 1968 disturbances sparked capital flight and a reimposition of French controls; at the same time Germany, the recipient of much of the flight capital, tightened its own barriers to capital inflows. Speculation continued into 1969:France resisted until the speculationtemporarily subsided, then surprised markets by devaluing in August. Speculation on a German revaluation reemerged in the same year in advance of parliamentary elections. Just prior to the election, the government closed the official foreign exchange market, then allowed the deutschemark to float. The new government of Chancellor Willy Brandt revalued the currency by just over 9 percent at the end of October. How could capital flows continue to undermine authorities’ efforts to defend exchange parities even in the face of tightened capital controls? Leads and lags in trade credits again provided an important conduit for speculative capital flows; indeed, Einzig (1968) characterized leads and lags as “the main cause of devaluation” in his book on the subject, although his broad definition of the phenomenon included changes in the timing of goods orders (not just payments) as well as forward currency trades. The growth of international trade after the early 1950s-in itself a prime desideratum of the Bretton Woods architects-ironically expanded the opportunities for disguised capital flows. The reopening of private foreign exchange markets and the emergence of the Eurocurrency markets in London in the 1960s further widened the scope for leakages from protected domestic financial systems. The growing tendency to delay realignments until the market forced the authorities’ hands, itself a result of increasing possibilities for speculation, ensured that a speculative attack might produce very large profits over a very brief period. Thus, even modest elasticities of trade credits, say, with respect to normal interest differentials, could translate into large flows of reserves in crisis episodes. The United States meanwhile had been facing its own problems since the end of the period of “dollar shortage” in the late 1950s. Growing U.S. balanceof-payments deficits were causing alarm. The counterpart of these deficits was

393

International Capital Mobility over the Long Run

a growing stock of short-term official dollar claims on the United States. Some of these claims were converted into gold, putting pressure on the American gold stock, but the bulk were held despite mounting anxiety that the dollar’s gold content might suddenly be reduced. In response, the United States took a number of measures to counter private capital outflows. Starting in 1961, an escalating sequence of dividend and interest taxes, voluntary guidelines, and mandatory limits were imposed on American capital outflows abroad (see Bordo 1993,58; Solomon 1982, chaps. 3 and 6). The ultimate effectiveness of these measures remains debatable even today. For example, New York banks, restricted from lending directly to foreigners, could legally set up London subsidiaries capable of taking dollar deposits and making the forbidden loans. Non-U.S. banks also competed for this business. Regulations meant to retain dollar inflows within U.S. borders therefore shunted these dollars into the London Eurodollar market, promoting that market’s spectacular growth at the expense of onshore U.S. banks.44 The dollar itself came under concerted attack in the early 197Os, a development due in part to President Lyndon Johnson’s escalation of military and domestic spending, in part to divergent productivity trends. Increasingly volatile capital flows set the stage for the ultimate collapse of fixed exchange rates in early 1973 (see Solomon 1982, chaps. 11-13; Yeager 1976, chap. 28). Several industrialized countries temporarily floated their currencies prior to the Smithsonian dollar devaluation of December 197 1, and several, including Germany, imposed restrictions on capital inflows (Bakker 1996, 122). When the new Smithsonian parities were attacked over 1972 and 1973, Japan, Switzerland, Germany, France, and the Netherlands all raised their barriers to capital inflows, including quantitative borrowing restrictions, interest taxes, and supplemental reserve requirements. Concerned by the disruptive effect of floating intra-European exchange rates on its common agricultural policy and on its ongoing drive for further economic integration, the European Community issued a general derogation from its May 1960 first directive on capital account liberalization and went further in directing members to develop or reinstate effective mechanisms for controlling capital flows and their effects on domestic money supplies (Bakker 1996, 116-18). The lira and sterling, like the dollar, were under selling pressure; Italy and Britain raised barriers to stem outflows as a result. In June 1972 the United Kingdom extended its exchange control system to apply to transactions within the Sterling Area and let the pound float downward. The pressure of speculation remained unbearable however. By March 1973, industrialized country currencies were floating against the U.S. dollar, with six EC currencies floating jointly within a “snake” while Italy and the Anglo-Irish currency union floated independently. 44. Concern about the U S . balance-of-payments deficit was not universal. For a cogent contrary position, see Kindleberger (1965). For some skeptical remarks on the importance of the “confidence problem’’ posed by an increasing ratio of official dollar liabilities to US.gold, see Obstfeld (1993, 211).

394

Maurice Obstfeld and Alan M. Taylor

11.3.5 The Process of International Financial Integration after 1973 Bretton Woods proved untenable in the end because its rules could not reconcile independent national policy goals, pegged exchange rates, and even the limited degree of capital mobility implied by an open world trading system. Once industrial countries had been forced to accept floating dollar exchange rates as an open-ended interim regime, however, at least some governments felt free to liberalize capital movements without sacrificing either their domestic policy priorities or an external currency commitment. Over the years 197475, the United States dropped its restrictions on capital outflows while Germany liberalized inflows. Germany would again deploy controls over inflows in the late 1970s, when dollar weakness threatened to enhance the reserve currency status of the deutsche mark, a development the Bundesbank has resisted. France and Italy retained and even tightened some controls in order to loosen the link between monetary and exchange rate policy. A strong motivation for these actions was the desire to limit intra-EC exchange rate fluctuations, first within the informal EC currency snake and later within the European Monetary System (EMS; Giavazzi and Giovannini 1989; Bakker 1996). The United Kingdom also tightened and retained controls until 1979. In that year, Thatcherite free market ideology, allied with a fear that North Sea oil would bring the “Dutch disease” of sterling appreciation, together led to suspension of the 1947 Exchange Control Act and full capital account liberalization. Japan largely opened its capital account in December 1980, the culmination of a series of steps beginning in 1974. Liberalization was typically undertaken to promote inflows or outflows that would counter yen depreciation or appreciation; only rarely were controls tightened. The liberalizing trend seems to have reflected pressures from the domestic business and financial communities (It0 1992, 316-21). Further measures to ease foreign asset exchanges were taken in 1984, partly as a result of U.S. pressure (Frankel 1984; Ito 1992, 329). Developing countries almost universally retained tight capital account controls throughout the entire Bretton Woods period, the most important sources of capital inflow being official loans and foreign direct investment (Cardoso and Dornbusch 1989). The two oil price shocks of the 1970s produced large and persistent surpluses for oil producers but only transitory deficits for the industrialized world. The oil surpluses were “recycled” to developing countries through industrialized country banks, so that by the early 1980s developing market borrowers owed a substantial debt to the banks, most of it government or government guaranteed. Most developing countries retained strict control over private exchange transactions. As of 30 April 1980, only 50 of 140 IMF members had formally ceased operating under the “temporary” Article XIV derogation from Article VIII (Dam 1982, lOl), although these countries accounted for most of world trade.

395

International Capital Mobility over the Long Run

Table 11.5

Net Capital Flows to Developing Countries, 1977-94 (bfflion U.S. dollars)

Category of Capital Inflow Foreign direct investment Portfolio investment Other (including bank lending) Total

1977-82

1983-89

1990-94

11.2

29.8

13.3 6.5 -11.0

39.1 43.6 22.2

30.5

8.8

104.9

- 10.5

Source: Folkerts-Landau and Ito (1995, 33).

The developing country debt buildup turned into a crisis in August 1982 under the pressure of a global economic downturn and sharply higher world interest rates. Bank lenders became unwilling to extend new loans or even roll over maturing debts, and generalized default loomed, as in 1931. Open default was avoided through concerted rescheduling orchestrated largely by the United States and the IMF. Only toward the end of the 1980s did U.S.-brokered debt workouts under the Brady plan begin to pave the way for renewed private lending to the developing world, which boomed in the early 1990s. Direct investment has grown significantly, but more strikingly, bank lending to governments has given way to portfolio investment in bond and equity markets (see table 11S ) .The shift in the composition of developing country liabilities is in part a reflection of wide-ranging financial sector restructuring in these countries. In the mid- 1970s several Latin American countries, notably Argentina, Chile, and Uruguay, opened their capital accounts as part of exchangerate-based stabilization programs. These programs, flawed by insufficient fiscal stringency, inadequate domestic financial supervision, and inconsistent wage indexation structures, all proved to be unsustainable and were followed by renewed capital account restrictions. More recently, numerous developing countries in East Asia have instituted domestic financial deregulation and at least partial capital account opening, often in the face of large external surpluses. Similar developments followed in Latin America against a background of serious fiscal reform, aggressive privatization, and successful inflation stabil i ~ a t i o nThese . ~ ~ reform efforts encouraged renewed capital inflows, although the decline in U.S. interest rates in the early 1990s is clearly an important additional causal factor. ' h o closely interrelated issues of debate, as yet unresolved, are the appropriate degree and form of exchange rate flexibility and capital account control for these countries. The durability of the developing world's return to the global capital market remains to be seen. Investor interest weakened when U.S. interest rates rose in 1994, and several industrializing countries faced pressure in the aftermath of 45. Edwards (1995, chap. 3) analyzes the forces behind the recent trend of economic liberalization in Latin America.

396

Maurice Obstfeld and Alan M. Taylor

the 1994-95 Mexican crisis. Markets have so far displayed greater resilience than in 1982; even Mexico has been able to borrow again, albeit with the aid of extraordinary financial backing from the U.S. Treasury and the IME Perhaps the most dramatic move toward full capital account liberalization occurred among the continental members of the European Union. Starting in the 1980s, these countries began moving toward the goal of free intraEuropean capital mobility foreshadowed in the Treaty of Rome-which in practice implied unrestricted mobility vis-a-vis the outside world as well, given Germany’s commitment to openness and the difficulty in any case of enforcing partial restrictions. France joined the trend after 1983, when President Frangois Mitterrand abandoned his socialist growth agenda in favor of the franc’s continued participation in the EMS exchange rate mechanism. Germany has consistently pushed its European partners toward capital account freedom, except while Bretton Woods was unraveling in the early 1970s. An important motive for this advocacy has been the belief that an open capital account would impose discipline over monetary and fiscal policies. Germany’s capital account was completely open by 1981; that of the Netherlands by 1986; Denmark’s by 1989; Belgium-Luxembourg’s and Italy’s by 1990; Spain’s, Portugal’s, and Ireland‘s by 1992; and Greece’s by 1994 (Bakker 1996,220). Austria, Sweden, and Finland, which joined the European Union in 1995, also had open capital accounts of fairly long standing by that time. Driving this broad liberalization was an acceleration in both commodity market integration within Europe and in plans for monetary union. The EMS currency crisis of the years 1992-93 illustrated once again the untenability of fixed exchange parities when capital is mobile and domestic economic conditions assume primacy over the exchange rate. However, calls to reinstate capital controls after the crises of the early 1990s have so far been rejected, and the European Union remains committed by treaty to full monetary and financial integration for a subset of members by 1 January 1999. If that momentous event comes to pass, even on the limited basis of a small core of countries including France and Germany, the utopian goal of European economic union first espoused by the United States in the late 1940s will be substantially achieved. Inevitably, that goal appears to coexist uneasily with internal political realities. Indeed, the late-1990s drive by EU governments to meet the Maastricht macroeconomic convergence criteria has allowed a considerable degree of exchange stability, but at the cost of rising unemployment. These recent developments are uncomfortably reminiscent of the coordinated global macropolicy contraction that brought on the Great Depression. Happily, most of the world is not participating this time.

11.4 Summary This paper has chronicled both the decline of the international capital market during the Great Depression and its gradual regeneration over the period

397

International Capital Mobility over the Long Run

since the Second World War. A major unifying theme in that story has been the basic incompatibility of open capital markets with a regime that aims to attain both exchange rate stability and domestic employment or growth objectives. Under the gold standard before 1914, exchange stability was the overriding goal of monetary policy and domestic objectives took a back seat. Thus, the monetary regime was consistent with considerable international capital mobility. The Great Depression discredited gold standard orthodoxy, propelled Keynesian economics to intellectual ascendancy,and, worldwide, solidified the already vocal political constituencies favoring high employment and govemment intervention over laissez-faire. The result was a postwar international monetary system based on capital account restrictions and pegged but adjustable exchange rates, one whose very success ultimately led to increasingly unmanageable speculative flows and floating dollar exchange rates among the industrialized economies. Floating rates, in turn, have allowed the industrialized countries to deregulate capital flows extensively while pursuing domestic macrogoals. In the European Union, where capital mobility is free, the tension between domestic political imperatives and the supranational goal of monetary union remains apparent only a year before the 1 January 1999 deadline for a single EU currency laid down by the Maastricht Treaty. In the developing world, there is continuing experimentation and debate over the optimal point on the trade-off among currency stability, freedom of capital movements, and other economic goals. Domestic financial deregulation,like capital account decontrol, also accelerated in the 1970s. In part, that development flows from the trend toward freer international financial trade. After the 1950s, countries increasingly allowed homegrown financial institutions to compete for international business within enclaves separated from domestic markets by a strict cordon sanitaire.As resident capital controls were lifted, however, domestic deregulation became a competitive necessity. Domestic deregulation and the consequent growth of the financial sector, in turn, have made it much harder to reimpose capital account restrictions effectively today. The potential threat to systemic stability that competitive financial deregulation poses has motivated the Basle Committee’s work since 1974 on supervisory collaboration among nations. At the start of the Great Depression the United States was the dominant economic power in world markets but had no appetite for a commensurate political leadership role. Even before World War I1 had ended, the U.S. government recognized America’s natural position of leadership and implemented several strategies to further a vision of a cooperative postwar international economic order. In the 1990s major components of that vision have largely been realized, though not within a policy regime the American postwar planners would recognize or, most probably, endorse. While the current system is not free of shortcomings, it has allowed sovereign nations to coexist and, where they have liberalized and opened their economies, to begin converging. Moreover, a replay of the Great Depression has been avoided so far. It is ironic that

398

Maurice Obstfeld and Alan M. Taylor

the current international system sprung directly from one based on disillusion with the interwar performance of free capital mobility and floating exchange rates alike. Thus Keynes’s remark in closing the Bretton Woods conference may still be appropriate: “How much better that our projects should begin in disillusion than that they should end in it!” (1980b, 103).

References Alesina, Alberto, Vittorio Grilli, and Gian Maria Milesi-Ferretti. 1994. The political economy of capital controls. In Capital mobility: The impact on consumption, investment and growth, ed. Leonard0 Leiderman and Assaf Razin. Cambridge: Cambridge University Press. Bakker, Age E P. 1996. The liberalization of capital movements in Europe: The monetary committee and$nancial integration, 1958-1994. Dordrecht: Kluwer. Bernanke, Ben S. 1995. The macroeconomics of the Great Depression: A comparative approach. Journal of Money, Credit, and Banking 27 (February): 1-29. Bernanke, Ben S., and Kevin Carey. 1996. Nominal wage stickiness and aggregate supply in the Great Depression. Quarterly Journal of Economics 11 1 (August): 853-83. Black, Stanley W. 1991. A Levite among the priests: Edward M. Bemstein and the origins of the Bretton Woodssystem. Boulder, Colo.: Westview. Boarman, Patrick M. 1964. Germany’s economic dilemma: Inflation and the balance of payments. New Haven, Conn.: Yale University Press. Bordo, Michael D. 1993. The Bretton Woods international monetary system: A historical overview. In A retrospective on the Bretton Woods system: Lessons for international monetary reform, ed. Michael D. Bordo and Barry J. Eichengreen. Chicago: University of Chicago Press. Bordo, Michael D., and Finn E. Kydland. 1995. The gold standard as a rule: An essay in exploration. Explorations in Economic History 32:423-64. Bordo, Michael D., and Hugh Rockoff. 1996. The gold standard as a “Good Housekeeping Seal of Approval.” Journal of Economic History 56 (June): 389-428. Bratter, H. M. 1939. Foreign exchange control in Latin America. Foreign Policy Reports 14:274-88. Brown, William Adams, Jr. 1940. The international gold standard reinterpreted, 19141934, 2 vols. New York: National Bureau of Economic Research. Cairncross, Alec. 1985. Years of recovery: British economic policy 1945-51. London: Methuen. Campa, JosC M. 1990. Exchange rates and economic recovery in the 1930s: An extension to Latin America. Journal of Economic History 50 (September): 677-82. Capie, Forrest, and Alan Webber. 1985. A monetary history of the United Kingdom, 1870-1982. London: Allen and Unwin. Cardoso, Eliana A,, and Rudiger Dornbusch. 1989. Foreign private capital inflows. In Handbook of development economics, vol. 2, ed. Hollis Chenery and T, N. Srinivasan. Amsterdam: Elsevier. Choudhri, Ehsan U., and Levis A. Kochin. 1980. The exchange rate and the intemational transmission of business cycle disturbances: Some evidence from the Great Depression. Journal of Money, Credit and Banking 12 (November): 565-74. Dam, Kenneth W. 1982. The rules of the game: Reform and evolution in the international monetary system. Chicago: University of Chicago Press.

399

International Capital Mobility over the Long Run

Davis, Lance E., and Robert J. Cull. 1994. International capital markets andAmerican economic growth, 1820-1914. Cambridge: Cambridge University Press. Deutsche Bundesbank. 1959. Foreign trade and the balance of payments. Monthly Report of the Deutsche Bundesbank 11 (January): 43-52. Diaz Alejandro, Carlos F. 1983. Stories of the 1930s for the 1980s. In Financialpolicies and the world capital market: The problem of Latin American countries, ed. Pedro Aspe, Rudiger Dornbusch, and Maurice Obstfeld. Chicago: University of Chicago Press. . 1984. Latin America in the 1930s. In Latin America in the 1930s: The role of the periphery in world crisis, ed. Rosemary Thorp. New York: St. Martin’s. Edelstein, Michael. 1982. Overseas investment in the age of high imperialism. New York: Columbia University Press. Edwards, Sebastian. 1995. Crisis and reform in Latin America: From despair to hope. Oxford: Oxford University Press. Eichengreen, Barry J. 1991. The comparative performance of fixed and flexible exchange rate regimes: Interwar evidence. In Business cycles: Theory, evidence and analysis, ed. Niels Thygesen, Kumaraswamy Velupillai, and Stefan0 Zambelli. New York: New York University Press. . 1992a. Golden fetters: The gold standard and the Great Depression, 19191939. Oxford Oxford University Press. . 1992b. Trends and cycles in foreign lending. In Capital flows in the world economy, ed. Horst Siebert. Tubingen: Mohr. . 1993. Reconstructing Europe’s trade and payments: The European Payments Union. Ann Arbor: University of Michigan Press. Eichengreen, Barry J., and Jeffrey D. Sachs. 1985. Exchange rates and economic recovery in the 1930s. Journal of Economic History 45 (December): 925-46. Einzig, Paul. 1934. Exchange control. London: Macmillan. . 1937. The theory of forward exchange. London: Macmillan. . 1968. Leads and lags: The main cause of devaluation. London: Macmillan. Ellis, Howard S. 1941. Exchange control in Central Europe. Cambridge, Mass.: Harvard University Press. Emminger, Otmar. 1977.The D-mark in the conflict between internal and external equilibrium, 1948-75. Princeton Essays in International Finance, no. 122. Princeton, N.J.: Princeton University, Department of Economics, International Finance Section. Epstein, Gerald A., and Juliet B. Schor. 1992. Structural determinants and economic effects of capital controls in OECD countries. In Financial openness and national autonomy, ed. Tariq Banuri and Juliet B. Schor. Oxford: Clarendon. Feldstein, Martin, and Charles Horioka. 1980. Domestic savings and international capital flows. Economic Journal 90 (June): 314-29. Fishlow, Albert. 1971. Origins and consequences of import substitution in Brazil. In International economics and development: Essays in honor of R a d Prebisch, ed. Luis Eugenio Di Marco. New York: Academic Press. Folkerts-Landau, David, and Takatoshi Ito, with others. 1995. International capital markets: Developments, prospects, and policy issues. Washington, D.C.: International Monetary Fund. Frankel, Jeffrey A. 1984. The yeddollar agreement: Liberalizing Japanese capital markets. Washington, D.C.: Institute for International Economics. Friedman, Milton. 1953. The case for flexible exchange rates. In Essays in positive economics. Chicago: University of Chicago Press. Gardner, Richard N. 1980. Sterling-dollar diplomacy in current perspective. New York: Columbia University Press. Giavazzi, Francesco, and Albert0 Giovannini. 1989. Limiting exchange rate flexibility: The European Monetary System. Cambridge, Mass.: MIT Press.

400

Maurice Obstfeld and Alan M. Taylor

Giovannini, Alberto. 1993. Bretton Woods and its precursors: Rules versus discretion in the history of international monetary regimes. In A retrospective on the Bretton Woodssystem: Lessonsfor international monetary reform, ed. Michael D. Bordo and Barry J. Eichengreen. Chicago: University of Chicago Press. Gordon, Margaret S. 1941. Barriers to world trade. New York: Macmillan. Goschen, G. J. 1861. The theory of theforeign exchanges. London: Effingham Wilson. Grilli, Vittorio, and Gian Maria h4ilesi-Ferretti. 1995. Economic effects and structural determinants of capital controls. IMF Staff Papers 42 (September): 5 17-5 1. Haberler, Gottfried. 1976. The world economy, money, and the Great Depression 19191939. Foreign Affairs Study no. 30. Washington, D.C.: American Enterprise Institute. Hallwood, C. Paul, Ronald MacDonald, and Ian W. Marsh. 1996. Credibility and fundamentals: Were the classical and interwar gold standards well-behaved target zones? In Modem perspectives on the gold standard, ed. Tamim Bayoumi, Bany J. Eichengreen, and Mark P. Taylor. Cambridge: Cambridge University Press. Hamilton, James D. 1988. The role of the international gold standard in propagating the Great Depression. Contemporary Policy Issues 6:67-89. Harrod, Roy F. 1951. The life of John Maynard Keynes. London: Macmillan. Helleiner, Eric. 1994. States and the reemergence of global finance: From Bretton Woodsto the 1990s. Ithaca, N.Y.: Cornell University Press. Horsefield, J. Keith, ed. 1969. The International Monetary Fund 1945-1965, vol. 3. Washington, D.C.: International Monetary Fund. Ito, Takatoshi. 1992. The Japanese economy. Cambridge, Mass.: MIT Press. Kaplan, Jacob J., and Giinther Schleiminger. 1989. The European Payments Union: Financial diplomacy in the 1950s. Oxford: Clarendon. Kapstein, Ethan B. 1994. Governing the global economy: Internationalfinance and the state. Cambridge, Mass.: Harvard University Press. Keynes, John Maynard. 1978. The collected writings of John Maynard Keynes. Vol. 22, Activities 1939-1945: Internal war jinance, ed. Donald Moggridge. London: Macmillan. . 1980a. The collected writings of John Maynard Keynes. Vol. 25, Activities 1940-1944: Shaping the post-war world: The Clearing Union, ed. Donald Moggridge. London: Macmillan. . 1980b. The collected writings of John Maynard Keynes. Vol. 26, Activities 1941-1 946: Shaping the post-war world: Bretton Woodsand reparations, ed. Donald Moggridge. London: Macmillan. . 1982. The collected writings of John Maynard Keynes. Vol. 21,Activities 19311939: World crises andpolicies in Britain andAmerica, ed. Donald Moggridge. London: Macmillan. Kindleberger, Charles P. 1965. Balance-of-payments deficits and the international market for liquidity. Princeton Essays in International Finance, no. 46. Princeton, N.J.: Princeton University, Department of Economics, International Finance Section. . 1984. A financial history of Western Europe. London: Allen and Unwin. . 1986. The world in depression, 1929-1939, rev. ed. Berkeley and Los Angeles: University of California Press. League of Nations. 1938. Report on exchange control. Geneva: League of Nations. Lewis, Karen K. 1996. Puzzles in international financial markets. In Handbook ofinternational economics, vol. 3, ed. Gene M. Grossman and Kenneth Rogoff. Amsterdam: Elsevier. Lothian, James R. 1995. Capital market integration and exchange-rate regimes in historical perspective. New York Fordham University, December. Photocopy. Lucas, Robert E., Jr. 1990. Why doesn’t capital flow from rich to poor countries? American Economic Review 80 (May): 92-96.

401

International Capital Mobility over the Long Run

Marston, Richard C. 1995. International jnancial integration. Cambridge: Cambridge University Press. McKinnon, Ronald I. 1979. Money in international exchange: The convertible currency system. New York: Oxford University Press. Mikesell, Raymond E 1954. Foreign exchange in the postwar world. New York: Twentieth Century Fund. Moggridge, D. E. 1971. British controls on long term capital movements, 1924-1931. In Essays on a mature economy: Britain after 1840, ed. Donald N. McCloskey. London: Methuen. Morgenstem, Oskar. 1959. International financial transactions and business cycles. Princeton, N.J.: Princeton University Press. New York Times. 1943. 30 March. Nurkse, Ragnar, with William Adams Brown, Jr. 1944. International currency experience: Lessons of the inter-war period. Geneva: League of Nations. Obstfeld, Maurice. 1993. The adjustment mechanism. In A retrospective on the Bretton Woodssystem: Lessonsfor international monetary reform, ed. Michael D. Bordo and Barry J. Eichengreen. Chicago: University of Chicago Press. . 1995. International capital mobility in the 1990s. In Understanding interdependence: The macroeconomics of the open economy,ed. Peter B. Kenen. Princeton, N.J.: Princeton University Press. Obstfeld, Maurice, and Kenneth Rogoff. 1996. Foundations of international macroeconomics. Cambridge, Mass.: MIT Press. Officer, Lawrence H. 1996. Between the dollar-sterling gold points: Exchange rates, parity, and market behavior. Cambridge: Cambridge University Press. Perkins, Edwin J. 1978. Foreign interest rates in American financial markets: A revised series of dollar-sterling exchange rates, 1835-1900. Journal of Economic History 38 (June): 392-417. Rees, David. 1973. Harry Dexter White: A study in paradox. New York: Coward, McCann and Geoghegan. Sachs, Jeffrey D., and Andrew Warner. 1995. Economic reform and the process of global integration. Brookings Papers on Economic Activity 1:1-1 18. Salera, Virgil. 1941. Exchange control and the Argentine market. New York Columbia University Press. Sayers, R. S. 1976. The Bank of England 1891-1944. Cambridge: Cambridge University Press. Solomon, Robert. 1982. The international monetary system, 1945-1981. New York: Harper and Row. Spalding, William F. 1915. Foreign exchange and foreign bills in theory and in practice. London: Pitman. Stewart, Robert B. 1938. Great Britain’s foreign loan policy. Economica 5 n.s. (February): 45-60. Svensson, Lars E. 0. 199I. The simplest test of target zone credibility. ZMF Staffpapers 38 (September): 655-65. Taylor, Alan M. 1996a. International capital mobility in history: Purchasing power parity in the long run. NBER Working Paper no. 5742. Cambridge, Mass.: National Bureau of Economic Research, August. . 1996b. International capital mobility in history: The saving-investment relationship. NBER Working Paper no. 5743. Cambridge, Mass.: National Bureau of Economic Research, August. Temin, Peter. 1989. Lessonsfrom the Great Depression. Cambridge, Mass.: MIT Press. Triffin, Robert. 1957. Europe and the money muddle: From bilateralism to nearconvertibility, 1947-1956. New Haven, Conn.: Yale University Press.

402

Maurice Obstfeld and Alan M. Taylor

Viner, Jacob. 1943a. Trade relations between free-market and controlled economies. Geneva: League of Nations. . 1943b. Two plans for international monetary stabilization. Yale Review 23177-107. Weill, N. E. 1903. Die Solidaritat der Geldmarkte. Frankfurt: Sauerlbder’s. Williamson, Jeffrey G. 1996. Globalization, convergence, and history. Journal of Economic History 56 (June): 277-306. Wood, Adrian. 1994. North-South trade, employment, and inequality: Changing fortunes in a skill-driven world. Oxford: Clarendon. Wyplosz, Charles. 1986. Capital controls and balance of payments crises. Journal of International Money and Finance 5 (June): 167-79. Yeager, Leland B. 1976. International monetary relations: Theory, history, and policy, 2d ed. New York: Harper and Row.

12

Implications of the Great Depression for the Development of the International Monetary System Michael D. Bordo and Barry Eichengreen

Understanding the impact of the Great Depression on the development of the international monetary system requires one to ask the counterfactual question: What would the world have been like had the Depression not occurred? In this paper we speculate about the evolution of the international monetary system in the last two-thirds of the twentieth century absent the Great Depression but present the major postdepression political and economic upheavals: World War I1 and the cold war.' We argue that without the depression the gold exchange standard would have persisted until the outbreak of World War 11. It would have been suspended during the war and for a period of postwar reconstruction before being restored in the first half of the 1950s. The Bretton Woods Conference would not have taken place, nor would a Bretton Woods system of pegged but adjustable exchange rates and restrictions on capital account convertibility have been established. Instead, an unreformed gold exchange standard of pegged exchange rates and unlimited international capital mobility would have been restored after World War 11. But this gold exchange standard would have collapsed even earlier than was Michael D. Bordo is professor of economics at Rutgers University and director of the Center for Monetary and Financial History at Rntgers. He is also a research associate of the National Bureau of Economic Research. Bany Eichengreen is the John L. Simpson Professor of Economics and Political Science at the University of California, Berkeley, a research associate of the National Bureau of Economic Research, and research fellow of the Centre for Economic Policy Research. The authors owe a great debt to Darren Lubotsky, Kris Mitchener, and Torsten Sl@kfor excellent research assistance. For helpful suggestions they thank Jim Boughton, Peter Kenen, Naomi Lamoreaux, Allan Meltzer, Ed Perkins, Anna Schwartz, and Mike Woodford. They thank Vittorio Grilli and Gian Milesi-Ferretti for sharing their data. The second author thanks the first for waiting until noon to call. 1. Alternatively, one might argue that these political upheavals would not have themselves occurred had they not been preceded by the depression. Here we ignore this more complicated counterfactual.

403

404

Michael D. Bordo and Barry Eichengreen

actually the case with Bretton Woods. The move toward floating exchange rates that followed would have taken place well before 1971. The remainder of the paper is organized as follows. Section 12.1 describes the counterfactual in more detail. Section 12.2 draws out its implications for the operation of the international monetary system. In section 12.3 we present a model of the international monetary system from 1928 to 1971 and simulate its implications for the determination of the world price level and the durability of the gold exchange standard. Section 12.4 examines the implications for economic growth and resource allocation of allowing 1920s-style international capital mobility after World War 11. Section 12.5 contemplates the implications for institution building and international cooperation of the “no Great Depression” scenario. Section 12.6 concludes with a summary evaluation of the impact of the depression on the international monetary system.

12.1 Constructing the Counterfactual An adequate counterfactual requires not only assuming that the Great Depression did not take place but also being explicit about how the macroeconomic disaster of the 1930s was averted. On the assumption that the depression resulted from contractionary monetary policies that were transmitted internationally through the operation of the gold exchange standard, our counterfactual is that the disaster of the 1930s was avoided by the maintenance of stable monetary policies. There exist two variants of this hypothesis, both with the same implications for our counterfactual analysis. In one, the depression was precipitated by monetary events in the United States. Restrictive monetary policies precipitated the 1929 downturn, and inept monetary policies aggravated the depth and duration of the slump.2Other countries, their monetary and credit conditions tied together by the fixed exchange rates of the gold exchange standard, imported these contractionary impulses; they too lapsed into depre~sion.~ The corresponding counterfactual is that the Federal Reserve maintained a stable monetary policy throughout the 1930s, averting the banking panics of 1930-33, and that the rest of the world did not import a deflationary shock from the United States? 2. Field (1984) and Hamilton (1987) show that stringent monetary conditions played an important role in the onset of the depression in the United States, while Friedman and Schwartz (1963) paint a detailed picture of the role of monetary policy in aggraiiating the severity of the slump. 3. This point is documented by Choudhri and Kochin (1980). 4. To the extent that bank failures would have occurred anyway, we assume that the Fed acted as lender of last resort to contain their spread. For simulations of the impact of a stable U.S. monetary policy on output and the price level in the interwar period, see Bordo, Choudhri, and Schwartz (1995) and McCallum (1990). This scenario presumes that the Fed had the understanding and policy tools to pursue monetary stability. Meltzer (1999, following Wheelock (1990), argues that Federal Reserve officials were wedded to the flawed Burgess-Riefler-Story doctrine and would have been unable to follow the correct course. Against this objection we cite two facts:

405

The Great Depression and the International Monetary System

In the other variant of the hypothesis, no one country was large enough and no one set of national policies was sufficiently powerful to precipitate a global slump. Rather, the depression was induced by the simultaneous adoption of inappropriately restrictive policies by several leading countries. Monetary restriction in the United States, which reflected the Federal Reserve Board‘s desire to damp down the Wall Street boom, coincided with restrictive monetary policies in France, as the French authorities refused to accommodate the increase in money demand that followed inflation stabilization and converted the country’s foreign exchange reserves into gold.5 Together, contractionary impulses emanating from the two countries that between them held more than half of global gold reserves forced monetary restriction on other nations and depressed output and employment worldwide. The corresponding counterfactual is that the depression was averted by better, coordinatedinternational monetary management. The countries concerned each could have reduced their discount rates and initiated expansionary open market operations without destabilizing their exchange rates or the gold exchange standard. The spread of deflation and banking crises under the fixed-rate gold standard would have been contained. This would have finessed the dilemma confronting the monetary authorities, of either staying on gold and deflating or following expansionary policies and being forced off (Eichengreen 1992).It would have prevented the Central European countries and the United Kingdom from being forced to abandon the gold standard in 193 1. Either way, the Great Depression would have been avoided, which is the starting point for our counterfactual analysis. Absent the depression, the gold exchange standard would have survived the 1 9 3 0 ~The . ~ devaluations of sterling in 1931, the dollar in 1933, and the franc in 1936 would not have occurred.’ The system then would have been suspended during World War 11, echoing the experience of World War I. We assume that the same fraction of wartime the connection between stable money and the real economy was known at the time, as noted by Laidler (1994), and other central banks had on previous occasions acted successfully as lenders of last resort. 5. This is the “international explanation” for the Great Depression. See Temin (1989), Eichengreen (1992), and Bemanke (1995). 6. This assumes that neither the decline in primary commodity prices in the second half of the 1920s nor German reparations transfers would have destabilized the exchange rates of major participants in the system. We provide analysis in defense of this assumption below. 7. This assumes that other problems with the operation of the gold exchange standard would not have precipitated its early collapse. Potentially,such problems included the tendency for countries to sterilize reserve flows (to violate the “rules of the gold standard game”), the liquidity problem (caused by insufficiently elastic supplies of gold and foreign exchange reserves), the confidence problem (caused by central banks’ wholesale liquidation of foreign exchange), and the failure of international cooperation in support of currencies in distress (Eichengreen 1990; Bordo 1993).We argue that none of these problems would have brought down the gold exchange standard in the 1930s and show that, absent the Great Depression, this assumption is reasonable.

406

Michael D. Bordo and Barry Eichengreen

expenditures would have been financed with inflation as actually was the case, so that the inflation rate in the United States, Britain, and other major countries would have been roughly the same as that actually observed. After the war the United States would have moved to restore the gold standard at the prewar parity (the prewar gold price of $20.67). It would have done so because this strategy had produced relatively satisfactory results after World War I. It would have been joined by the major continental European countries, since their experience with the interwar gold standard, to which they had returned (at devalued parities) following the inflation of the early 1920s, would have been broadly satisfactory in the absence of the depression.* The distribution of the monetary gold stock would have been similar to what it was; the United States would have held the lion's share. While the increase in U.S. gold holdings that resulted from the 1933-34 devaluation of the dollar would not have occurred, there still would have been gold flows to the United States as a result of political uncertainty in E ~ r o p eAnd . ~ even if the European Allies had retained more gold in 1939, they would have used more of it to purchase war mattriel from the United States following the outbreak of hostilities. The implication is that the dollar shortage that plagued postwar Europe would have been little different in the absence of the Great Depression.'O The Marshall Plan and other postwar aid still would have been needed. Assuming that resumption would have occurred in the first half of the 1950s with the U.S. price of gold at $20.67 and other currencies realigned as they were with respect to the dollar, countries other than the United States would have had to deflate in order to acquire the reserves needed to restore convertibility. Given the lessons learned from the 1920-21 deflation, there is reason to think that there would have been resistance to a very radical deflation like 8. Of the major countries, only Britain might have dissented. Although the 1931 devaluation of sterling would not have occurred under our counterfactual, the high unemployment of the 1920s would. We assume that unemployment would have persisted at approximately the same rate in the 1930s. Combined with the loss of reserves and the obligations to the United States and the Commonwealth incurred during World War 11, this would have led Britain to resist returning to the gold exchange standard at the prewar parity a second time. Instead, Britain would have devalued the pound to approximately its 1949 level. An alternative assumption is that Britain would have floated the pound in response to these difficulties. Against it one might argue that British policymakers were unwiIIing to countenance floating after World War II (the possibility was rejected in the context of the ROBOT Plan in 1952) and that they would have been even less willing to contemplate such radical measures absent the experience of the Great Depression. On the other hand, one might conjecture that absent the perceived problems with floating exchange rates in the 1930s, policymakers would have been more willing to toy with the idea of floating after World War 11. Our own view is that contemporary perceptions of the operation of floating exchange rates in the 1930s were heavily conditioned by the instability and unsatisfactory performance of floating in the first half of the 1920s (Eichengreen 1992). The interlude of floating in the early 1920s would still have occurred under our counterfactual; hence, we assume that there would have been a continued aversion to floating after World War II. 9. Romer (1993) ascribes the largest part of the gold and capital flows from Europe to the United States in the second half of the 1930s to politically motivated capital flight. 10. For a sampling of writings on the dollar shortage, see Balogh (1946), Williams (1952), and MacDougall(l957).

407

The Great Depression and the International Monetary System

that of 1920-21. The 1920-21 deflation had caused severe dislocations: with the decline of wage indexation in Britain (the “sliding scale agreements”-see Thomas 1994) and, more generally, the increasingly structured nature of labor markets after World War 11, a repeat of that experience would have given rise to even higher levels of unemployment, something that powerful trade unions would have been unwilling to accept.” The only way the system could then have been resurrected would have been for the United States to provide sufficient liquidity to enable the European countries to restore convertibility at the somewhat higher level that prices had scaled.’* In this case and assuming that actual and counterfactual world money supplies were the same, the United States would have had to transfer $26 billion to its trading partners, almost double what it provided under the Marshall Plan, to permit the gold exchange standard to be reestablished in the early 1950s. This might seem like a pipe dream, given the opposition that existed in the U.S. Congress to the Marshall Plan, but if we add to Marshall Plan transfers the Anglo-American loan of 1945 and the U.S. quota in the International Monetary Fund (since we will argue that this last institution would not have existed in the absence of the Great Depression), we get a total that equals the liquidity required to restart the global gold exchange standard in the early 1950s. Given the imperatives of the cold war, which would have remained under our counterfactual, a transfer of this magnitude is not impla~sib1e.l~

12.2 Implications of the Counterfactual Assuming that the major countries all resumed current and capital account convertibility in the early 1950s and that other significant provisions of the gold exchange standard were maintained, the post-World War I1 international monetary system would have differed from that which actually prevailed in four important respects. 12.2.1 The Bretton Woods Institutions It follows from our counterfactualthat the Bretton Woods Conference would not have been convened. Most of the problems of concern to its participants would not have arisen in the absence of the depression: these include bilater11. Eichengreen (1992, chap. 3) discusses the macroeconomiceffects of the 1920-21 deflation. Bayoumi and Eichengreen (1996) discuss the literature, suggesting a decline in wage and price flexibility after World War II. 12. Eichengreen (1993) concludes that this transfer was not necessary because it assumes the maintenance of capital controls, which limited the impact of shocks on the balance of payments and hence the need for reserves. 13. This, of course, is only one of several more or less equally plausible assumptions. Readers for whom this stretches credulity will want to refer to model D below, where we undertake a sensitivity analysis of this assumption, assuming that the world instead initiated a radical deflation after World War II, implying that the post-World War II world started off with a money supply some 45 percent below the actual, obviating the need to increase the magnitude of the Marshall Plan to get the gold exchange standard restarted.

408

Michael D. Bordo and Barry Eichengreen

alism, exchange controls, competitive devaluations, destabilizing speculation, hot money flows, and the international transmission of deflation. There would thus have been no International Monetary Fund (IMF) to provide surveillance and conditional assistance and to supplement other sources of international reserves. Cooperation and policy coordination would been arranged directly among the major powers as in the 1920s. Governments would have found it difficult to resort to parity changes in the event of balance-ofpayments problems, as authorized by the IMF Articles of Agreement in the event of fundamental di~equilibrium.’~ And there would have existed no systemwide controls on capital flows. This would have made it more difficult for countries to follow independent monetary and fiscal policies. 12.2.2 The Supply and Demand for Reserves Because post-World War I1 central banks would have been bound by national convertibility statutes, which required them to back their liabilities with gold and (often, limited amounts of) foreign exchange, gold would have been a more important component of international reserves than under Bretton Woods. The world demand for monetary gold would have been greater than was the case in fact. But the supply would have been different too. Gold production would have been less between 1929 and 1933 because, with less deflation, the relative price of gold would have been lower, and the new flow supply would have been 1 e ~ s . lProduction ~ would have risen at best slowly in the 193Os, as it had in the 192Os, reflecting improvements in mining technology. It would have fallen during World War 11, as it did in the latter stages of World War I, reflecting paper-money-induced inflation and governments’ efforts to limit exchange rate fluctuations.16These changes would have reduced the world monetary gold stock in the postwar period below what it actually was. As a result, there would have been an even greater demand for foreign exchange to supplement gold as international reserves. Most of that additional 14. In contrast, this practice was uncommon under the gold exchange standard of the 1920s. Eichengreen (1995) discusses why this “escape clause” provision, which had also been invoked (in the form of temporary suspensions of convertibility followed by resumption at the previous rate) under the pre-World War I gold standard (see Bordo and Kydland 1995) had become increasingly difficult to invoke under the reconstructed gold standard of the 1920s. 15. Production also would have been less after 1933 because the United States would not have raised the nominal gold price from $20.67 to $35.00 per ounce. 16. Fiat money inflation would normally be expected to raise the nominal price of gold and the prices of other commodities commensurately, with no implications for the real price of gold. During World War I, however, European governments sought to peg their currencies to gold and the dollar at only slightly depreciated rates; they prevented the price of gold from rising at the same rate as other commodity prices, eroding the incentive to devote resources to its production. And the United States maintained the $20.67 peg through the war (with the exception of the gold embargo of 1917-19). Our assumption is that broadly similar policies would have been pursued during World War II, especially since the United States never abandoned the $35.00 per ounce peg and the Tripartite Agreement of 1936 attempted to stabilize the franc and the pound. In addition, in the face of less deflation, the nonmonetary gold stock would have increased relative to the monetary gold stock in the 1930s and more so during the World War I1 inflation.

409

The Great Depression and the International Monetary System

demand would have been met by dollars rather than sterling, since Britain’s post-World War 11economic difficulties would not have been any less severe and governments would have been reluctant to hold the currency of a country in dire straits. The stock of dollars held as official foreign balances would have grown more quickly than was actually the case.” This would have pushed the world toward a dollar standard in which the United States as center country held gold as its reserves and the rest of the world held dollars. For political reasons, however, a number of countries would have resisted giving up their gold reserves and going onto a dollar standard. They would have likely precipitated its collapse even sooner than was actually the case.18

12.2.3 The Capacity for Adjustment and Intervention The reconstructed gold exchange standard would have been free of restrictions on capital movements, as had been the case before World War I and in the 192Os.I9Countries therefore would have faced an even tougher balance-ofpayments constraint than under Bretton Woods. Sterilized intervention would have been the only instrument available for insulating economies from the policies required for the maintenance of external balance. The same evidence that suggests that sterilized intervention has offered governments only limited room for maneuver in the high-capital-mobility environment of recent decades suggests that this would have been true after World War I1 under our counterfactual.*O Under Bretton Woods, countries could alter parities in response to a fundamental disequilibrium. In the counterfactual this is no longer the case. The only channel of adjustment for deficit countries (other than breaking the link to gold) would have been deflation and, in the face of sticky wages and prices, 17. This assumes that countries had been able to acquire these balances at the desired rate. As explained above, this would have required that the Marshall Plan transfer to the European countries to have been larger than the $13 billion granted. 18. Had the price of gold been increased afrer World War II, as was suggested in the 1960s by Jacques Rueff, among others, then the pressure would have been less and the ultimate collapse would have been later. But we think that there were good reasons why schemes to adjust the domestic price of gold under a gold-exchange-standard-likesystem were problematic.An altemative point of view (advocated by, inter aha, Meltzer 1991 and McKinnon 1969) is that had the United States continued to follow stable money policies, the rest of the world would have been willing to continue to use dollars as international reserves instead of gold. The gold exchange standard would have evolved into the kind of dollar standard advocated in the Bretton Woods era by McKinnon and others.We argue below, however, that historical factors grounded in countries’ earlier experiences with the operation of the gold standard rendered this outcome unlikely. 19. In addition, the IMF would not have been present to advance liquidity to countries facing balance-of-paymentsdifficulties, as already noted. 20. Evidence for the interwar period based on offset coefficients (Kwiecinska-Kalita 1996) similarly suggests that most countries (other than the United States) had very little leeway for independent monetary policy action. The evidence for the Bretton Woods period, when capital controls were pervasive, stands in contrast: Kouri and Porter (1974) and Obstfeld (1982), among others, demonstrate a role for sterilized intervention for the major European countries. Pasula (1994, 1996), in contrast, finds that the offset to monetary policy under Bretton Woods was complete.

410

Michael D. Bordo and Barry Eichengreen

depression. This would have created pressure to break with gold in favor of greater exchange rate flexibility. 12.2.4 The Efficiency of Resource Allocation Absent the capital controls of the Bretton Woods system, the reintegration of world capital markets would have occurred more rapidly. Resource allocation would have been better, accelerating growth and convergence in countries that started the postwar period with real incomes below that of the United States. How much faster convergence would have been and what countries would have benefited is unclear, however. Abramovitz (1989) and Wolf (1995) show that there was already rapid convergence among the high-income countries, notwithstanding low levels of international capital mobility. It could be that other channels, including international trade and technology transfer, transmitted to the members of the “convergence club” the main gains from openness. At the same time, Saint-Paul (1995) suggests that Europe and Japan would have grown even faster had international financial liberalization allowed them to more quickly augment their capital stocks. Sachs and Warner (1995), on the other hand, can be read as suggesting that the main benefits from financial liberalization would have accrued not to the high-income countries that were already members of the convergence club but to low-income developing countries less able to import expertise through other channels.

12.3 The Gold Exchange Standard, 1925-60 In this section we develop a model of the gold exchange standard on the assumption that, rather than collapsing during the Great Depression, it continued operating until the outbreak of World War I1 and was reestablished following the conclusion of hostilities. We extend a generic model of the gold standard to incorporate some special features of interwar monetary arrangements. In calibration we use data for a composite of 21 countries that account for 75 percent of the world monetary gold stock and a comparable share of economic activity in 1928.

12.3.1 The Model

A simulation model of the global gold standard was developed by Bordo and Ellson (1983, building on the theoretical model in Barro (1979). The model contains a money market and a gold market. The former determines the world price level, given the world monetary gold stock determined in the latter. The gold market takes the price of gold as fixed by the authorities and, given the price level from the money market, determines the real price of gold and the world monetary gold stock. Equation (1) is the money supply:

411

The Great Depression and the International Monetary System

(1)

M s = XP,G,,

where M" is the world money supply in dollars, X is the money-gold multiplier (the ratio of currency plus deposits to the monetary gold stock), pG is the fixed nominal price of an ounce of gold, and G, is the world monetary gold stock (in ounces). Equations (2a) and (2b) are two variants of the income velocity of circulation. Equation (2a) assumes that velocity is a logarithmic function of the nominal interest rate:*l

v

(24

=

Via.

Equation (2b), following Bordo and Jonung (1987), takes trend velocity as a function of institutional factors that evolve at a rate p:

v

(2b)

= Vewr.

The nominal interest rate is

(3)

i = r + n ,

where r is the real rate of interest and ~r is the expected rate of change of the price level. Following Mundell(l970) the real interest rate depends negatively on expected inflation: r = F - an.

(4)

We assume perfect foresight so that actual and expected inflation are equal:22 (5)

Tr

=

(r: - &)Jr:-*,

Money market equilibrium requires (6)

P = XVP,G,Jy.

Given 1,PG, perfect foresight, and an exogenous level of output, y, the price level is determined by the monetary gold stock. Equations (7) to (9) determine the gold market equilibrium and, together with equations (1) to (6), the money supply and the price level. Gold production is characterized by increasing costs; the supply of new gold is (7)

g = gcbeYt,

21. We depart from Barro (1979), who assumes a constant real interest rate and makes the demand for money a function of the expected rate of change in the price level. But following Barro, eqs. (2a) and (2b) assume the real income and price elasticities of real money demand to be one. 22. Bordo and Ellson (1985) also use an adaptive expectations scheme. In their simulations, the adjustment path of the model differed under the two schemes but the long-run equilibrium values of the endogenous variables are of course unaffected.

412

Michael D. Bordo and Barry Eichengreen

where g is gold production, P, is the real price of gold ( P ,/ P ) ,and y is the exogenous rate of technological progress in the mining The flow demand for nonmonetary gold is (8)

6,

= (E

+

S)(G,Y - G,) ,

where G, is the net change in the nonmonetary gold stock, G,* is the target stock of nonmonetary gold, and G , is the actual stock. G,* is defined as G,Y =

CN p-0 Ti-+ g Y

,

E is a partial adjustment factor, and S is the depreciation rate or normal replacement flow. Since the monetary authorities maintain a fixed price of gold, the change in the monetary gold stock is

(9)

G, = g - G N ,

where GMis the net change in the monetary gold stock. Taking logs, and solving equations (1) to (9) simultaneously, steady state solutions in terms of growth rates are P = GM = G N = O . This implies g = g,*: that, in equilibrium, gold production equals the depreciation rate multiplied by the desired nonmonetary gold 12.3.2 Incorporating the Gold Exchange Standard A key difference between the gold standard and the gold exchange standard is A, the money-gold multiplier. After World War I, most nations adopted a form of gold exchange standard in which foreign bonds and bills of exchange supplemented monetary gold stocks. To capture this feature of the gold exchange standard, we decompose A into the ratio of broad money to highpowered money (MIH),the ratio of high-powered money to gold and foreign exchange reserves (HIR), and the ratio of international reserves to gold (RIG) (Bernanke 1995). MIH depends on the development of the banking system, while HIR is determined by sterilization policies and by the laws specifying gold backing for liabilities; in the United States, for example, the Federal Reserve before 1945 had a gold cover requirement of 40 percent against notes and 35 percent against deposits, which implied a minimum value for HIR of 23. Barro does not account for technological progress but discusses the implications of incorporating it. We assume that technological progress is exogenous. In fact, there is evidence that major technological changes in the gold industry were both induced and exogenous (Rockoff 1984). 24. While Bordo and Ellson (1985) account for the fact that gold is a durable exhaustible resource, we eliminate this aspect of the model to simplify the analysis. Our hypothetical gold standard will exhibit less deflation than if we accounted for depletion.

413

The Great Depression and the International Monetary System

2.63. RIG reflects the substitution of foreign exchange for gold. For the center countries of the interwar gold standard, the United States and the United Kingdom, RIG = 1. The same is true of the United States after World War 11. Figure 12.1A shows the evolution of the three ratios from 1925 for our aggregate of 21 countries.z5MIH is stable prior to the onset of the depression, when it declines in response to the global banking crisis. HIR begins falling with the establishment of the gold exchange standard in 1925, reflecting sterilization policies in the United States and France. RIG rises up to the onset of the depression, reflecting the substitution of foreign exchange for gold by participating countries. It then declines as the participants, fearing speculative attacks on the reserve currencies, converted their foreign exchange to gold. The multiplier A can be broken into two components:

M = -M H A(1) = R H R and A(2) =

R G

-

12.3.3 The Distribution of Gold The adequacy of the global supply of monetary gold and its distribution between deficit and surplus countries dominated discussions of the interwar gold exchange standard and the Bretton Woods system. The first concern is readily addressed using our model, while addressing the second requires us to specify one further relationship. We assume that the demand for reserves by the center country(ies) and the rest of the world depends on levels of economic activity, a measure of the opportunity cost of holding reserves, and factors such as their openness (see Eichengreen 1990). The center country(ies) holds only gold reserves, while the rest of the world holds both gold and foreign exchange, with the breakdown between the two determined by the return on short-term foreign assets and legal gold cover requirements. We assume that the center country's share of total reserves (and of the world monetary gold stock) paralleled the evolution of the share of its output in the world's total. 12.3.4 Simulating the Model Table 12.1 lists the parameters used in the simulation.26Table 12.2 shows the initial values and data definitions. We developed world aggregates for the endogenous variables (money supply, monetary base, international reserves, 25. Complete data for our 21 countries are not available after 1935. 26. The elasticities in the gold market equations are derived from regression estimates using data for the period 1880-1928. In the sensitivity analysis reported below we also used elasticities estimated in the U.S. Gold Commission Report (1982) using post-World War II data.

A

..

4.0-

- 10

3.5-

-9

3.0-

-8

2.5-

2.01.5

*-

-

1.0-

-

0.5

___----

- _ _ -. -.

. - - - ---*.; - -_ 5_ - _

I

I

I

I

I

---

R/G - M/H Lambda (2) - - - H/R

6 6

-6 -

. I .

I

-

--

-7

* - - -

I

I

1

4

I

Lambda(1) (right axis)

M/R (right axis)

1

r-10

5

4 3

2 1 0

- Lambda(1) (right axis) _- -_ -_ M/H R/G - M/R (right axis)

- - * *

Fig. 12.1 Ratios for the world, 1925-35 (A) and 1950-51 ( B )

H/R Lambda ( 2 )

Note: International reserves and gold reserves have been adjusted to take out the IMF, as described in the text.

415

The Great Depression and the International Monetary System Parameters of the Simulated Models

Table 12.1

B. 1950-71

A. 1928-38

-

V = 2.26 a = 0.25 g = 3.31 (Y = 0.5 f3 = 0.6 G , = 0.08 0 = 0.3 q = 0.7 K = (q; - ) : q

-

V = 1.65 a = 0.25 g = 5.04 (Y = 0.5 fi = 0.6 -

C. Gold Commission Elasticities

p = 0.6 y = 0.03 e = 1.2 q = 1.0 4 = 0.1

G , = 0.04

e = 0.3

y = 0.133*0.0394

q = 0.7 K = ($

-9 ): y = 0.55*0.0428 8.5 h(1) = (M/R),950 = 10.7 (R/G)1950 = 1.2 A(2) = (WG),,,,e"'; (WG),928= 1.3 h(2) = (R/G)1950eK'; = 0.03 = 0.03 r = 0.045 7 = 0.0145 p = -0.026 = 0.00285 y = 0.018 y = 0.018 E = 0.5 E = 0.5 6 = 0.01 6 = 0.01

A( 1) =

+

=

+

Source: Panel C from US.Gold Commission Report (1982).

gold reserves, price level, real income) based on data for the 21 countries considered in Bordo and Schwartz (1996).27 These 21 countries account for roughly 75 percent of the world monetary gold stock before World War I1 and 85 percent thereafter, and for the vast majority of world economic activity.28 We construct X(l) and X(2) assuming that the shares of the 21 countries in world gold and international reserves remained constant after two benchmark years (1928 and 1950) for which world data are a ~ a i l a b l eThis . ~ ~ allows us to connect the data for our 21 countries to the world gold market totals in the US. Gold Commission Report (1982).30 We simulate the model over the period 1929-38 on the assumption that the 27. The countries are United States, United Kingdom, France, Germany, Japan, Italy, Canada, Netherlands,Belgium, Sweden, Norway, Denmark, Finland, Portugal, Spain, Switzerland,Greece, Australia, Argentina, Brazil, and Chile. 28. For the postwar period, we dropped the three Latin American countries from the sample because they followed unstable monetary and exchange rate policies atypical of the Bretton Woods experience and no longer played a significant role in the internationalmonetary system. 29. Whenever we refer to 21 countries, we should be understood as meaning 18 countries (excluding the three Latin Americans) after World War II. We also adjusted the world monetary gold stock series taken from League of Nations (1931), which ends in 1931, to the series world central bank gold reserves from the US.Gold Commission Report (1982), which extends from 1913 to 1980. The two series became virtually identical after the United States nationalized private gold holdings in 1934. 30. We used U.S. interest rates as representative of the world and assumed US. total factor productivity growth (y) as representative of technologicalprogress in world gold production.

416

Michael D. Bordo and Barry Eichengreen

Table 12.2 Initial Value

P, = 20 P = 1 y = 191

M S = 85

H = 23

R = 10 G , = 7.7 G , = 535 G , = 489 g = 20 V = 2.26 i = 0.045

P , = 20 (20) P = 1.51 (1.61) y = 276 (351)

M S = 254 (317) H = 64 (81) R = 24 (25) G , = 19 (20)

G , = 917 (928) G , = 566 (605) g = 24 (24) V = 1.65 (1.87) i = 0.0145 (0.0252)

Initial Values and Definitions of Variables Definition A. 1928-38 Price of gold; dollars per ounce. World price index: weighted sum of 21 countries’ GNP deflators 1928 = 1, weights = shares of each country’s GNP in total GNP in current dollars. World output; billion dollars. Sum of 21 countries’ real GNP in current dollars, it is assumed to grow at approximately 3.9 percent, the growth rate in 1921-28. Money supply; billion dollars. M2 definition (currency plus total deposits), sum of 21 countries in current dollars. High-powered money (currency plus reserves): sum of 21 countries in current dollars. International reserves (central bank gold reserves plus foreign exchange reserves): sum of 21 countries in current dollars Central hank gold reserves: sum of 21 countries in current dollars. World monetary gold stock; millions of ounces. World nonmonetary gold stock; millions of ounces. World gold production first period; millions of ounces. World velocity in 1928. U S . nominal interest rate in 1928 (short-term commercial paper rate). B. 1950-71 (1953-71) Fixed price of gold; dollars per ounce. World price index: weighted sum of 18 countries’ GNP deflators 1928 = 1, weights = shares of each country’s GNP in total GNP in current dollars. World output; billion dollars. Sum of 18 countries’ real GNP in current dollars, it is assumed to grow at approximately 4.3 percent, the growth rate in 1950-71. Money supply; billion dollars. M2 definition (currency plus total deposits), sum of 18 countries in current dollars. High-powered money (currency plus reserves): sum of 18 countries in current dollars. International reserves (central bank gold reserves valued at $20 per ounce plus foreign exchange reserves less reserve position in IMF plus IMF quota), sum of 18 countries in current dollars. Central bank gold reserves plus IMF quota: sum of 21 countries in current dollars. World monetary gold stock; millions of ounces. World nonmonetary gold stock; millions of ounces. World gold production first period; millions of ounces. World velocity in 1950. U.S. nominal interest rate in 1950 (short-term commercial paper rate).

arrangements in place in the second half of the 1920s were not interrupted by the depression. We assume that the system was then suspended with the onset of World War I1 (as it had been following the outbreak of World War I). In the period 1939-49, we treat the world as on a fiat monetary standardthat is, as if central banks closed their gold windows. Our first counterfactual

417

The Great Depression and the International Monetary System

Table 12.3

Assumptions behind the Simulated Models

Model Model A: benchmark

Assumptions 1. Constant velocity at average level of 1921-28 for

interwar simulation; 1950 level for postwar simulation 2. Ratio of world money supply to international reserves (h(1))fixed at 1928 level for interwar simulation; 1950 level for postwar simulation

3. Ratio of international reserves to gold (X(2)),a function of the difference between the income elasticities of

Model B: Velocity trends

Model C: Velocity as function of nominal interest rate Model D: Post-World War II resumption similar to post-World War I Model E: Sensitivity analysis of model A using postwar gold market elasticities Model F: Sensitivity analysis of model A with the postwar gold exchange standard

demand for total international reserves and the demand for gold reserves (0.133interwar; 0.55 postwar) times the trend growth rate of world real income (3.9percent interwar; 4.4percent postwar) 4. Gold market elasticities (p, 8, q, +), based on regressions for the period 1880-1928 5. Real output and total factor productivity grows at the 1921-28 rate for interwar simulation; 1950-71 rate for postwar simulation Same as model A except velocity declines at 1921-28 trend in the interwar period (-2.6 percent) and rises at the 1950-71 trend in the postwar period (0.3percent). Same as model A except velocity varies with simulated interwar and postwar trends in nominal interest rates. Same as model A except money supply and price level start at much lower levels in 1950. Same as model A except for the use of postwar elasticities (p, 8, q. +) in the gold production and nonmonetary demand for gold equations. Same as model A except post-World War I1 resumption occurs in 1953 after the Marshall Plan.

reinstates the gold exchange standard in 1950 with the U.S. price of gold at $20.67 and other currencies realigned as they were vis-A-vis the dollar.31 Table 12.3 presents the assumptions underlying the simulations described below. Results are summarized in table 12.4. Model A is the simplest variant of the model; it fixes velocity at its 1928 level and allows output and total factor productivity to continue growing at their 1921-28 rates. The multiplier X(l), the ratio of world money supply to international reserves, is fixed at its 1928 level (see fig. 12.1A). This assumes that neither the United States nor France followed the restrictive monetary pol31. We choose 1950 to allow for reconstruction and the reestablishment of prewar financial relationships, and because the major devaluation of sterling and 23 other currencies in 1949,most authorities believe, reestablished the pre-World War II parities. In a simulation reported below we also tried 1953 as the starting period.

418

Michael D. Bordo and Barry Eichengreen

Table 12.4

Period and Variable

Annual Growth Rates of Different Variables (actual values and simulated values) Model

Actual

A

-1.0 -3.7 2.8 1.8 2.4 7.0 -22.0

-1.3 2.6 2.1 2.0 2.0 2.6 -2.6

B

C

D

E

F

1928-38

P M MGS NMGS WGS g i

1939-49 P M MGS NMGS WGS g i 1950-7 1" P M MGS NMGS WGS g 1

5.9

-3.3 3.2 2.7 1.5 2.2 3.8 -3.4

-2.3 2.9 2.4 1.8 2.1 3.2 -3.5

- 1.3

- 1.3

2.6 2.1 2.0 2.0 2.6 -2.6

2.6 2.0 2.0 2.0 2.6 -2.1

-1.3 2.6 2.0 2.0 2.0 2.6 -2.1

-

5.9

5.9

5.9

1.1

5.9

-

-

-

-

-

-

2.7 0.7 1.9 -5.2 6.8

2.7 -0.3 1.4 -1.6 6.8

2.7 0.0 1.6 -1.6 6.8

2.7 -0.1 1.5 -1.6 6.8

2.7 -0.0 1.3 6.8

2.7 -0.3 1.4 -1.6 6.8

2.2 0.4 1.4 -1.6 6.8

2.9 7.1 0.9 3.2 1.9 3.4 6.4

-0.8 3.4 1.1 2.6 1.7 2.3 -2.1

-0.5 3.5 1.2 2.7 1.8 2.1 -2.0

-0.9 3.7 1.4 2.6 1.9 2.8 -1.0

-0.3 4.0 1.6 2.8 2.1 2.0 -1.8

-1.0 3.3 0.9 3.0 1.8 2.8 -1.8

-1.0 3.2 0.9 2.8 1.7 2.8 -1.2

1.5

5.9

Note: P = price level, M = money supply, MGS = monetary gold stock, NMGS = nonmonetary gold stock, WGS = world gold stock, g = production, and i = interest rate. "Model F is simulated for 1953-71 (see text).

icies that led them to accumulate a growing share of global gold reserves and plunged the world economy into the depression. We assume that the ratio of reserves to gold evolved as a function of the difference between the income elasticities of the demand for total international reserves and the demand for gold reserves (equal to 0.133) multiplied by the trend rate of growth of world real income, that is,

(see fig. 12.1A).32Based on initial values for the 21-country aggregate, we simulate the model to obtain the money supply, price level, gold production, 32. These elasticities were estimated from cross-sectional regressions using data for 24 countries in 1929 in Eichengreen (1990). The sample includes most of the 21 countries included in our

419

The Great Depression and the International Monetary System

world total gold stock, world nonmonetary gold stock, world monetary gold stock, and interest rates shown in figures 12.U-12.2F. For 1939-49,the money supply process is determined not by gold supplies but by the exigencies of war finance. We take prices as exogenous and assume that they followed the actual pattern of U.S. inflation.33World gold production depends on the real price of gold, which is driven by U.S. price-level movements and productivity in the mining industry. For the postwar period we fix velocity and X(1) at their 1950 levels (Fig. 12.1B)and allow world output and U.S. total factor productivity to grow at their 1950-71 rates. The multiplier X(2), which depends on the difference between the income elasticities of demand for reserves and gold in the interwar years, turns out to be too low to provide an equilibrium solution to the model, given higher postwar growth rates. We experimented with different values for the difference in these elasticities before settling on 0.55 as the benchmark case.34(See fig. 12.1B.) We start our simulations in 1950using the hypothetical world monetary gold stock but actual international reserves, money supplies, prices, and output. This assumes that the gold exchange standard was restored without a radical postwar deflation like that of the 1920s. Given a hypothetical monetary gold stock of $21 billion and actual reserves of $61 billion, foreign exchange reserves would have had to be $13 billion larger than the actual value of $27 billion in 1950.The United States would have had to double its Marshall Plan transfer of $13 billion to get the gold exchange standard restarted.35

12.3.5 Simulation Results

Gold production increases in our simulation by less than it actually did in the 1930s (fig. 1224).This is because the model generates less deflation in the world aggregate. By using the estimates for 1929 we are therefore eliminating from our counterfactual scenario the effects of the contractionary shift from gold to foreign exchange undertaken by the Bank of France during the depression. 33. Adequate data for our 21-country aggregate are not available for the period of the war. We add simulated gold production to the 1938 simulated world gold stock and adjust for depreciation to obtain a new hypothetical stock. We assume that the world monetary gold stock increased from its hypothetical 1938 level following the actual 1939-50 trend. Given the monetary gold stock we then derive the nonmonetary gold stock. 34. The difference between the income elasticities of reserves and gold based on simple ordinary least squares regressions over the 1950-71 period was 0.47. 35. As we note above, one can imagine that the $8.8 billion of gold that member countries transferred to the IMF could have been used for other purposes, reducing the increase in Marshall aid to slightly more than $4 billion. The Anglo-American loan of 1945 of $3.75 billion by the United States and $1.5 billion by Canada as well as U.S.loans to France and Germany together could have made up the difference. Interestingly, the sum of the Marshall Plan transfer plus the hypothetical transfer that is required to restart the gold exchange standard in our scenario just equals the $26 billion that Keynes advocated as necessary to start the International Clearing Union. Below, we undertake some sensitivity analyses, varying this assumption to see how much difference is made by different degrees of postwar deflation.

A

Millions of fine OUtlCeS 50 -

45 40 35

-

30 25 20 15708 28

I I I I

33

I

I

I

- Actual

I

~

I

38

I

I

I

43

I

Simulated 50-71

B

Mil lions of f i n e ounces

1400

1

I

I

I

~~ I

I

48

I

~

~

58

- Simulated

~

I I

63

,~ I I I I I I I I I

68

I

38-50

/

llzLlzl 38

~

/ /

600

33

I

- - Simulated 28-38

I

1200i

400 28

~ I

53

43

48

j_jl 53

58

63

68

- Actual _

. ~ Simulated 28-38 and 38-50 Simulated 50-71

- Actual Simulated

50-71

- Trend"

3 8- 5 0 Simulated 2 8 - 3 8 I'

~~~

Fig. 12.2 Model A: A, world gold production; B, nonmonetary gold stock; C,monetary gold stock

D

$ Millions

1200000

1000000

1

I

800000

600000 400000 200000 0

I

i

43

48

53

58

63

68

Simulated 50-71 Actual 28-38 Simulated 28-38

1928=1

E 3.0 2.5 2.0

1.5

-----

1.0 0.5

68

_ _ _ Simulated 50-71 - Actual 28-38

-

Simulated 28-38 and Actual 3 8 - 5 0 Actual 50-71

-

Actual Simulated 50-71

- Simulated

38-50

- - - Simulated 28-38

Fig. 12.2 (cont.) D, actual and simulated money supply; E, actual and simulated prices; F, total world gold stock

422

Michael D. Bordo and Barry Eichengreen

absence of the depression and because without the 1933 devaluation of the dollar the United States would not have raised the nominal price of gold. The nonmonetary gold stock grows faster, reflecting the fact that gold is cheaper relative to other commodities in the absence of deflation (see fig. 12.28); the world monetary gold stock therefore grows more slowly (fig. 12.2C). M2 rises instead of falling as it did during the depression (fig. 12.20), and the price level (fig. 12.2E) declines at an annual rate of 1.3 percent (in contrast to the sharp deflation and rebound that actually occurred-see table 12.4). Gold production during World War I1 is below actual levels because the gold price is lower in the absence of dollar devaluation, reducing the world gold stock (fig. 12.2F). The monetary gold stock follows its actual trend but starting from the lower simulated 1938 level. The nonmonetary gold stock, derived as a residual from the total stock, is also below its historical level. The price level evolves very differently under our hypothetical standard than under Bretton Woods. Simulated deflation is about 1 percent a year, compared to an actual annual inflation rate of 2.9 percent.36Though the monetary gold stock grows rapidly in the postwar simulations and money supply expands by 3.3 percent per annum, this growth is not rapid enough to offset the effect on the price level of the 4.3 percent per annum rate of output growth. Below we summarize the results of sensitivity analyses of our model, varying assumptions on the behavior of velocity and using alternative resumption scenarios. (See tables 12.3 and 12.4 for the assumptions and results of the different scenarios.) Model B assumes that M2 velocity, rather than remaining constant, declined in the 1930s at the same 2.6 percent trend as in 1921-28 before rising at a rate of 0.3 percent per annum (as it did) from 1950 to 1971.37 This aggravates deflation between the wars but ameliorates it after World War 11. Model C assumes that velocity is a function of the nominal interest rate (which incorporates expected inflation on the assumption of perfect foresight).38This produces less of a decline in velocity in the interwar period and more of a decline after World War 11 (see fig. 12.3). Deflation is less than in model B in the interwar period and about the same as in model A in the postwar years (see table 12.4). Model D simulates the effects of deflating before restoring the gold standard after World War 11, that is, repeating the patterns followed by the United Kingdom and other countries after World War I. We start with model A but assume 36. The tendency for deflation to reduce the nonmonetary gold stock would have been offset by the growth of real income. While deflation would have stimulated gold production, the net effect would still have been growth in the monetary gold stock at rates below the actual. 37. A justification for this assumption is that trends in velocity reflected the institutional factors documented by Bordo and Jonung (1987, 1990). 38. We use the U.S. three-month commercial paper rate to proxy for the nominal interest rate. To avoid difficulties encountered in making the model converge we initially solved model A for the price level, substituted it into eqs. (4)to (6) and then solved for velocity. We then substituted the resultant trend of velocity into model A and solved for the endogenous variables.

423

The Great Depression and the International Monetary System

2.4 2.2

2.0

1.8 1.6

1.4 Model A Model C - Model B

7

- Model B A Model C

- Model

Fig. 12.3 Underlying velocity trends in models A, B, and C

that when World War LI ended the United States deflated vigorously to restore the price level to its 1938 level and other countries did likewise. Rather than allowing international liquidity to increase by $13 billion by 1950, we assume that monetary authorities maintained international reserves at their actual levels in relation to gold reserves in 1950 and that money supplies were kept in the same proportion to reserves as actually prevailed. This implies a money supply 45 percent below the actual.39 Finally, we test for sensitivity to alternative gold market elasticities. In place of elasticities based on pre-Great Depression data, we use elasticities from the US. Gold Commission Report (1982) based on data €or the postwar period. Postwar gold production and nonmonetary gold demand functions are somewhat more elastic than their prewar counterparts (see table 12.2), but simulations in model E produce patterns for the endogenous variables not that different from those in the benchmark model A (see table 12.4).40 39. Although money supplies and price levels start out below those of model A (fig. 12.4), the lower price level stimulates gold production and increases the nonmonetary gold stock. The net effect would have been to raise the monetary gold stock. This would have produced a path for the price level not dissimilar from that of model A. 40.This reinforces our belief in the robustness of the model. We tried some additional experiments. Instead of using the postwar gold market elasticities from the U.S. Gold CommissionReport (1982). we estimated our own, using the 18-country data set. The elasticities and simulations were quite similar to those reported here. Second, we ran the simulations using the pre-Great Depression elasticities for the interwar period and our postwar elasticities for the postwar period. Again the model was robust. We regard the simulations based on pre-Great Depression data as closest to the spirit of our counterfactual scenario, on the grounds that, had the gold exchange standard been preserved, the structure of the gold market would not have changed dramatically.

424 A

Michael D. Bordo and Barry Eichengreen $ Millions 1200000 -

1000000 800000

600000 400000

-_ -_ -- _ _ --

200000 0

/

1 1 1 1 1 1 1 1 1 1 1 , 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

Simulated 50-71 Actual 28-38 Simulated 28-38

B

1928=1

1

1-8

1.61.4 1.2 -

___

Simulated 50-71

__ Actual 28-38

Simulated 28-38 and Actual - Actual 50-71

38-50

Fig. 12.4 Actual and simulated (A) money supply and (B) prices, model D

12.3.6 Viability of the Gold Exchange Standard and the Triffin Dilemma

We now address two key questions about the operation of the hypothetical gold exchange standard. We ask whether there would have existed adequate gold reserves in the 1930s and how the Triffin dilemma would have played itself out after World War 11.

425

The Great Depression and the International Monetary System

1.00.8 0.6-

0.4 0 .S02 O

i 28

29

30

- Simulated

Actual

31

32

33

34

35

36

37

38

- - - 0.32

Fig. 12.5 Actual and simulated G/H for the world, 1928-38

Note: For 1936-38 data were available for only a limited number of countries.

The first “Interim Report of the Gold Delegation” (League of Nations 1930) devoted considerable attention to whether there would be enough gold to allow the system to operate for another decade. The league’s experts warned of a gold shortage but observed that it could be ameliorated by policies to encourage governments and central banks to economize on use of precious metal. We address this question by comparing ratios of gold reserves to the monetary base obtained from simulations with the league’s estimates of the legal minimum ratio mandated by gold standard statutes?’ Figure 12.5 plots the counterfactual reserve ratio for model A, along with the actual ratio and the minimum legal ratio of 31.7 percent. While the actual ratio rose in the 1930s (reflecting the collapse of price levels and the rise in the real price of gold), the simulated ratio declines slightly but never approaches the legal minimum before World War 11. These simulations suggest that a gold shortage would not have been an insurmountable obstacle to the persistence of the system. Even had this interwar problem been surmounted,Triffin (1947) warned that there would be insufficient gold to finance the growth of world output and trade in the postwar period. The substitution of foreign exchange, primarily dollars, for gold might postpone the problem, but as foreign dollar holdings increased 41. League of Nations (1930, annex 13, table 5, p. 96). To calculate the hypothetical ratio we multiplied the simulated ratio of the world monetary gold stock to the world money supply by the 1928 ratio of world money supply to the monetary base. This calculation is based on the assumption that in the absence of the Great Depression and its banking panics, the money supply multiplier (MIH)would not have declined as it did.

426

Michael D. Bordo and Barry Eichengreen

35000 30000 25000 20000 15000- _ 100005000 -

#

*-\

# \

\ #

I

\

- - @

0-

-5000

1

1

1

1

1

1

1

1

1

~

I

I

I

I

I

I

I

I

I

l

I

relative to the U.S. monetary gold stock, a point would be reached where the United States would be unable to satisfy demands for conversion. Official holdings of U.S. dollars did in fact surpass the U.S. monetary gold stock in 1965 (Bordo 1993, fig. l.lO), leading the U.S. government and Federal Reserve System to adopt gold-conserving policies and to support the development of special drawing rights (SDRs; effectively a form of paper gold).42 We can use our simulations to see whether the Triffin problem would have also arisen under the hypothetical postwar gold exchange standard. Figure 12.6, based on model A, depicts official holdings of dollars and monetary gold in the United States and the rest of the ~ o r l d . 4Starting ~ with the actual 1950 ratio of U S . gold to total reserves, we allow this ratio to then move with the ratio of U.S. to world output.44 Figure 12.6 shows that the U.S. monetary gold stock, rather than declining, 42. Total foreign dollar holdings (both private and official) surpassed the U.S. monetary gold stock in 1960. 43. We derived official dollar holdings by subtracting the simulated world monetary gold stock from simulated international reserves to give us foreign exchange holdings. We then used the actual ratio of dollars to total foreign exchange to hack out dollars held by the rest of the world in the form of foreign exchange. This calculation assumes that sterling would have declined as a reserve asset under the postwar gold exchange standard much as it in fact did under Bretton Woods. 44. To derive the U.S. monetary gold stock in 1950, we assumed that between 1929 and 1949 the U S . monetary gold stock had the same share in the hypothetical world monetary gold stock as it did in fact. This presumes that the political factors that led to capital flight and gold flows in the 1930s and 1940s were the key determinants of the rising U S . share.

427

The Great Depression and the International Monetary System

35000 0 0 *

30000 -

0

**

25000 -

****

\ \

I

I

I

‘

20000 -

15000 -

10000 5000 -

0

I

I

l

I

I

I

I

I

I

I

I

I

I

I

I

,

I

I

l

Simulated US Monetary Gold Stock Actual FE in USD Actual US Monetary Gold Stock

Fig. 12.7 “nifsn crisis,” model F

Note: FE = foreign exchange; ROW = rest of world.

would have increased throughout the period, while the monetary gold stock of the rest of the world would have declined to zero in 1970. This result is driven by the fact that the demand for reserves increases faster than the demand for gold (that h(2) = 0.55), together with the fact that the demand for reserves by the United States (the reserve currency country) could be satisfied only by gold. The dollar balances of the rest of the world would have exceeded the U.S. monetary gold stock in 1955, pointing to the possibility of a crisis at a relatively early date. Model B, based on actual trends in velocity, produces similar paths for monetary gold stocks, although the levels are somewhat higher, reflecting the inheritance of a larger monetary gold stock from the interwar period (a consequence of the extra deflation produced by the falling trend in velocity). This pushes the date of the confidence crisis out to 1959. Model C, based on endogenous velocity, presents a pattern similar to model A. Model D, the post-World War I resumption scenario, reveals that dollar holdings would have surpassed the U.S. monetary gold stock in 1955, as in model A. In our final exercise in sensitivity analysis, we posit that European countries postponed resumption until 1953 to better cope with the dollar shortage. This assumes that the inflow of Marshall aid would have sufficed at that point to provide the liquidity needed to start up the system. We label these simulations model F. Figure 12.7 shows that the world would have immediately

428

Michael D. Bordo and Barry Eichengreen

entered the zone in which U.S. foreign monetary liabilities exceeded U.S. gold reserves.45 One can imagine four possible scenarios once the world entered the “crisis zone” (in the words of Kenen 1960) where foreign holdings of dollars exceeded the US. monetary gold stock. First, the gold exchange standard could have unraveled into a pure gold standard, as countries unwilling to shift to a pure gold exchange standard followed deflationary policies to restore their original gold reserves. We think this scenario is implausible. Just as the decline in the share of foreign exchange in global reserves from 37 percent to 11 percent between 1928 and 1931 aggravated the interwar problem of deflation and depression, ultimately leading countries to abandon the system, it is unlikely that they would have had more stomach for a deflationary crisis in the 1950s. Alternatively, in the face of an unraveling gold exchange standard, the major countries might have negotiated something like the Bretton Woods agreement. This would have involved creation of an institution like the IMF to provide a reserve asset as a substitute for dollars. Whether it would also have encouraged the use of capital controls and the adjustable peg is questionable. This would likely have depended on whether our hypothetical gold exchange standard, when it was operating, induced asymmetric adjustment and destabilizing capital flows as in the 1930s, or smooth adjustment and stabilizing capital movements as in the pre-1914 period. This in turn would reflect the credibility of commitment to gold convertibility. But we are skeptical that an IMF would have been established absent the events of the 1930s but present the perception of events of the 1920s-the perception that floating exchange rates led to destabilizing capital flows and the deflationary consequences of Britain’s return to gold at the original parity-since in our hypothetical gold exchange standard world the perceived problems associated with a gold exchange standard would not have materialized. Finally, the system could have been transformed into a pure dollar standard. Some authors (e.g., McKinnon 1969; Meltzer 1991) posit that had the United States followed stable monetary policies under Bretton Woods-had it geared policy to the maintenance of price stability-the gold-dollar standard would have evolved into a pure dollar standard that could have lasted indefinitely. There would have been no “Triffin crisis” because countries other than the United States would have been willing to hold unlimited quantities of dollars. An examination of the composition of countries’ gold reserves during the Bretton Woods period in fact suggests that most advanced countries, with the principal exceptions of Germany, Italy, and Japan, in fact strongly preferred to 45. This is not surprising, since between 1950 and 1953 U.S. gold reserves had already begun their precipitous decline and, combined with the Marshall Plan aid and the Korean War inflationinduced balance-of-payments deficits, official outstanding dollar liabilities would have exceeded the hypothetical monetary gold stock.

429

The Great Depression and the International Monetary System

hold gold throughout the 1950s and 1960s (see Kenen 1963; IMF 1972). To the extent that the preference for gold over interest-earningdollar-denominated assets reflected memories of capital losses due to the devaluations of the 1930s (and not simply the monetary policies the United States followed in the 1960s), one could argue that, in our scenario of a preserved and restored gold exchange standard, this preference would have been less; absent the inflationary shocks of the Vietnam War and the shift to expansionary monetary policy, the system might have prevailed considerably longer than we suggest above.&The counterargument, which we find compelling, is that the preference of some countries-notably France-for gold over foreign exchange reserves predated the Great Depression. France had begun converting its gold into foreign exchange as early as May 1927, more than two years before the onset of the slump (Clarke 1967). The Bank of France relied heavily on reserve inflows and outflows for balance-of-payments adjustment even before World War I (Bloomfield 1959), and the country’s experience with high inflation in the first half of the 1920s reinforced French officials’ commitment to the operation of a pure gold standard. It is unlikely that, in the absence of the depression, French policymakers would have been more willing to hold foreign exchange re~erves.4~ Thus, as U.S. foreign monetary liabilities began to rise, countries like France with the strongest preference for gold over foreign exchange would have presented the liabilities to the Federal Reserve for conversion (as they in fact did in the mid-1960s). This would have created a problem of collective action: other countries might have been willing to hold U.S. monetary liabilities so long as the central banks and governments of countries like France did the same, but once the latter showed a desire to convert their foreign exchange to gold, all the governments concerned had an incentive to get out of dollars before U.S. gold reserves were exhausted. Although the precise timing of the crisis remains difficult to pin down, the final denouement is the same. Following the collapse of the gold exchange standard, it is likely the world would have moved toward the managed float with capital mobility that we have today. Rather than attempt to create a world central bank and a new international monetary system, or to put up with a dollar standard, the major nations would have preferred cutting the link with gold and following independent financial policies. Given the problems associated with the above three options, we believe that this one would have been most likely. 46. Indeed, in a world without the Great Depression and the aversion to deflation it spawned, the United States, as center country of the gold exchange standard, may have followed the type of monetary policy that Britain did when it was the center country of the classical gold standard-a policy of maintaining convertibility at any price. 47. One can argue that the depth of the depression in France, which was itself perceived to have resulted from commitment to the maintenance of the gold standard (Mow5 1991,274-77), worked to weaken the country’s preference for holding gold, which would therefore have been even stronger in the absence of the slump.

430

Michael D. Bordo and Barry Eichengreen

12.4 Implications for Growth and Resource Allocation after World War I1 Having established that the interwar gold standard would have persisted for the duration of the 1930s and been resurrected after World War 11, we now consider the implications of this arrangement for growth and resource allocation. Perhaps the most important implication of the postwar international monetary system for resource allocation followed from the prevalence of capital controls, which were pervasive after the war but would have been absent under our counterfactual. 12.4.1 The Problem We assume that controls on capital and current account transactions would have been absent following a relatively limited transitional period on the grounds that these were incompatible with the operation of the gold standard. Freedom to convert currency and token coin, however obtained, into gold at the statutory rate and to import and export that gold was a keystone of the gold standard The question then becomes how the post-World War I1 world would have evolved in the absence of controls. Here it is important to distinguish between the early and not-so-early postwar periods. One reason is that data are more partial and less complete for the early postwar years. Only from the mid-1960s does the IMF begin providing quantitative indicators of the incidence of capital controls; comparable indices for earlier years (of our own construction) may be less reliable. Even for the advanced industrial countries, macroeconomic data for the immediate postwar years are incomplete. Data gaps are an even greater problem, of course, for developing countries, many of which did not exist as independent states before the 1960s. While it unambiguously follows from our counterfactual that the early postwar period-say through the 1950s-would have been free of capital controls, it is not so clear how to characterize the subsequent counterfactual regime. Had the reconstructed gold exchange standard collapsed in, say, 1960 and countries accepted greater exchange rate flexibility, would that flexibility have been accompanied by the imposition of controls, as in the 1930s, or by the mainte48. To be sure, central banks and governments used a variety of measures that resembled capital controls. They used the gold devices-paying out clipped and worn coin, or accepting gold imports only at inland cities rather than at central bank offices located close to a port-as a way of discouraging reserve inflows and outflows and mimicking some of the effects of controls. Even Britain, that most faithful adherent to the gold standard, had embargoed foreign lending in the second half of the 1920s. In the early 1930sthe United States, under the provisions of the Johnson Act, had prohibited lending to countries in default on their outstanding debts. But notwithstanding these exceptions, the essence of the gold standard remained free international capital mobility. Most of the restrictions on capital flows imposed in the 1930s were adopted in response to the collapse of the international monetary and financial system, an event that we assume away in our counterfactual.

431

The Great Depression and the International Monetary System

nance of open capital markets? For reasons given in section 12.3, we assume the latter and thus contrast the impact of the actual regime of current and capital account restrictions with a counterfactual of open capital markets. The obvious hypothesis is that the absence of restrictions on international capital flows would have made for a more efficient allocation of resources and, ultimately, faster economic growth. Capital would have flowed from countries where it was abundant to those where it was scarce. The stimulus to investment and growth would have been most apparent in those countries on the receiving end, but a more efficient allocation of funds could also have redounded favorably on the creditor countries, insofar as they received a higher return on their investments and felt the favorable repercussions of higher growth worldwide. Financial repression that distorted the intersectoral allocation of resources and depressed domestic savings rates in the countries in which it was practiced would not have been possible in the face of open international financial markets. Counterarguments include (1) that capital controls were never entirely effective and therefore that one should not expect to discern a large impact on economic outcomes and (2) that policies of industrial targeting, in East Asia for example, helped to solve coordination problems and internalize externalities that would have held back the growth process otherwise and that these interventions would not have been possible in the absence of controls. The effect of controls on growth and resource allocation is ultimately an empirical question. 12.4.2 The Literature The literature contains two notable attempts to answer this question. Alesina, Grilli, and Milesi-Ferretti (1994) consider the experience of 20 industrial countries since 1950, relating a dummy variable for the presence or absence of capital controls constructed from the IMF’s Annual Report on Exchange Arrangements and Exchange Restrictions to the accumulation of public debt, the rate of inflation, the real interest rate, and the GDP growth rate. They find that countries with controls accumulate more public debt, have lower real interest rates, and run higher inflation rates. This is consistent with the view that capital controls, by bottling up international financial flows, facilitate collection of the inflation tax and reduce the costs of servicing the public debt. The implications for the rate of GDP growth are unclear: the authors find only a small and statistically insignificant impact of controls on the growth of GDP per capita. Grilli and Milesi-Ferretti (1995) extend the analysis to a panel of 61 industrial and developing economies. They distinguish three measures of controls: restrictions on payments for capital transactions (as in Alesina et al. 1994); separate exchange rates for capital transactions or invisibles (a proxy for multiple currency practices, which theory suggests should have much the same effect as capital controls); and restrictions on payments for current transactions (which are often used in the attempt to evade restriction on capital transac-

432

Michael D. Bordo and Barry Eichengreen

ti0ns).4~They too find that the presence of capital controls, current account restrictions, and multiple exchange rate practices are associated with higher rates of inflation and lower real interest rates (where the analysis of real interest rates is limited to the industrial countries). There is some evidence that capital controls are positively associated with growth, while current account restrictions are negatively associated, although neither correlation is robust. In this section we extend and reconsider this evidence. We extend the data back from the mid-1960s to 1959, which is important given the focus of our counterfactual analysis. We consider intermediate variables linking controls to growth, such as the investment rate, the savings rate, and export growth. We control for problems of unobserved heterogeneity. While Grilli and MilesiFerretti test for the endogeneity of capital controls and find that the results are little affected by this correction, our concern is different, namely, that countries with and without controls are not drawn from the same underlying population-that they differ in ways that are difficult to observe. We implement an econometric correction for this problem. And we use principal components to construct a summary measure of the incidence of controls that removes some ambiguity about the effects of alternative measures.

12.4.3 Methodology Governments impose controls when they believe the policy will achieve a desired end (such as reducing inflation or limiting external deficits). Hence, controls tend to be observed where they are likely to have the largest effect. This creates problems of selectivity, in whose presence least squares estimation will deliver coefficient estimates of the average effect of such policies. To obtain unbiased estimates we condition on the propensity score.5oThe effect we seek to measure is the difference in, say, the current account of a particular country as a function of whether it imposes capital controls. While we only have data on the current account balance associated with the policy that was actually in place, we are interested in the counterfactual, in what would have happened had the other policy been followed. If controls were imposed randomly, least squares would suffice. But countries with controls are not identical to countries that shun the policy. In the language of experimental design, we lack a control group that is otherwise identical to the treatment group. If xzis a vector of factors determining the likelihood of capital controls, the joint distribution of x, for countries that have capital controls is different from the joint distribution for countries that do not. We address this problem by using the information in xt to sort countries in our sample by the likelihood of their adopting controls. We form a subgroup 49. As before, they consider the impact on inflation (excluding inflation rates above 80 percent per annum to prevent the results from being dominated by outliers), real interest rates, and per capita GDP growth. 50. For a general treatment of propensity score methods see Rosenbaum and Rubin (1983, 1984). For a related example in the economics literature, see Angrist (1995).

433

The Great Depression and the International Monetary System

with a similar likelihood of imposing controls-with a similar propensity score-and compare the current accounts of countries with and without controls within that subgroup, thereby minimizing the bias owing to heterogeneit~.~' In a small sample we are unlikely to have countries with identical propensity scores. Instead, we group the sample into quintiles, where the first group comprises the 20 percent least likely to have controls, the last group the 20 percent most likely to have them. Within each of the five groups we then compare the current accounts of countries with and without The propensity of a country-year pair to have imposed any form of capital control (restrictions on payments for capital transactions, separate exchange rates for capital transactions or invisibles, and restrictions on payments for current transactions) is modeled as a function of the number of executive changes in the country from 1950 to 1982, the legal independence of the central bank, the average turnover rate of central bankers per year, a dummy indicating a majority coalition government is in power, a dummy indicating a leftof-center government is in power, a dummy indicating a fixed or managed exchange rate arrangement is in place, the log of real per capita GDP, and the following variables lagged one year: the ratio of the current account deficit to GDP, the ratio of government consumption to GDP, and the ratio of trade (the sum of imports and exports) to GDP. For each dependent variable six regressions are run. The first two include only industrialized countries, for which we have relatively few missing observations. The first of these regressions has no controls for time or country effects, while the second includes year controls. The next four regressions em5 1. We thus implement a logit model designed to predict whether capital controls are in place as a function of the components of x, and use the coefficients from this model to derive the fitted probability of a country's having capital controls in a given year. This fitted probability is the propensity score. The proposition underlying the methodology is that conditional on the estimated propensity score the joint distribution of x, for countries with capital controls is identical to the joint distribution for countries without them. We would like to take a group of countries with the same values of the x, vector and compare the current accounts of countries that imposed capital controls with the current accounts of those that did not. More generally, we can compare countries with the same value of some function of x,, say b(x,).In the present context, b(x,)is the propensity score estimated from the logit model. Comparing the current accounts of countries with the same value of b(xJ should then deliver unbiased estimates of the effect of capital controls. 52. This comparison is undertaken using a standard least squares model with the current account (and other measures of economic policy and performance) as the dependent variable and the independent variables employedby Grilli and Milesi-Ferretti (1995). Since we are comparing countries with similar but not necessarily equal propensity scores, not all of the bias will be eliminated. However, Rosenbaum and Rubin (1984) argue that this method is likely to remove some 90 percent of the bias due to heterogeneity.Finally, to find the average effect of capital controls in the population, we take the averageof the five estimated coefficientsfrom the quintileregressions. In computing the standard error of this average, we assume the propensity score for each observation is measured without error, placing all observations in the correct quintile. Assuming that the coefficients from the five quintiles are independent of each other, we can compute the standard error of the weighted average of coefficients using the simple formula for the variance of a weighted sum of independent random variables.

434

Michael D. Bordo and Barry Eichengreen

ploy data for all countries in the sample, industrial and developing. The first has no time or year controls, the second has year controls, the third has country controls, and the last has country controls and includes the lagged value of the dependent variable as an additional regressor.53The range of years for each regression is 1959 to 1989.54Each of these regressions includes the three alternative measures of capital controls. In addition, however, we use principal component methods to construct a single measure of controls from our three dummy variables. Each of the regressions includes as independent variables, in addition to the measures of capital controls, the log of the level of real per capita GDP in 1959, the ratio of government consumption to GDP, the ratio of trade to GDP, a dummy variable for nondemocracies, the average turnover rate of central bankers, and an index of the legal independence of the central bank.55 12.4.4 Results The full set of results is available from the authors on request. Here we summarize our principal findings. We concentrate on the regressions with no country or year controls (the reader should assume this except where stated to the contrary), although we discuss the other results where the estimates are particularly sensitive to the inclusion or exclusion of controls. The results in table 12.5 suggest that countries with controls run significantly higher inflation. This is consistent with the notion that such countries limit capital mobility as a way of making more active use of the inflation tax.56 For the industrial countries, most of the explanatory power appears to reside with the measure of capital account restrictions; for the developing countries, multiple currency practices are particularly important. According to the ordinary least squares (OLS) regressions on the full sample, both capital account restrictions and multiple currency practices affect inflation for the sample of both industrial and developing countries, but the correction for heterogeneity eliminates the first of these effects. While our heterogeneity correction does not much modify the overall conclusions, there are some interesting differences from the OLS regressions. In 53.This last specification is designed to mitigate autocorrelationproblems. 54. While we would have liked to extend the data set back to 1953 or 1955 to coincide with the end of the counterfactualtransition period we describe above, this turns out to be difficult, as data constraints on our exercise become increasingly binding the further back we go. The year 1959 is quite close to the beginning of our counterfactual period of capital account convertibility, though. 55. To conserve space, we do not report the coefficients on these variables, which are broadly similar to those reported and discussed by Grilli and Milesi-Ferretti (1995). The regressions for the real interest rate include also the average number of executive changes per year, a dummy variable indicating that a left-wing government is in power, a dummy variable indicating that a coalition government is in power, and the lagged value ofthe ratio of the government budget deficit to GDP. 56. Grilli and Milesi-Ferretti (1995) note the possibility that capital controls, by enlarging the inflation tax base, may cause the optimal rate of inflation to fall, but they find the same positive coefficient on a more limited sample.

Table 12.5

Effect of Capital Controls on Inflation: Ordinary Least Squares and Heterogeneity-Corrected Estimates

4

5

Average of Subsample Regressions

2.292 (3.258) -0.618 (1.291) -3.795 (2.144) 1.335 (2.484)

2.532 (2.339) - 1.487 (1.827) 0.133 (1.915) 1.891 (2.084)

2.602 (0.906) -0.213 (0.772) -0.625 (0.854) 2.43 1 (0.759)

-5.827 (2.925) - 1.882 (2.299) 2.544 (2.442) -4.595 (2.185)

5.337 (3.052) -5.14 (2.448) 7.174 (1.926) 2.538 (1.877)

0.987 (1.026) 0.802 (0.909) 4.730 (0.865) 2.430 (0.727)

Subsample Regressions by Quintile Variable Capital account restrictions Current account restrictions Multiple currency practices Principal component

S.E. R2

N

Full Sample 2.344 (0.519) 0.187 (0.595) -0.059 (0.654) 2.151 (0.455)

t-Statistic

1

2

3

A. Inflation Rate, Industrial Countries Only 3.504 1.977 2.705 4.516 ( 1.672) (1.008) (0.778) 1.442 -0.188 0.314 (1.371) (1.208) 0.013 1.148 -0.090 (1.471) (1.120) 4.727 3.009 3.849 2.072 (1.462) (1.107) (0.722)

r-Statistic 2.873 -0.295 -0.732 3.204

5.044 0.148 581

B. Inflation Rate, All Countries Capital account restrictions Current account restrictions Multiple currency practices Principal component

S.E R2

N

2.008 (0.855) 0.224 (0.872) 5.329 (0.811) 2.802 (0.623)

2.349 0.257 6.57 1 4.498

0.996 (1.155) 7.445 (1.818) 6.629 (1.984) 5.114 (0.998)

5.670 (1.662) 0.501 (1.786) 2.932 (1.670) 5.217 (1.286)

11.277 0.265 1,244

Notes: S.E. = standard error of regression. Numbers in parentheses are standard errors.

-1.243 (2.090) 3.092 (1.694) 4.371 (1.519) 3.874 (1.508)

0.961 0.883 5.468 3.341

436

Michael D. Bordo and Barry Eichengreen

the industrial countries, capital account restrictions are significantly associated with inflation only for the first and second quintiles (those countries least likely to have controls). An interpretation is that in industrial countries most likely to impose controls, market participants discovered means of evading them and that only where controls were unexpected did they have major effects. These results are likely to be important for economic activity if they affect relative prices. Consider, for example, the (ex post) real interest rate on government bonds. Like Grilli and Milesi-Ferretti (1995), we find that capital controls are associated with significantly lower real interest rates, as if countries used controls to bottle up financial capital domestically (table 12.6). In the equations not corrected for heterogeneity, all three measures of controls border on statistical significance at conventional levels, and all enter negati~ely.~’ Insofar as controls are associated with significantly lower real interest rates, we might expect investment rates to be higher. This is what we find for the industrial countries (table 12.7), although different kinds of controls seem to have different effects. Capital account restrictions are associated with higher investment rates, while current account restrictions and multiple currency practices have the opposite effect.58For the full sample including developing countries, the signs are the same but the net effect is negative, since the depressing impact of current account restrictions and multiple currency practices is larger. An interpretation is that capital account restrictions worked to channel domestic savings into domestic investment, but current account restrictions increased the shadow price of imported capital goods and reduced the marginal efficiency of investment. It makes sense that these last effects should be largest in developing countries. We can explore this hypothesis further by looking at the determinants of the current account as a share of GDP (since domestic saving is the sum of investment, considered above, plus the current account). For the industrial countries, the regression with no country or year dummies suggests that all three measures of capital controls have a statistically significant effect on the current account (table 12.8). Capital account restrictions enter with the largest coefficient and negative sign (in the OLS regression), and the summary measure of capital controls suggests that these worked to depress the current account surplus and, by implication, domestic saving. The OLS results for the full sample similarly indicate a negative effect of capital account restrictions, but significance is eliminated by the correction for heterogeneity. While the evidence is 57. However, the heterogeneity correction eliminates the effect of current account restrictions and multiple currency practices in the equations for industrial countries, and that for current account restrictions in the equation for the full sample of industrial and developing economies. This is notable insofar as Grilli and Milesi-Ferretti (1995) lay considerable stress on the current account restrictions variable, which they interpret as a proxy for the extent of financial repression and the intensity of capital conuols. 58. The effect of the latter is eliminated by the correction for heterogeneity.

Table 12.6

Effect of Capital Controls on Real Interest Rates: Ordinary Least Squares and Heterogeneity-Corrected Estimates Subsample Regressions by Quintile

Variable

Full Sample

t-Statistic

4

5

Average of Subsample Regressions

-1.322 (0.958) 1.577 (0.941) 2.885 (0.878) -1.214 (0.931)

-8.971 (2.692) -0.606 (0.8 13) 0.468 (1.086) -4.904 (1.557)

-4.431 (2.273) 0.1192 (1.277) -0.040 (1.518) -3.364 (1.512)

-3.530 (0.763) 0.549 (0.517) - 1.033 (0.562) -2.580 (0.524)

-4.625

B. Real Interest Rate, All Countries -1.985 -0.943 0.948 (0.745) (0.757) (1.092) -0.238 -0.481 (0.915) (0.803) -2.078 -1.871 (0.866) (0.865) -2.232 - 1.849 - 1.732 (0.831) (0.649) (0.923)

-8.132 (4.780) -0.9020 (1.816) -9.847 (2.537) 2.324 (3.448)

-6.878 (2.590) - 1.289 (1.362) -2.142 (1.712) -5.232 (1.743)

-3.777 (1.129) -0.728

-3.345

1

2

3

t-Statistic

A. Real Interest Rate, Industrial Countries Capital account restrictions Current account restrictions Multiple currency practices Principal component

S.E. R2

N Capital account restrictions Current account restrictions Multiple current practices Principal component

S.E. R2

N

-2.25 1 (0.359) - 1.260 (0.435) -0.728 (0.424) -2.619 (0.329)

-6.270

-2.280 (0.924)

-2.897 -1.717 -7.960

-2.445 (0.981)

-0.645 (0.614) 0.03 1 (1.046) - 1.673 (0.898) -0.975 (0.575)

1.062 - 1.836

-4.922

3.398 1.582 545 -2.659 (0.529) - 1.267 (0.560) -2.583 (0.568) -3.175 (0.452)

-5.026 -2.263 -4.548 -7.024

5.159 0.134 63 1

Nores: S.E. = standard error of regression. Numbers in parentheses are standard errors.

(0.644)

-3.985 (0.824) -2.674 (0.822)

-1.130 -4.835 -3.253

Table 12.7

Effect of Capital Controls on Investment: Ordinary Least Squares and Heterogeneity-Corrected Estimates Subsample Regressions by Quintile

Full Variable Capital account restrictions Current account restrictions Multiple currency practices Principal component

S.E. R2

N

Capital account restrictions Current account restrictions Multiple current practices Principal component

S.E. R2 N

Sample

3.837 (0.394) - 1.485 (0.450) -1.143 (0.496) 2.548 (0.361)

t-Statistic

1

2

3

4

5

Average of Subsample Regressions

A. Investment as a Fraction of GDP, Industrial Countries 9.739 2.018 3.711 1.604 3.062 (0.948) (0.811) (1.096) (2.348) -3.300 -1.003 -1.731 -3.313 (1.235) (0.866) (0.953) -2.304 -2.313 -0.511 -0.483 (1.170) (0.959) (1.576) 7.058 2.223 2.584 -0.574 -2.632 (1.034) (0.775) (1.001) (1.837)

8.533 (1.897) 0.995 (1.424) 0.082 (1.457) 7.792 ( 1.730)

3.786 (0.689) -1.761 (0.571) -0.806 (0.656) 1.879 (0.601)

B. Investment as a Fraction of GDP, All Countries 1.215 6.131 2.677 4.147 3.952 (1:464) (0.731) (0.895) (1.089) -2.585 -6.159 -1.291 -7.674 -4.924 (1.163) (1.117) (0.892) (0.943) -0.452 -4.650 -1.196 -1.360 -0.580 (1.094) (0.898) (1.110) (0.808) -0.383 -2.792 -0.216 0.766 -2.441 (0.941) (0.676) (0.783) (0.603)

6.857 (2.079) -6.939 (1.785) -6.255 (1.173) -3.400 ( 1.265)

3.770 (0.600) -4.380 (0.547) - 1.969 (0.459) -1.135 (0.396)

t-Statistic

5.494 - 3.085

- 1.228 3.124

3.819 0.443 585 3.041 (0.496) -3.814 (0.497) -2.060 (0.443) -0.980 (0.351) 6.67 0.365 1,419

Notes: S.E. = standard error of regression. Numbers in parentheses are standard errors.

6.287 -8.011 -4.290 -2.866

Table 12.8

Effect of Capital Controls on Current Account: Ordinary Least Squares and Heterogeneity-CorrectedEstimates Subsample Regressions by Quintile Full Sample

Variable Capital account restrictions Current account restrictions Multiple currency practices Principal component

S.E.

A. Current Account as a Fraction of GDP, Industrial Countries -6.410 -1.390 -1.153 0.611 -1.474 (0.692) (0.447) (0.789) (1.532) 0.884 2.113 1.114 -0.461 (0.694) (0.647) (0.607) 3.289 1.510 0.720 1.462 (0.643) (0.694) (1.008) -5.452 -1.526 -0.118 1.605 -1.816 (0.753) (0.449) (0.702) (1.159)

-2.531 (1.369) 1.507 (1.069) 2.470 (1.232) -3.956 (1.281)

-1.187 (0.470) 1.068 (0.388) 1.541 (0.463) -1.162 (0.412)

B. Current Account as a Fraction of GDP, All Countries -1.917 -1.693 -1.763 -2.290 0.561 (0.564) (0.550) (0.772) (0.866) 1.287 0.408 1.188 1.031 -0.152 (0.866) (0.586) (0.615) (0.654) -0.147 3.848 0.785 0.721 -1.195 (0.884) (0.532) (0.577) (0.665) -0.717 -0.639 -0.333 -0.222 -0.009 (0.493) (0.422) (0.512) (0.617)

5.338 (1.669) -1.316 (1.274) -2.548 (1.044) 0.948 (0.938)

0.03 1 (0.436) 0.232 (0.375) 0.322 (0.342) -0.051 (0.279)

t-Statistic

1

2

3

4

r-Statistic -2.526 2.752 3.328 -2.820

2.704 0.161 581

R= N

Capital account restrictions Current account restrictions Multiple current practices Principal component

S.E.

-0.673 (0.351) 0.453 (0.352) -0.048 (0.326) -0.180 (0.251) 4.730 0.100 1,323

R2

N

Notes: S.E.

- 1.782 (0.278) 0.282 (0.319) 1.151 (0.350) - 1.363 (0.250)

5

Average of Subsample Regressions

=

standard error of regression. Numbers in parentheses are standard errors.

0.070 0.618 0.941 -0.183

440

Michael D. Bordo and Barry Eichengreen

not robust, it is consistent with the interpretation of the effect of controls on savings, investment, and capital formation offered above. Exports were the other engine of growth according to most studies of the post-World War TI growth process.59We find little evidence that the growth of exports (in constant dollars) was affected by capital controls (table 12.9). The basic regression for the industrial countries, with and without corrections for heterogeneity, suggests a small positive effect of multiple currency practices, but the measures of capital controls are consistently insignificant as a group, and the addition of year and country controls eliminates even the multiple currency regime effect. The same is true for the full sample including the developing countries: while there are some sign reversals on the individual measures of controls, our three capital control proxies as a group have no significant impact on export growth. Our regressions for the industrial countries show little effect on economic growth (table 12.10). The basic regression with no year or country controls suggests that capital account restrictions had a small, statistically significant, positive effect on growth. The quintile regressions suggest that this effect is concentrated mainly in countries relatively unlikely to have controls. But the three measures of controls are insignificant as a group, and the addition of year effects eliminates even that impact. Interestingly, the effects on growth of our three measures of controls have the opposite sign when the sample includes also developing countries. This result is driven by a larger (negative) effect of multiple currency practices, which-the quintile regressions correcting for heterogeneity suggest-significantly depress growth in these countries. On balance, our three measures of controls just border on significance at conventional confidence levels when the industrial and developing countries are combined. 12.4.5 Summary of Results Our results tell a consistent story. They suggest that controls had a significant effect on international financial flows and were associated with higher inflation and lower ex post real interest rates. Lower interest rates were associated with higher domestic investment but discouraged saving. Additional investment did not translate into faster economic growth, however, due perhaps to the higher cost of imported capital goods associated with current account restrictions and a lower marginal efficiency of investment. There is some evidence that current account restrictions and multiple currency practices tended to depress growth in developing countries, although the effect is marginal statistically and economically. These results suggest that the greater international capital mobility associ59. These studies tend to disagree about whether investment and exports were dual engines of growth (as in Eichengreen's 1996 model of the advanced industrial countries), whether rising investment induced the increase in exports (as in Rodrik's 1995 model of the East Asian growth miracle), or whether exports induced the rise in investment (as in models of export-led growth).

Effect of Capital Controls on Growth of Exports: Ordinary Least Squares and Heterogeneity-Corrected Estimates

Table 12.9

Subsample Regressions by Quintile Full Sample

Variable Capital account restrictions Current account restrictions Multiple currency practices Principal component

5

A. Growth Rate of Exports, Industrial Countries 2.811 -0.043 0.039 -0.013 (0.029) (0.017) (0.031) 0.004 -1.604 -0.034 (0.027) (0.025) 2.873 0.049 0.027 (0.025) (0.027) 1.956 -0.047 0.038 0.006 (0.032) (0.016) (0.027)

0.062 (0.061) -0.048 (0.024) 0.033 (0.040) -0.042 (0.047)

0.053 (0.046) -0.001 (0.036) 0.045 (0.038) 0.019 (0.041)

0.020 (0.018) -0.020 (0.014) 0.039 (0.017) -0.005 (0.015)

B. Growth Rate of Exports, All Countries 0.004 0.028 0.029 (0.017) (0.021) (0.032) - 1.746 0.004 -0.036 -0.027 (0.026) (0.022) (0.026) -0.055 0.012 -0.008 -2.338 (0.026) (0.020) (0.024) -1.321 -0.004 0.001 -0.008 (0.014) (0.016) (0.021)

0.007 (0.043) -0.034 (0.033) -0.008 (0.033) -0.026 (0.03 1)

0.089 (0.084) -0.022 (0.064) -0.136 (0.053) -0.026 (0.047)

0.03 1 (0.021) -0.023 (0.017) -0.139 (0.015) -0.013 (0.013)

&Statistic

1

2

3

t-Statistic

1.101 -1.384 2.325 -0.339

0.090 0.049 581

S.E.

R2 N

Capital account restrictions Current account restrictions Multiple current practices Principal component

0.017 (0.015) -0.026 (0.015) -0.032 (0.014) -0.014 (0.01 1)

1.133

0.199 0.039 1,324

S.E.

R2 N

Notes: S.E.

0.026 (0.009) -0.017 (0.011) 0.034 (0.012) 0.016 (0.008)

4

Average of Subsample Regressions

=

standard error of regression. Numbers in parentheses are standard errors.

1.504 1.376 -2.617 -0.990

Table 12.10

Effect of Capital Controls on Economic Growth: Ordinary Least Squares and Heterogeneity-CorrectedEstimates Subsample Regressions by Quintile

Variable Capital account restrictions Current account restrictions Multiple currency practices Principal component

Full Sample

0.007 (0.003) -0.009 (0.003) 0.004 (0.004) 0.003 (0.003)

S.E. R’ N

0.028 0.078 581

Capital account restrictions

0.007 (0.004)

Current account restrictions Multiple current practices Principal component

S.E R2

N

-0.011

(0.004) -0.009 (0.004) -0.005 (0.003)

&Statistic

1

2

3

4

5

Average of Subsample Regressions

t-Statistic

A. Growth Rate of Real Per Capita GDP, Industrial Countries 2.273 0.006 0.013 0.019 0.034 (0.009) (0.006) (0.008) (0.022) -2.576 -0.001 0.008 -0.018 (0.009) (0.007) (0.009) 1.226 0.010 -0.001 -0.002 (0.009) (0.007) (0.014) 0.870 0.006 0.014 0.019 -0.007 (0.009) (0.006) (0.007) (0.017)

0.001 (0.015) -0.008 (0.011) 0.008 (0.012) -0.008 (0.013)

-0.005 (0.005) 0.004 (0.005) 0.005 (0.005)

-.1.068

B. Growth Rate of Real Per Capita GDP, All Countries 0.004 0.015 1.721 0.005 0.013 (0.012) (0.010) (0.006) (0.006) -0.019 -0.006 -0.007 -2.698 -0.012 (0.010) (0.009) (0.106) (0.006) O.OO0 -0.028 -0.040 -2.598 -0.032 (0.009) (0.009) (0.008) (0.006) - 1.776 -0.009 0.002 -0.006 -0.004 (0.008) (0.007) (0.005) (0.004)

-0.021 (0.017) -0.012 (0.015) 0.0 17 (0.010) -0.013 (0.009)

0.003 (0.005) -0.01 1 (0.005) -0.010 (0.004) -0.006 (0.003)

0.648

0.052 0.045 1,324

Notes; S.E. = standard error of regression. Numbers in parentheses are standard errors

0.014 (0.N)

2.441

0.737 0.941

-2.379 -2.499 - 1.955

443

The Great Depression and the International Monetary System

ated with a post-World War I1 gold exchange standard would not have much affected the rate of economic growth. The literature features two interpretations of the connections between capital mobility and growth. In one, the postwar growth process (especially in developing countries) had to be initiated by government interventions to boost savings and investment rates and solve coordination problems. In the other, such interventions depressed the efficiency with which savings and investment were deployed. Our results suggest that the two effects canceled out in the industrial countries, while the latter may have depressed the rate of economic growth on balance in the developing world.

12.5 Implications for Institution Building and International Cooperation A final implicationof the depression for the development of the international monetary system concerns the institutions of international monetary cooperation, specifically the IMF. The design of the Bretton Woods institutions was powerfully shaped by the experience of the 1930s: Keynes, White, and their colleagues sought a structure that would prevent any recurrence of the international monetary instability of that decade. In addition to providing for capital controls and parity adjustments in the event of a fundamental disequilibrium, they established the IMF to oversee the operation of their new international system. Had the interwar gold standard operated smoothly up to the outbreak of World War 11, there would have been no comparable impetus to establish Bretton Woods-like institutions thereafter. The post-World War I1 gold exchange standard would have operated without an IMF to provide exceptional liquidity, apply policy conditionality, and encourage international cooperation. The question is how much difference the Fund’s absence would have made. One conceivable answer is “not very much.” The interwar gold exchange standard operated without the support of a comparable institution, and our counterfactual analysis is predicated on the assumption that this system functioned smoothly throughout the 1920s and 1930s without the supervision of an international monetary institution. If so, why should the 1950shave been different? They would have been different, first of all, because of the problem of collective action posed by the Tkiffin dilemma. As we saw in section 12.3, the supply of monetary gold was sufficient after World War I for the interwar gold standard to operate through the end of the 1930s, assuming the absence of the Great Depression,without the official foreign liabilities of the principal reserve currency country, the United States, ever exceeding America’s holdings of monetary gold. No problem of collective action would have arisen in which other countries had to agree not to cash in their dollar reserves, since there was no question of the convertibility of those dollars into gold. In the 1950s, in contrast, the wartime rise in the price level and the rapid postwar growth of the

444

Michael D. Bordo and Barry Eichengreen

world economy would have so augmented the demand for international reserves that U.S. official foreign liabilities would have quickly come to exceed the country’s gold reserves had a dollar-based gold exchange standard been restored in the aftermath of the war. Countries holding dollars as reserves would have continued to do so willingly only if they were confident of the equal willingness of others. This same problem existed under Bretton Woods, of course, although it brought down that system of pegged but adjustable exchange rates only later. One explanation for the lag, and for the length of the interlude of relative stability that preceded it, is that the IMF provided a partial solution to the problem of collective action. The Fund offered a forum for the exchange of information and opinion among governments; as in models of “cheap talk,” this helped to facilitate cooperation. The Fund assembled information on central bank policies and monitored international financial transactions, limiting the incentive for governments to renege on their agreement to cooperate by increasing the likelihood of rapid detection. There is reason to think that the Bretton Woods system would have been brought down even earlier by the Triffin dilemma had there been no IMF. It follows that the absence of an IMF-like entity in our postwar gold exchange standard scenario would have brought Triffin’s chickens home to roost even sooner. There are also other, related grounds for arguing that the absence of the Fund would have weakened the operation of our hypothetical postwar gold exchange standard. The IMF played a role under Bretton Woods in lending to countries experiencing exceptional balance-of-payments difficulties. Its loans helped such countries accommodate temporary shocks without having to abandon their Bretton Woods pegs. No comparable institution existed in the interwar period, of course, and when the 1931 financial crisis struck the gold exchange standard, countries had to negotiate foreign financial assistance government to government, on an ad hoc, bilateral basis. The absence of established procedures for the extension of such loans and the inevitable politicization of the intergovernmental process hindered extension of the requisite loans. France hesitated to assist Germany, for example, because Berlin had reportedly concluded a customs union agreement with Austria in violation of the Versailles Treaty and because of the belief that Germany was secretly rearming. One can imagine that political and diplomatic obstacles would have similarly arisen in the 1950sand 1960shad balance-of-payments loans had to be negotiated on an ad hoc basis directly between governments. It is thus likely that the extension of balance-of-payments support for post-World War I1 currencies in distress would have been less in the absence of the Fund. The absence of IMF lending would have been felt most strongly by smaller countries of less consequence for the stability of the international monetary system-by the countries and currencies of least concern to the major industrial powers who would have been least likely to receive exceptional bilateral support. One can think of instances to the contrary where the advanced industrial countries have provided financial support for the currency of a smaller, less developed economy: the US.loan to Mexico and the French loan to the

445

The Great Depression and the International Monetary System

Communaut6 Financikre Africaine (CFA) countries in 1995 are recent cases in point. But in both instances the larger country was galvanized into action by exceptional circumstances: in the case of the United States by Mexico’s geographical proximity, by the need to support the North American Free Trade Agreement and by Mexico’s status as a benchmark for emerging market investors; in the case of the CFA by France’s long-standing colonial ties with the region and Paris’s foreign policy ambitions. In many other instances where the Fund has provided balance-of-payments support, the advanced industrial countries presumably would not have bothered. Hence, the absence of an IMFlike entity under our no-depression counterfactual would have had negative consequences for the external financial support that such countries could expect. The Fund has encouraged adjustment in developing countries through its policy conditionality as well, making the negotiation of a mutually acceptable adjustment package a prerequisite for the disbursal of finance. Dominguez (1993) argues that the IMF has functioned as a commitment technology, using its funds and stature to promote course corrections and lock in policy reform. Rodrik (1996) and Gilbert et al. (1996) similarly argue that multilateral lending by the IMF and the World Bank has carried out a function that intergovernmental lending could not. Not only did the IMF and the World Bank provide monitoring and signaling services and threaten sanctions if policy went awry, but the markets had reason to believe that, because the Ih4F’s own purse was at risk, it would take its information-gathering and monitoring functions seriously. And because the policy conditionality that served as a commitment device was applied by a multilateral agency one step removed from national capitals, this intervention in national affairs was more politically palatable. One can argue that policy reform in developing countries would have proceeded more slowly after World War I1 in the absence of the Fund. Initiatives to coordinate macroeconomicpolicies internationallywere hardly pervasive under Bretton Woods, but it is likely that they would have been even less frequent and less successful after World War I1 had there been no IMF. Policy coordination must overcome costs of assembling and evaluating information, negotiating mutually acceptable course corrections, and monitoring countries’ compliance with the terms of their agreement. An international institution can facilitate this process in a number of ways. Because information is a nonrival good, an international institution can presumably assemble it at reduced cost; the role of the IMF in gathering and publishing balance-ofpayments and government finance statistics can be thought of in this light. Assessing cross-border spillovers of policy can be thought of as a central function of IMF surveillance. The institutionalized exchange of information and views can facilitate the formation of a consensus. The regular meetings of the IMF Executive Board provide precedents and shape agendas, delineating the policy domain that is fair game for discussion. Staff analyses provide terms of reference, statistics provided in background reports serve as focal points directing officials toward pressing policy problems, and the written record and institu-

446

Michael D. Bordo and Barry Eichengreen

tional memory of staff lend continuity to a process that would otherwise be disrupted by changes in government and cabinet composition. The absence of the IMF would thus have dealt a setback to efforts to coordinate policies after World War 11. The failure of interwar monetary and economic conferences at Brussels in 1920, Genoa in 1922, and London in 1933, where attempts were made to coordinate policies without the mediation of an international institution, lends support to this view. It suggests that macroeconomic policy coordination would have been minimal, due to in part to the absence of the Fund, under our postwar gold exchange standard, as it had been in the second part of the 1920s. If no IMF had been created after World War 11, perhaps another, already existing institution would have stepped in to fill the void. The obvious candidate is the Bank for International Settlements (BIS), established in 1930 to oversee the transfer of German reparations. The BIS never played a leading role in the management of the post-World War I1 international economic order. In part this role was usurped by the IMF, but in part the legitimacy of the BIS was damaged by its inability to mount a concerted response to the financial crises of the early 1930s, events that would not have occurred in our counterfactual. Yet the fact that the BIS was involved in the reparations dispute and that it was accused of abetting the Nazis during World War II led a number of Western countries to oppose granting it a more prominent role (and, in the case of the Dutch, to actively plump for its abolition). This makes it unlikely that the BIS would have filled the opening left by the absence of the IMF.

12.6 Conclusion How would the international monetary system have evolved in the absence of the Great Depression? Our conclusion is that the depression interrupted but did not permanently alter the development of international monetary arrangements. As a result of the international monetary instability of the 1930s-the unsatisfactory experience with output instability, hot money flows, exchange rate variability, and beggar-thy-neighbor policies-the depression prompted the construction of a very different post-World War I1 monetary and financial regime characterized by pegged but adjustable exchange rates, highly regulated domestic financial markets, and pervasive controls on international capital flows. The gradual recovery of domestic and international financial transactions from the disruptions of depression and war eventually resulted in the growing porousness of controls, mounting difficulties with operating pegged but adjustable rates, and ultimately the collapse of the Bretton Woods system in 1971. In the absence of the depression, this interlude of pegged but adjustable rates and restrictions on capital mobility would not have occurred; the perception that the interwar gold standard had functioned reasonably smoothly in the 1920s and 1930s would have encouraged the restoration of similar arrangements after World War 11. The postwar international monetary system

447

The Great Depression and the International Monetary System

would have been characterized by very infrequent parity changes and high capital mobility. The world would have experienced mild deflation in the 1930s and after the war, in contrast to the actual roller coaster of sharp deflation and inflation in the 1930s and secular inflation in the 1950s and 1960s. But the same factors that brought down the Bretton Woods system-the failure of the flow supply of gold to match the bouyant growth of the world economy and hence of governments’ demands for international reserves, leading to an overhang of U.S. official foreign liabilities and questions about the convertibility of the dollar into gold-would have brought down this hypothetical postwar gold exchange standard as well, and probably at an earlier date. In the absence of the Great Depression, there would have been little impetus for the creation of an organization like the IMF, and there would have been little institutionalization of international monetary cooperation. Under these conditions, the problems of collective action that had to be solved to successfully navigate the transition from the gold exchange standard to a dollar standard would have been insurmountable.The most likely scenario for subsequent events would have been a transition to freer floating. Somewhat to our own surprise, we conclude that the depression slowed but did not permanently alter the development of the internationalmonetary system; it only delayed the transition to the kind of system with which we live today. How much of a difference this change in timing made for the development of the world economy is difficult to say. The connections between financial arrangements and economic growth are among the most difficult for economists to analyze; it is not surprising that our findings on this question are less than clearcut. Our best guess is that freer capital mobility in the wake of World War I1 would have had little effect on economic growth in the advanced industrial countries, across which capital-labor ratios and productivity did not differ greatly, but that it would have permitted a more efficient allocation of resources in the developing world, accelerating at least slightly the process of economic growth and development there. The Great Depression was a watershed in many respects, as the other chapters in this volume show. But so far as the long-term development of the international monetary system is concerned, it may have made less of a difference than is commonly supposed.

Appendix A Data Sources Part I: Data Sources for Initial Values The first part of this appendix explains how the initial values given in table 12.2 and used in the simulations were created and gives the sources.

448

Michael D. Bordo and Barry Eichengreen

Values for the 1928-38 Simulation Prices: Calculated as a GDP-weighted average for the following countries: Argentina, Australia, Brazil, Chile, Netherlands, Portugal, Spain, Belgium, Canada, Denmark, Finland, France, Norway, Sweden, Switzerland, Gerniany, Greece, Italy, Japan, United Kingdom, and United States. GDP weights, price indexes, and exchange rates versus U.S. dollars are from the Bordo-Schwartz database; see Bordo and Schwartz (1996). Nominal GDP: Same source as prices. Money supply, M2: Same source as prices. Money base: Notes and coins in circulation plus central bank deposits from League of Nations, Monthly Bulletin of Statistics (Geneva, 1932-39) and Mitchell (1992, 1993, 1995). International reserves: “Gold and Foreign Reserves,” in League of Nations, Statistical Yearbook (Geneva, 1926, 1931-32, 1940-41). Central bank gold reserves: Same source as international reserves. Gold production: World gold production from U.S. Gold Commission Report (1982). World monetary gold stock: Same source as gold production. World nonmonetary gold stock: Same source as gold production. Interest rate: U S . short-term interest rate (three months) from BordoSchwartz database; see Bordo and Schwartz (1996). Assumed growth rates in the simulation period. Real GDP: 3.9 percent, from Bordo and Schwartz (1996). Velocity: -2.6 percent, from Bordo and Schwartz (1996). Productivity: 1.8 percent, from Kendrick (1961, tables A-XXII and AXXV).

Assumptions for 1938-50 Assumed growth rates in the simulation period. Monetary gold stock: 2.7 percent, from U S . Gold Commission Report (1982). Productivity: 2.0 percent, from Kendrick (1961, table 3-3).

Values for the 1950-71 Simulation Prices: Calculated as a GDP-weighted average for the following countries: Australia, Belgium, Netherlands, Portugal, Spain, Canada, Denmark, Finland, Switzerland, France, Norway, Sweden, Germany, Greece, Italy, Japan, United Kingdom, and United States. GDP weights, price indexes, and exchange rates versus U.S. dollars are from the Bordo-Schwartz database; see Bordo and Schwartz (1996). Nominal GDP: Same source as prices. Money supply, M2: Same source as prices. International reserves: IMF (1972). Central bank gold reserves: IMF (1972).

449

The Great Depression and the International Monetary System

High-powered money: IMF (1972). IMF quota: IMF (1972). SDRs (1970-71 only): IMF (1972). Gold production: World gold production from U.S. Gold Commission Report (1982). World monetary gold stock Same source as gold production. World nonmonetary gold stock: Same source as gold production. Interest rate: U.S. short-term interest rate (three months). Assumed growth rates in the simulation period. Real GDP: 4.4 percent, from Bordo and Schwartz (1996). Velocity: 0.3 percent, from Bordo and Schwartz (1996). Productivity: 2.2 percent, from Kendrick (1973, table 3-3).

Part 11: Data Sources for the Capital Control Regressions The data on capital and current account restrictions, multiple currency practices, and exchange rate arrangements are from elaborations on IMF, Annual Report on Exchange Rate Arrangements and Exchange Restrictions (various issues). Capital account restrictions are defined as “restrictions on payment on capital transactions.” Current account restrictions are defined as “restrictions on payments for current transactions.” Multiple currency practices are defined as “separate exchange rate(s) for some or all capital transactions and/or some or all invisibles.” Data on the inflation rate, nominal interest rates on government debt, the ratio of the current account deficit to GDP, annual exports measured in U.S. dollars, and the ratio of the government budget deficit to GDP are from IMF, International Financial Statistics (various issues), and national sources. The inflation rate is the annual rate of change of the consumer price index. The real ex post interest rate on government debt is the nominal rate less actual inflation. Data on the growth rate of real per capita GDP, the ratio of government consumption to GDP, the ratio of investment to GDP, and the ratio of the sum of imports and exports to GDP are from Summers and Heston (1991) and Penn World Table 5.5 update. The index of legal central bank independence and the average yearly turnover rate of central bankers are from Cukierman et al. (1992). Higher values of the index of central bank independence correspond to more bank independence. Dummy variables indicating a coalition government is in power, a majority government is in power, a left-of-center government is in power, and a nondemocratic government is in power are from elaborations on Banks (various issues). Data on the number of government changes between 1950 and 1982 is from Taylor and Jodice (1983).

450

Michael D. Bordo and Barry Eichengreen

References Abramovitz, Moses. 1989. Thinking about growth. In Thinking about growth. Cambridge: Cambridge University Press. Alesina, Alberto, Vittorio Grilli, and Gian Maria Milesi-Ferretti. 1994. The political economy of capital controls. In Capital mobility: The impact on consumption investment and growth, ed. Leonard0 Leiderman and Assaf Razin, 289-321. Cambridge: Cambridge University Press. Angrist, Joshua. 1995. Conditioning on the Probability of Selection to Control for Selection Bias. NBER Working Paper no. 181. Cambridge, Mass.: National Bureau of Economic Research. Balogh, Thomas. 1946. The United States and the world economy. In Bulletin of the Oxford Institute of Statistics 8:309-323. Banks, Arthur S. Various issues. Political handbook of the world. New York: McGraw-Hill. Barro, Robert. 1979. Money and the price level under the gold standard. Economic Journal 89 (March): 13-33. Bayoumi, Tamim, and Barry Eichengreen. 1996. The gold standard and the evolution of the international monetary system. In Modern perspectives on the classical gold standard, ed. Tamim Bayoumi, Barry Eichengreen, and Mark Taylor. Cambridge: Cambridge University Press. Bernanke, Ben. 1995. The macroeconomics of the Great Depression: A comparative approach. Journal of Money Credit and Banking 27 (1): 1-28. Bloomfield, Arthur I. 1959. Monetary policy under the international gold standard. New York: Federal Reserve Bank of New York. Bordo, Michael. 1993. The Bretton Woods international monetary system: An historical overview. In A retrospective on the Bretton Woods system, ed. Michael Bordo and Barry Eichengreen, 3-98. Chicago: University of Chicago Press. Bordo, Michael, Ehsan Choudhri, and Anna J. Schwartz. 1995. Could stable money have averted the Great Depression. Economic Inquiry 33 (3): 484-505. Bordo, Michael, and Richard Ellson. 1985. A model of the classical gold standard with depletion. Journal of Monetary Economics 16: 109-20. Bordo, Michael, and Lars Jonung. 1987. The long-run behavior of the velocity of circulation: The international evidence. Cambridge: Cambridge University Press. . 1990. The long-run behaviour of velocity: The institutional approach. Journal of Policy Modelling 12 (2): 165-97. Bordo, Michael, and Finn Kydland. 1995. The gold standard as a rule: An essay in explorations. Explorations in Economic History 32 (4): 423-64. Bordo, Michael, and Anna J. Schwartz. 1996. Operation of the specie standard: Evidence for core and peripheral countries 1880-1990. In Historical perspectives on the gold standard: Portugal and the world, ed. Barry Eichengreen and Jorge Macedo, 11-83. London: Routledge. Choudhri, Ehsan, and Kochin Levis. 1980. The exchange rate and the international transmission of business cycle disturbances: Some evidence from the Great Depression. Journal of Money, Credit and Banking 12:65-74. Clarke, S. V. 0. 1967. Central bank cooperation 1924-31. New York: Federal Reserve Bank of New York. Cukierman, Alex, et al. 1992. Measuring the independence of central banks and its effect on policy outcomes. World Bank Economic Review 6 (September): 353-98. Dominguez, Kathryn. 1993. The role of international organizations in the operation of the Bretton Woods system. In A retrospective on the Bretton Woods system, ed. Michael D. Bordo and Barry Eichengreen, 357-404. Chicago: University of Chicago Press.

451

The Great Depression and the International Monetary System

Eichengreen, Barry. 1990. The gold exchange standard and the Great Depression. In Elusive stability: Essays in the history of international finance 1919-1939, 239-70. Cambridge: Cambridge University Press. . 1992. Golden fetters: The gold standard and the Great Depression I91 9-1 939. New York: Oxford University Press. . 1993. Reconstructing Europe 5 trade andpayments. Manchester: University of Manchester Press. . 1994. International monetary arrangements for the 21st century. Washington, D.C.: Brookings Institution. . 1995. Central bank cooperation and exchange rate commitments: The classical and interwar gold standards compared. Financial History Review 2, pt. 2 (October): 99-1 17. . 1996. Institutions and economic growth: Europe after World War 11. In Economic growth in Europe since 1945, ed. Nicholas Crafts and Gianni Toniolo, 38-72. New York: Cambridge University Press. Field, Alexander J. 1984. A new interpretation of the onset of the Great Depression. Journal of Economic History 44:489-98. Friedman, Milton, and Anna J. Schwartz. 1963.A monetary history of the United States, 1867-1960. Princeton, N.J.: Princeton University Press. Gilbert, Christopher L., Raul Hopkins, Andrew Powell, and Amlan Roy. 1996. The World Bank: Its functions and its future. Global Economic Institutions Working Paper no. 15. London: Economic and Social Research Council. Grilli, Vittorio, and Gian Maria Milesi-Ferretti. 1995. Economic effects and structural determinants of capital controls. IMF Staff Papers 42:5 17-5 1. Hamilton, James. 1987. Monetary factors in the Great Depression. Journal of Monetary Economics 13: 145-69. InternationalMonetary Fund. Various issues. Annual report on exchange arrangements and exchange restrictions. Washington, D.C.: International Monetary Fund. . 1972. International financial statistics: Supplement. Washington, D.C.: International Monetary Fund. Kendrick, John W. 1961. Productivity trends in the United States. Princeton, N.J.: Princeton University Press. . 1973. Postwarproductivity trends in the United States, 1948-1969. New York: Columbia University Press. Kenen, Peter. 1960. International liquidity and the balance of payments of a reserve currency country. Quarterly Journal of Economics 94 (November): 572-86. . 1963. Reserve asset preferences of central banks and stability of the goldexchange standard. Princeton Essays in International Finance, no. 10. Princeton, N.J.: Princeton University, Department of Economics, International Finance Section. Kouri, Pentti J. K., and Michael G. Porter. 1974. International capital flows and portfolio equilibrium. Journal of Political Economy 82:443-67. Kwiecinska-Kalita,Halina. 1996. Monetary independence under a fixed exchange rate: The interwar experience 1925:5-1931:4. Unpublished Ph.D. diss., University of Toronto, Toronto. Laidler, David. 1994. The golden age ofthe quantity theory. Princeton, N.J.: Princeton University Press. League of Nations. 1930. Interim report of the Gold Delegation of the Financial Committee. Geneva: League of Nations. . 1931. Second interim report of the Gold Delegation of the Financial Committee. Geneva: League of Nations. McCallum, Bennett. 1990. Could a monetary base rule have prevented the Great Depression? J o u m l of Monetary Economics 26 (1): 3-26. MacDougall, Donald. 1957. The world dollarproblem. New York St. Martin’s.

452

Michael D. Bordo and Barry Eichengreen

McKinnon, Ronald. 1969. Private and official international money: The case for the dollar. Princeton Essays in International Finance. Princeton, N.J.: Princeton University, Department of Economics, International Finance Section. Meltzer, Allan. 1991. U.S. policy in the Bretton Woods era. Federal Reserve Bunk of St. Louis Review 73 (MayIJune): 54-83. . 1995. Why did monetary policy fail in the thirties? Pittsburgh, Penn.: Carnegie Mellon University. Mimeograph. Mitchell, B. R. 1992. International historical statistics: Europe, 1750-1988. New York: Stockton. . 1993. International historical statistics: The Americas, 1750-1 988. New York: Stockton. . 1995. International historical statistics: Afiica, Asia, and Oceania, 17501988. New York: Stockton. Mourk, Kenneth. 1991. Managing the franc Poincart!: Economic understanding und political constraint in French monetary policy 1928-1936. Cambridge: Cambridge University Press. Mundell, Robert. 1970. Monetary theory. Goodyear. Obstfeld, Maurice. 1982. Can we sterilize? Theory and evidence. American Economic Review 72:45-50. . 1993. Destabilizing effects of exchange rate escape clauses. NBER Working Paper no. 3603. Cambridge, Mass.: National Bureau of Economic Research. Pasula, Kit. 1994. Sterilization, Ricardian equivalence and structural and reduced form estimation of the offset coefficient. Journal of Macroeconomics 16 (4): 663-99. . 1996. Monetary independence under Bretton Woods: Perspectives from a stochastic maximizing model. Canadian Journal of Economics 29 (3): 643-64. Rockoff, Hugh. 1984. Some evidence on the real price of gold: Its costs of production and commodity prices. In A retrospective on the classical gold standard 1821-1931, ed. Michael D. Bordo and Anna J. Schwartz, 613-50. Chicago: University of Chicago Press. Rodrik, Dani. 1995. Getting interventions right: How South Korea and Taiwan grew rich. Economic Policy 20:53-101. . 1996. Why is there multilateral lending? In Annual World Bank conference on development economics, ed. Michael Bruno and Boris Pleskovic, 67-93. Washington, D.C.: World Bank. Romer, Christina. 1993. What ended the Great Depression? Journal of Economic History 52~757-84. Rosenbaum, Paul, and Donald Rubin. 1983. The central role of the propensity score in observational studies for causal effects. Biometrica 70 (1): 41-55. . 1984. Reducing bias in observational studies using subclassification of the propensity score. Journal of the American Statistical Association 79 (September): 16-524. Sachs, Jeffrey D., and Andrew Warner. 1995. Economic reform and the process of global integration. Brookings Papers on Economic Activity 1:1-1 18. Saint-Paul, Gilles. 1995. The franc: Real and monetary aspects of French exchange rate policy under the Fourth Republic. In Europe S postwar recovery, ed. Barry Eichengreen, 292-3 19. Cambridge: Cambridge University Press. Summers, Robert, and Alan Heston. 1991. The Penn world table (mark 5): An expanded set of international comparisons 1950-1988. Quarterly Journal of Economics 106 (May): 327-68. Taylor, Charles Lewis, and David A. Jodice. 1983. World handbook of political and social indicators, 3d ed. New Haven, Conn.: Yale University Press. Temin, Peter. 1989. Lessonsffom the Great Depression. Cambridge, Mass.: MIT Press.

453

The Great Depression and the International Monetary System

Thomas, Mark. 1994. Wage behavior in inter-war Britain: A sceptical enquiry. In Labour market evolution, ed. George Grantham and Mary MacKinnon, 245-69. London: Routledge. Triffin, Robert. 1947. National central banking and the international economy. Postwar Economic Studies 7:46-8 1. U S . Gold Commission Report. 1982. Washington, D.C.: Government Printing Office. Wheelock, David. 1990. The strategy and consistency of Federal Reserve monetary policies 1924-1931. Cambridge: Cambridge University Press. Williams, John H. 1952. Economic stability in the modem world. London: Athlone. Wolf, Holger. 1995. Postwar Germany in the European context: Domestic and external determinants of growth. In Europe 5 postwar recovery, ed. Bany Eichengreen, 32352. Cambridge: Cambridge University Press.

This Page Intentionally Left Blank

Contributors

Katherine Baicker Department of Economics Harvard University Cambridge, MA 02138

Claudia Goldin Department of Economics Harvard University Cambridge, MA 02138

Michael D. Bordo Department of Economics New Jersey Hall Rutgers University New Brunswick, NJ 08904

Douglas A. Irwin Department of Economics Dartmouth College Hanover, NH 03755

Charles W. Calomiris Graduate School of Business Columbia University 3022 Broadway Street, Uris Hall New York, NY 10027

J. Bradford De Long Department of Economics University of California, Berkeley 549 Evans Hall Berkeley, CA 94720

Lawrence F. Katz Department of Economics Harvard University Cambridge, MA 02138 Gary D. Libecap Karl Eller Center College of Business McClelland Hall University of Arizona Tucson, AZ 85721

Barry Eichengreen Department of Economics University of California, Berkeley 549 Evans Hall Berkeley, CA 94720

Jeffrey A. Miron Department of Economics 270 Bay State Road Boston University Boston, MA 02215

Richard B. Freeman National Bureau of Economic Research 1050 Massachusetts Avenue Cambridge, MA 02138

Wallace E. Oates Department of Economics University of Maryland College Park, MD 20742

455

456

Contributors

Maurice Obstfeld Department of Economics University of California, Berkeley 549 Evans Hall Berkeley, CA 94720

David N. Weil Department of Economics Box B Brown University Providence. RI 029 12

Hugh Rockoff Department of Economics New Jersey Hall Rutgers University New Brunswick. NJ 08903

David C. Wheelock Research Department Federal Reserve Bank of St. Louis P.O. Box 442 St. Louis, MO 63166

Alan M. Taylor Department of Economics Northwestern University 2003 Sheridan Road Evanston, IL 60208

Eugene N. White Department of Economics New Jersey Hall Rutgers University New Brunswick, NJ 08903

John Joseph Wallis Department of Economics University of Maryland College Park,MD 20742

Name Index

Abbott, Edith, 138 Abramovitz, Moses, 410 Abramson, A., 145 ACSS. See Advisory Council on Social Security (ACSS) Adams, James D., 247n49,255 Advisory Council on Social Security (ACSS), 299,300,305,307,312 Advisory Council on Unemployment Compensation (ACUC), 229 Ahearn, Daniel S., 45,47 Alesina, Alberto, 355111, 431 Alston, Lee J., 195n47 Alston, Richard M., 150, 1681113 Altmeyer, Arthur, 227,229,236n21,24On31, 246n46,3 18n22 Anderson, Gary M., 25 Anderson, Patricia M., 234n16.246n45, 254n59,255,256n64 Angell, James W., 140-41 Angrist, Joshua, 432n50 Armstrong, Barbara, 3 11 Auerbach, Alan, 68, 83 Bailey, Michael, 342, 349n24 Bailey, Stephen, 79 Bain, G. S., 272t, 2731,2791 Baker, Gladys L., 21Gn77 Bakija, Jon M., 314t, 315,3181119 Bakker, Age E I?, 391,392,393,394,396 Baldwin, Marc, 234 Balke, Nathan S., 7f Ball, Robert M.,307, 312

457

Balogh, Thomas, 406n10 Barber, William J., 126, 133n2 Barnett, George, 278 Barro, Robert J., 410,41In21,412n23 Barsky, Robert B., 247 Barth, James R., 115 Bauer, J., 137 Bauer, Raymond A., 348 Bayoumi, Tamim, 407nl1 Beck, Adam, 137 Becker, Joseph M., 227n2,236nn21,22 Benedict, Murray R., 181n1, 185, 186nn10, 11, 1871113, 1881117, 189nn19,20,21, 190, 1911129, 193n39, 194n45,195, 196n49, 197n52, 199n58 Bennett, Wallace F., 110 Berkowitz, Edward D., 133, 135 Bernanke, Ben, 9,377n27,379n28,405n5, 412 Bernstein, Edward M., 377-78n27 Bernstein, Irving, 271, 279, 287 Bernstein, Michael A., 1481118 Biggar, Emerson, 136, 137 Black, John D., 190n24 Black, Stanley W., 377-78n27 Blaisdell, Donald C., 181n1, 191nn32.33, 192n36, 196n50 Blank, Rebecca, 234n17 Blaug, Mark, 133 Blaustein, Saul J., 236n21, 238n27, 239, 2441143 Blinder, Alan, 60,72-73 Block, F., 1681113

458

Name Index

Bloomfield, Arthur I., 429 Blum, John M., 39,401116.58 Boarman, Patrick M., 391n43 Boehm, W. T., 197n53 Bomar, Thomas R., 114 Bordo, Michael, 8n7,46,53n22,54, 611125, 357,393,404n4,405n7,408n14,410, 411,412n24,415,422n37,426,448,449 Boskin, Michael, 315n17 Bothwell, Robert, 185n7, 216 Boyd, John H., 116 Bradford, David F., 175 Bradley, Edward S., 1I1 Brand, Charles J., 188 Brandeis, Elizabeth, 239 Brandeis, Louis, 239, 243 Brandt, John, 190n24 Brandt, Willy, 392 Bratter, H. M., 370, 371n15, 375, 376 Brechling, Frank, 256n64 Breimeyer, Harold F., 190n28, 195n48 Bremer, William W., 168n13 Brewer, Elijah HI, 88t, 113 Brewer, John, 71n2 Breyer, Stephen, 200 Bricker, John W., 95 Britnell, George E., 185n7,216 Brock, William R., 1681113 Brooks, H. Phelps, Jr., 114 Brown, Charles C., 177 Brown, Douglass V., 73 Brown, E. Cary, 78 Brown, J. Douglas, 301n5, 306, 316, 320 Brown, Josephine Chapin, 166 Brown, William A,, Jr., 368 Brownlee, W. Elliot, 4 Bmnner, Karl, 26,501119 Bryce, James, 155, 163 Buchanan, James, 83 Bureau of the Census, 8f Bums, Arthur, 80n8 Bums, Helen M., 90 Bums, James McGregor, 279 Butler, Hugh, 95-96 Cairncross, Alec, 388, 389 Callahan, Colleen, 334n6, 337n9 Calorniris, Charles W., 38.46, 59n24, 89, 90, 118, 142,385 Campa, JosC M., 377n27 Capie, Forrest, 361n7 Card, David, 234n17, 246n45, 255 Cardoso, Eliana, 372,394

Carey, Kevin, 377n27, 379n28 Carter, Susan B., 308 Caskey, John, 52n21,61 Cates, Jerry R., 299 Celler, Emanuel, 112 CES. See Committee on Economic Security WES) Chandler, Lester V., 28-29, 34, 35, 36, 37, 48 Chari, V. V., 52n21 Chase, Salmon, 69 Choudhri, Ehsan U., 377n27,404nn3,4 Christiano, Lawrence J., 52n21 Clark, John Maurice, 137 Clarke, S . V. O., 429 Cloward, R. A., 168n13 Cochrane, Willard, 181111, 190n23, 195n48 Committee on Economic Security (CES), 299, 303,304,305,307,309,311 Commons, John R., 142, 143,236-38, 246n47,267 Cooke, T., 142 Coolidge, Calvin, 188 Corni, Gustavo, 185n7.215.216 Corson, Walter, 234 Costa, Dora, 308, 319 Council of Economic Advisers, 7f, 8f Crowley, Leo, 96 Crucini, Mario J., 335 Cukierman, Alex, 449 Cull, Robert J., 363118 Currie, Lauchlin B., 38 Curtis, Frederic, 36 Darn, Kenneth, 384,385,394 Davis, Horace B., 280 Davis, J. Ronnie, 133 Davis, Joseph, 190n24,211nn79, 80 Davis, Lance E., 363118 Deere, Donald R., 255,256, 258t Deering, Ferdie, 205n67 Delano, Preston, 96 De Long, J. Bradford, 10,59,60,71n2,80, 82 Derber, Milton, 279 Deny, Kathleen, 138-39 Derthick, Martha, 300n4,306,308,3 181122 Dewey, R. L., 137 Dexter, Lewis A., 348 Diaz Alejandro, Carlos F., 375, 376n23, 377n27,381 Director, Aaron, 133, 246n47, 247 Dominguez, Kathryn, 445 Dorfman, Joseph, 133, 134, 135, 143, 145 Dombusch, Rudiger, 335n7, 372, 394

459

Name Index

Douglas, Paul, 133, 137118, 138-39, 142, 236n21,237-38,242,246n47, 247 Drozdowski, R., 101n8 Drummond, Ian, 185117,216 Dubois, Ben, 96 Dunlop, John, 267,274 Eccles, Marriner S., 31-32, 37-41,43 Edelstein, Michael, 358114 Edwards, Sebastian, 395x145 Eichenbaum, Martin, 52n21 Eichengreen, Bany, 8n7,9,44n17,53n22, 61n25,334n6,335,337,357,358, 365n11,368,369,370,372,374,377, 380,39On42,405,406n8,407nnll, 12, 408n14,413,418n32,44On59 Einzig, Paul, 361,367,368, 369, 370, 371, 372,374,389,392 Eisenhower, Dwight D., 348 Eliot, Thomas H., 241n33,242n39,299n2 Ellenbogen, Henry, 3051110 Ellis, Howard S., 372,374,375 Ellson, Richard, 410,41ln22,412n24 Ely, Richard T., 145-46 Emminger, Otmar, 3911143 Engerman, Stanley, 249n53 English, John, 185117,216 Epstein, Gerald, 25 Erhard, Ludwig, 391 Eriksson, Erik McKinley, 241-42n36 Evans, Sam, 210n77,211 Ewalt, Josephine H., 105 Ezekiel, Mordecai, 190n24 Faber, J. F., 318 Farquharson, J. E., 185n7,216 Feinberg, Robert M., 89 Feis, Herbert, 139 Feldman, Herman, 227n2 Feldstein, Martin, 255,321, 356, 358 Ferguson, Thomas, 25 Feme, J. P., 168n13 Field, Alexander J., 404n2 Filene, Edward, 135 Fischer, Stanley, 335n7 Fisher, Irving, 9118, 25 Fishlow, Albert, 376n23 Fite, Gilbert C., 1881117, 189n19 Fleming, Robben Wright, 287 Flood, Mark D., 90 Fogel, Robert W., 151 Folkerts-Landau, David, 395t Folwell, William W., 143

Fordney, Joseph, 334 Fowke, V. C., 185n7,216 Frankel, Emil, 141 Frankel, Jeffrey, 394 Frederick, John H., 141 Freeman, Richard B., 271 Friedel, Frank, 279 Friedman, Milton, 9, 10, 25-26, 30118, 36, 38n13,45,76,89, 147, 149,383n33, 387,404n2 Friedman, Rose, 147 Friend, Emil, 140 Frost, Peter A., 38 Galenson, Walter, 271,284t, 285, 289-90 Garber, Peter M., 441117 Gardner, B. Delworth, 181n1, 182n2, 19711.54, 2141187 Gardner, Bruce, 181n1, 182n2, 190n27, 192nn34.35, 193n39, 194n43, 1971153, 198111155.57, 199-200,214n87 Gardner, Richard N., 386,388,389 Gertler, Mark, 116 Giavazzi, Francesco, 394 Giebisch, Robert, 282t Gilbert, Christopher L., 445 Giovannini, Alberto, 358,360n5.394 Glass, Carter, 28-29,38 Goldenweiser, E. A,, 35,37,41 Goldfield, M., 279 Goldin, Claudia, 2491153 Goldstein, Judith, 342 Golembe, Carter H., 87111, 89, 90 Goodfriend, Marvin, 52n21.61 Gordon, Margaret S., 370 Gordon, Robert J., 7f, 76,82 Goschen, G. J., 358n5 Graebner, William, 308,3 11 Gramlich, Edward, 175 Gratton, Brian, 310 Green, David A,, 235 Green, Robert C., 21 11182 Grigg, David, 185n7.216 Grilli, Vittorio, 35nl,431,433n52,434nn55, 56,436 Grossman, Richard S., 105 Groves, Harold, 142, 236n22 Gruber, Jonathan, 232n12,255n62 Haber, Carole, 3 10 Haber, William, 236n21 Haberler, Gottfried, 73, 377-781127, 383-84 Hadwiger, Don E, 21On77,211,212n84

460

Name Index

Haggard, Stephan, 338,341,3421117 Hallwood, C. Paul, 358,360n5 Halpin, Terrence, 255n61 Hamermesh, Daniel S . , 233n14 Hamilton, Alexander, 69 Hamilton, David E., 187n16 Hamilton, James D., 377n27,404n2 Harl, Maple T., 96 Harrison, George, 37,42 Harrod, Roy, 383n34 Hawtrey, Ralph, 76, 377-781127 Hayek, Friedrich, 73 Hayes, Alfred, 54, 55 Hefford, R. K., 185n7, 216 Heid, Walter G., 210n77 Heier, William D., 114 Helleiner, Eric, 355n1 Heller, Walter, 79, 80n8, 82 Hendricks, Gisela, 185117, 215, 216 Heston, Alan, 449 Higgs, Robert, 2f,4n2, 127, 132, 150, 151, 184, 185n8, 186nll. 190n27, 197n52, 203,214n86,216 Hillman, Arye, 350n25 Hines, James R., 175, 176 Hoekman, Bernard, 89 Hoffman, Elizabeth, 186n12, 187, 1881117, 189n19, 191n31, 192n36,193nn39,40, 199n58 Hofstadter, Richard, 5n4 Holcombe, Arthur, 138 Holland, David S., 87nl Holt, John Bradshaw, 185n7 Hoover, Herbert, 4, 83, 187, 334; balanced budget policy, 67,74-75; as head of U.S Food Administration, 186; on tariff protection, 338; view of Federal Reserve, 75-76 Hopkins, Hany, 166, 167-69 Horioka, Charles, 356,358 Horsefield, I. Keith, 382n32, 384, 385, 386 Howard, Donald S., 166n9, 170nn14, 15, 16 Hughes, Jonathan R. T., 2111, 127, 132, 150 Hughes, Sidney J., 95 Hulbert, L. S . , 190n24 Hull, Cordell, 338, 339-40, 342 Hutcheson, Bill, 285 Iorio, Karl, 255n62 Irons, Peter H., 190n27 Irwin, Douglas A., 334, 335, 344, 345, 346 Isaacs, Arthur, 345 Ito, Takatoshi, 394, 395t

Jackson, Andrew, 70 James, William, 145-46 Jenkins, Thomas, 347 Jodice, David A,, 449 Johnson, D. Gale, 181n1, 183n5, 199,207n71 Johnson, G. Griffith, 30.3 1 Johnson, Lyndon B., 49-50,393 Johnson, Ronald L., 169 Jones, Joseph M., 337 Jonung, Lars, 411,422n37 Joy, Aryness, 139 Kahn, Alfred, 149 Kahn, James, 335 Kane, Edward J., 115 Kaplan, Jacob J., 388, 390n42 Kaplan, Seth, 89 Kapstein, Ethan B., 355n1 Kearl, J. R., 150 Keeler, John T. S., 216n89 Kelley, William B., Jr., 331n4, 332, 333, 348n23 Kelly. Florence, 1381110 Kendnck, John W., 448,449 Kenen, Peter, 428, 429 Kenkel, Joseph F., 328111, 334 Kennickell, Arthur, 95 Keynes, John Maynard, 8, 72,76, 83,355, 368,381-86,388,398,443 Kidd, Charles V., 227n2 Kindleberger, Charles, 368, 372n16.393n44 Kinky, David, 143 Klare, Karl, 279 Knutson, R. D., 197n53 Koch, Karin, 346n19 Kochin, Levis A,, 377n27,404n3 Kolko, Gabriel, 6n5 Kotlikoff, Laurence, 3 17 Kouri, Pentti J. K., 409n20 Krooss, Herman E., 441117, 57 Kroszner, Randall S., 334n6, 345 Kwast, Myron L., 95 Kwiecinska-Kalita, Halina, 409n20 Kydland, Finn, 46, 357, 4081114 Laidler, David, 404-5114 Lake, David A,, 350n25 Larger, William, 95 League of Nations, 337, 368, 372, 374, 375, 376,377,425 Leddy, John M., 347 Lee, Frederick P., 190 Leff, Mark, 306-7

461

Name Index

Leidy, Michael P., 89 Lester, Richard A., 227n2 Leuchtenberg, William, 279 Levine, Phillip B., 246n45, 255 Lewis, John L., 278,279,280-81,283,285 Lewis, Karen K., 364 Lewisohn, Sam A., 246n47 Libecap, Gary, 6115, 169, 185-86n9, 1861112, 187, 188n17, 1891119, 191n31, 192n36, 193nn39,40, 199n58 Liepmann, H., 337n10 Lindsay, Donald P., 113-14 Lindsay, Samuel L., 143 Litan, Robert E., 115 Lloyd, A. G., 185n7,216 Lothian, James R., 365 Lowi, Theodore, 181 Lubove, Roy, 309 Lucas, Arthur F., 139 Lucas, Robert E., Jr., 60,356 Lumsdaine, Robin, 319 McAdoo, William G., 69 McCabe, Thomas, 43,44 McCallum, Bennett, 404114 McCloskey, D. N., 134 McCraw, Thomas, 149 McCune, Wesley, 192n35 McDonald, Judith, 334n6,337n9 MacDonald, Ronald, 360115 McDougal, James, 36 MacDougall, Donald, 4061110 Macey, Jonathan R., 87nl McHugh, Richard, 234 McKellar, Kenneth, 338 McKinnon, Ronald, 381n29,409n18,428 McQuaid, Kim, 133, 135 Maisel, Sherman, 5 1, 60 Marsh, Ian W., 3 5 8 , 3 6 0 ~ 5 Marshall, Alfred, 136 Marston, Richard C., 358115 Martin, William McChesney, 44, 48-5 1, 54-56 Marvell, Thomas B., 107n10, 108nll Matthews, John 0.. 112 Maybank, Burnet R., 95 Meigs, A. James, 501119 Mellon, Andrew, 74-75 Meltzer, AIIan H., 9n9, 10n10, 26, 27n4, 5On19,404n4,409n18,428 Meyer, Bruce D., 234n16,246n45,254n59, 255 Mikesell, Raymond, 382, 387

Milesi-Ferretti, Gian Maria, 355n1,431, 433n52,434nn55,56,436 Miller, Adolph, 35, 36 Miller, Arthur Selwyn, 241n36 Miller, E. T., 140 Miller, Geoffrey P., 87nl Miron, Jeffrey, 76, 247 Mitchell, B. R., 448 Mitterrand, FranGois, 396 Moen, Jon R., 308 Mofsky, James S., 111 Moggridge, D. E., 3691112 Mondschean, Thomas H., 88t, 113 Morgenstern, Oskar, 358115 Morgenthau, Henry, Jr., 38-41,385-86 Moser, Peter, 3501125 Moss, David A,, 134n3 Moulin, Annie, 185n7, 216 Mour6, Kenneth, 429n47 Mundell, Robert, 41 1 Murphy, P. L., 190n26 Murray, Merrill G., 236n21 Murray, Philip, 283,289-90 Muth, Hans Peter, 216n89 Myers, Robert J., 299,311,312, 313, 316nn18, 19 Nelson, Daniel, 236n21 Nelson, Douglas, 342n15 Nicholson, Walter, 234 Nixon, Richard M., 18 North, Douglass C., 151 Norwood, Janet L., 347 Nourse, Edwin D., 190nn24,25, 191nn30.31, 33, 192n34, 193n39, 195, 196n50 Nurske, Ragnar, 371,372;373,375,380 Oates, Wallace E., 3f, 175, 177 O’Brien, Anthony, 334n6,337n9 Obstfeld, Maurice, 358,364, 391n43,393n44, 409n20 Officer, Lawrence H., 358115 Organization for Economic Cooperation and Development (OECD), 272t Parker, J. S., 300.3181122 Pasour, E. C., Jr., 182n2, 1971153, 198x57 Pastor, Robert, 3411114,345 Pasula, Kit, 409n20 Patman, Wright, 108 Patterson, James T., 5n4 Peacock, Alan T., 184,216 Peck, George N., 342

462

NameIndex ~

Peck, H. W., 136-37 Penn, J. B., 1971153 Pepper, Claude, 95 Perkins, Edwin J., 361 Perkins, Frances, 167,236n21,237,240-43, 304 Perkins, Van L., 1861112, 189nn20,22, 190nn24,26, 192n37, 193nn39,41, 1941142, 1991158 Perren, Richard, 185117.215-16 Pigou, A. C., 134, 139-40, 141-42 Piven, Frances F., 1681113 Pool, Ithiel de Sola, 348 Pope, David, 21 1 Porter, Bruce D., 127 Porter, Michael G., 409n20 Portus, G. V., 139 Potts, Frederick A., 96 Price, Robert, 272t, 273t, 279t Puffert, D., 315117 Ransom, Roger, 308,3091113,318 Rasmussen, Wayne D., 210n77 Raushenbush, Paul, 236n22,239 Reagan, Ronald, 68 Redish, Angela, 53n22 Rees, David, 386n38 Riddell, W. Craig, 235 Riefler, Winfield, 48 Ripley, Wiliam Z., 141 Robbins, Lionel, 73, 377-78n27 Rockoff, Hugh, 10-11, 133, 1971152,357, 412n23 Rodrik, Dani, 4401159,445 Rogoff, Kenneth, 364 Romer, Christina D., 8n7, 10,406n9 Romer, Paul, 305,306n I1 Roosevelt, Franklin D., 4, 39, 76, 83, 90, 167, 169,237, 240,243; appointments to Supreme Court, 241; balanced budget policy, 67; campaign trade policy (1930), 338; gold policy (1933), 29; as governor of New York, 237; on old-age pensions, 298; position on tariff protection, 338; on social security program, 15; on trade policy (1934), 340 Rooth, Tim, 185n7, 216, 343 Rosenbaum, Dan T., 2551163 Rosenbaum, Paul, 432n50.4331152 Rubin, Donald, 432n50, 4331152 Rubinow, Isaac M., 142, 144,2381126 Rucker, Randy, 1951147 Ryan, Mary, 181n1, 190n23, 1951148

Sachs, Jeffrey D., 353, 377n27, 410 St. Gennain, Fernand, 113-14 Saint-Paul, Gilles, 410 Salant, Walter, 76 Salera, Virgil, 376n24 Saloutos, Theodore, 1871113 Sawer, Geoffrey, 156111 Sayers, R. S., 373,380 Schattschneider. Elmer E., 89, 329113, 334n6 Schleiminger, Gunther, 388, 390n42 Schlesinger, Arthur M., Jr., 15, 147-48, 236n21,239,279,304,305 Schneider, Michael, 272t Schnietz, Karen E., 340, 342 Schultz, Theodore W., 190n27, 193n39, 195, 1991158 Schumpeter, Joseph, 73, 136n6, 148.3701113 Schwartz, Anna J., 9, 10, 25-26, 30nn8, 9, 36, 38n13,45,59n24,76,89, 149,404nn2, 4,414, 448, 449 Seligman, Joel, 111, 112 Shaw, A. G. L., 185117,216 Sheflin, Neil, 252-53t, 267,269, 270t, 289-90 Shepherd, William G., 200 Shergold, Peter, 2 11 Shideler, James, 1861112 Shleifer, Andrei, 71n2 Shover, John L., 190n26 Shugart, William F., 25 Simons, Henry, 133 Skocpol, Theda, 134 Slesnick, Daniel T., 233n14 Smith, Adam, 71 Smith, Donald M., 227n2 Snyder, John W., 43,96 Sobel, Robert, 111 Solomon, Robert, 393 Sowards, Hugh L., 111 Spalding, William F., 361n6 Sprague, Irvine H., 116 Sproul, Allen, 48 Squires, B. M., 144n15 Starr-McCluer, Martha, 95 Steagall, Henry, 90 Stedman, Alfred D., 190n24 Stein, Herbert, 68, 72, 75116, 79, 84, 133n2 Steuerle, C. Eugene, 314t, 315, 318n19 Stewart, Bryce M., 236n21.238 Stewart, Robert B., 3731117 Stewart, William, 156111 Stigler, George J., 126, 149 Stine, Oscar, 181nl Stockman, David, 150

463

Name Index

Strong, Benjamin, 25-27, 34-35, 37, 47-48, 52 Stuhler, Elmar A,, 185n7, 216 Summers, Lawrence H., 80, 82 Summers, Robert, 449 Sutch, Richard, 308,309n13,318 Svensson, Lars E. O., 358,360n5 Swope, Gerald, 135 Taft, Philip, 279, 287 Taft, William Howard, 330 Tasca, Henry J., 338,340,341n13,344 Taussig, Frank W., 135, 136, 143-44,328n2, 329,334 Tavlas, George S., 133 Taylor, Alan M., 357, 358, 364 Taylor, Charles Lewis, 449 Taylor, Myron, 283 Temin, Peter, 8nn6,7,9,76, 377n27,405n5 Teweles, Richard, 111 Thaler, Richard H., 175, 176 Thomas, Mark, 407 Tobey, Charles W., 95 Tocqueville, Alexis de, 155 Tolles, Newman A,, 144 Tollison, Robert D., 25 Topel, Robert, 2461145,255 Tracy, Michael, 185n7.215, 216 Triffin, Robert, 387n39,388,390n42,425 Troy, Leo, 252-53t, 267,269,27Ot, 284t, 285t, 286t, 289-90 Truman, Harry S, 44,58,347 Tugwell, Rexford, 134-35 Tuttle, Charles, 143, 145 Udell, Gillman G., 182, 200 Underwood, Oscar, 330 United States Tariff Commission, 345, 346 US.Department of Health and Human Services (HHS), 297,299 U.S. Department of Labor, 288; Bureau of Labor Statistics (BLS), 289-90 Vandenberg, Arthur, 96,345,347,349 Vaughan, Michael B., 150 Verdier, Daniel, 349 Viner, Jacob, l33,383n33,387n39 Vinson, Fred, 386

Visser, Jelle, 272t, 273t Volcker, Paul, 58 Wagner, Richard, 83 Wagner, Robert, 3 11 Walker, Forrest Q., 1411111 Wallace, Henry A., 187 Wallace, Henry C., 190, 195 Wallis, John Joseph, 3f, 164t. 165116, 167nn10, 12, 171 Warner, Andrew, 353,410 Watson, Richard A., 3481121 Weaver, Carolyn L., 310 Webb, Sidney, 138 Webber, Alan, 361n7 Weill, N. E., 358n5 Weingast, Barry R., 342 Wheelock, David C., 9n9, 26, 27, 29, 38, 57n23,89,385,404n4 White, Eugene N., 89, 90, 118, 142 White, Harry Dexter, 355, 382, 384-86,443 White, Lawrence J., 87n1, 115, 116 Wicker, Elmus, 9n9, 26 Willcox, Alanson, 306 Williams, E. A,, 167 Williams, John H., 4061110 Williamson, Jeffrey G., 357 Willie, Frank, 113-14 Wilson, Beny, 38 Wilson, Woodrow, 330,332-33 Wise, David, 3 19 Wiseman, Jack, 184,216 Witmer, Helen Leland, 1451116 Witte, Edwin E., 236n21 Wolf, Holger, 410 Wolman, Leo, 144-45.289 Wood, Adrian, 357 Wood, Donna, 185n8 Wyplosz, Charles, 389n41 Yeager, Leland, 54, 372, 374, 386, 387n39, 388,391,393 Yoder, Dale, 1451116 Young, Edwin, 279 Young, James H., 186n10 Young, Roy, 30 Zevin, B. D., 298 Zinn, Howard, 3051110

This Page Intentionally Left Blank

Subject Index

AAA. See AgriculturalAdjustment Act (1933) AFL. See American Federation of Labor Agricultural Adjustment Act (1933), 27; Agricultural Adjustment Administration under, 189-9 1;CCC price-support loans (1938), 194, 199; expansion of earlier regulation (1938), 194, 196, 199,211; Farm Credit Administration, 189; funding of price supports and programs, 165; intent, 201; marketing agreements and orders, 1911131, 192n36, 196nn49.50; presidential trade authority under, 342; Thomas Amendment provisions related to Federal Reserve, 3 1-32, 39; unconstitutionality of tax provision, 195, 240, 299n2. See also Commodity Credit Corporation (CCC); US.Department of Agriculture (USDA) Agricultural cooperatives: antitrust exemption, 1871114; CooperativeMarketing Act (1926), 187; cooperative marketing associations (1920s), 187; Farm Board loans to, 187 Agricultural Marketing Act (1929): Cotton and Grain StabilizationCorporations, 189; promotion of cooperatives under, 187; supply controls under, 189 Agricultural sector: categories of legislation and regulation, 201-5; crisis (1920s). 186-87; Ever-Normal Granary policy, 195, 196; Freedom to Farm Act (1996), 182; guaranteed-loan programs, 182; impact of New Deal policy, 205; legacy of

465

Great Depression regulation, 213-20; New Deal laws linked to 1962 regulation, 217-1st; New Deal regulation of, 189-97; pre-New Deal regulation, 185-89; price support policy, 12, 165, 194, 199; regulation (1940-70), 197-200 Agricultural Stabilization and Conservation Service (ASCS), 198 Amalgamated Clothing Workers (ACW), 286 American Farm Bureau Federation, 186 American Federation of Labor (AFL), 285 American Stock Exchange (AMEX) trust fund. 111 Bailey v. Drenel Furniture Co. (1922), 240 Balance of payments: capital account controls under Bretton Woods, 394; capital account liberalizationby EU members, 396; capital and current account restrictions in capital control model, 434-43; constraints in absence of Bretton Woods system, 409-10; current account as measure of borrowing abroad, 357; current and capital account convertibility, 381-91; definitions of capital and current account restrictions, 449; Federal Reserve response to deficits (1950s), 53-54. See also Capital flows; Capital markets Bank for International Settlements (BIS), 446 Banking Act (1935): commercial bank reserve requirements under, 39; provisions related to Federal Reserve, 31n10,32-34, 38, 60;temporary and permanent federal deposit insurance, 90-91

466

Subject Index

Banking Acts (1934, 1935). 7, 27 Banking system: in absence of New Deal, 117-19; collapse (1930-33). 90; demand deposit requirements, 33-34; excess reserves (1934-35), 32-35; federal deposit insurance in, 91-104; New Deal legislation related to, 12-13; panics and crises during depression, 8-9, 27; reserve requirements (1936-37). 32-34,37-38. See also Deposit insurance; Federal Reserve System; Reserves, free; Savings and loan associations (S&Ls) Benefits. See Unemployment insurance (UI) Bretton Woods system: in absence of Great Depression, 403,407-8; agreement (1944), 367; capital account controls under, 394; collapse (1971-73), 54; conditions for collapse of, 24; effect of, 45, 52-56; failure of, 394; hypothetical effect of Triffin dilemma, 444; influence of, 61; influences on design of, 443; pegged exchange rate regime of, 354; role of IMF under, 444-45 Broker-dealer insurance, 13, 109-12 Budget, federal: automatic stabilizers built into, 78; peacetime surpluses, 70. See also Debt, federal Business cycle: countercyclicalfiscal policy arguments, 70-74; overinvestment theories related to, 73, 76 Capital: conditions for and types of misallocation of, 356-57; with interwar exchange controls, 370-72; World War I exchange controls, 367-70 Capital account. See Balance of payments Capital controls: under Bretton Woods, 17; effect of, 43 1-32; logit model of country’s propensity to impose, 432-43 Capital flows: advantages of international mobility of, 356-57; depression years (1931-39), 372-81; EEC liberalization of movements, 391-92; historical international, 357-60; measures of, 357; mobility under interwar gold standard, 430n48; nominal interest rates as indicator of mobility, 358, 360-63; with post-1973 financial system integration, 394-96; postWorld War II plans, 384-86; private international (1943 and after), 386; relation to post-World War II currency inconvertibility, 386-91; relevance of UIP and PPP to mobility of, 364; with trade growth (1950s), 392; trends of foreign,

358-60; with World War I exchange controls, 367-70 Capital markets: in absence of Bretton Woods system, 410; with capital mobility, 356-57; integration of international, 363-66; post-1973 opening, 394-96; reintegration (1960s). 17 CCC. See Commodity Credit Corporation (CCC) Census of governments, 159 CES. See Committee on Economic Security (CES) Chas. C. Steward Machine Co. v. Davis (1937), 243n40 CIO. See Congress of Industrial Organizations (CIO) Civilian Conservation Corps, 145-46, 171 Committee on Economic Development (CED), 68 Committee on Economic Security (CES), 167, 240-41, 300; economic effects of depression on the elderly, 309-10; estimates of workers with pension coverage, 309; forecast of over-65 population, 309 Commodities: as focus of AgriculturalAdjustment Act, 191-92; government stockpiles, 198, 199t; quotas in 1938 agricultural law, 196; regulation (1940-70), 197; supply management programs, 198 Commodity Credit Corporation (CCC): guaranteed-loanprograms, 182; nonrecourse loans to farmers, 194-95, 199, 206; purchases of “excess” commodities, 197-98, 199-200.206-7; storage facilities, 194; wheat program loans and purchases, 209-13,219-20t Commodity markets: effect of New Deal regulatory programs on, 196-97; pre-New Deal regulation of, 185-89; provisions of regulatory policy, 200 Commodity prices: impact of wheat program on, 209-13,219-20t Congress of Industrial Organizations (CIO): auto and steel organizing drives (1930s and 1940s). 283; formation of, 265; money given to unions (1935-41). 283-84; relation of money given to union membership (1935-41), 283,285 Council of Economic Advisers (CEA), 79, 80n8 Credit unions: insurance for, 109-1 3; legislation for federal charters (1934). 109; National Credit Union Share Insurance Fund (NCUSIF), 110

467

Subject Index

Currency: in absence of Great Depression, 405; crises (1931), 372-73; EC float within a snake, 393-94; hard-currency scarcity (1943 and after), 386; multiple currency practices, 49. See also Exchange rates; Triffin dilemma Currency convertibility: absent during and after World War 11,386-87; during Great Depression,9-10; IMF Article VIII provisions related to, 387-88 Current account. See Balance of payments Data sources: analysis of hypothetical absence of Great Depression, 447-49; analysis of New Deal impact on wheat prices, 21 1-12; capital control regressions, 449; estimates of government fiscal activity, 159-60; to measure seasonality before and after UI, 247-50; relation of interest rates to capital mobility, 360-63 Debt, federal: deficit spending of Keynesian theory, 72; depression-era,67-68; during first Roosevelt administration,77-78; with full employment, 78; during Hoover administration,77-78; Hoover’s policy related to, 74-75,77; predepression views, 70-74; reduction of wartime, 67-69-70; U.S. interwar period (1920-40), 77; during World War II, 68 Deposit insurance: for credit unions, 109-1 3; economists’ support for idea of government-sponsored, 142; higher coverage, 113-15; of New Deal regulations, 13, 87; predepression proposals, 13, 89-90; public acceptance of, 116-17; scenario without, 117-19 Disinflation costs (198Os), 61 Economic regulation. See Regulation Economic Security Act (ESA), 167, 169 Edwards K United Stares (1937), 196n50 EEC. See European Economic Community (EEC) Emergency Banking Act (1933). 28,29 Emergency Relief AppropriationsAct (ERAA, 1935), 169 Emergency TariffAct (1921), 333 Employee Retirement Income Security Act (1974). 112 Employment Act (1946), 1, 18,79-80 Employment exchanges, government-run, 141-42 EPU. See European Payments Union (EPU) ESA. See Economic Security Act (ESA)

European CooperationAdministration, United States, 390 European Economic and Monetary Union (EMU), 62n26 European Economic Community (EEC): creation of (1957), 391; liberalizationof capital movements, 391-92 European Payments Union (EPU), 354,390 Exchange rates: under bilateral financial agreements, 386; controls during and after World War I, 367-72; controls during Great Depression, 372-77; covered interest arbitrage (dollar-sterling),361-63; Keynes’s view of adjustments, 383-84. See also Foreign exchange markets Exchange rates, fixed: under Bretton Woods, 16-17.53; convertibility during Great Depression, 9-10 Exchange rates, floating: of Friedman, 387; with hypothetical collapse of gold exchange standard, 429-30; Keynes’s view of, 383 Exchange StabilizationFund, 27,30, 32 Experience rating system: based on reserve ratio, 245-46; defined, 245; seasonality in formulation of UI, 246-47 Fair Labor StandardsAct (1938), 4 Farm Block lobby. See Interest groups Farm Board, 187, 189 Farmers’ National Relief Conference, 186 Farmers’ Union, 186 Federal Credit Union Act: creation of National Credit Union Administration (1970), 110; to establish federal charters (1934), 109 Federal Deposit Insurance Corporation (FDIC), 90-100 Federal Emergency Relief Administration (FERA), 165-67 Federal Home Loan Bank Board (FHLBB), 105-7 Federalism: banking system under, 13; influence of New Deal on, 11-12, 155-56; U.S. fiscal (1902-92), 156-58 Federal Open Market Committee (FOMC): actions during World War II,42-45; under Banking Act (1935). 38-39; intervention in foreign exchange markets, 56; liquidationistpolicy related to, 35-37; new structure (1935). 38; operation during free reserves policy (1950s), 48-52; under Treasury pressure (1936-37), 38-42

468

Subject Index

Federal Reserve Act (1913): limits to monetization of debt, 57; 1917 amendment, 27; real bills principle under, 59n24 Federal Reserve System: accommodationist view, 35; authority to adjust bank reserve requirements (1935), 27; departure from real bills principle, 57, 59n24; effect of operating strategy (1950s and 1960s). 45, 47-52; free reserves policy, 34-35; under Glass-Steagall,27-28; independence with 1951 accord, 24,44-45; as lender of last resort (1929-33), 59n24, 89; liquidationist view, 35-37; monetary policy (1960s and 1970s),52-59; monetary policy during Great Depression, 25-26; open market operations before and during depression, 25-26; operations in postaccord era, 47-52,57-58; policy under gold standard, 29-30; pre-Great Depression policy, 18; reduced independence in New Deal era, 18.23-24; reforms and structural changes, 3 1-32; reorganization of Open Market Committee (1935), 28; reserve requirement increases (1936-37), 33-34, 37-38; Strong’s leadership, 25-27; World War I1 monetary policy, 42-45. See also Banking system; Federal Open Market Committee (FOMC); Reserve requirements; Reserves, free Federal Savings and Loan Insurance Corporation (FSLIC), 104-9 FERA. See Federal Emergency Relief Administration (FERA) Financial system: in absence of New Deal, 117-19; bilateral agreements during and after World War II, 386; federal insurance of intermediaries in, 87-89; of New Deal era, 13; post-1973 integration of international, 394-96. See also Banking system Fiscal policy: arguments against countercyclical predepression, 70-74; changes in national revenue structure (1960s). 178; discretionary, 82-83; effect of depression on, 17-18; during Great Depression, 74-80; post-World War 11, 80-82; predepression, 67, 69-74, 80-81; as stabilization tool, 83; of unbalanced budget, 79. See also Revenues; Taxation Fiscal system: grant programs as part of (1960s), 178; New Deal transformation of federal, 163-72; post-New Deal federal, 172-74; unemployment compensa-

tion, 228,232-33,236-37,239-40, 243-50 Florida v. Mellon (1927). 240, 243 FOMC. See Federal Open Market Committee (FOMC) Fordney-McCumber tariff (1922). 331, 333-35 Foreign exchange markets: during depression years ( 1931-39), 372-8 1; Federal Reserve intervention in, 56-59; interwar period, 370-72; during World War I, 367-70 Freedom to Farm Act (1996), 182 Free reserves. See Reserves, free General Agreement on Tariffs and Trade (GATT): acceptance by Congress (1947), 326; Annecy Round (1949), 346; Article 19 escape clause, 347; under Bretton Woods agreement, 382; derived from IT0 commercial policy articles, 346; DilIon Round (1961-62), 346; Geneva Rounds (1947, 1955-56), 346; Kennedy Round (1962-67), 346; Torquay Round (1950-51), 346 Glass-Steagall Act (1932): effect on monetary policy, 27-29,42,59n24; effect on monetization of debt, 45-46,57; intent, 29n6 Gold: convertibility under Bretton Woods, 52-56; during depression years (193139), 372-81; effect of Emergency Banking Act on (1933), 28-29; effect of Gold Reserve Act (1934), 27-30; ideology of, 46; inflows after contraction, 32-42; prices and production in model of gold exchange standard (1925-60), 410-29 Gold exchange standard: in absence of Great Depression, 403-5; counterfactual hypothesis, 17,412-14, 447; effect of 1931 financial crisis on, 444; interwar, 367; model of continued (1925-60), 410-29; operation of interwar, 53,443; reconstructed hypothetical,407, 409; Triffin dilemma in hypothetical postwar, 424-29 Gold Reserve Act (1934), 27-30.56 Gold standard: before and after World War I, 71-72,367-68.443; capital mobility in interwar period under, 430n48; as cause of Great Depression, 9-10; collapse of, 61; difference from gold exchange standard, 412; as discipline for U.S. monetary policy, 45,53, 56; international suspension of (1931, 1933), 18, 24, 27, 29,

469

Subject Index

46,367,372; in model of gold exchange standard (1925-60), 410-29; Nixon’s closing of gold window (1971), 18,54; predepression, 24,62; as principle for Federal Reserve, 57; U.S. return to (1934), 29-30 Gold stock, monetary: in absence of Great Depression, 406,408-9; in hypothetical postwar gold exchange standard, 424-29; in model of gold exchange standard (1925-60), 410-29. See also Triffin dilemma Government: cooperation among levels of, 178; growth in New Deal era, 10; growth of federal, state, and local (1902-92), 157-59; New Deal pattern of intergovernmental relations, 178-79; public expenditure as share of GNP (1902-92), 157-58 Government, federal: civilian spending (1900-1990), 130-32; commodity Stockpiles, 198, 199t; fiscal role (1900), 15759, 162-63; New Deal effect on growth of, 125-32; pre-New Deal regulation of commodity markets, 185; purchase of farm surplus production, 193-94; role in unemployment compensation, 228-29; transformation of fiscal role under New Deal, 163-72 Government, state and local: administration of New Deal programs at level of, 168-72; autonomy related to unemployment compensation, 228-29.244; autonomy under New Deal programs, 178-79; erosion of UI eligibility, 234-35; fiscal role (1900). 157-59, 162-63; precedents for social security, 303-4; setting UI income replacement rate, 232-33 Government intervention: in agricultural sector (1930s), 185; in commodity markets, 206; economists’ support of, 11, 136-40; during Great Depression, 7-8; New Deal scope and scale, 12,215-16; predepression ideas related to, 125-26, 132-47; public perception of, 10-1 1; to reduce agricultural supplies, 192-93 Government role: call for increased, 5-6; continuation of New Deal regime (1946-76). 147-48; created by New Deal, 18-19, 172; economists’ predepression ideas of, 136-47; ideological shift related to, 12526, 148-50; liberal ideology associated with, 148; perception during and after World War I, 4; perceptions from Great

Depression, 4-5; predepression UI actions and legislation, 236-43 Grant programs: outside New Deal pattern (1960s). 178 Great Depression: constructing counterfactual to, 404-7; contraction and recovery (1929-39), 7; effect of hypothetical absence of, 446-47; Federal Reserve discretionary policy as effect of, 52-54; influence on US. economy, 6; influences on inflation, 45-47; as international event, 326; legacy for agricultural regulation, 213-20; monetary explanation, 8-10; origins and persistence of, 6-10; spurt in union growth during, 278-87; U.S. fiscal policy during, 74-79 Hoover administration:federal deficits during, 77-78; outlays (FY 1933), 164-65 Humphrey-Hawkins Act (1978), 79 IMF. See International Monetary Fund (IMF) Income: introduction of income tax (1913), 14; maintenance programs, 178; redistribution of New Deal programs, 178; subsidies to farmers (194O-70), 1971153; transfer provision in agriculture regulation (1884-1970), 200-205 Inflation: disinflation (1980s), 61-62; Federal Reserve policy (1950s), 48-49; without gold standard, 52-56; under Keynesian ideology, 60; levels (1965-66), 45, 50-52; persistence of inflationary policy (1960s and 1970s),61 Institutions: 1930s changes in, 27-32; spread of insurance among financial, 87-89 Insurance: for broker-dealers, 13, 109-12; for pension funds, 13, 109-13; for savings and loans, 13, 104-9; social security as form of, 319-20. See also Deposit insurance; Old-age insurance (OAI); Unemployment insurance (UI) Interest groups: effect of New Deal legislation on, 12, 147; farm lobby, 12, 186-93, 197, 199; mobilization to promote commodity purchases, 207; pressure in SmootHawley and RTAA legislation, 341; support for extension of RTAA, 348-49 Interest Rate Control Act (1966). 108 Interest rates: convergence of real, 363-64; Federal Reserve’s Operation ’hist, 55-56; in free reserve operations, 62; under Keynesian ideology, 60; nominal

470

Subject Index

Interest rates (con?.) rates as indicator of capital mobility, 358, 360-63; real interest parity, 365; real rates as indicators of capital mobility, 363-66; Regulation Q ceilings, 107-8; uncovered interest parity, 364 International Clearing Union (ICU): Keynes’s proposed (1944), 383 International Ladies’ Garment Workers’ Union (ILGWU), 286 International Monetary Fund (IMF): Article VIII currency convertibility, 385, 387-88; effect of absence in analysis of hypothetical interwar events, 443-46; EPU countries accept Article VIII convertibility provisions, 390; real-life role of, 444-46; role under Bretton Woods system, 382, 384,444-45; solution to collective action problem, 444 International Trade Organization (ITO), proposed, 346 Keynesian theory, 45,46,60; deficit spending, 72-73; factors reducing multiplier, 78; fall of (1970s). 149; fiscal revolution, 18; influence on government programs debate, 133-34; post-World War 11capital flows, 384-86 Korean War: Treasury borrowings during, 58 Labor force: coverage under social security, 311-13 Labor force, federal: increases by type of agency (1934-38), 127-29; levels of employment in (1908-90), 127-28 Legislation, New Deal. See New Deal legislation; and spec@ laws by name Lever Food Control Act (1917), 186 McNary-Haugen legislation, proposed, 18889,210,213-15 Macroeconomic policy, international: in hypothetical absence of IMF, 445-46; trilemma, 354-55,384 Macroeconomic theory: dominance of Keynesian, 63; rise of neoclassical, 149 Marketing agreements and orders: under AAA (1933), 191n31, 192n36, 196nn49,50 Marshall Plan, 349,354,390,406-7 Massachusetts v. Mellon (1923), 163 Medicaid program, 168 Medicare program, 168

Miller-Tydings Act (1937). 4 Minimum wage concept, 138-40 Monetarism, 149 Monetary policy: under Bretton Woods, 45, 52-56; as cause of Great Depression, 355; after contraction (1933-38), 32-42; effect of institutional reforms (1932-35), 27-32; exchange rate controls (1914-39). 367-81; free reserves to measure monetary conditions, 34-35; without gold standard, 58; gold standard discipline for, 45; before and during Great Depression, 25-32; influence of New Deal legislation on, 59-62; influences on outcomes in (1960s and 1970s), 56-59; in Keynesian ideology, 60; New Deal provisions related to, 38; Treasury Department dominance of (1930s), 39-42,46 Monetary system, international: in absence of Great Depression, 407-10,446-47; gold standard, pre-World War I, 71-72; influence of Great Depression, 17 Monetization: of debt (1930s), 42; of debt (1950s and 1960s), 45-47,58,62-63 Money-gold multiplier: international monetary system without Great Depression, 412-14 Money supply: under Bretton Woods, 52-56; decline during Great Depression, 8; effect of Glass-Steagall Act on growth of, 45-46; growth related to free reserves fluctuations, 49-52; lack of institutional check on growth, 62 Most favored nation (MFN) policy: under GATT rules, 346; in RTAA (1934). 14, 341; unconditional U S . (1923), 33 Munn v. Illinois (1877), 2411136 National Defense Highway Act (1956), 172 National Grange, 186 National Housing Act (1934), 105 National Industrial Recovery Act (NIRA, 1933): AFL union exploitation of opportunity provided by, 286; as basis for New Deal legislation, 4; presidential trade authority under, 342 National Labor Relations (Wagner) Act (1935), 4, 16,265,278-82,287-88 National Milk Producers Federation, 186 New Deal: alteration of financial system by, 117; financing of, 178; impact on agricultural regulation, 183-85; influence on fed-

471

Subject Index

eralism, 155-56; programs of, 4; transformation of federal fiscal role under, 163-72 New Deal agencies: civilian spending (1930s), 130-32; growth after World War II, 10; growth during New Deal, 127-29; in U.S. Department of Agriculture, 183-85 New Deal era: centralizationof public sector, 178; state and local autonomy during, 178-79; union growth during, 280-82 New Deal legislation: analysis of agricultural, 200-205; deposit insurance, 13, 87-100, 104-19, 142; effect of, 12; effect on savings and loan institutions, 105-6; extensions (1940-70), 197; influence on Federal Reserve behavior, 38,59; link to 1962 agricultural regulation, 217-1st; related to agriculture, 181nl New Deal programs: agricultural sector, 182-85; FERA spending, 165-67; grant programs, 163-72, 174-76; influence on federalism, 156-57; post-World War I1 expansion of some, 172; spending for relief and highway programs, 165-66; state government role in, 12 New York Stock Exchange (NYSE): trust fund (1964), 111 NIRA. See National Industrial Recovery Act (NIRA, 1933) Nixon administration:expansion of New Deal programs, 173; general revenue sharing, 173 N U B v. Jones and Laughlin Steel Co. (1937). 196n50,241,281n2 OASI. See Old-age and survivors insurance (OASI) Old-age and survivors insurance (OASI), 301 Old-age assistance (OAA): benefits, 307; at federal government level, 242n39; replaced by federal Supplemental Security Income program, 3 16; under Social Security Act, 298-99; state-level precedents, 303-4,307 Old-age insurance (OAI): absence of means testing, 307; conditions for receiving benefits under, 300-303,307-8; at federal government level, 242n9; financing of, 299; precedents for, 304; under Social Security Act, 15,298-99. See also Old-age and survivors insurance (OASI) Operation mist, 55-56

Pension Benefit Guaranty Corporation (PBGC), 112 PPP. See Purchasing power parity (PPP) President, U.S.: reciprocal trade agreement negotiating authority, 14, 349; role in tariff setting, 329, 331-32, 350; tariff reduction agreements under RTAA, 325 Prices, agricultural: income subsidies to support, 197n53, 198; New Deal impact on wheat prices, 209-13,219-20t Propensity score (capital controls), 432-33 Protectionism: Hoover's position, 338; 1929-32 period, 337; post-Civil War Republican position, 328-3 1, 344-49; Smoot-Hawley Tariff Act (1930). 333-37 Public sector: centralization during New Deal, 178; fiscal centralization (1902-92), 157-59; growth with intergovernmental grants, 174-76; transformation under New Deal, 163-72 Purchasing power parity (PPP), 364 Reagan administration: cutback in New Deal grant programs, 173; deficits under, 68; Reagan Revolution, 150 Real bills doctrine, 57 Recession (1937-38), 37-38 Reciprocal Trade Agreements Act (RTAA, 1934). 325-26; Collier bill, 337-38; implementation and effect of, 342-44; passage of, 340; political opposition to, 340, 344-45; provisions, 341 Reconstruction Finance Corporation, 4 Regulation: of agricultural sector, 181-85; cascading, 89; deregulatory policy of Alfred Khan, 149; growth of agricultural regulation under New Deal, 200-207 Regulation Q: ceilings of, 107-8; savings and loans under, 108 Reserve ratio. See Experience rating system Reserve requirements: Federal Reserve increase of (1936-37), 32-34,37-38; reserves during free reserve policy period, 35. See also Reserves, free Reserves, free: levels (1950s and 1960s), 49-52; policy guide (1920s and 1930s), 34-35 Resource allocation: in absence of Bretton Woods, 410; in counterfactualof postWorld War 11,430-43 Revenues: federal-state revenue sharing, 173, 174, 178; growth of state and local

472

Subject Index

Revenues (con?.) (1927-92). 179; historical generation by tariffs, 333; during recession, 78 Robinson-Patman Act (1936), 4 Roosa bonds, 56 Roosevelt administration: deficits during first, 77-78; idea of unbalanced budget in second, 79; New Deal, 2; NIRA influence on union growth, 286; outlays in New Deal years, 164-65 RTAA. See Reciprocal Trade Agreements Act (RTAA, 1934) Savings and loan associations (S&Ls): collapse of industry, 115-16; federal deposit insurance for, 104-9; under Federal Home Loan Bank System, 105-7; governed by Regulation Q (1966), 108; pressure for higher insurance coverage, 113 Schecter Poultry Corp. v. United States (l935), 240,281n2 Seasonality: before and after unemployment insurance, 247-50 in UI experience rating system, 245-47 Securities Act (1933), 7 Securities and Exchange Commission (SEC): authority under Securities Investor Protection Act, 112; economists’ support for idea of, 140-41 Securities Exchange Act (1934), 7; selfregulation under, 110-1 1 Securities Investor Protection Act (SIPA, 1970), 111-12 Silver Purchase Act (1934), 28, 3 1 Smoot-Hawley Tariff Act (1930). 8; effect of, 333-37; effect on tariff escalation (192932), 326-27; RTAA as amendment to, 341; specific duties in, 347; tariffs under, 325 Social programs: expansion (1960s and 1970s). 12; of New Deal, 5, 14. See also Agricultural Adjustment Act (1933); Social Security Act (1935); Unemployment insurance (UI); Workers’ compensation (WC) Social Security Act (1935): amendments (1937). 147; amendments (1939). 300303, 307; amendments (1977), 314-15; development of unemployment compensation provisions, 239-45; earnings test for OASI, 315-16,320 economic factors at time of enactment, 308-1 1; extensions

of OASI coverage (1950), 3 11-15; factors influencing changes in, 3 1 1-2 1; intent of unemployment compensation, 228,230; legislated benefit changes, 3 13-15; Medicaid program, 168; Medicare program, 168; old-age assistance (OAA), 298-99; old-age insurance (OAI), 298-99; Supplemental Security Income program (SSI), 316; tax increases for OASI, 312-13,316-17; trust fund accumulation, 306-7, 317-18; unemployment compensation under, 243-45. See also Old-age and survivors insurance (OASI); Old-age assistance (OAA); Oldage insurance (OAI); Unemployment Insurance (UI) Social security system: expanded form, 168; welfare programs of, 167-68 Soil Conservation and Domestic Allotment Act (1936), 195, 211 Soil conservation policy, 195, 196n49 Spending, government: Agriculture Department (1921-70), 184; by federal New Deal agencies, 131-32; federal outlays to civilians ( 1900-90), 130-3 1; influence of intergovernmental grants on state and local, 174-76; for regulated agricultural programs, 205-7; share of federal, state, and local (1902-92). 157-62; under social security system (1992), 168; USDA and CCC civilian (New Deal period), 208-9 Stabilizers, automatic: built into federal budget, 78; of Committee on Economic Development, 68, 82; post-World War 11, 80-84 Stock exchange funds, 1I1 Taft-Hartley Act (1947), 265 Tariff Act: 1897, 331-32; 1913, 328, 332 Tariffs: actions related to reform, 330; congressional setting of, 328-32; Dingley tariff (1894). 328; Emergency Tariff Act (1921), 333; escalation (1929-32), 326-27; Fordney-McCumber tariff (1922), 331, 333-35; legislation with limited reciprocity, 33 1-32; McKinley tariff (1890), 332; political issue after Civil War, 327-32; post-Civil War rates, 328; predepression and depression era political debates, 327-50; reductions under GATT negotiations, 346-47; under

473

Subject Index

RTAA mechanism, 326-27.347; under Smoot-Hawley Act (1930), 325, 333-37; Underwood tariff (1913), 328,332.334 Taxation: OASI rates legislated and actual, 3 12-13; for old-age insurance, 299-300; payroll tax financing of social security, 304-5,316-17; state and local levels (1930s), 178; UI experience rating system, 245-46 Tennessee Valley Authority (TVA) concept, 136-37 Thrifts. See Savings and loan associations (S&LS) Trade: during depression (193 1-39), 372-81; effect of post-World War I1 currency inconvertibility, 386-91; with post-1973 liberalization, 394-96; post-World War I1 arrangements, 345-46; under SmootHawley Tariff Act (1930), 335-37 Trade agreements: participants under RTAA, 342-44; post-World War I1 GATT, 346-49; under RTAA (1934). 14,325, 341-44.350 Trade policy: influence of Great Depression on, 350; partisan political debates, 328-49; public opinion (1945). 348-49; under RTAA (1934), 14.325-26.350. See also General Agreement on Tariffs and Trade (GATT); Protectionism; Reciprocal Trade Agreements Act (RTAA, 1934); Smoot-Hawley Tariff Act (1930); Tariffs Treasury Department: Exchange Stabilization Fund, 27, 30, 32, 38,56; under Gold Reserve Act (1934). 30-31; influence on Federal Reserve policy making, 38-42, 46; wartime borrowings (1950s), 58; World War I1 monetary policy, 42-45 Treasury-Federal Reserve Accord (1951), 24, 444557 Treaty of Rome (1957), 391 Triffin dilemma: collective action posed by, 387,443-44.447; in hypothetical postwar gold exchange standard, 424-29, 443,447 Truman administration, 346 UIP. See Uncovered interest parity (UIP) Uncovered interest parity (UIP), 364 Unemployment: Humphrey-Hawkins Act focus on, 79; ideas to decrease depressionera, 145-46; rates during U S . interwar

period (1920-40), 77; as rationale for urging depression-era retirement, 3 10-1 1 Unemployment insurance (UI):benefits and experience rating at state level, 251-60; counterfactual explanation, 242; coverage under, 15, 230; criticism of U.S. system, 260; early national- and state-level legislation, 14, 236-39; economists’ support for idea of, 142-45; eligibility, 233-35; experience rating feature in US., 15, 227-29.244.245-50; extended benefits provisions, 231n10; intent of legislation, 228, 230; legal replacement rate schedule, 232-33; limited duration of benefits under, 228, 23 1; as New Deal program, 14, 165, 168; under Social Security Act, 243-45; U S . system compared to OECD countries, 235 Union growth: AFL unions, 285; CIO unions, 283; density estimates, 289-93; during depression, 15-16; models generating spurts in, 273-78; spurts in developed countries, 271-73; spurts in US., 267-71, 278-87 United Mine Workers (UMW), 286 UnitedStates v. Butler (1936), 195, 240 U S . Congress, Joint Economic Committee, 79 U S . Department of Agriculture (USDA): Agricultural Adjustment Administration (AAA), 190; Bureau of Agricultural Economics (1922). 189; Dairy Bureau, 189; Division of Cooperative Marketing (1926), 187; impact of New Deal on, 183-85, 208-9; pre-New Deal regulation (1884-1932), 185 U.S. Food Administration, 186 U.S. Gold Commission Report (1982), 413n26,415,423 U.S. Tariff Commission (USTC): creation (1916), 330-31; investigations (192229), 331n4 Vietnam War, 58 Wagner Act (1935). See National Labor Relations (Wagner) Act (1935) Wagner-Lewis bill (1934), 239-40 War Finance Corporation, 4 War on Poverty (1960s), 12,172,178 Welfare programs. See Social programs West Coast Hotel Co. v. Parrish (1937), 2411135

474

Subject Index

Wilson administration, 80 Workers’ compensation (WC), 235 Works Progress Administration (WPA), 16566, 169-71 World Bank, 382 World War I: agricultural policy during, 12, 182, 186; controls on trade, 328; exchange rates and price levels during,

367-70; influence on 1930s changes, 4 World War 11: cementing of New Deal programs, 4; exchange controls, 381-82; Federal Reserve monetary policy during, 42-45; spurt in U.S. union growth, 278-87; trade agreements programs during, 345

This Page Intentionally Left Blank