Politics and Economics in the Eighties 9780226012827

Citation preview

Politics and Economics in the Eighties

~~~

~

A National Bureau of Economic Research Project Report

Politics and bconomcs in the Eighties

Edited by

Albert0 Alesina and Geoffrey Carliner

The University of Chicago Press

Chicago and London

ALBERTOALESINAis the Paul Sack Associate Professor of Political Economy at Harvard University. GEOFFREY CARLINER is executive director of the National Bureau of Economic Research.

The University of Chicago Press, Chicago 60637 The University of Chicago Press, Ltd., London 0 1991 by the National Bureau of Economic Research All rights reserved. Published 1991 Printed in the United States of America 00999897969594939291 5 4 3 2 1 Library of Congress Cataloging-in-Publication Data Politics and economics in the eighties I edited by Alberto Alesina and Geoffrey Carliner. p. cm.-(A National Bureau of Economic Research project report) Includes bibliographical references and indexes. ISBN 0-226-01280-8 (cloth : alk. paper).-ISBN 0-226-01281-6 (pbk. : alk. paper). 1. United States-Economic policy-1981-Decision making. 2. Economics-Political aspects-United States. I. Alesina, Alberto. 11. Carliner, Geoffrey. 111. Series. HC 106.8.P65 1991 338.973’009’048-dc20 9 1-3928 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 239.48-1984.

National Bureau of Economic Research Officers Geoffrey Carliner, executive director Charles A. Walworth, treasurer Sam Parker, director offinance and administration

George T. Conklin, Jr., chairman Paul W. McCracken, vice chairman Martin Feldstein, president and chief executive oficer

Directors at Large John H. Biggs Andrew Brimmer Carl F.Christ George T. Conklin, Jr. Kathleen B . Cooper Jean A. Crockett George C. Eads Morton Ehrlich

Martin Feldstein George Hatsopoulos Lawrence R. Klein Franklin A. Lindsay Paul W. McCracken Leo Melamed Michael H. Moskow James J. O’Leary

Robert T. Parry Peter G. Peterson Robert V. Roosa Richard N. Rosett Bert Seidman Eli Shapiro Donald S. Wasserman

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

James L. Pierce, California, Berkeley Andrew Postlewaite, Pennsylvania Nathan Rosenberg, Stanford Harold T. Shapiro, Princeton Craig Swan, Minnesota Michael Yoshino, Harvard Arnold Zellner, Chicago

Directors by Appointment of Other Organizations Rueben C. Buse, American Agricultural Economics Association Richard A. Easterlin, Economic History Association Gail Fosler, The Conference Board A. Ronald Gallant, American Statistical Association Robert S . Hamada, American Finance Association David Kendrick, American Economic Association

Ben E. Laden, National Association of Business Economists Rudolph A. Oswald, American Federation of Labor and Congress of Industrial Organizations Dean P. Phypers, Committeefor Economic Development Douglas D. Purvis, Canadian Economics Association Charles A. Walworth, American Institute of Certified Public Accountants

Directors Emeriti Moses Abramovitz Emilio G. Collado Frank W. Fetter

Thomas D. Flynn Gottfried Haberler Geoffrey H. Moore

George B. Roberts Willard L. Thorp William S. Vickrey

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 the 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

Preface

Introduction Albert0 Alesina and Geoffrey Carliner 1.

2.

3.

Elections and the Economy in the 1980s: Shortand Long-Term Effects Moms P. Fiorina Comment: William D. Nordhaus Leaning into the Wind or Ducking out of the Storm: U.S. Monetary Policy in the 1980s James E. Alt Comment: Benjamin M. Friedman Party Governance and U.S. Budget Deficits: Divided Government and Fiscal Stalemate Mathew D. McCubbins Comment: Robert J. Barro

ix 1

17

41

83

4.

Changes in Welfare Policy in the 1980s John A. Ferejohn

5.

The Politics of Tax Reform in the 1980s Charles H. Stewart I11 Comment: David F. Bradford

143

6.

Political Foundations of the Thrift Debacle Thomas Romer and Barry R. Weingast Comment: Robert E. Litan

175

123

viii

7.

8.

Contents

The Spatial Mapping of Minimum Wage Legislation Keith T. Poole and Howard Rosenthal Comment: Charles Brown

215

U.S. %ade Policy-making in the Eighties 1. M. Destler Comment: Anne 0 . Krueger

25 1

Contributors

285

Name Index

287

Subject Index

292

Preface

We have tried to accomplish something unusual in this volume: an interdisciplinary analysis of economic policy-making that balances new empirical research with a coherent review of the most important policy changes that occurred in the eighties. Our experience in organizing the conference and editing the book suggests that this volume may generate more passionate reactions than other NBER volumes for several reasons. First, most of us who lived through the 1980s in America are likely to have strong views on taxes, welfare spending, budget deficits, the savings and loan debacle, and other aspects of the decade’s economic policy. Many readers will probably disagree with the arguments or the emphasis of some of the papers. However, these readers may find themselves in agreement with the comments on the papers, which sometimes offer dissenting views. In our introduction, we tried both to summarize the basic arguments of each paper and comment and to offer our own point of view. Second, this book is interdisciplinary: leading political scientists have written the papers, and leading economists have written comments. We believe that it is important to have both disciplines represented in a study of economic policy-making. The economists bring a deeper understanding of the significance of changes in these policies and their effects on the economy. The political scientists bring a deeper knowledge and understanding of the political institutions and forces that led to these changes. Especially in this area, we think that the strengths of the two disciplines are an essential complement of each other. The background, language, modeling strategy, and choice of emphasis are often quite different between the two disciplines. Therefore, we asked the participants in this project to write for a wide audience, which includes both members of the two disciplines and a wider public of interested readers.

ix

x

Preface

Finally, these papers include descriptions and explanations of the most important policy changes that occurred in the eighties, along with some technical modeling and hypothesis testing. Different authors have chosen different combinations of these ingredients. We hope that readers who are familiar with the policy changes in the eighties will find enough in the way of new research results, and that readers looking for descriptions and explanations of policy developments will not be put off by the technical sections of the papers. We would like to thank the Andrew Mellon Foundation for financial support. Kirsten Foss Davis and Ilana Hardesty made sure that the conference went smoothly, and Mark Fitz-Patrick helped prepare the manuscript for publication. We would especially like to thank Candace Morrissey for her assistance in the entire process.

Introduction Alberto Alesina and Geoffrey Carliner

When Ronald Reagan won the presidential election of 1980 and the Republicans gained control of the Senate for the first time since 1954, did American voters reject past policies and indicate a permanent shift to the right? Or was the 1980 election a predictable response to the perceived mismanagement of the economy in the late seventies with no long-run implications? Did the economic policies that were adopted during the 1980s reflect a long-term shift toward conservatism, or were they merely extensions of past policies with, at most, a short-lived shift to the right during the two years (1981-82) when Republicans had effective control of the White House and Congress? More generally, how did politics influence the economic policy decisions in the 1980s? The eight chapters included in this volume examine the evidence. They study voting patterns, monetary and fiscal policies, welfare spending, tax reform, minimum-wage legislation, the savings and loan debacle, and international trade policy. Taken together, they indicate that a sharp temporary shift to the right followed the 1980 election, as evidenced by the policies adopted during the early years of the first Reagan administration. Subsequently, the Democratic gains in the congressional elections of 1982, 1984, and especially 1986, contributed to moderate the administration’s policies.

voting Morris P. Fiorina’s analysis of voting patterns in the 1980s supports this view. He notes that Jimmy Carter faced the 1980 election with double digit Alberto Alesina is the Paul Sack Associate F’rofessor of Political Economy at Harvard University. Geoffrey Carliner is executive director of the NBER. This work was supported by a grant from the Andrew Mellon Foundation.

1

2

Alberto Alesina and Geoffrey Carliner

inflation and a stagnant economy. Carter probably would have lost in 1980 on economic issues alone, even leaving aside the problem of the hostages held by Iran. In fact, several studies of previous elections have emphasized that slow growth and (to a lesser extent) high inflation significantly hurt incumbent presidents.’ Thus, Carter’s defeat does not imply a sharp realignment of the electorate toward the right. The Democrats’ loss of the Senate in 1980 was more unusual than their loss of the presidency, but again, it was a reaction to the specific events of the late 1970s rather than the result of a long-term shift in party allegiance. In fact, the large number of Republican victories in the Senate was not repeated in subsequent elections. In particular, the elections of 1982 and 1986 continued the pattern of midterm voting cycles in which the party holding the White House loses votes and seats in Congress. The presidential elections of 1984 and 1988 are also consistent with historical patterns. Reagan’s reelection and Bush’s victory confirm that voters favor incumbents and look at the rate of change in income and in prices in the election year when choosing presidents.2 In summary, Fiorina finds that the election results of the 1980s can be largely explained by the short-run performance of the economy in election years, and by the midterm cycle. In his comments to this paper, William D. Nordhaus notes that the apparent popularity of Republicans and their grip on the presidency could evaporate at the first reappearance of recession or inflation. However, Fiorina also reports a longer-term shift in voters’ attitudes in favor of the Republican party, especially in the upper middle class. The percentage of voters who identified themselves as Republicans in surveys rose from about 33 percent in 1962-82 to about 40 percent in 1984-88. Measures of party identification are obtained by asking voters in surveys whether they perceive themselves as Democrats (weakly or strongly), Republicans (weakly or strongly), or independents. Party identification reflects the voter’s ideological bias in favor of a party and is influenced by several economic and noneconomic factors. Thus, voters may identify themselves as Democrats, even if they occasionally vote for a Republican candidate: party identification reflects long-run attitudes of the electorate and may differ from the actual voting behavior in each election. For instance, Jimmy Carter in 1980 was defeated when many voters who identified themselves as Democrats voted for Ronald Reagan as a reaction to Carter’s perceived failures. Fiorina’s results suggests that some of these voters may now consider themselves “permanently” Republican. Voting in the eighties, particularly in the second half, also confirms a pattern that is becoming increasingly common in American political history: divided government-that is, a situation in which the same party does not control the presidency, the House, and the Senate. One explanation, formalized elsewhere by Fiorina and others, suggests that divided government is the result of a conscious attempt by the voters to achieve moderate policy.3Accord-

3

Introduction

ing to this view, the American electorate has to choose between two parties that offer different policy options. Voters in the middle of the political spectrum desire policies in between those advocated by the median members of the two parties. These voters may prefer divided government rather than unified government: if different parties hold the presidency and Congress, they balance each other, and the resulting policy outcome is more centrist than would be with a unified government. This desire for moderation may explain both split-ticket voting when Presidential and Congressional elections are held simultaneously, and the midterm voting cycle. Both phenomena occurred in the 1980s; in this respect the last decade was far from unusual. In fact, a recurring theme of this book is the constraint placed on the ability of the Reagan administration to pursue conservative policies after the 1982 elections increased Democratic strength in the House, and especially after the Democratic party regained the majority in the Senate after the 1986 mid-term elections.

Monetary Policy James E. Alt’s study in this volume of monetary policy in the 1980s also emphasizes the importance of the balance of power between Congress and the president. He suggests that the Federal Reserve can be viewed as an agent with three principals: the president, Congress, and the financial community. The chairman of the Federal Reserve has a certain amount of independence, which he wants to preserve for himself and for the institution. Alt emphasizes that an agent with multiple principals enjoys a fair amount of autonomy when the principals disagree, which may be the case when the presidency and Congress are held by different parties. In addition to the goal of independence, Fed chairmen may want to enhance their chances of reappointment. Alt examines how different politicoeconomicmodels can explain the Fed’s behavior, given this principal-agent relationship and the chairmen’s goals of independence and reappointment. Earlier studies have suggested two explanations for political influence over monetary policy. The first suggests that Democratic administrations are more expansionary than Republican administrations because the former care relatively more about unemployment (and less about inflation) than the latter.4 The second emphasizes that presidents are almost exclusively concerned with reelection and thus engage in preelectoral manipulations of monetary policy to boost g r ~ w t h . ~ Alt blends these two approaches by arguing that, at the beginning of a new administration, Fed chairmen tend to accommodate the president’s partisan goals. When an election approaches, the Fed tends to follow a prudent course of action for two reasons. On one hand, the Fed wants to avoid preelectoral contractions in order not to jeopardize the incumbent’s performance at the polls. On the other hand, the Fed avoids policies that are too clearly expan-

4

Albert0 Alesina and Geoffrey Carliner

sionary and thus favorable to the incumbent. In fact, the Fed wants to preserve its reputation for independence and avoid displeasing the challenger who, after all, may be the next president. Finally, the Fed may need to balance these political influences with the financial community’s preference for stability in financial markets. Monetary policy in the eighties was largely consistent with traditional Fed behavior, as described by these models. By 1979, inflation was perceived as the most serious economic problem, and the Carter administration and the Federal Reserve were held at least partly responsible for it. In the fall of 1979, one year before the presidential elections of 1980, Carter needed to support anti-inflationary policies to avoid losing control of the economy and of the election.6Therefore, he appointed Paul Volcker as chairman of the Fed, who immediately shifted monetary policy to target the money supply instead of interest rates and began to slow the growth of the money supply to reduce inflation. In 1981 and 1982, with a new Republican president and a conservative Congress, the Fed received full political support for its tough, anti-inflationary policy. In these years, the United States experienced the deepest recession since the thirties, but inflation was quickly reduced. The weak Democratic contingent in Congress could do little to oppose this course of action. In the summer of 1982, the Fed loosened its policy and started to rely less on monetary targeting. This change of policy can be explained by the feeling that inflation had been reduced substantially and the economy needed help to recover. Also the threat of financial crises due to the less developed countries’ (LDCs’) debt problem and the difficulties of the savings and loan institutions (S&Ls) may have influenced the Fed’s decision. After 1982, Democratic gains in Congress reduced the political support for further anti-inflationary policies. In his comments on Alt’s paper, Benjamin M. Friedman notes that the Fed decreased its reliance on monetary targeting after 1982 because of the collapse of the relation between money growth and the growth of nominal income. In addition, Friedman emphasizes how the Fed’s independence is enhanced not only by the multiplicity of principals, but also by the fact that each of the principals (Congress, the administration, and the financial community) may not have homogeneous preferences.

Fiscal Policy: Taxing, Spending, and the Deficit Some of the most dramatic changes in economic policy during the 1980s involved fiscal policy, particularly in the first two years of the Republican administration. Military spending grew substantially, many social programs were cut back; tax rates fell sharply and the federal budget deficit soared. Some of these policies reflect traditional Republican views; in addition, some of these changes, like an increase in military spending, were initiated at the end of the Carter administration. However, Reagan’s fiscal policies in

5

Introduction

1981-82 viewed together represent a substantial departure from the seventies, including previous Republican presidents. One dramatic piece of evidence of Reagan’s departure from previous fiscal policy is the budget deficit. The most economically relevant measure of accumulated deficits is the debt/GNP ratio. This ratio shows a downward trend in the post-World War I1 period. Starting from a peak of 1.17 in 1945, this ratio was 0.23 in 1980. The large deficits in the eighties clearly reversed this trend: the debt/GNP ratio was 0.35 in 1985 and about 0.30 in 1989.’ The deficit’s share of GNP peaked at 6.3 percent in 1983, and then fell gradually for the next six years, particularly after the defense buildup slowed after 1986. By 1989, the deficit was 2.9 percent of GNP, about its share of output at the beginning of the decade. Republican presidents traditionally have vigorously opposed budget deficits. In contrast, Reagan spoke out most strongly against tax increases and used his political power to oppose legislation that would raise tax rates and revenue, even if that opposition meant continued high deficits. Congressional Democrats, traditionally less concerned with deficits than Republican presidents, refused to cut spending, even though that also meant high deficits. Mathew D. McCubbins in his contribution to this volume suggests that the increase in budget deficits was the result of “divided government,” particularly of the control of the Senate and the House by different parties. He suggests that such divided control leads to fiscal deadlocks and prisoners’ dilemmas, which resulted in spending in excess of tax revenues. Republicans and Democrats in Congress favor different spending programs. If the two legislative branches are controlled by different parties, they reconcile their differences by increasing spending on both parties’ favored programs. The result is higher spending than would occur if either party had undivided control of the legislature. A different explanation is that the large deficits in the early eighties were the result of miscalculations. First the supply-siders in the administration vastly overestimated the incentive effects of the 1981 tax cut. Second, the recession of 1982 turned out to be the worst in post-World War I1 history and worse than was predicted. Third, inflation fell faster than anticipated, further decreasing revenues below their expected levels. The indexation of the individual income tax, which took effect in 1985, permanently eliminated the automatic source of increased revenues brought by inflation. When revenues fell sharply below planned expenditures, political and legislative inertia and the combined effects of various pressure groups made it difficult to correct the mistake. That is, politically costly decisions needed to stop the growth of the debt were postponed because of the difficulty of reaching a compromise over budget cuts or tax increases of the size needed to balance the budget. A third political explanation, which does not rely on large miscalculations, is that the debt accumulation of the early eighties was a strategic attempt by the Reagan administration to constrain spending by future Democratic Con-

6

Albert0 Alesina and Geoffrey Carliner

gresses or future Democratic presidents. By creating large deficits whose effects would persist after the Republicans lost control of the Senate, the Reagan administration reduced the flexibility available to Democratic legislators to increase nonmilitary spending. With a large deficit and a rapidly increasing interest bill, it is difficult to justify major increases in spending programs, particularly if tax increases are viewed as economically and politically costly. Thus, the debt is the legacy of the Reagan administration to the f ~ t u r eAc.~ cording to this argument, Reagan accepted large deficits because he was more committed to a reallocation of spending priorities than previous Republican presidents. In his comments on McCubbins, Robert J. Barro argues that a large part of the history of the U.S. budget can be explained by the tax-smoothing model.lo This theory suggests that it is optimal to finance unusual increases in spending, such as those occurring during wars or severe recessions, primarily with debt rather than with tax revenues. Smooth tax revenues and fluctuating deficits (and surpluses) distort the economy less than balanced budgets and fluctuating taxes. Indeed this theory is broadly consistent with the steady decline, with minor deviations for recessions and local wars, in the ratio of debt to GNP since its peak at the end of World War 11. Whether or not the tax-smoothing model explains the deficits in the 1980s is debatable: the debt/GNP ratio almost doubled in a decade without major wars. The military buildup of the eighties can be viewed as once-and-for-all effort to win the Cold War, for which it would have been optimal to run temporary deficits. The collapse of the Soviet bloc lends some support to this view, but recent events in Eastern Europe have deeper roots than the American military buildup. Furthermore, increased military spending accounts for only a fraction of the increase in the deficit. The tax cut of 1981 played a larger role: the tax-smoothing model may explain why deficits may have grown to finance the military buildup in a cold war period, but it cannot explain why taxes were cut. One of the recurrent themes in these explanations of the deficits is the politicoeconomic struggle over allocation of government expenditure to domestic versus defense programs. Thus, it is important to analyze whether this allocation has substantially changed in the eighties. Although the military buildup was a major theme in Reagan's agenda, the buildup was in fact started by President Carter. Military spending in real terms fell steadily from 9.6 percent of GNP in 1968 to 4.8 percent of GNP in 1977. After two years of constant real spending, military spending increased in response to the Soviet invasion of Afghanistan. The Reagan administration sharply accelerated this buildup by raising military spending from 5.3 percent of GNP in 1981 to 6.5 percent in 1986, when Democrats regained control of Congress and ended the increases. By the end of the decade, before the startling events of the fall of 1989, military spending had fallen to less than 6 percent of GNP. The military buildup was accompanied by reduction of spending in various

7

Introduction

social programs, particularly in the first half of the first Reagan administration. As John A. Ferejohn observes in his contribution to this volume, the programs that were cut the most or eliminated entirely, were those narrowly directed at the poor, such as CETA, Job Corps, Head Start, and other education programs. Spending on broadly based social insurance programs that benefited primarily the middle-class elderly, such as Social Security and Medicare, continued to grow untouched during the 1980s. Ferejohn attributes this difference in treatment to the structure of the legislative process. He argues that the structure of Congressional committees insulated middle-class programs from attacks; thus, advocates of those programs could resist administrationpressure for cuts more effectively than could supporters of programs beneficial to the poor. Ferejohn also notes that the administration, to a certain extent, managed to circumvent congressional opposition by tightening eligibility standards administratively. In addition to congressional committees, the pressure of public opinion created great obstacles to any attempt of reducing Social Security benefits. There was no similar outcry of voters when the administration proposed cutting poverty programs. Ferejohn observes that, after the high water mark of Republican power in 1981-82, Democrats resisted further cuts in social programs. Even before they regained control of the Senate in 1986, they were able to stop further efforts by the Reagan administration to change federal spending patterns. He concludes the social spending cuts of Reagan administration were less widespread than commonly perceived and concentrated in a two-year period. Ferejohn also emphasizes the importance of the Congressional committee structure in protecting middle-class programs from further cuts advocated by the administration. It is interesting to note that once the Democrats regained control of Congress, they made few attempts to introduce new poverty programs or to return existing ones to their former size. In part, this may have been due to the constraints in spending imposed by the accumulated deficits of the eighties. On the other hand, it may indicate a changing attitude: American voters and politicians were willing to support such spending under both Republican and Democratic presidents from the early 1960s until 1980, but during the 1980s this support diminished. This is in sharp contrast to the continued strong support for middle-class entitlement programs, which remained essentially untouched throughout the decade. Tax policy also underwent substantial and frequent change during the 1980s. Although the changes that were passed into law, and the debates associated with them, reflect the traditional views of Republicans and Democrats, the emphasis of the Reagan administration was somewhat different from those of Eisenhower, Nixon, and Ford. Charles H. Stewart 111, in his contribution to this volume, describes the major tax changes that occurred during the decade. The largest was the Eco-

8

Albert0 Alesina and Geoffrey Carliner

nomic Recovery Tax Act of 1981 (ERTA), passed in the first months of the new administration at the peak of Reagan’s power. ERTA cut marginal tax rates, substantially increased tax incentives for corporate investment, and indexed the personal income tax for inflation. The following year, Congress had second thoughts about the size of the 1981 tax cut, and repealed many of the business tax incentives that had been introduced in ERTA. Excise taxes on cigarettes and telephone calls were also raised. Then, in 1983, taxes were raised again, as new federal employees were subject to the Social Security payroll tax, and the payroll tax rate and the maximum income subject to this tax were increased. Finally, the 1986 Tax Reform Act (TRA) continued the trend to lower marginal tax rates but eliminated many of the tax incentives for business investment that had been favored by Republicans for many years. For instance, the 1986 reform eliminated the investment tax credit and raised the maximum tax on capital gains from 20 percent to 28 percent. TRA was designed to be revenue neutral: its significantly reduced personal income tax revenues and offset this loss with increased corporate income tax receipts. The end result of all this legislation was a major change in the structure of federal taxation. The marginal tax rate on the highest income bracket fell from 70 to 28 percent, many low-income individuals were dropped from the income tax roles, payroll tax rates and ceilings rose, some personal income tax deductions were scaled back, and many corporate tax benefits were eliminated. A recurring theme in the debates over tax policy during the eighties was over the need to raise tax revenues after the very large reductions put in place by ERTA. If these reductions had not been so large, the debate over tax policy probably would not have been so sharp. Stewart emphasizes the decreased control by Congressional leaders in setting tax policy as an explanation for the size of the 1981 tax cut. When Republican and Democratic leaders in Congress reasserted their control in later years, tax favors to special interests were reduced. Stewart also notes that in previous decades inflation pushed taxpayers with constant real incomes into higher tax brackets and thus automatically raised real tax revenues. Congress could then decide to spend these revenues or to take credit with constituents by voting to decrease taxes. When ERTA eliminated this bracket creep by indexing the personal income tax, revenues no longer increased automatically. Any change in taxation that raised total revenue or cut the taxes of one group required an explicit action to raise taxes on another group. Thus, indexing the personal income tax, along with Reagan’s opposition to tax increases, also intensified the debate over tax policy. David F. Bradford’s comment on Stewart’s paper discusses the paradox of a Republican administration introducing tax changes in ERTA that were extremely generous toward business investment and then supporting their elim-

9

Introduction

ination only five years later in TRA. The explanation for this reversal in policy may lie in Reagan’s commitment to lowering marginal tax rates. He was willing to abandon the investment incentives traditionally favored by Republicans in order to bring down the tax rates. This decline in the top marginal tax rates coupled with the change in spending priorities indicate a shift of emphasis in fiscal policies toward “efficiency” at the expense of more “inequality.”

Deregulation An important theme of the Reagan administration’s agenda was the need to deregulate the economy. At best, the Reagan administration claimed, government interference imposed costs on workers and consumers that far outweighed the benefits to society. More often, government regulations simply created red tape that confused the efficient operation of market forces while restricting personal freedom. Cutting regulation, along with cutting social spending and marginal tax rates, was thus an area of major policy change during the 1980s. Like other policy shifts examined in this volume, the deregulation of the Reagan administration was preceded by actions taken by earlier presidents. In particular, the Carter administration eliminated long-standing government regulation of airlines, railroads, trucking, and telecommunications. This policy was designed to encourage competition in industries where regulation was perceived as benefiting the regulated producers. The purpose of President Carter’s deregulation was to reduce prices and improve service for consumers by increasing competition. In contrast, the deregulation of the Reagan years involved the loosening of health, safety, environmental regulations which were considered unnecessary and inefficient, more generally constraining “market forces.” Much of this deregulation was accomplished through changes in enforcement rather than changes in law. An important and long lasting legacy of the eighties is the collapse of much of the savings and loan industry and the cost to the government of bailing out depositors. The industry’s problems began in the 1970s, when inflation and nominal interest rates rose sharply. Since most S&L assets were in long-term fixed-interest residential mortgages, a large number of firms in the industry suffered a loss of all their equity. Even after interest rates fell in the 1980s, the market value of the liabilities of many S&Ls exceeded their assets. In most industries, insolvent firms go out of business. Because of federal deposit insurance, insolvent S&Ls could continue to operate. With their equity already gone, they could “gamble for resurrection” by lending money to very risky projects. If the projects were successful, the S&Ls won. If the projects failed, the government insurance fund lost. To prevent this type of gambling, as well as the outright fraud that also

10

Albert0 Alesina and Geoffrey Carliner

contributed to the S&L debacle, regulators needed to close down “the walking dead.” This required tighter regulations, additional staff for the regulatory agencies, and infusions of cash to pay off the insured depositors. Thomas Romer and Barry R. Weingast, in their contribution to this volume, examine how a problem that started out as a loss of a few billion dollars for the government insurance fund exploded into a debacle that may eventually cost much more.I2 They emphasize the role of Congress in responding to interest-group pressure from the S&Ls. By offering campaign contributions to members of Congress, especially those with key positions on committees concerned with regulatory agencies, the S&L industry was able to persuade these Congressmen to intervene on their behalf with these agencies. During the mid-1980s Congress also blocked, delayed, or watered down measures to raise the insurance premiums on S&Ls, increase the amount of capital required of owners of S&Ls, and increase the staff at the regulatory agencies. These steps would have allowed the regulators to limit the costs of the S&L collapse. The administration also played a role in the S&L debacle. Robert E. Litan’s comment on Romer and Weingast underscores the importance of 1982 legislation, supported by the administration, that allowed S&L owners to use federal insurance guarantees to expand rapidly without putting in any of their own money. He argues that loosening the capital requirements for S&Ls “invited high rollers and crooks into the industry.” In addition, the administration, given its antiregulation policy, encouraged further loosening of regulations and opposed increases in staff and budgets of the regulatory agencies. Romer and Weingast argue that congressional behavior on this issue was simply business as usual: that is, support for narrow interest groups unless there is widespread pressure from other parts of society. They observe that the president typically represents the interests of the society at large. However, the ideology of the Reagan administration favoring deregulation meant that regulations were loosened when problems first arose, and that measures to correct the industry’s problems were delayed until they were extraordinarily costly. An administration less opposed to government interference in the economy might not have allowed the crisis to grow so large before taking steps to correct it. Like health, safety, and environmental deregulation, the deregulation of the S&L industry during the early 1980s represents a substantial shift from the policies of previous administrations. Another aspect of regulation policy is government intervention in setting wages. The history of the federal minimum wage during the 1980s reflects very closely the political trends of the decade. Legislation passed under President Carter raised the nominal value of the minimum to $3.35 an hour on 1 January 1981, just as the Reagan administration was about to enter office. For most of the decade, Congress took no steps to raise the minimum wage. As a result, inflation eroded its purchasing power, so by 1986 its value in 1981 dollars had declined to $2.69.

11

Introduction

As Keith T. Poole and Howard Rosenthal observe in their contribution to this volume, this is entirely consistent with history: no legislature with a Republican majority has ever voted to increase the minimum wage. By examining roll call voting, Poole and Rosenthal argue that the minimum-wage issue is in fact, highly ideological. Legislators who can be classified as “right wing” according to their overall voting record, tend to be consistently against increases in the minimum wage. Ideology explains voting on the minimum wage better than the socioeconomic composition of the legislator’s constituencies. It is quite interesting to note that when the Democrats regained control of the Senate in January 1987, they failed to pass any minimum-wage legislation for more than two years. When Reagan left office, the real value of the minimum wage in 1981 dollars had declined to $2.41. Even after two years of Democratic control of the legislature, the minimum wage in real terms was lower than in the sixties and seventies. Once Reagan left office, the Democratic Congress wasted no time in negotiating a higher minimum wage with President Bush. Even this increase, however, supports the hypothesis that there was an ideological shift to the right. Poole and Rosenthal show that the roll call votes in 1989 followed the same right-left pattern that explained earlier votes on this issue, but they find that the $3.85 minimum wage passed in 1989 provided for a significantly smaller increase than past experience would have predicted. In his comments, Charles Brown expresses some doubts over Poole and Rosenthal’s calculation showing a decline in the real minimum wage desired by Congress. Their conclusion relies upon comparisons of congressional votes in the sixties and in 1989. Brown argues that these votes were on bills that were too different to use as the basis for inferences about Congressional views on the minimum wage. Unlike the minimum wage, which can be described with one number and a few measures of coverage, policies that regulate U.S. international trade include a wide range of decisions that are often difficult to quantify. I. M. Destler’s paper describes the major trade issues of the 1980s-the formal or informal protection for autos, steel, textiles, motorcycles semiconductors, and machine tools; a free trade agreement with Canada; legislation that created new reasons for restricting imports; and more generally harsh rhetoric favoring “fair trade” rather than “free trade.” Do these changes constitute a basic shift in trade policy? Destler’s answer is no. He argues that the fundamental pattern of trade policy remained steady over the decade, in spite of unprecedented pressures for protection. This pattern includes demands by Congress for import protection for specific industries, but a willingness to let the administration make most specific decisions on trade policy, and an administration that expresses a firm commitment to free trade but occasionally gives in to political pressure and approves import restraints.

12

Albert0 Alesina and Geoffrey Carliner

Perhaps the biggest change in trade policy in the 1980s was the intensity of the pressure for protection. As Destler shows in table 8.1, the U.S. trade deficit rose from about 1 percent of GNP in 1977-82 to 3.5 percent of GNP in 1987, and then fell to 2.1 percent of GNP in 1989. At the same time that imports were rising sharply in the early 1980s, the United States experienced its deepest recession in over 40 years. Although the loss of sales due to the recession was often larger than the loss due to imports, the combination of the two was enough to cause several politically powerful industries to seek import protection from the government. Such demands continued even after several years of recovery, since the trade imbalance did not disappear. Another longer-term source of political pressure for protection was the decline in U.S.dominance in the world economy. As the U.S. technological lead over other developed countries has eroded, there has also been a long-term erosion in support for free trade. Although outright protectionism is still unpopular, political pressures have grown for “fair trade” and for unilateral actions by the United States that go against the spirit if not the letter of international trade agreements. This long-term pressure is likely to continue even if the trade deficit shrinks to pre-1980s levels. Anne 0. Krueger, in her comment on Destler’s paper, questions how many small changes are necessary to add up to a substantial change in U.S. policy. She cites, in particular, the import quotas and VERs on autos, the U.S.Canadian free trade agreement, and the bilateral negotiation over Super-301 status as steps that undermined the open multilateral trading system. Whether these actions constitute a substantial shift in policy is of course a matter of judgment. Certainly the Reagan administration did not come into office in 1981 urging that the United States abandon free trade philosophies espoused since World War 11. There was no ebbing of protectionist sentiment after the elections of 1982 or after the Democrats regained control of Congress in 1986. Rather, the Congressional Democrats were the ones urging a change in trade policy, and the Reagan administration defended, in rhetoric if not always in practice, the policies of past administrations.

Conclusion As the 1980s began, prices were soaring, growth was stagnant and unemployment was rising. In reaction to these events, American voters turned against the incumbent party and gave the Republicans control of the White House and the Senate. Conservative Democrats in the House who were sympathetic to the goals of the Reagan administration gave Republicans effective control of the entire Congress. This sharp reaction against the Democratic party lasted only until the recession and midterm election of 1982, which returned effective control of the House to the Democrats and eroded Republican power in the Senate. However, this two-year period at the beginning of the decade was enough time for

13

Introduction

the Reagan administration to make substantial changes in several areas of economic policy. As this introduction has summarized, and the papers in this volume examine in more detail, during 1981-82 the president proposed and Congress passed a large income tax cut that reduced marginal rates and set the stage for an unprecedented increase in the budget deficit, sharp cuts in social spending for the poor, and an increase in the military buildup begun under President Carter. The Reagan administration considerably weakened enforcement of health and safety regulations that had become law during the 1970s and discontinued most antitrust enforcement. In addition, the administration supported the Fed’s anti-inflationary effort. In many areas then, the early eighties saw the implementation of policies traditionally advocated by the right. After the 1982 election, and especially after the 1986 election that returned full control of Congress to the Democrats, the Reagan administration had difficulty implementing conservative policies. However, there may also have been a longer-term turn to the right that fasted beyond 1982 and beyond 1986. One indication of this is the increasing percentage of voters who identify themselves as Republicans. Another is the failure of the Democrats in Congress to introduce new social spending programs once they had majorities in both houses of Congress. One area in which the Reagan Administration clearly departed not only from the seventies but also from traditional conservative values is that of budget deficits. These deficits are, in part, the result of an inability of liberals and conservatives to agree on tax and spending policies, but they are also the result of a significant change in the attitudes of conservatives and Republicans on the relative importance of low deficits versus low taxes and low social spending. Republican presidents before the 1980s chose to reach compromises with Democratic Congresses that produced smaller deficits, perhaps because the distance between the two sides was smaller or because deficits were considered more harmful. In either case, persistent large budget deficits and the accumulation of debt that they produced are the gift of the 1980s to the future.

Notes 1 . For instance, Kramer (1971) and Fair (1978, 1988). For a discussion on whether the effect of preelection income growth on voting is rational or myopic, see Alesina, Londregan, and Rosenthal(1991). 2. On the advantage of incumbents, see Kramer (1971), Fiorina (1981) and Fair (1978, 1988). 3. See Fiorina (1988, 1991) and Alesina and Rosenthal(1989a,1989b). An interesting question is why divided government usually has Republican control of the executive and Democratic control of Congress, rather than the other way around. One answer might be that voters prefer a Republican president in charge of foreign policy and

14

Albert0 Alesina and Geoffrey Carliner

a Democratic Congress setting domestic spending policies. Furthermore, given the incumbency advantage, once a large number of seats in the House is held by Democrats, it may take a long time to reverse this situation. Thus, if voters prefer divided governments, they will elect Republican presidents because of the Democratic dominance in the House. 4 . Seminal work in this area i s by Hibbs (1977, 1987). See also Alesina (1987). 5. This is the approach followed by Nordhaus (1975) and Tufte (1978). For two recent surveys that provide somewhat different views regarding the literature on the politics of monetary policy, see Alesina (1988) and Nordhaus (1989). 6. Note that the timing of Carter’s macroeconomic policy is opposite from the “political business cycle” hypothesis (Nordhaus 1975). Rather than expanding the economy immediately before the election, Carter followed an anti-inflationary policy in 1980. This does not imply that Carter was not interested in reelection: at that time, inflation was perceived by the voters as the most pressing economic problem. 7. This measure of the debt does not include debt held by the Federal Reserve and other Government agencies. McCubbins (in this volume) uses a different measure. 8. The “miscalculation theory” is suggested by Stockman (1987). For a critical discussion of this view, see Friedman (1988). 9. For a nontechnical exposition of this argument, see Friedman (1988). For a formal treatment, see Persson and Svensson (1988) and Alesina and Tabellini (1990). 10. For a more extensive treatment of this theory see Barro (1985, 1986, 1987). 11. Viscusi (forthcoming) notes that the Reagan administration sharply reduced the enforcement of health and safety regulations during its first two years but made little attempt to reform or rationalize the regulations that they considered objectionable. After this initial period, according to Viscusi, regulatory policy “resembled that of the pre-Reagan era.” 12. Kane (1989, 87) estimates that the cost to the federal government of paying off insured depositors at insolvent thrifts rose from $3 billion in 1982 to $63 billion in the middle of 1988. Since then, the cost of the bailout has undoubtedly risen considerably.

References Alesina, A. 1987. Macroeconomic Policy in a Two-Party System as a Repeated Game. Quarterly Journal of Economics 102 (August):65 1-70. . 1988. Macroeconomics and Politics. NBER Macroeconomics Annual, edited by Stanley Fischer, 11-55. Cambridge, Mass.: MIT Press. Alesina, A , , and G. Tabellini. 1990. A Positive Theory of Fiscal Deficits and Government Debt. Review of Economic Studies 57 (July):403-14. Alesina, A., J. Londregan, and H. Rosenthal. 1991. A Political-Economy Model of the United States. NBER Working Paper no. 361 1. Cambridge, Mass., February. Alesina, A . , and H. Rosenthal. 1989a. Partisan Cycles in Congressional Elections and the Macroeconomy. American Political Science Review 83 (June):373-98. . 1989b. Moderating Elections. NBER Working Paper no. 3072. Cambridge, Mass., August. Barro, R. 1985. Government Spending, Interest Rates, Prices and Budget Deficits in the United Kingdom, 1730-1918. University of Rochester working paper. . 1986. U.S. Deficits since World War I. Scandinavian Journal of Economics 88:193-222. . 1987. Macroeconomics, 2d ed. New York: Wiley.

15

Introduction

Fair, R. 1978. The Effects of Economic Events on Votes for Presdient. Review of Economics and Statistics 60 (May): 159-72. . 1988. The Effects of Economic Events on Votes for President: 1984 Update. Political Behavior 10:168-79. Fiorina, M. 1981. Retrospective Voting in American National Elections. New Haven, Conn.: Yale University Press. . 1988. The Reagan Years: Turning Point to the Right or Creeping Towards the Middle. In The resurgence of conservatism in Anglo-American democracies, ed. B. Cooper et al., 430-59. Durham, N.C.: Duke University Press. . 1991. An Era of Divided Government. In Developments in American politics, ed. B. Coin and G. Peele. London: Macmillan, forthcoming. Friedman, B. 1988. The Day ofReckoning. New York: Random House. Hibbs, D. 1977. Political Parties and Macroeconomic Policy. American Political Science Review 71 (December):1467-87. . 1987. The American Political Economy: Electoral Policy and Macroeconomics in Contemporary America. Cambridge, Mass.: Harvard University Press. Kane, E. J. 1989. The S&L Insurance Mess: How Did It Happen? Washington, D.C.: Urban Institute. Kramer, G. 1971. Short Term Fluctuations in U.S. Voting Behavior, 1896-1966. American Political Science Review 65: 131-43. Nordhaus, W. 1975. The Political Business Cycle. Review of Economic Studies 42 (April): 169-90. . 1989. Alternative Approaches to Political Business Cycles. Brookings Papers on Economic Activity, no. 2. Persson, T., and L. Svensson. 1989. Why a Stubborn Conservative Would Run a Deficit: Policy with Time-Inconsistent Preferences. Quarterly Journal of Economics 104 (May): 326-46. Stockman, D. 1986. The Triumph ofPolitics: How the Reagan Revolution Failed. New York: Harper & Row. Tufte, E. R. 1978. Political Control of the Economy. Princeton, N.J.: Princeton University Press. Viscusi, W. K. Forthcoming. The Mis-specified Agenda: The 1980s Reforms of Health, Safety, and Environmental Regulation. In American Economic Policy in the 1980s, ed. M. Feldstein. Chicago: University of Chicago Press.

This Page Intentionally Left Blank

1

Elections and the Economy in the 1980s: Short- and Long-Term Effects Morris P. Fiorina

In 1980 Ronald Reagan led the Republican party into the Promised Land. That the Reagan-Bush ticket carried 49 states was noteworthy enough, but the party also scored unanticipated victories in numerous Senate races, giving it control of that body for the first time since 1954, and a respectable gain of 33 seats in the House of Representatives gave the president a “working conservative majority” in that Democratic stronghold. All of this set off talk of a “turn to the right,” a “Reagan Revolution,” and a(nother) new Republican majority. After the rhetoric cleared however, research pronounced a less sweeping verdict. The 1980 elections were just another example of the rejection of failed leadership. In particular, Americans found the Carter administration wanting in two major respects (Schneider 1981). First, there was national frustration with the course of international affairs, especially with America’s apparent helplessness in the face of terrorism. Second, there was deep dissatisfaction with the course of economic affairs. A Democratic president with comfortable Democratic majorities in Congress had presided over a “stagflation” culminating in double-digit inflation and interest rates combined with moderate unemployment and low growth. The succeeding elections of the 1980s confirmed the view that traditional, off-the-shelf explanations still applied. In 1982, coming out of the worst recession since the Great Depression, the Republicans lost 26 House seats. But with inflation crushed and the economy growing, Republican fortunes rebounded, and in 1984 Reagan enjoyed a sweeping reelection victory. To the surprise of many economists and the consternation of many Democrats, the recovery continued, and in 1988 George Bush profited by leading the Republican party to its third straight presidential victory and fifth out of the past six. Moms Fiorina is professor of government at Harvard University.

17

18

Morris P. Fiorina

Some commentators argue that the story of the Reagan elections is little more than the story of the economy (Kiewiet and Rivers 1985). Others argue that the story is more complicated, but no one denies that the economy was a major element of the story. This paper briefly recounts that story. Part of it is as straightforward as these introductory sentences suggest: the economy had the expected effects on presidential approval, the presidential vote, and the distribution of congressional seats. What political scientists refer to as the “short-term” effects of the economy operated much as models and methods developed with data from the 1960s and 1970s predicted-low inflation and rising incomes are political goods, while the opposites are “bads.” From this standpoint the 1980s simply gave us more observations and a bit more variance. But the economy had a deeper, more subtle effect as well. During the 1980s the balance of partisan affiliations shifted toward the Republicans. This shift involves what political scientists refer to as long-term effects, the basis of statements about the “majority party,” the “emerging Republican majority,” and the end of the “New Deal party system.” The two major sections of this paper describe the short- and long-term effects of the economy in the 1980s. In the next section I will survey the short-term effects. Then, in the more original part of the paper, I will report some preliminary analyses that suggest deeper, more lasting effects that will be felt in elections yet to come.

1.1 Economic Conditions, Public Opinion, and Voting In the past two decades few topics have received more scholarly attention that the relationship between economic conditions on the one hand and public opinion and voting on the other. First, political scientists attempted to match fluctuations in economic time series with fluctuations in the congressional vote (Kramer 1971; Tufte 1975)and presidential approval (Mueller 1970;Kernell 1978; Monroe 1978). Then, seeing easy pickings, economists improved on our methods and models (Fair 1978; Frey and Schneider 1978;Golden and Poterba 1980) and even endogenized economic conditions themselves via models of the “political business cycle” (Nordhaus 1975).The literature is too vast to even attempt to review here.2 I will simply survey its principal findings by taking a closer look at the electoral politics of the 1980s.

1.1.2 Economic Conditions and Reagan Approval Studies of presidential approval have a standard design though they differ in numerous details. Gallup presidential approval figures (often quarterly averages) are regressed on measures of economic conditions (typically variations in unemployment, inflation, and real income), variables designed to capture the effects of wars (Korea, Wet Nam), dummy variables representing important events (Watergate, the Iranian hostage crisis), and variables that capture changes in presidential administrations. Analysts have reached no

19

Elections and the Economy

firm consensus about the lagged effects of economic variables (Golden and Poterba 1980; Hibbs 1982; Norpoth 1985) and whether approval follows welldefined cycles or trends (Stimson 1976). Within administrations, however, recent economic conditions have clear and reasonably precise effects on presidential approval. Casual consumers of 1980s political commentary will be surprised by the thrust of research findings in this area. From the standpoint of approval models, there was no “teflon” president, at least in the economic realm. When economic dirt hit, it stuck. The Reagan administration took harsh action to halt the inflation of the previous decade. As an economic result, the country entered into a serious recession. As a political result, Reagan’s approval figures plummeted more than 30 points to a low of 35 percent (as a benchmark consider that Richard Nixon’s ratings bottomed out at 23 percent just before resigning). As the economy recovered Reagan’s approval ratings recovered with it. By the 1984 election he had gained 20 points. After careful analysis Kiewiet and Rivers conclude that “differences between Carter’s and Reagan’s levels of popularity are satisfactorily explained by the differences between the respective economic records and rally points of the two administrations. . . . Reagan’s popularity at reelection was almost solely a function of the performance of the economy after the 1982 midterm elections” (1985, 81-82). The most recent analyses using data extending to 1987 agree. The pattern of Reagan’s popularity-not just one but two recoveries after declines-may have been unprecedented, but the underlying causes were not.) Ostrom and Simon (1989) make a heroic effort to augment the standard analyses with measures of the Reaganauts’s purported flair for the dramatic-the heart-tugging prime-time speech and the well-covered presidential trip. They conclude that such public relations efforts had little effect. When the economy faltered, Reagan’s approval figures weakened; when the economy gathered steam, Reagan’s figures perked up.4 He left office the most popular president since Eisenhower because he left office with a strong economy.

1.1.3 Economic Conditions and the Presidential Vote The frequency with which pollsters inquire about presidential approval makes time-series analysis a natural choice for students of presidential popularity. The infrequency with which presidential elections actually occur has the opposite effect on students of the presidential vote. Only Niskanen (1975) and Fair (1982) have carred out analyses and votes analogous to those for approval. Niskanen examined the 20 elections between 1896 and 1972, regressing the log of the incumbent party’s vote on economic variables, the previous party vote, and incumbency. A noteworthy feature of his analysis is that changes in the economy were measured across the four-year interval between elections. While there is some disagreement in the literature, the modal analysis sup-

20

Morris P. Fiorina

ports a shorter frame of reference (i.e., myopic voters). Nonetheless, Niskanen’s estimates indicate that voter support for presidents is significantly related to real per capita net national product since the previous election. Fair develops a general model that includes Kramer’s backward-looking voters as a special case, but with so few observations-the 22 elections between 1892 and 1976-he is not able to analyze the full model. Constrained analyses indicate that the presidential vote responds to the growth rate of real per capita GNP, change in unemployment, and perhaps to change in the price level. Developments in the election year appear to be most important, except in the case of prices where a two-year change is better. The 1980s provided Fair (1982, 1988) the opportunity to update the basic model. With the 1980 and 1984 elections included in the estimation the negative electoral effect of inflation becomes more apparent. And the evidence of voter myopia grows stronger, as the growth rate of GNP during the second and third quarters of the election year is more important than change measured over a longer interval. For purposes of this discussion the most interesting question is how well the original Fair model predicts the elections of the 1980s. Coefficient estimates based on the 1892-1976 elections predict that Reagan would get about 53 percent of the vote in 1980, an underestimate of 2 percent, and 55 percent of the vote in 1984, an underestimate of about 4 percent. Bush was predicted to get only 49 percent of the vote in 1988, an underestimate of 5 percent.’ Do the underestimates for the 1980s elections reflect the vaunted Reagan personality factor? Probably not. Fair’s equations do not take account of foreign relations, which were working against the Democrats in all three elections.‘j Thus, these underestimates are perfectly comprehensible. In view of the data limitations and theoretical difficulties encountered by analyses like Niskanen’s and fair'^,^ and given the existence of high-quality election year surveys, political scientists have concentrated on cross-sectional analysis of the effects of the economy and economic issues on the presidential vote. The existence and importance of economic influences has never been much in question; rather, the variety of such effects and the manner in which they operate have been the concerns of political scientists. Numerous analyses demonstrate that voting reflects individual perceptions of both one’s own economic circumstances (Fiorina 1978), and of the broader economic climate (Kinder and Kiewiet 1979). Judgments of the economic performance of the government are most important of all, but such judgments reflect factors such as partisanship and candidate attractiveness as well as pure economic performance (Fiorina 1981b). To some extent perceptions of the condition of their group (eg., blacks, farmers) shapes how people react to real economic conditions (Kinder, Adams, and Gronke 1989). In general, the political science cross-sectional studies suggest that the individual behavior underlying aggregate election results is more heterogeneous (Rivers 1988) and more complicated than might appear from an examination of aggregate time-series anal-

21

Elections and the Economy

yses. Voters do not simply look at their wallets and vote accordingly; rather, they make a more complex judgment that reflects individual, sectoral, and national conditions, both those already realized and others only expected. Complications aside, however, it was these cross-sectional studies that established that 1980 was largely a rejection of Jimmy Carter’s performance rather than an endorsement of supply-side economics (Markus 1982; Miller and Wattenberg 1985). Similarly, cross-sectional studies established that in 1984 voters chose Reagan despite closer agreement with Mondale on issues such as defense spending, Central America, and abortion. They chose Reagan because they thought he had performed well as president, and one of the reasons they thought he had performed well was because the economy was thought to be strong (Abramson, Aldrich, and Rohde 1986). At the time of this writing, cross-sectional studies are establishing that George Bush’s victory in 1988 was not just an artifact of the diabolical cleverness of Republican media wizards. Rather, he won because the economy continued strong and people who approved of Reagan’s performance transferred their approval to Bush (Shanks and Miller 1990; Weisberg 1989).

1.1.4 Economic Conditions and the Congressional Vote Kramer’s 1971 article is clearly the seminal piece in the modem study of economics and elections. Taking the House elections between 1896 and 1964, Kramer regressed the aggregate Republican vote on a series of economic variables including unemployment, inflation, and real income. The estimates indicate that variations in real income carried the explanatory weight in the equation, with a 1 percent decline in real income producing a .5 percent decline in the House vote share of the administration. Kramer’s model assumed identical economic effects in presidential and offyear elections. Noting that, with the exception of 1934, the incumbent administration had lost House seats in every midterm election since the Civil War, Tufte (1975) argued that off-year elections should be treated separately as referenda on the performance of the incumbent President. Taking the eight elections between 1946 and 1974, Tufte regressed (the logit of) the aggregate congressional vote on the rate of change in real per capita disposable income during the election year, presidential approval at the time of the election, and a measure of the “baseline” vote. His estimate of the vote consequence of real income change (.35 percent) is smaller than Kramer’s but this estimate is net of changes in presidential popularity that, as noted above, also varies with changes in real income. Forecasting the results of congressional elections quickly developed into a cottage industry, with producers such as Jacobson and Kernel1 (1983), LewisBeck and Rice (1984), and Campbell (1985). Like the Tufte model, all of these refinements include measures of presidential approval and economic conditions and in some way or another take account of differences between midterm and presidential-year election^.^ The models differ in their assump-

22

Morris P. Fiorina

tions about the lags with which presidential approval and economic conditions affect the election results. Oppenheimer, Stimson, and Waterman (1986) add an “exposure” variable that represents the number of seats held by a party in excess of its “normal” holding. A party that is highly exposed (the 1966 Democrats) is in greater danger than one that is minimally exposed (the 1986 Republicans). In a sense the exposure variable provides a substantive explanation for an observed regression to the mean. How well did such models perform in the 1980s? Each election tends to yield a new winner, but as a group they do quite well.Io The Marra-Ostrom model (1989) misses the 1984 outcome by one seat.I’ The 1986 case was particularly instructive. Early in the election year journalists speculated about the “six-year itch.” Since 1932 the average midterm seat loss for the party of a reelected president was more than 50 (Cook 1985). So, in a classic case of naive forecasting some pundits anticipated a Republican disaster. In actual fact the Republicans lost only five seats, one of the smallest midterm losses ever, but exactly as predicted by Marra and Ostrom and very close to the prediction of Oppenheimer et al. (1986) (seven) for a minimally exposed party. There have been some disastrous sixth year showings, but they reflect conditions (the 1938 and 1958 recessions, the 1966 city and campus riots, and the 1974 recession and Watergate crisis) that were not present in 1986. Although the accuracy of the congressional models is impressive, their forecasting performance again exceeds our capacity to describe the underlying behavioral processes. For one thing, the early studies of Kramer (1971) and nfte (1975) sought to predict the House vote, whereas the later generation models focus directly on House seats. The justification for the shift is that the analyst is interested in a system-level response (control of Congress) to a system-level condition (state of the economy). This is obviously a specious argument. Seats are not affected directly by economic conditions; seats do not experience employment or inflation; seats do not vote. The United States has a single-member simple plurality (SMSP) electoral system, not a proportional representation (PR) system, and, as is well known, SMSP systems translate seats into votes in a nonproportional and variable manner (Gudgin and Taylor 1979). Moreover, there is evidence that the translation of votes into seats underwent a structural change in the 1960s, a regime shift not incorporated in existing midterm models (Ansolabehere, Brady, and Fiorina 1988). Nevertheless, such models forecast rather accurately. Apparently aggregation saves the forecasting models from the consequences of their logically questionable specifications. A second area of uncertainty again involves the microbehavior that underlies the effects of aggregate economic conditions. Cross-sectional studies using survey data have found no effect of individual financial condition on the congressional voting, at least after 1960 (Fiorina 1978). This has stimulated two alternative theories. First, Kinder and Kiewiet (1979, 1981) argue that voting behavior is based less on individual economic circumstances than on

23

Elections and the Economy

individual perceptions of collective circumstances, that voters are “sociotropic” rather than individually self-interested. Second, Jacobson and Kernell (1983) maintain that the effect of economic conditions on elections is partially a self-fulfilling prophecy, as strong candidates decline to run and contributors decline to give in “bad” years for their party. For this reason the Jacobson and Kernel1 forecasts utilize presidential approval and economic conditions in the spring of the election year, since that is when the deadlines for candidate filing occur. Kramer (1983) mounted a vigorous attack on these lines of work, arguing that cross-sectional variation in real income does not much reflect government policies or actions, whereas a substantial portion of temporal variation does. Thus, cross-sectional studies of economic conditions and voting are essentially useless. Markus (1988) and Rivers (1990) utilize pooled time-series cross-sectional designs to refute Kramer partially and to identify some individual basis for the aggregate results. At this time, scholarly guns are quiet, but the matter is still open. Finally, several scholars recently have questioned the very existence of a direct link between aggregate economic conditions and midterm election results. They argue that although economic conditions affect the presidential vote two years prior, the midterm loss reflects a “presidential penalty” (Erikson 1988, 1990) or “moderation” of the president (Alesina and Rosenthal 1989) that is not a direct effect of the economic circumstances prevailing between the presidential and midterm elections. Jacobson (1990) rebuts these analyses, questioning their specifications, and suggesting alternative specifications under which recent economic conditions do affect the congressional outcome. For present purposes, the answer to this question is not important. Even if Erikson, Alesina, and Rosenthal are correct, they do not suggest that the 1980s are in any way different; rather, the implication of their work is that analysts misinterpreted the data all along. In sum, whatever the resolution of the remaining puzzles and controversies about the effects of the economy on public opinion and national voting, it is clear that they derive from the normal progress of a research program. Nothing about the politics of the 1980s called into question models first developed in the 1970s. From the standpoint of the short-term effects of the economy, the 1980s have been politics as usual.

1.2 Economic Conditions and the Party Balance The work discussed in the preceding section focuses on the short-term effects of the economy, that is, the impact of economic conditions on particular decisions, like whom to support in presidential and congressional contests. Such analyses implicitly view elections as determined by the particular circumstances surrounding them. The apparent statistical evidence for voter my-

24

Morris P. Fiorina

opia further reinforces that presumption. But while the particular circumstances surrounding elections obviously are important, political scientists have long been aware that election outcomes also reflect long-term factors that are relatively constant from election to election. One such long-term factor is party image, the popular view of a party based on a history of policy and performance extending considerably beyond the previous two quarters. For more than a generation Americans viewed the Democratic party as the party of prosperity.L2From the time Gallup began asking the question in the early 1940s until 1981, the Democrats trailed the Republicans only three times.I3 The Republicans pulled ahead in 1981, fell back in 1982-83, and pulled ahead for good in 1984. Most students of elections considered this development as important as the actual election results of the 1980s, for the simple reason that many voters are innocent of the particular candidates and issues in an election and vote on the basis of these general, long-standing party images. For election analysts the quintessential long-term force is party identification, called partisanship or party ID for short. Gallup has never taken a poll in which more respondents identified themselves as Republicans than as Democrats. The same is true for the American National Election Studies (NES). Every postelection survey, including those in which Republicans embarrassed the Democratic opposition (1956, 1972, 1980, 1984, 1988), has found a plurality of citizens classifying themselves as Democrats. In panel studies (interviews with the same respondents at two or more times) no survey item shows greater stability, and movements in aggregate party ID typically are described as “glacial” (table 1.1 and fig. 1.1). So impressed by its stability were the first students of party identification that they likened it to popular religious affiliations-learned in childhood, devoid of doctrinal underpinnings, and impervious to change in later life (Campbell, Converse, Miller, and Stokes 1960). Not only was it a preexisting “bias” that most voters carried into the voting booth, but it also served as a “perceptual screen” through which voters selectively perceived the candidates, issues, and conditions of the time. Election analyses of the 1960s pronounced party ID to be the single most important factor in American elections. Developments of the late 1960s and early 1970s led to revisions in the prevailing view First, there was the much-discussed rise in self-identified independents. By the late 1970s pundits regularly referred to independents as the second largest “party,” ahead of the Republicans.I4Second, there was an erosion of “strong” partisans, as fewer respondents admitted to an unconditional affiliation with either party. Third, there was a weakening of the link between professed party identification and the presidential vote, as self-identified Democrats blithely chose Republican presidential candidates. Stimulated by such anomolies younger researchers began to contemplate the possibility that party ID was not an unmoved mover and took the heretical step of putting it on the left-hand side in their analyses. Jackson (1975) showed that party ID

25

Elections and the Economy Party Identification, 1952-88 (in %)

Table 1.1 Year

Strong DEM Weak

DEM

Leaning DEM

INDEP

10 9 6 7 6 7 9 9 10 10 11 13 12 14 11 11 11 10 12

6 7 9 7 10 8 8 12 11 13 13 15 15 14 13 11 11 12 10

Leaning REP Weak REP Strong REP

~

1952 1954 1956 1958 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988

22 22 21 27 20 23 27 18 20 20 15 18 15 15 18 20 17 18 17

25 26 23 22 25 23 25 28 25 24 26 21 25 24 23 24 20 22 18

7 6 8 5 7 9 6 7 9 8 11 9

10 10 10 8 12 11

13

14 14 14 17 14 16 14 15 15 15 13 14 14 13 14 14 15 15 14

14 13 15 11 16 12 11 10 10 9 10 8 9 8 9 10 12 11 14

Source: National Election Studies. Note: DEM = Democrat; INDEP = independent; REP = Republican.

was partly a function of the issue positions held by voters. Fiorina (1981a) demonstrated that party ID moved with judgments of party performance. Such analyses did not deny that party ID was “sticky”-it clearly has a strong inertial component. But these analyses did establish that aggregate stability masked politically explicable individual movement. Today, the prevailing view is that partisanship continues to be one of the most important factors in how people vote. No cross-sectional analysis of voting could be published without including it.I5 But partisanship responds, albeit slowly, to evaluations of party positions and judgments of government performance. Thus, when electoral analysts search for a Reagan “legacy,” they look beyond his personal victories to shifts in the underlying distribution of party ID that would indicate the end of the New Deal Democratic majority and the emergence of a new Republican majority. Nothing so grand appears to have occurred, but change on a smaller scale has become increasingly apparent. Between the mid-1960s and mid-1970s the Democrats lost ground among white Southerners and blue-collar workers, losses that surely owe much to the party’s stands on racial and social issues (Petrocik 1987; 1989). But Republican ID showed no commensurate gain during this period. In the mid-1980s however, Republican ID did move upward especially among the young (Norpoth 1987; Norpoth and Kagay 1989).16 Given the erosion of the Democratic party-of-prosperity image and the at-

26

Morris P. Fiorina 55 50 45 -

40

~

* C

g

2

35

-

30

-

25

-

20 1 5 1 ' 1952

' 1954

'

1956

' 1958

'

1960

' 1962

'

1964

+Democratic

Fig. 1.1

' 1966

'

1968

' 1970 Year

'

1972

' 1974

'

1976

' 1978

'

1980

' 1982

'

1984

' 1986

'

1988

+Independent +Republican

Party identification in the United States, 1952-88

tribution of recent Republican electoral successes to economic good times, it is somewhat surprising that few analysts have focused on economic developments as the basis for recent changes in party ID. The inflation set in motion by Lyndon Johnson and uncontrolled by Carter ate away at nominal wage gains and pushed workers into higher tax brackets. Meanwhile, Democratic identification eroded from its 1964 high. Then, an electorate that yelled "No!" when queried whether they were better off today than four years ago elected Ronald Reagan, who stopped inflation and presided over a sustained recovery. Meanwhile, Republican identification strengthened. Coincidence? Probably not. In an important recent contribution MacKuen, Erikson, and Stimson (1989) analyze a Gallup party ID series (1953-87). They find that, like presidential popularity, fluctuations in quarterly party ID averages are significantly associated with fluctuations in the economy, in this case consumer confidence as measured by the University of Michigan Surveys of Consumer Attitudes. The remainder of this paper extends and refines the MacKuen-EriksonStimson analysis. Using National Election Studies (NES) data, losses in Democratic identification and gains in Republican identification can be located in the 33d to 95th percentiles of the income distribution, with somewhat offsetting gains in Democratic identification in the bottom sixth of the distribution. At all income levels expressed party identification varies with economic con-

27

Elections and the Economy

ditions, though the poor and the better-off show differential sensitivity to unemployment and growth. 1.2.1 Data The NES surveys have been carried out after each national election since 1952.” Their national samples range from a low of 1,139 in 1954 to a high of 2,705 in 1972, with the number being a joint function of available resources and the larger intellectual themes underlying each survey. The common content in these surveys is available in a major collection called the Cumulative Data File (CDF) that reorganizes the data so that variable numbers, codes, and so on, are identical from year to year. The CDF now contains 19 observations on national party identification.I8 The CDF classifies respondents very roughly into five income categories. These are not quintiles, however, but represent instead a more sociological interpretation of income.19The lowest category, whom I will call the “poor,” runs through the 16th percentile of the income distribution. The next category, the “lower-middle” runs through the 33d percentile. The “middle” category includes the third of the distribution from the 34th to 67th percentiles. The “upper-middle” category runs from the 68th to 95th percentiles. Finally, the “rich” are the top five percent (noneconomists will be surprised and sobered to learn that in 1988 the 96th percentile was a bit less than $90,000.)20 Table 1.2 reports a preliminary examination of trends in party ID within the income categories. Because of the controversy surrounding the classification of independent leaners, I have examined both ways: “broad” Democrats and Republicans include the Democratic and Republican leaning independents as partisans, while “narrow” Democrats and Republicans exclude Democrat and Republican leaners. Within each income category I have regressed the measures of partisanship on a constant, on the election year (1952-88), and on Table 1.2

’Lkendsin Party ID by Income Category NalTOW DEM

Lower Lower-middle Middle Upper-middle Upper

- .02 ~23) -.16 (2.38) - .38 (5.71) - .42 (8.29) -.ll (.78)

Broad DEM .24 (2.38) .03 (.35) - .22 (3.40) - .36 (6.37) - .07 (.43)

NalTOW REP

Broad REP

- .41 (4.37) - .08 (.88) - .05 (.78) .01 ~24) -.I6 (.97)

- .23 (2.48) - .04 (.43) .09 (1.25) .26 (3.58) .05 (.26)

Note: Entries are regression coefficients of Party ID (% in each category) on time (1952-88) and dummy variables for 1964 and 1974. Absolute values of ?-statisticsare in parentheses. DEM = Democrat; REP = Republican.

28

Morris P. Fiorina

dummy variables for Goldwater (1964) and Watergate (1974), both of which were thought to have “shocked” partisanship. To summarize: 1. The poorest segment of the population shows a clear loss in Republican partisanship, narrowly and broadly defined, and a gain in Democratic partisanship, broadly defined (i.e. independent, but leaning Democratic). 2. The lower-middle category has the most stable party affiliations of all the categories, showing only a slight decline in narrowly defined Democratic ID. 3. The middle third of the distribution shows a significant decline in Democratic partisanship, but no commensurate gain in Republican partisanship. 4. The upper-middle sixth shows a clear decline in Democratic ID along with a clear rise in Republican ID, broadly defined (i.e., independent, but leaning Republican). 5. Because of the small number of rich respondents, partisanship figures fluctuate greatly. None of the coefficients are significant in the four regressions. These preliminary regressions are unimpressive, with low R2s and unsatisfactory Durbin-Watson (D-W) statistics, but they suggest that the party loyalties of the population have not moved in unison during the past generation. Rather, different parts of the income distribution show different movements. The question now arises as to whether the addition of suitable economic variables can improve on the statistical qualities of the regressions and add to our substantive understanding of the observed movements in party identification. The answer is a clear yes. Table 1.3 reports regressions that augment those just reported in two ways. First, previous party ID was added to the equations in order to capture the notion of party ID as a running tally of party performance that is continuously updated as the world unfolds (Fiorina 1981a). Second, various economic variables were added to the equations. With only 19 observations, collinearity often precluded including more than one or two economic variables in an equation; in such cases I kept the variable that produced the best overall equation.*l Usually such decisions were not difficult, but in a few cases two alternative specifications were so close in their performance that I present both in table 1.3. In the interest of efficiency I omitted the Goldwater, Watergate, lagged Partisanship, and trend terms when they fell far short of significance, except in a few cases where their omission greatly detracted from the quality of the overall regression or the performance of other variables.22Table 1.3 provides the details. Again, I will summarize the findings rather than proceeding seriatim through 25 regressions. 1. Poor. Within the lowest income category there continues to be a trend away from the Republicans and toward the Democrats, broadly defined. Goldwater’s candidacy added about 10 points to Democratic ID and detracted at least that much from Republican ID. Of most interest, movements in party ID, however measured, are associated with changes in the level of unemployment. Democratic ID responds most strongly to unemployment change since

29

Elections and the Economy

Table 1.3 Category and Variable

Economic Conditions and Party ID Change NDEM

BDEM

NREP

10.78 (2.62)

20.93 (1.98) - .20 (2.02) .74 (4.01) 13.90 (4.25)

AUEl

...

22.89 (2.24) .30 (2.97) .20 (2.97) 11.91 (2.93)

AUE2

1.37 (2.19)

1.71 (2.78)

- .98 (2.30)

.35 1.99

.54

.82

A. Low income: 47.66 Constant (48.78)

EY

...

Lag ID GW

RZ D-W

...

...

NDEM

NREP

BREP

19.66 (1.89) - .21 (2.09) .84 (4.53) - 16.49 (4.81) - 1.38 (2.32)

35.22 (3.19) - .24 (2.60) .38 (1.74) - 10.44 (2.78)

...

- 1.75 (3.43) .83

BDEM

.60

BREP

33.13 (2.97) - .28 (2.92) .59 (2.50) - 15.16 (3.65) - 2.53 (3.34)

.59

NREP

BREP

56.75 (66.79)

7.26 (1.28)

13.71 (2.W

11.69 (3.83)

9.94 (2.79)

.72 (2.84) -9.06 (2.11)

AUE2

.69 ( I .45)

.98 (1.82)

- .92 (1.76)

.60 2.48 - 12.03 (2.66) - 10.15 -2.83 - 1.34 (2.51)

Rz D-W

.56 1.83

.33 1.58

.35

B. Lower-middle income: 56.72 Constant (12.31) EY - .I4 (2.12) Lag ID ... GW W

...

BDEM

C. Middle income: 64.80 Constant ( 17.94) - .31 EY (6.24) Lag ID ...

63.50 (18.78) -.14 (2.99)

9.33 (3.98)

...

...

NDEM

GW

...

...

...

9.41 (4.16)

...

.47

BDEM

NREP

BREP

64.71 (21.24) -.I5 (3.71)

12.29 (2.36)

25.45 (4.49)

.49 (2.21) ...

.26 (1.51) -7.69 (2.96)

...

9.68 (4.89)

...

...

30

Morris P. Fiorina

Table 1.3 Category and Variable

W AGNPI AGNP2 UE

Continued NDEM

BDEM

...

...

- .77 (3.99)

- .81 (4.35)

.44

.39 (3.22)

...

(3.50) AUE2 R 2

D-W

...

AGNPI AGNP2

D-W

E. Upper income: Constant EY Lag ID W

BREP

-4.02 (1.50)

- 6.24 (2.39)

- .09 (1.40)

...

...

1.39 (4.50)

...

- 1.51 (3.84) .64

.79 1.91

.84 2.02

.32

NDEM

BDEM

BDEM

NREP

BREP

46.28 (3.80) - .17 (2.49) .30 (1.95)

47.78 (4.11) -.19 (2.95) .31 (2.09)

20.74 (3.55)

27.24 (7.57) .17 (3.38)

-5.08 (2.40) - .33 (2.44)

...

AUE2 R 2

NREP

.86 2.14

D. Upper-middle income: 65.84 Constant (19.50) - .38 EY (7.93) Lag ID ... W

...

BDEM

.90 (3.12)

- .22 (4.09) .91 (3.35) .91

.12 (1.42) - .84 (2.33)

.33 (2.14)

- 1.25 (3.34)

.86 2.44

...

...

...

.41

.75 1.94

NDEM

BDEM

NREP

NREP

BREP

24.68 (5.28)

33.25 (5.05)

- .07 .43 - 14.16 (2.96)

64.09 (5.98) - .28 (2.09)

50.36 (3.86)

(2.97) - .02 (.I3 - 5.85 (1.12)

66.29 (5.87) - .31 (2.02) 9.29 (1.31)

5.99 (35)

...

.90

.27 1.28

...

- .40 (3.71)

...

...

...

...

...

(2.60) .12 (. 74) 5.67 (35)

31

Elections and the Economy

Table 1.3 Category and Variable AGNPl AGNP2 R1

D-W

Continued NDEM

- 1.04 (3.61)

...

.54

...

BDEM

- 1.40 (4.56) ... .53

...

NREP

1.13 (2.52)

.24 1.93

NREP

BREP

...

...

.58 (2.65)

.80 (3.61)

.26 1.88

...

.44

Note: EY = election year; GW = Goldwater (1964 = I); W = Watergate (1974 = 1); Lag ID = percentage in party ID category 2 years earlier; AUE1 = change in unemployment during election year; AUE2 = change in unemployment since last election year; AGNPl = growth rate of GNP during election year; AGNP2 = growth rate of GNP since last election year; UE = unemployment level in election year. Absolute values of ?-statisticsare in parentheses.

the lust election, while changes in Republican ID are predicted equally well by one-year or two-year changes. Looking at the two-year change equations, a one-point rise in unemployment goes along with a two-point net swing in narrowly defined ID and a three- or four-point net swing in broadly defined ID. 2. Lower-middle. Within the lower-middle income category the trend away from narrowly defined Democratic ID becomes clearer. Goldwater’s candidacy had roughly a 10 point impact on all categories; Watergate affected broadly defined Republicanism only. While the estimates are more tenuous than those for the poor, movements in all four measures of party ID show some responsiveness to two-year changes in unemployment, with estimated magnitudes about three-fourths as large as in the lowest income category. 3, Middle. Within the large middle-income category the previously identified trend away from the Democrats remains intact. The Goldwater candidacy gave the Democrats almost 10 points and Watergate took half as much away from the Republicans. The economic impacts on partisanship are different from those in lower income categories. Democratic ID shows a highly significant relationship to the growth rate of GNP and to the unemployment level during the election year. Thus, the 1982 unemployment level of almost 10 percent added about 4 points to Democratic ID, and the 6.75 percent growth in GNP in 1984 took away 5 points of Democratic ID. In contrast, movements in narrowly defined Republican ID show no relation to economic variables, and movements in broadly defined Republican ID correspond only to movements in two-year changes in ~nemployment.~~ 4. Upper-middle. Within the upper-middle income category the Democrats have clearly been losing and the Republicans less clearly gaining. Although all categories except narrow Republicans show a significant relationship with growth of GNP, the relationship is only half as strong as in the middle category. Moreover, narrowly defined Democrats and broadly defined Republi-

32

Morris P. Fiorina

cans show a response to one-year changes, while narrowly defined Republicans show a response to two-year changes, and broadly defined Democrats show a response to both. All categories except narrow Democrat show a significant relation to two-year changes in unemployment, with each one-point increase in unemployment taking away one point of narrow Republican ID and creating a net swing of two points in broad ID. In this group Watergate took away a few points of narrow Democratic ID. 5 . Rich. Among the rich, Democratic ID rises and falls with the growth rate of GNP, though Democrats show more of a response to the short term and Republicans to the longer term. Each one-point increase is associated with a two-point net shift in the narrow ID balance. Only in this blessed category does party ID show no sensitivity at all to unemployment. Watergate was costly to the Democrats within this category with a particularly strong impact on narrowly defined Democratic ID. Overall, these simple analyses support two arguments. First, party identification responds not only to perceived economic conditions as established by MacKuen et al. (1989). The finding is stronger: party ID shows a clear relationship to fluctuations in actual economic conditions. Second, the party loyalties of the population do not move in unison as the economy moves; different income levels respond in different ways. Not surprisingly, the less affluent show a greater sensitivity to changes in unemployment, whereas general economic expansion has greater importance among those who enjoy higher income levels. Finally, I emphasize that nothing in the foregoing analyses conflicts with the extensive political science literature on the subject. The adjustment of previous party ID as new performance information becomes available is consistent with micromodels already supported by cross-sectional data (Fiorina 1981a)." And the economic effects that we have found do not detract from the discussions of race and social issues as sources of the Democratic party's current disarray. Most of the regressions for the lower-middle, middle, and upper-middle income groups show a significant anti-Democratic trend that may well reflect the party's estrangement from middle America on issues of race and culture. But whatever other issues are at work, the economy continues to influence the underlying balance of party affiliationsjust as it did during the New Deal and as it will undoubtedly continue to do in the future. 1.3 Conclusion

The economic developments of the 1980s had an impact on the electoral politics of the decade, an impact not only on the outcomes of the elections held between 1980 and 1988, but also on the elections that will occur in the 1990s and possibly beyond. Economic conditions affected the immediate election outcomes to the Democrats' dismay in 1980, 1984, and 1988, and to their joy in 1982. But beyond those short-term impacts, economic conditions

33

Elections and the Economy

left an imprint on the distribution of party identification among the citizenry. Whatever the economic conditions that existed in 1988, the conditions that existed in 1984 and earlier had been incorporated in voter partisanship to the general good fortune of George Bush. Of course, economic bad times under Bush would have the reverse effects. The hard-won gains in Republican ID during the 1980s could be dissipated by economic misfortunes in the years ahead. Fortunately for the Republicans they have been profiting from Democratic policies and performance in other issue domains such as raciaVcultural issues and foreign policy; economic success was the third ace in their hand. Thus, less-than-stellar economic performance need not cost them presidential elections, nor even all their gains in party identification. But continued economic success would buttress the gains they already have made. And attention to employment, in particular, would enable them to continue to make inroads in the lower ranges of the income distribution that are traditionally viewed as “natural” Democratic territory.

Notes 1. Of course, the Republicans did lose control of the Senate in 1986. Economic distress in farm states was often cited as a partial explanation. But the 1980 Senate victories that were reversed in 1986 were something of a fluke to begin with (Fiorina

1984). 2. A number of excellent reviews are available. On the effects of economic conditions on voting and presidential approval see Monroe (1979) and Kiewiet and Rivers (1984). For an excellent general review that deals with sociological as well as economic aspects of the topic see Weatherford (1986). And for a review of the recent political business cycle literature see Alesina (1988). 3. Generally, when a president’s ratings plummet, they never fully recover. Reagan’s, however, dropped from 67 percent in the flush of his 1981 legislative victories to 35 percent after the 1982 elections. But by late 1983 he was back over 50 percent and, following his reelection, hovered near 65 percent through most of 1985-86.

While analysts were studying this unusual recovery, the Iran-Contra scandal dropped him back to 40 percent, but again he recovered and left office with approval ratings over 60 percent. 4. Ostrom and Simon (1989) attribute more than three-fourths of the 1981-82 decline and the 1983-84 resurgence to economic factors. Particular events (Lebanon bombings, Grenada, etc.) also were important. Speeches had no significant impact, presidential trips were of minor import. 5. For the 1980 and 1984 elections I utilized eq. 4 and the data reported in Fair (1982). (This is the original 1892-1976 equation reestimated with national accounts data revised in 1980.) For the 1988 prediction I substituted NBER data generously provided by Gerald Cohen. 6. According to numerous survey studies, the Iranian hostage crisis severely damaged Carter in 1980. Conversely, in 1984 the Grenada invasion, destruction of the Libyan MIGs, and the capture of the hijackers of the Achille-Lauro buttressed Reagan’s image as a no-nonsense leader. The administration had a major arms control

34

Morris P. Fiorina

treaty by 1988. A recent update of Fair’s model (1988) adds a variable that measures change in the size of the armed forces relative to the population. While this captures major wars, it will not pick up incidents and developments such as those just mentioned. 7. To get enough data, time-series analysts must venture across what political scientists and historians refer to as different party systems (where structural changes are hypothesized). For example, Fair (1978) observes that the model performs badly before 1916, though that is not a date ordinarily identified with a change in party system. He also cautions (1988) that the positive-trend term for the Democrats appears questionable in light of 1980s political developments. Political historians would argue that the Democratic base was higher in the New Deal party system (1932-64, approximately) than either before or since. While Fair could dummy in such considerations, he naturally worries about being accused of “mining” his data. 8. These findings do not necessarily contradict the time-series finding that voters are myopic. If voters are heterogeneous, with varying time horizons, different foci (national vs. local), differential sensitivities to growth, unemployment, and inflation, etc., it may be difficult to find evidence of nonmyopic behavior inasmuch as different voters will incorporate different information in different ways. For a discussion of voter rationality that touches on time horizons among other things, see Nordhaus (1989). 9. Some, like Jacobson and Kernel1 (1983) and Campbell (1985), use only midterms for their estimations. Others, like Lewis-Beck and Rice (1984) use a dummy variable for midterms. 10. The one exception was 1982 when all the models overestimated the Republican losses. Jacobson and Kernel1 (1983) argue that in the euphoria following Reagan’s 1980 victory, the Republicans were able to recruit good candidates and raise considerable money, which cushioned their losses when the economy turned sour in late 1981. 11. This model incorporates all the others. It predicts across presidential and midterm years, includes the exposure variable, and also includes measures of major events (as in approval models) and party identification. 12. With somewhat less regularity the Republicans were viewed as the party of peace. 13. April 1943, December 1955, and September 1972. 14. This turns out to be a highly controversial matter. Unlike Gallup, academic surveys ask self-identified independents whether they lean toward either party. The rise in independents occurred almost entirely among these “leaners,” whose presidential votes are often more loyal than those of weak partisans of the same party. See (Keith et al. 1987). 15. Assuming no interactions with other variables, time-series analyses will pick up the average level of party ID in the constant term. 16. There is some disagreement on this point, with some subscribing to the “young Republican” thesis (Norpoth 1987; Miller 1990), while other argue that Republican gains are more evenly scattered across the age distribution (Petrocik 1989). The CBS/ NYT data appears to give somewhat different answers than NES data, and classifying independent leaners as partisans gives somewhat different answers than classifying them as independents (see n. 14 above). 17. Since 1952 election year surveys have been carried out under the auspices of the Institute for Survey Research at the University of Michigan. Since 1978 these surveys have been funded by the National Science Foundation and carried out under the supervision of an academic governing board. 18. Although the Gallup series is longer, the NES series has two advantages. The first is simple convenience. Take the income variable, for instance. The NES staff

35

Elections and the Economy

codes each respondent according to national income percentile. To use Gallup data, one would have to take each Gallup survey and do the calculations and recoding oneself. Even if bountiful research assistance were available, however, the NES collection would still be preferable for present purposes. Gallup makes only a simple threecategory classification of party ID, whereas the NES differentiates between strong and weak partisans and pure and learning independents. As noted in n. 14 above, these finer differences are consequential. 19. Personal conversation with Warren Miller, principal investigator of the NES surveys, April 13, 1990. 20. Since respondents are not asked their exact incomes, but only a range, these proportions are only approximate. In 1988 the upper boundaries of the first four categories were $9,999, $14,999, $34,999, and $89,999. 2 1 . In all cases the economic variables are multiplied by a binary variable for control of the Presidency (Democrat = - 1 , Republican = 1). Good economic conditions under Republican presidents are expected to enhance Republican partisanship while detracting from Democratic partisanship, and poor economic performance should have the opposite effect. I did consider the alternative hypotheses that the parties “own” different issues, so that inflation always helps Republicans and unemployment Democrats even if they are in office when it occurs. Fortunately (in view of the perverse incentives posited by these hypotheses) I found no support for them. Keech and Swain (1990) argue however, that particularly sensitive subpopulations-e.g., blacks-behave somewhat in line with the alternative hypotheses. At this level of aggregation I can not take account of their argument. 22. One might object that the dummy variables for 1964 and 1974 act as proxies for the economic conditions (quite good in 1964, quite poor in 1974) that prevailed in those years and thus detract from the impact of the economic variables in the regressions. On the contrary, eliminating the Goldwater and Watergate dummies when significant generally detracts greatly from the overall regressions and produces weaker impacts for the economic variables. In short, there was more to the 1964 and 1974 experiences than good and bad economic times, respectively. 23. Note that an alternative specification for broadly defined Democrats also suggests the importance of two-year changes in unemployment. 24. Some readers have asked why few of the lagged party ID terms attain significance when they are invariably highly significant in cross-sectional analyses. The answer appears to be the aggregate level of analysis. In the absence of a time trend, aggregate ID levels fluctuate around the constant baseline in accord with variations in economic variables.

References Abramson, Paul R., John H. Aldrich, and David Rohde. 1986. Change and Continuity in the 1984 Elections. Washington, D.C.: Congressional Quarterly Press. Alesina, Alberto. 1988. Macroeconomics and Politics. NBER Macroeconomics Annual, edited by Stanley Fischer, 3:13-51. Cambridge, Mass.: MIT Press. Alesina, Alberto, and Howard Rosenthal. 1989. Partisan Cycles in Congressional Elections and the Macroeconomy. American Political Science Review 83:375-98. Ansolabehere, Stephen, David Brady, and Moms P. Fiorina. 1988. The Marginals Never Vanished. Research Paper no. 970. Stanford University, Graduate School of Business.

36

Morris P. Fiorina

Campbell, Angus, Philip E. Converse, Warren E. Miller, and Donald E. Stokes. 1960. The American Voter. New York: Wiley. Campbell, James E. 1985. Explaining Presidential Losses in Midterm Congressional Elections. Journal of Politics 47: 1140-57. Cook, Rhodes. 1985. Will the “Six-Year I t c h Strike Again in 1986? Congressional Quarterly Weekly Report 43:1284-86. Erikson, Robert S. 1988. The Puzzle of Midterm Loss. Journal of Politics 50: 101 129. . 1990. Economic Conditions and the Congressional Vote: A Review of the Macrolevel Evidence. American Journal of Political Science 34:168-79. Fair, Ray C. 1978. The Effect of Economic Events on Votes for President. Review of Economics and Statistics 60:159-73. . 1982. The Effect of Economic Events on Votes for President: 1980 Results. Review of Economics and Statistics 64:322-25. . 1988. The Effects of Economic Events on Votes for President: A 1984 Update. Political Behavior 10:163-79. Fiorina, Moms P. 1978. Economic Retrospective Voting in American National Elections: A Micro-Analysis. American Journal of Political Science 22:426-43. . 1981a. Retrospective Voting in American National Elections. New Haven, Conn.: Yale University Press. . 1981b. Short- and Long-term Effects of Economic Conditions on Individual Voting Decisions. In Contemporary Political Economy, edited by Douglas Hibbs, Heino Fassbender, and R. Douglas Rivers. Amsterdam: North Holland. . 1984. The Presidency and the Contemporary Electoral System. In The Presidency and the Political System, edited by Michael Nelson. Washington, D.C.: Congressional Quarterly Press. Frey, Bruno S., and Friedrich Schneider. 1978. An empirical Study of PoliticoEconomic Interaction in the United States. Review of Economics and Statistics 60:174-83. Golden, David G., and James M. Poterba. 1980. The Price of Popularity: The Political Business Cycle Reexamined. American Journal of Political Science 24:696-7 14. Gudgin, Graham, and Peter Taylor. 1979. Seats, Votes, and the Spatial Organization of Elections. London: Pion Press. Hibbs, Douglas A., with R. Douglas Rivers and Nicholas Vasilatos. 1982. On the Demand for Economic Outcomes: Macroeconomic Performance and Mass Political Support in the United States, Great Britain, and Germany. Journal of Politics 43 :426-62. Jackson, John E. 1975. Issues, Party Choices, and Presidential Votes. American Journal of Political Science 19:16 1-85. Jacobson, Gary C. 1990. Does the Economy Matter in Midterm Elections? American Journal of Political Science 34:400-404. Jacobson, Gary C., and Samuel Kernell. 1983. Strategy and Choice in Congressional Elections. 2d ed. New Haven, Conn.: Yale University Press. Keech, William R., and Carol M. Swain. 1990. Party, Race and Electoral Decision Rules Regarding Economic Performance. Paper presented at the meetings of the Public Choice Society, March 16-18, 1990, Tucson. Keith, Bruce E., David B. Magleby, Candice J. Nelson, et al. 1987. The Myth of the Independent Voter. Typescript. University of California, Berkeley. Kernell, Samuel, 1978. Explaining Presidential Popularity. American Political Science Review 72506-22. Kiewiet, D. Roderick, and Douglas Rivers. 1984. A Retrospective on Retrospective Voting. Political Behavior 6:369-92

37

Elections and the Economy

. 1985. The Economic Basis of Reagan’s Appeal. In New Directions in American Politics, edited by John E. Chubb and Paul E. Peterson. Washington, D.C.: Brookings. Kinder, Donald R., Gordon S. Adams, and Paul W. Gronke. 1989. Economics and Politics in the 1984 American Presidential Election. American Journal of Political Science 33:491-515. Kinder, Donald R., and D. Roderick Kiewiet. 1979. Economic Discontent and Political Behavior: The Role of Personal Grievances and Collective Economic Judgments in Congressional Voting. American Journal of Political Science 23:495-5 17. . 1981. Sociotropic Politics: The American Case. British Journal of Political Science 11: 129-61. Kramer, Gerald H. 1971. Short-Term Fluctuations in US. Voting Behavior, 18961964. American Political Science Review 5 :13 1-43. Kramer, Gerald H. 1983. The Ecological Fallacy Revisited: Aggregate- versus Individual-Level Findings on Economics and Elections and Sociotropic Voting. American Political Science Review 77:92-111. Lewis-Beck, Michael S., and Tom W. Rice. 1984. Forecasting U S . House Elections. Legislative Studies Quarterly 9:475-86. MacKuen, Michael B., Robert S. Erikson, and James A. Stimson. 1989. Macropartisanship. American Political Science Review 83: 1125-42 Mama, Robin F., and Charles W. Ostrom, Jr. 1989. Explaining Seat Change in the U.S. House of Representatives, 1950-86. American Journal of Political Science 33~541-69. Markus, Gregory B. 1982. Political Attitudes during an Election Year: A Report on the 1980 NES Panel Study. American Political Science Review 76538-60. . 1988. The Impact of Personal and National Economic Conditions on the Presidential Vote: A Pooled Cross-sectional Analysis. American Journal of Political Science 32:137-54. Miller, Arthur H., and Martin P. Wattenberg. 1985. Throwing the Rascals Out: Policy and Performance Evaluations of Presidential Candidates, 1952-1 980. American Political Science Review 79:359-72. Miller, Warren E. 1990. The Electorate’s View of the Parties. In The Parties Respond, edited by L. Sandy Maisel. Boulder, Colo.: Westview Press. Monroe, Kristen R. 1978. Economic Influences on Presidential Popularity. Public Opinion Quarterly 42:360-69. . 1979. Economic Analyses of Electoral Behavior: A Critical Review. Political Behavior 1 :137-73. Mueller, John E. 1970. Presidential Popularity from Truman to Johnson. American Political Science Review 64: 18-34. Niskanen, William A. 1975. Economic and Fiscal Effects on the Popular Vote for the President. Graduate School of Public Policy Working Paper No. 25. University of California, Berkeley, May. Nordhaus, William D. 1975. The Political Business Cycle. Review of Economic Studies 42:169-90. . 1989. Alternative Approaches to the Political Business Cycle. In Brookings Papers on Economic Activity, edited by William C. Brainard and George L. Perry, 1-68. Washington, D.C.: Brookings Institution. Norpoth, Helmut. 1987. Under Way and Here to Stay: Party Realignment in the 1980s? Public Opinion Quarterly 51:376-91. Norpoth, Helmut, and Michael R. Kagay. 1989. Another Eight Years of Republican Rule and Still No Partisan Realignment? Paper presented at APSA annual meeting, Atlanta, GA.

38

Morris P. Fiorina

Oppenheimer, Bruce I., James A. Stimson, and Richard W. Waterman. 1986. Interpreting U.S. Congressional Elections: The Exposure Thesis. Legislative Studies Quarterly 11:227-47. Ostrom, Charles W., Jr., and Dennis M. Simon. 1989. The Man in the Teflon Suit: The Environmental Connection, Political Drama, and Popular Support in the Reagan Presidency. Public Opinion Quarterly 53:353-87. Petrocik, John R. 1987. Realignment: New Party Coalitions and the Nationalization of the South. Journal of Politics 49:347-75. . 1989. Issues and Agenda: Electoral Coalitions in the 1988 Election. Paper presented at APSA annual meeting 1989. Rivers, Douglas. 1988. Heterogeneity in Models of Electoral Choice. American Journal of Political Science 32:737-57. . 1990. Microeconomics and Macropolitics: A Solution to the Kramer Problem. American Political Science Review. Schneider, William. 1981. The November 4 Vote for President: What Did It Mean? In The American Elections of 1980, edited by Austin Ranney. Washington, D.C.: American Enterprise Institute. Shanks, J. Memll, and Warren E. Miller. 1990. Policy Direction and Performance Evaluation: Complementary Explanations of the Reagan Elections. British Journal of Political Science 20: 143-235. Stimson, James A. 1976. Public Support for American Presidents: A Cyclical Model. Public Opinion Quarterly 40: 1-21. Tufte, Edward R. 1975. Determinants of the Outcomes of Midterm Congressional Elections. American Political Science Review 69:812-26. Weatherford, M. Stephen. 1986. Economic Determinants of Voting. Research in Micropolitics 1:219-69. Weisberg, Herbert F. 1989. Some Perspectives on the 1988 Presidential Election: The Roles of Turnout and Ronald Reagan. Presented at APSA annual meeting, Atlanta, GA.

COl'llmeIlt

William D. Nordhaus

The paper by Morns Fiorina is an informative and measured survey of the literature on the impact of economic events upon political variables. In particular, the first part provides a review of the existing literature that will be most helpful for those who are looking for a political scientist's view of recent developments in this area. He argues that the 1980s did not bring a major revolution or realignment in politics in America, and that it was more a combination of past trends and good economic performance that led to the Republican ascendancy of this period. If he is right, one would predict that, at the first whiff of recession and inflation, the popularity of Republicans would revert the mean. The major new research in the paper is the data on party identification by income sextile. This is an important approach, as we clearly need to go to panel data to resolve some of the unanswered questions about political behavWilliam D. Nordhaus is professor of economics at Yale University and a research associate of the National Bureau of Economic Research.

39

Elections and the Economy

ior. Among the important questions that cannot be resolved in aggregate data are distributional issues such as, Does aggregate or individual or perhaps class fortune determine individual voting behavior? The panel data should help us sort out these kinds of questions. I will not repeat the major results of this path-breaking study, Rather, as befits a discussant, I will carp on a few points that need more attention, particularly some areas of specification and statistical methodology. Among the issues that arise are the following. The classification is pretty clearly not independent of the state of the economy. Whether you are in the bottom sextile may well depend upon whether you have been thrown out of work, although this lottery probably does not apply to the top brackets. This will make bottom sextile particularly sensitive to unemployment or bad economic states and may bias the unemployment rate coefficients. To illustrate, some of the people in the bottom sextile will be people who in the prior period were in the fifth sextile and subsequently lost their job and were thrown into the bottom. We know that people hold the government responsible for economic events (the so-called responsibility hypothesis), which suggests they will become displeased with the government. This will tend to raise the coefficients on unemployment in the bottom sextiles and lower them in the top sextiles. Even though the sample size appears respectable, breaking it down so finely may introduce significant sampling error in individual cells. For example, say we have a sample size of 1,800 broken into 6 income sextiles, which are then broken into 6 political groups, which yields a total of 50 persons in each of the 36 cells. Because the unemployment rate is a small number, the actual sampling error of unemployment will be very large from period to period for short unemployment spells. For example, for a sample of 50 workers with an average unemployment rate of 5 percent, I ran a Monte Car10 for the unemployment of the sample. Observations for a representative sample were [4, 2, 7, 1, 3, 0, 1, 2, 2, 2, 1 , 3, 3, 3, 4, 6, 51. The standard deviation of this run is 1.82; the theoretical standard deviation for this binomial distribution is 1.54; the standard deviation of the aggregate annual unemployment rate for the postwar period is 1.2 percentage points. This suggests that the estimated coefficients may be dominated by sampling error. The lag structure was highly surprising. The movement in party ID was labeled “glacial,” by which I presume we mean that there are large elements of habitat and persistence and that the lags of party Republican to shocks would be quite long. For example, we might suppose that the events of the Great Depression cast a long shadow over party alliliations; if so, the response of ID to the Depression would have a large lag term.If we thought that half of the group who became Democrats because of the Great Depression left the party after 20 years, the lag coefficient would be .85 per two-year period. Or

40

Morris P. Fiorina

in the extreme, if the change in party ID responds to various shocks, the lag coefficient would be one. In fact, the lags are short or nonexistent. It is as if there is in effect no memory to the system. Given the firm belief that party affiliation is indeed very persistent, these results suggest that the data have a high degree of shortrun white noise that swamps the lower level of low-frequency, red noise of persistence in party ID. In the end, this effort does not succeed very well in extracting much information about trends in party ID. Perhaps an alternative specification would improve the results. One approach would be to pool the data and fit fewer coefficients. Another possibility is to track the median party ID in an income class, which probably would give sharper results. Overall, issues of party affiliation and of the influence of economic events on political affiliation are an area that clearly calls for true panel data, where the experience of individuals can be tracked over time. From a research perspective, the NBER should collaborate with other organizations in promoting the collection of such true panel data. We have learned a great deal from this paper. With further work on the statistical specification, and with the use of true panel data, we could learn even more.

2

Leaning Into the Wind or Ducking out of the Storm? U.S. Monetary Policy in the 1980s James E. Alt

2.1 Introduction

American monetary policy in the 1980s contained two experiments separating periods of normality. Debates continue about the exact purpose, nature, and duration of these episodes. The first one, which began in late 1979, involved at least a technical change in the procedure used by the Federal Reserve (the body charged with managing the nation’s money) to control money growth. Its broader consequences included periods of double-digit interest rates and a recession of extraordinary severity. The second, smaller, experiment lasted for a year or so from mid-1985 and involved international monetary summits and accords that orchestrated an orderly reduction in the real exchange value of the dollar of about a third in a little over a year. Neither experiment really continues. The 1979 procedures were discontinued sometime in late 1982. While intermittent summits continue, they have not created a situation in which the monetary policies of the United States and other countries are routinely coordinated. Just how novel were these episodes? Did they mark lasting departures in the way monetary policy is made? Can existing theories of presidential and congressional influence and partisan politics in monetary policy accommodate them, or do we need new theory? And of course-though it is a little like asking Mrs. Lincoln how she felt about the play-how well do the existing theories explain the rest of monetary policy in the 1980s? It is trite but always worth repeating that monetary policy is made in a political context. Its politics should be as subject to systematic explanation as any other. Models of monetary policy-making range from the economists’ rational maximizer of social welfare to the organizational process theorist’s James E. Alt is professor of government at Harvard University.

41

42

JamesE. Alt

muddler who stumbles incrementally from crisis to crisis. In between lies the view taken in this paper of the monetary authority as an agent, who maximizes his own welfare in the institutional context of incentives set by another (the principal) so that the agent’s maximizing serves the principal’s purposes. This agency perspective has two immediate consequences. First, there is little purpose in asking whether the agent’s behavior serves his or the principal’s purposes. It should serve both. Instead, we seek to understand how the structure of the institutions affects the agent’s performance and, at the same time, given the principal’s desires, why such institutions should have evolved. Second, the Federal Reserve is an agent-in spite of its “independent” status-with several principals (Goodfriend 1988). Focusing on the multiplicity of principals and the agent’s motives (particularly, the chairman’s desire to be reappointed) integrates findings that otherwise seem contradictory. So while we do not understand the politics-or the economics-of monetary policy perfectly, the argument goes, there does not, on the whole, seem to be much need to jettison all our existing theory.

2.1.1 %o Questions, Three Theories However, “the existing theory” is not a tight, unified corpus of empirically testable, deductively linked propositions. Even leaving aside technical questions, it includes the economic theory of regulatory capture, the politicaleconomic theory of regulatory design, political business-cycle theory, what has come to be called partisan theory, the theory of bureaucracy, and the economics of international coordination. Throughout this rich mklange of ideas, however, most who write on the political economy of monetary policy want to answer one of two basic questions. First, what does the Fed do and why? And, given the answer to the first, should we change the design of our institutions for making monetary policy, and how? The answers that exist come largely from three directions. One is called “partisan theory.” Applied to monetary policy it predicts finding politically created effects coming ufer elections, as presidential influence is used to bring about policies beneficial to the supporters of the president’s party. Evidence supports the expectation that partisan presidential influence should be reflected in regular swings in policy, that Democrats favor easier, and Republicans tighter, policies. Second, the theory of the “political business cycle” predicts easing of monetary policy before elections, attributed to (usually presidential) political influence motivated by the demands of reelection politics. There is supporting evidence, but it also appears (as much in the announced targets as in the actual policy outcomes) that the Fed generally avoids visible policy changes (that is, ducks out of the storm) during election years. In fact, our model of the president-Fed relationship unifies these first two approaches. Considering the Fed chairman’s desire for reappointment along with the president’s desire for reelection provides predictions of different sorts of behavior at different times. Whether pre- or post-election shocks to the

43

Ducking out of the Storm: U.S. Monetary Policy

money supply occur depends on whether an incumbent is running, how probable his reelection is, and how long the chairman and the president have served, among other things. Third, “congressional influence” is also claimed to exist in monetary policy. While presidential influence is rooted in the power of appointment, congressional influence arises from the ability to amend the Federal Reserve Act. However, there is more congressional involvement in financialregulatory policy then monetary policy, probably because the distributive aspects of monetary policy do not generally translate neatly into the geography of congressional districts.

2.1.2 One Theme: Agency Relationships While the Federal Reserve is nominally an independent agency within the government, it is not autonomous. It lacks the ability to set rules of communication and establish the structure of incentives to which others must reactin short, to design the game. Rather, it should be expected at most to act optimally within a set of incentives created by others, its principals. It is an agent with several principals: the president, Congress, and the interest groups that make up the financial community. Introducing principal-agent models has two main consequences. First, in general, in multiple-principal problems the inability of principals to coordinate strategies may dominate their ability, given a joint strategy, to coordinate agents’ behavior optimally (Myerson 1982). Without simple, unique equilibria in such models, the posited standards that economists like to use to specify optimal economic policy in normative models cannot be rooted in rational collective action. However, absence of unique equilibria need not prevent stable interactions if principals bargain over policy and choose an agent who has strong preferences for the bargaining outcome. Then, however, in the political context of monetary policy, things that affect the bargaining outcomesay, divided partisan control of government or ideological conflict within administrations-will affect the agent’s behavior as well. Second, thinking of monetary policy this way focuses attention on the interesting feature of principal-agent problems, the agent’s private information. The essence of these problems is that, while principals seek incentive systems in which agents optimally carry out principals’ intentions, the agent’s behavior is typically difficult to monitor, his level of effort impossible to measure, or his performance imperfectly subject to verification, any of which means that the agent knows something about the situation that is hidden to the principal. The agent exploits the private information, at a cost to the “optimal” execution of the principal’s goals. The gains from specializing the agent’s expertise or ability induce principals to tolerate this agency cost. The “independence” of the Federal Reserve, that is, the space it has to pursue its own preferences, depends both on the heterogeneity of preferences of its principals and the extent of its private information. Much of the literature

44

JamesE. Alt

on monetary policy reflects the importance of private information. How does the Fed operate? What determines voting outcomes in the Federal Open Market Committee (FOMC)? All the work on estimating monetary policy reaction functions represents an attempt to figure out what the Fed’s strategy is and what, if any, rules it might have been following. Does the Fed do what it says it is doing? And does it have the effect it is supposed to? These concerns make sense when the agent has considerable private information about the economy that the principals seek to discover through appointment and reappointment power, forced disclosure, and various threats of legislative intervention, major themes in the relationship between the president, Congress, and the Fed in the 1970s.

2.1.3 Implications for Institutional Design Studies of central bank laws and economic outcomes suggest some connection between at least partial central-bank independence of government and lower long-run inflation (Banaian, Laney, and Willett 1986; Parkin and Bade 1985). The laws set a basic structure for interaction, but practically, in the model discussed above, the Fed’s independence reflects both divergence of the principals’ goals and the agent’s private information. I model the Fed’s announced targets to find how the goals of monetary policy get chosen, whether or not there is discretion, what the structure of choice is, what rules the Fed appears to follow, and how complex they are.2 I find no evidence of systematic deception, though one can see how strategies of ambiguity and obscurity on the Fed’s part serve to maintain the Fed’s at least partial independence. The final part of the paper discusses institutional design and relates independence to accountability and effectiveness. Many economists seem disillusioned with maintaining Federal Reserve independence as a goal of policy. For instance, Meltzer (1989) concludes his review of Federal Reserve policy with a plea for institutional reform to make the Fed more accountable, particularly to the president. But empirical evidence and the model of an agency relationship make it clear that such a change would have significant and undesirable consequences beyond the ones he seeks. Similarly, Friedman (1985) seeks to coordinate presidential and congressional economic intentions better to guide monetary policy. But reducing the heterogeneity of the principals’ preferences may reduce Fed independence even more than directly altering accountability procedures. Before taking up accountability, institutional design, and the consequences of different arrangements, I review major economic and policy developments of the 1980s, bureaucracy models (which treat the Fed as autonomous), and then agency models of Fed behavior with Congress, interest groups, and the president as principals. We look first at Congress for an understanding of the institutional form of the Fed, then at the financial community to see its advantages in monitoring, and finally at the president, to unify the evidence of pre-

45

Ducking out of the Storm: U.S. Monetary Policy

election and partisan cycles in policy. I characterize the targets of monetary policy directly from the records and indirectly by observing Federal Reserve actions and policy outcomes and consider whether the balance shifted between domestic and international sources of policy choice, and why.

2.2 The Economic and Political Context To make the topic manageably narrow, I treat the study of American monetary policy as a matter of explaining the causes and consequences of decisions taken by the FOMC. The FOMC comprises the chairman and other members of the Board of Governors of the Federal Reserve System and the presidents (only some of whom vote at any time) of the regional banks of that system. Other agencies are also involved in aspects of monetary policy (for example, exchange rate management is really the responsibility of the Treasury) and the Federal Reserve has other responsibilities (often regulatory) that involve it in politics. Nevertheless, say “monetary policy” and people think first of the management of interest rates and the supply of money. With this view of monetary policy, the natural place to begin a survey of the 1980s is actually in July 1979, with the appointment of Paul Volcker as chairman of the Federal Reserve Board. He served for eight years. I provide a description of the procedures and outcomes of monetary policy, starting with the first policy experiment. Then I place monetary policy in its domestic and international economic context and look at the instability of some important economic relationships underlying monetary policy.

2.2.1 The Policy Experiment and the Economy in the 1980s The nature of the experiment was as follows. Briefly, in the 1950s and 1960s, monetary policy had been oriented toward providing stable money market conditions with little systematic attention paid either to monetary aggregates or quantitative measures of broader economic conditions. Starting just before the appointment of Arthur Bums as chairman in 1970, in response to developments in research and the economic record of the 1960s, the FOMC began a policy of manipulating the federal funds rate with a view to exercising some control over the growth of the money stock. In this period, both before and after a variety of congressional mandates to do so, the FOMC began publishing “targets” of monetary policy, quantitative ranges within which they hoped to steer both interest rates and the money stock. Essentially, the procedure was one in which the FOMC intermittently chose long-term targets for money growth consistent with its broader objectives (e.g., stable prices), and then, at each subsequent meeting, chose a short-term target for the federal funds rate and related targets for monetary aggregates believed to be consistent with the long-term target. Figures 2.1 and 2.2 show the outcomes and published short-term targets (as available) from 1970-88. For the period up to 1979, the federal funds rate

46

James E. Alt Per cent

20

t rate

0

, 1970

1

1980

I

1990

Month and year

Fig. 2.1 Announced targets and actual federal funds rate in per cent, 197089; data have been slightly smoothed

(fig. 2.1) was closely targeted, while the money stock (fig. 2.2) generally missed even its short-term targets by a good deal, usually on the high side, apart from a brief period in the mid-1970s. Meltzer (1989) reviews several accounts of this, including underestimates of the demand for money as well as the effects of other shocks, but concludes that the Fed simply emphasized the interest rate targets and ignored the monetary targets. Hence the 1979 experiment. Announced on a Saturday to great confusion in money markets (and some within the Fed-see Melton 1985), the reform gave greater emphasis in Fed decisions to the supply of (nonborrowed) bank reserves and less to interest rates.3 The idea was that a change in demand for money, which previously was satisfied at an unchanged interest rate by a change in bank reserves, would now be satisfied at unchanged reserve levels by a change in interest rates. Attention focused on the narrow money supply M1, and the goal of policy was to bring down the annual rate of growth of M1 by one percentage point a year for several years from the 9 per cent growth rate prevailing in 1980.4The narrow money supply M1 grew exactly as hoped in 1981 and 1982, but the experiment ended some time in the late summer of 1982. Then monetary growth leapt, with M1 growing by 11 per cent in 1983 before returning to 7 per cent in 1984.5

47

Ducking out of the Storm: U.S. Monetary Policy Per cent

15 -

+++

Announced targets

-Actual growth rate of M2 10 -

5-

O

t

I

1970

I

1980

I

1990

Month and year

Fig. 2.2 Announced targets and actual growth rate of money supply M2 in per cent, 1970-89; data have been slightly smoothed

Figure 2.1 shows that both the level and the variability of the federal funds rate and its targets increase between 1979 and 1982. The targets for money supply M2 (see fig. 2.2) bounce around a lot in the period, and there are significant errors in outturns v i s - h i s targets, but clearly the variability of M2 also declines during the experiment. For more or less the whole period it remains within the only published long-term target growth rate range, 6-9 per cent per annum. Since 1982, the federal funds rate has resumed tracking its targets almost as perfectly as before, and huge swings-again, mostly on the up side-have reappeared in money growth. For the 1980s as a whole, M1 growth is just under 8 per cent per annum, two points higher than in the 1970s, while M2 growth averagesjust over 8 per cent, the same as it was in the 1970s. As figure 2.2 showed, M2 growth is persistently higher in the first half of the 1980s, but lower in 1985-86. In terms of broader economic aggregates, table 2.1 shows that the policyinduced increase in unemployment in the early 1980s was underpredicted in the 1982 Economic Report of the President of the Council of Economic Advisers (reprinted in Tobin and Weidenbaum 1988) and that the ultimate reduction was achieved about two years later than initially expected. The inflation rate, by contrast, fell more rapidly than expected, though the long-run level

48

James E. Alt Outcomes and Forecasts for the 1980s

Table 2.1

Year 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989

Per Capita Disposable Income Growth

Forecast

.5

.I - 1.2 2.4 4.9 2.3 2.7 2.1 3.6

4.3 4.1 2.7 4.6 4.0 4.0

Inflation Rate 13.5 10.4 6.2 3.2 4.4 3.6 1.9 3.7 4.1 4.8

Forecast

6.6 5.1 4.7 4.6 4.6 4.4

Unemployment Rate 7.2 7.6 9.7 9.6 7.5 7.2 7.0 6.2 5.5 5.3

Forecast

8.9 7.9 7.1 6.4 5.8 5.3

Nore: Source of forecasts is the 1982 Economic Reporr of rhe Presidenr (in Tobin and Weidenbaum 1988). Other data are from Citibase.

seems to be about what was anticipated. However, the 1982 Report was not alone. Michigan consumer surveys show that people substantially overestimated the inflation rate in the early 1980s, and surveys of financial market participants show that, after an initial steep drop in the long-run expected inflation rate between the end of 1980 and early 1982, expectations for the next decade stabilized at 6.75 per cent, about the value of trend inflation to that time. Indeed, after the easing of monetary policy (the federal funds target was reduced from 12.5 to 9 per cent in August 1982), leading economists believed that the higher rates of money growth in 1983 would produce 9 per cent inflation in 1984. By far the most optimistic forecasts of the Economic Report of the President dealt with fiscal policy, output, and investment.‘j Only in the election years of 1984 and 1988 (nearly) did per capita disposable incomes reach the projected levels of the 1982 report (see table 2. l).’ In the longer term, real per capita output growth has been level at just above 2 per cent per annum since World War 11, and (averaging the tax-cut-induced boom of 1984 with the recession of 1981) the 1980s are no exception. After the big surge in the early 1980s, the cyclically adjusted, constant price federal fiscal deficit is now just a little larger than one would have projected a decade earlier from 1950-79 data. Similarly, hopes that tax policy would durably affect investment proved illusory. Neither gross nor net private investment departs substantially from long-run trends (projected forward from 1979) across the 1 9 8 0 ~ . ~ 2.2.2

International Context

The second “experiment,” if it existed, corresponds to the period of decline in the exchange value of the dollar in and after 1985. It was certainly not the original intention of the Reagan administration: expressing confidence that an

49

Ducking out of the Storm: US. Monetary Policy

orderly domestic monetary policy would curb disorder in international exchange markets (1982 Report, p. 173), intervention to support the dollar was to be eliminated, or at most “minimal” (Sprinkel, cited in Melton 1985, p. 172). However, traces of a second “experiment” appear in figures 2.1 and 2.2 in the surge in money and easing of interest rates. But did the Fed bring down the dollar? The conjunction of a small and short-lived shift in domestic interest rate targets with a larger decline in the dollar’s exchange value (and surge in the money supply) suggest that policy change was minimal, geared to signaling and orchestration of market-led developments. However, the prominence in public political discussion of international policy coordination suggests that we should try to determine more systematically whether, for how long, and to what extent, the period after 1985 reflected unprecedented attention in monetary policy to concerns of international origin. However, international concerns in U.S. monetary policy are not new. The exchange value of the dollar intruded on domestic politics in the 1960s and 1970s as well as the 1980s. Indeed, some argue that the 1979-82 experiment had international origins, occurring only after Volcker was convinced that without an increase in American interest rates the dollar and the international financial system faced a grave crisis. Others believe that the 1982 policy easing was brought about by the possibility of a banking collapse caused by the effects of high real interest rates on Third World debt. Both these episodes are international aspects of domestic problems. Neither case really shows the Fed doing something for international reasons that it would not have done for domestic reasons. Thus, whether there really was a second experiment depends on whether international concerns became goals of policy after 1985 in an unprecedented way, and whether monetary policy was actually coordinated (as opposed to being the subject of public discussion and negotiation) among nations. What was different in the 1980s? First, it was a period of dollar strength and capital inflows. Figure 2.3 charts the real effective exchange rate of the dollar since 1967.9 The real dollar declines by something like 30 per cent between 1970 and 1978, bottoms out in 1979-80, and, after 1980, commences a huge upward surge that lasts half a decade before coming down (even quicker) and stabilizing around its 1973 levels in 1989-90. Movements in net foreign investment, the capital inflow left after imports and net international transfer payments are subtracted from exports (the lower series in fig. 2.3), closely follow the real exchange rate in the 1980s. In fact, they only share common trends; Granger-type tests show that neither “causes” the other. Moreover, from the end of 1981 to 1985 American real interest rates (longterm bond yields minus expected inflation) were above-often significantly above-the range of comparable long-term real interest rates of the two principal substitutes for the dollar, the yen and the deutschmark. In the period of dollar weakness in the late 1970s the real return on dollars is below the substitutes’ range, but only at the end of the period in 1978 and 1979. And this

50

James E. Alt 1973 = 100

150

...' .

Real exchange rate

++Ad

50 -

O

++++

Net foreign investment

:&++%+=$&$

-50 -

$ Billions

A

+ .,*

?@+ -

0

'....+++y+++ +

,

100

%t

+

++

+" w+

I

"+ 2

,

I

I

I

,

,

Fig. 2.3 Real trade-weighted relative U.S. exchange rate, left scale, March 1973 = 100 and net foreign investment in billions of 1982 dollars, right scale, 1967-89

takes place in a context of growing dollar deposits outside the United States: the eurodollar market had grown enormously, from one-fortieth of M2 in 1965 and one-tenth in 1970 to equality by 1980. However, the relationship between the dollar's real exchange value and American real relative interest rates was not new in the 1980s. Possibly the direction of effects changes, from interest rates set for domestic purposes affecting exchange rates to interest rates set to affect the exchange rate having other domestic consequences. Moreover, other factors influence real exchange rates, and some of them are political. We analyze these, and also describe the (limited) impact of international factors on American monetary policy, below. 2.2.3 Changing Economic Relationships in the 1980s While output and investment remained close to long-run trend values in the 1980s, some economic relationships involving money ceased to hold. For one, there is the well-known decline in the velocity of money. The steady decline in the ratio of M1 to gross domestic product (a stable rate of growth in the velocity of money) through the 1970s changes abruptly in the 1980s. Without predictable velocity, there is no point in trying to control money

51

Ducking out of the Storm: U.S. Monetary Policy

growth. The cause of the change, still disputed, could be technical changes in the definition of M1, unanticipated complexities in the money-income relationship, or the impact of foreign demand (not pumped back into U.S. output) on the growth of M1 in the 1980s. Second, the relationship between money and inflation, loosely established over the 1960s and 1970s, also vanishes. A simple monetarist inflation model using quarterly M1 money growth plus its last four lags to project annual inflation in the quarter one year ahead, as shown in figure 2.4, does a good job of capturing broad contours of consumer price inflation from 1960 to 1979 (though the peaks of inflation, particularly in the 1970s, tend to overshoot the cumulated effects of money growth). But forecast forward from 1979-that is, fix the coefficients estimated for 1960-79 and simulate the effects of observed money growth in the 1980s-and the model utterly fails to predict the inflation rate. In fact, the parameter estimates from the 1960s and 1970s predict 20 percent inflation in 1987, and nothing in the model predicts the fall in inflation in the early 1980s. A parallel forecast (not shown) using M2 is less extreme, and does pick up the decline of inflation from 1979 to 1982, but also predicts double-digit inflation after 1985. Neither model predicts the lack of inflationary impact of the surge in money growth after 1985.

20

-

10

-

Actual inflation O t

-lo/,

,

1960

,

,

,

I

,

,

,

,

1970

,

,

1980

,

,

,

I ,

1990

Quarter and year

Fig. 2.4 Inflation rate based on consumer price index and forecast one year ahead from money supply M1, quarterly, 1960-89

52

James E. Alt lnilation rate in per cent

15 -

5-

-5

,

1

2

4

,

1

1

1

6

1

1

1

1

8

1

1

1

1

10

Unemploymentrate in per cent

......... tt+

1947-1968 1969-1978

- 1979-1989

Fig. 2.5 Inflation rate based on consumer price index and unemployment rate, monthly, 1947-89; data have been smoothed

Finally, figure 2.5 provides a historical chart of inflation and unemployment. It shows the broad contours of the trade-off in the years before 1969, with the recession loops of 1949, 1958, and 1961 visible along with the inflationary surge of the Korean War. Starting in 1969 there is first the Nixon loop, ending in the vertical inflation acceleration of inflation in 1973-74, and the Ford-Carter loop, ending in the vertical acceleration of 1978. The counterinflation policies of the 1980s, neither immediately credible in the financial world nor the beginning of an era of stable money growth, had significant employment costs, but their extreme values lead ultimately back to values in the middle of the Nixon loop of 20 years ago. But are we a 4-point blip up in inflation away from starting the 1980s over again? Or a 2-3 point reduction away from getting back to that golden age, the trade-off of the 1950s?

53

Ducking out of the Storm: U.S. Monetary Policy

2.3 Political Models of Monetary Policy-Agent

and Principals

2.3.1 Inside the Fed The Federal Reserve has a broad mandate to manage money “in the public interest.” Most of its decisions lack controversy and are made in secret, its Governors serve long terms, it controls its own budget. Thus, many economists model the Federal Reserve as a unitary actor formulating monetary policy to maximize social welfare. Even such an actor maximizes subject to beliefs about how the economy works, so understanding its choices in the 1980s requires taking account of the influence of prolific scholarship by monetarist economists in the 1970s and later those interested in international coordination of policy on policymakers’ views of fundamental relationships between money, inflation, interest rates, and exchange rates. Economists are active pamphleteers on this supply-side of the market for ideas, but each policy experiment also shows how the demand for “solutions” was increased by an unprecedentedly prolonged problem, inflation in the 1970s and the high real dollar exchange rate in the 198Os.’O Diverse research staffs of the regional banks ensures that new ideas get inside representation somewhere within the system, even if monetarism itself was resisted by many bank officials. Leaving aside the beliefs, whether the Fed is agent or independent, its objectives need to be characterized. “Bureau” theorists treat the Fed as autonomous and explain its behavior as self-interest on the part of Fed officials. For instance, the Fed earns revenue from interest on securities purchased in open market operations. It could be that the Fed maximizes revenue to maximize spending (Toma 1986) or employment (Shughart and Tollison 1986). However, while the Fed consumes more when it earns more, neither paper shows that it generates excessive revenues. In fact, the Fed turns over most of its excess revenue to the Treasury. It could also be that the Fed seeks to increase its influence (Beck 1988), an idea that is particularly important when the Fed’s regulatory powers are at issue. Finally, authors as diverse as Chant and Acheson (1986) and Woolley (1984) stress the Fed’s goal of maintaining or increasing discretion, independence, or flexibility. Others suggest that “avoiding crisis” is its goal. But, by themselves, flexibility, independence, and crisis avoidance are Iess goals than operating procedures. Some independence comes with private information, and naturally the agent does not give it up. Moreover, some goals are being pursued even while attention is paid to the risk that things might get out of hand: crisis avoidance says nothing about what these goals are, beyond the ideas of bureau theory. Since a good deal of policy disagreement exists within the Fed, its policy goals may not be consensual. Voting behavior in the FOMC can be analyzed to describe the Fed’s “welfare function.” Though the chairman dominates Fed

54

James E. Alt

decision making (see Kettl 1986 in leadership), even a strong chairman like Volcker was overruled on at least two occasions, once in 1984 and once in 1986,” and dissents from FOMC directives are common. Woolley (1984,6264)pointed out that presidents of regional banks were more likely to dissent for tighter policy, while governors were more likely to dissent for easier policy. Moreover, he shows that over 80 percent of those governors’ dissents favoring tightness were cast by governors appointed by Republican presidents, a point which takes on more significance in the context of the relationship between Fed and president (below).‘* 2.3.2 Congressional Politics and Oversight The traditional view of the relationship between the Federal Reserve and Congress is that, while the Fed is technically subject to congressional control and is an agency created by the Federal Reserve Act, which can be amended by Congress, the relationship remains largely symbolic. Of course Congress (with the president’s consent) could change the Fed’s status, and legislation is frequently introduced with that purpose. However, this legislation rarely passes, and, on the whole, while Congress has sought more information about monetary policy and complains vocally when interest rates are high, it has never moved positively to take direct control of monetary policy. Moreover, Congress does not appropriate the Fed’s budget, and thus the Fed avoids normal congressional oversight. In fact, the Fed is an example of what McCubbins and Schwartz (1984) call “fire-alarm oversight.” The Fed’s mandate is broad and vague. Such a mandate promotes the possibility of regulatory capture by the affected industry, whose activities facilitate the reelection of congressional representatives. The costs of such capture to the public are widely dispersed. As long as the industry is content, Congress is as well, and hence Congress attends to policy only when it gets complaints from interest groups-the fire alarms. When divisions appear wifhin the financial community, not only the Federal Reserve but also representatives of different sectors within the financial industry become directly involved. Financial deregulation legislation in the 1980s-“nonbank banks”-was a case where a wide division of interest between smaller and larger institutions coupled with changes to and fro in composition of congressional committees produced years of legislative initiative and frustration (Woolley 1988b). Does Congress influence monetary policy anyway? Grier (1988) argues that politicians who especially need to serve constituents with policies requiring monetary outcomes will join the committee that controls monetary policy. If the committee has veto power, then there will be an equilibrium in policy reflecting the committee median policy position. If the committee chairman has a veto, it will be the chairman’s policy that dominates. Thus changes in committee chairs-or, more specifically, changes in the relative distance of chair from committee median or of the committee from the chamber me-

55

Ducking out of the Storm: U.S. Monetary Policy

dian-produce changes in policy. Direct congressionalhearings or control are not required, since ex post threats to reduce the Fed’s independence (that is, to amend the Federal Reserve Act) are sufficient to extract compliance. Three testable empirical propositions follow from this model of Fed as agent and Congress as principal. There should be (1) self-selection onto the banking committees of representatives with strong banking interests in their constituencies, (2) attempts by Congress to reduce the Fed’s private information to improve monitoring and verification of performance, and (3) monetary policy changes corresponding to changes in the preferred policies of congressional chambers, committees, and chairs. Bank PACs concentrate donations to Banking Committee members (Woolley 1984, 135). Woolley (1988b) demonstrates self-selection, with banking committee members’ districts overrepresentingthe national average in savings and loans and real estate presence by about 2: 1. These representatives are the “high demand” preference outliers that populate committees. Krehbiel(l991) cites Shepsle’s (1978) evidence that the best predictor of requests for Banking and Currency is “financial or real estate occupational background” to support information-economizingin committee appointments, as Congress organizes efficiently to discover the agent’s private information but waits for specialist interest groups to raise the fire alarms. Second, there was a flurry of congressional activity over information in the 1970s. Ultimately the GAO was authorized to audit the Fed’s administrative practices, if not its monetary policies. Kettl(l986) reports that the Fed cut in half the time it waited to release FOMC directives in 1975 “to deflect congressional pressure for full disclosure.” Confronted with an adverse court decision in a Freedom of Information Act suit in 1976, the Fed reduced the delay still further, but at the same time cancelled its long-standing policy of keeping (but publishing with greater delay) minutes of its meetings. Other legislation compelled the Fed to publish its targets and the chairman to report to Congress on them. (However, Fed chairmen can generally cloak themselves in vagueness, and there has frequently been little understanding of monetary policy in Congress.I3)Peterson and Rom (1988) report that chairmen of the Fed frequently take the opportunity of a visit to Congress to criticize fiscal policy. While a partially successful effort was made to lower the costs to Congress of monitoring Fed behavior, the Fed was able to maintain the secrecy with which its decisions had traditionally been taken. Woolley (1984) gives an excellent history of the debates over Congressional attempts to increase scrutiny of (if not responsibility for) monetary policy. In fact, while the toothless measures that ultimately passed led scholars to conclude that Congress was largely irrelevant to monetary policy, if Congress was trying to preserve fire-alarm oversight, but increase its-or even more, financial interest groups’-ability to monitor the Fed, this conclusion is misleading. Grier (1988) claims that significant changes in monetary policy coincide with ideological changes in Senate Banking Committee chairs but changes in

56

James E. Alt

1980 make this ambiguous.I4In a recent paper I have been unable to obtain, Ferejohn and Shipan report a similar finding for the House. Belden (1989) reports a consistent but statistically insignificant difference in the likelihood of bank presidents’ dissents when Proxmire was chair compared to Sparkman and Garn. All this is possible, but not conclusive. A note of caution: Grier closes his paper with the prediction that the change from Garn back to Proxmire after the Democrats regain the Senate in 1986 should produce monetary easing. A glance at figure 2.1 shows that it does not. One other thing is clear. Widespread distress and anger over high interest rates gets Congress to take public positions decrying monetary policy. The flurry of congressional activity in the mid-1970s was triggered by the 1975 recession. During the 1979-82 experiment, many bills were introduced in Congress, including proposals condemning high interest rates, requiring the president to assure adequate affordable credit to small borrowers, requiring the Fed to abandon money targets and reintroduce targeting of interest rates, and even one to impeach Volcker (Woolley 1984). Kettl(l986) produces more systematic data on congressional attention, which are measured as the total of congressional bills and resolutions each year addressed to the Federal Reserve or monetary policy. I extended his data through the 1980s, and over half the variance in congressional attention, year to year from 1950 to 1988, can still be explained by a simple regression model involving only the current shortterm interest rate.I5 From the point of view of individual constituents (unlike interest groups), this congressional position taking and blame shifting dominates Grier’s “distributive politics” view of monetary policy: the geographic basis of the distributional consequences of policy are insufficiently clear to justify the sort of specialization to committee oversight required by his theory.

2.3.3 The Financial Community The Fed’s broad mandate along with heavy congressional involvement in intrasectoral disputes and blame shifting when interest rates are high are all consistent with the model of financial community as coprincipal. The financial community has all the usual resources discussed when an industry “captures” its “regulators.” It has information that the Fed needs, and it supplies personnel through circulation and recruitment of staff. Many staffers and FOMC members come from financial industry backgrounds and expect to return. For example, when Robert Heller quit the Federal Reserve Board of Governors (citing low salaries), he became vice president of a bank. The financial community is clearly able to monitor Fed behavior. The Fed and the financial community interact frequently. Goodfriend (1988) points out that business representatives serve as reserve bank directors. There is regular contact with the financial community through the New York Fed. Formal contact through the Federal Advisory Council and informal contact through meetings coexist. The regional bank presidents provide direct representation on the FOMC. Bank representatives’ public statements indicate general happi-

57

Ducking out of the Storm: U.S. Monetary Policy

ness about Fed policy (Woolley 1984). Since the interest groups rarely go public (except on regulatory matters)-that is, the fire alarms do not go off too often-they must have alternative strategies for monitoring. What the financial community seeks is generally taken to be stability and predictability in market conditions (Poole 1976). Fed performance will be easily verifiable, since both open market operations and the stability of interest rates are routinely observed. In fact, the main argument against the view that the financial community is the Fed’s principal (though not against their desire for stability) is the community’s broad opposition to the monetary targeting experiment (Woolley 1984). Does the experiment show that the financial community ceased to obtain interest rate smoothing from the Fed? Hardly. First, a glance at figures 2.1 and 2.2 shows that interest-rate smoothing seems far more important than monetary stabilization. There is more variance in interest rates 1980-82 (and less than usual in monetary targets) but closer inspection shows how briefly this lasts and how interest-rate smoothing has reasserted itself. The Fed’s interestrate targets vary a little more in the 1980s than in the 1970s, and of course since 1980 it has only published a 4 percentage point range, not an individual target value. Nevertheless, the impact of the “experiment” was smaller than one might believe from some of the contemporary descriptions. Table 2.2 formalizes this with estimates of an error-correcting model (Davidson et al. 1978) of the target-setting function. The error-correcting model is of the form

AT =

(Y

+ PAX + y ( T - , - 6X

where AT is the latest change in the announced target, AX is the latest observed change in the federal funds rate or money supply, and T - and X- are most recent observed levels. The parameters to be estimated include any secular trend (the constant, a),the short-run effect of the last observed change in outcomes(p), and the effect of the long-run, cast as a reequilibration rate (y) of the deviation of the last target from its equilibrium level fix-,. The results for the federal funds rate are easy to understand. For the whole period, they show that 70 percent of the last change in interest rates is accommodated in the current change in the target. That is, changes in the target follow changes in the outcome, but damp the extent of the changes. Coefficients hardly change in the short period of the experiment, though the mean change is larger by a factor of six and the standard error of the regression is bigger by a factor of two, indicating greater target variability in this period. The reduction in the 6 coefficient suggests that the equilibrium level of interest rates was now below the observed level during the period, which is believable given doubledigit interest rates. With respect to monetary targeting, no monetary variable-base, free reserves, or M2 changes-ever appears significantly in the target equation. The overall structure remains the same in the policy experiment period as in the

58

James E. Alt Error-CorrectionModel of Monetary Targets, 1974-88

Table 2.2

Whole Period

1979-82 Experiment

Equation (l), federal funds rate: a

P

Y

F

R2

Observations SE

.05

.70*

- .35* 1 .O* .54 176 .58

Equation (2), money supply M2: a 2.42* P .08 Y - .28* 6 .03 R2 .12 Observations 176 SE 1.1

.70 .69* - .40* .80*

.60 48

.90 5.68* .24 - .52* - .25 .27 48 1.2

Note: Data are from Citibase and the Federal Reserve Bulletin, various issues. Targets in months in which the FOMC did not meet are given their previous values. ‘Coefficient is significant at .05 level or better.

whole period, one of return-to-normality targeting around an exogenous, unchanging long-run target level. The reequilibration coefficient is twice as high in the short period, so the pull back to equilibrium target level from deviations is stronger. Hence, targeting appears to follow orthodox monetarist recommendations: a single long-run target for money growth with much closer attention paid to complementary short-run money growth targets during the policy experiment. That, with the reestablishment of interest-ratesmoothing after 1982 suggests that the financial community continues to get what it wants most, which is stable and predictable operating conditions.

2.4 The Political Monetary Cycle with the President as Principal 2.4.1 Cyclical theories There are two political-cyclical theories of macroeconomic policy. One, the “political business cycle,” predicts that reelection desires of politicians lead them to create desirable economic conditions immediately before elections, even if these policies require costly adjustment later. If a short-sighted electorate rewards governments for such behavior, the electorally induced cycle creates unnecessary costs (Nordhaus 1975). The other, “partisan theory,” predicts that partisan governments deliver benefits to their core constituencies. Economic policy choices have different distributional consequences for these con-

59

Ducking out of the Storm: U.S. Monetary Policy

stituencies, and politicians deliver these redistributions through economic policy. Modifications have been proposed since Hibbs’s (1977) original work on partisan preferences for unemployment and inflation, but all predict policy changes of a redistributive nature in post-election periods. While Hibbs predicted policy changes persisting across entire incumbencies, others have proposed that post-election effects should be transitory. For one, rational economic models imply that private-sector adjustments will offset political shocks, at least once the “surprise” of discovering which party wins the election has worn off (Alesina and Sachs 1988). Alternately Chappell and Keech (1988) propose that rising inflation may raise the cost of reducing unemployment (and vice-versa) beyond what a support-maximizing government would pursue. Also, the need to build new coalitions dictates policy changes once the “debt” to one’s supporters has been paid off (Alt 1985). The evidence suggests that post-election policy shocks-when observed-have had transitory effects, but the reasons have not been tested against each other. However, some find the theory inapplicable to American monetary policy, citing the Fed’s “independence’’or the neutrality of anticipated monetary policy. The empirical evidence-which should really be decisive-is divided. Scattered findings support the pre-election theory in monetary policy (e.g., Grier 1984; Haynes and Stone 1988) while others contradict it (Beck 1987). Chappell and Keech (1988) show that post-election effects appear to exist but have limited impact on the real economy. I unify these two theories in a simple model in which both pre-election and post-election effects could appear, but in which the timing of the chairman of the Fed’s appointment has a critical role in determining monetary policy.’6 This model also helps us understand some important, but more anecdotal, questions raised by the first policy experiment, given our political theories of monetary politics. For one, the experiment employed monetarist procedures and targeted money and inflation to an unprecedented extent. Partisan theory makes the fact that Volcker, who spearheaded the experiment, was appointed by a Democratic president seem anomalous. Moreover, the apparent beginning of a credit squeeze in an election year (1980), with an incumbent running for reelection, seems to contradict political business cycle theory. We will take up these episodes after discussing the model and the other evidence.

2.4.2 Reappointment Politics If the president is the Federal Reserve’s principal, his ability to set incentives has to derive ultimately from one of three sources: the power of appointment, the president’s influence over nonfinancial economic policy, or the strength of moral suasion, which inheres in elected representatives with popular mandates to change policy. The last exists, but it wears off quickly after elections. Influence over nonfinancial policy is most important when coordination with other executive agencies is most useful or necessary for successful execution of monetary policy, as, for example, when an internationally coor-

60

James E. Alt

dinated intervention is an object of policy. Beck (1982) points out that appointment power is limited. Fed governors, after all, serve 14-year terms, and the presidents of the regional banks are not subject to presidential removal. There is indirect evidence of presidential influence on FOMC voting. Recall that Woolley (1984) showed that voting for easier or tighter policies varied with partisanship of the appointment. Moreover, in the FOMC governors vote differently from regional bank presidents. Gildea (1987) shows that Federal Reserve Board Governors’ political affiliation affects how likely they are to vote for noncontractionary policies: self-described Democrats are more likely to favor easier policies. (Surprisingly, he does not check whether political affiliation covaries with party of the appointing president.) He also shows that whether governors’ (and again, other FOMC members’) vote for monetary ease apparently depends on the U.S. president’s current approval or disapproval rating, though, of course, this disapproval rating is itself a weighted combination of current economic conditions (among other things). The president’s power to appoint the Fed chairman is more important. Political business cycle theory makes the president’s desire for reelection a prime policy motive, but why should the president’s desire for reelection be so much stronger than the Fed chairman’s desire for reappointment? Obviously, the chairman’s problem is different: reappointment requires the approval of both the president and the financial community, each of which acts as a constraint on his desire to promote the interests of the other. Moreover, the interaction of the electoral cycle and the appointment cycle have clear implications for the strategy of a chairman seeking reappointment. Consider first the post-election period. The president has been elected, and he is aware that his early-term actions constitute an important signal to the electorate about policy. The Fed chairman, formerly up for appointment in year 2 of the presidential term, recently in year 3, has a clear incentive to give the president what he wants, at least initially.” If what the president wants is consistent with what the financial community wants, he has even more incentive to do this. The less common history the president and chairman share, the larger this same incentive. So one would expect to observe early-term monetary policy following the desire of the elected president, more so when the president is new since first impressions count relatively more then, and more so when the president is a Republican, if there is closer convergence of preferences of the financial community and the Republican Party. This treats the interaction as an iterated game of incomplete information in which the president tries to discover the chairman’s “type,” before reappointing him, introducing elements discussed in reputational models of monetary policy (Rogoff 1985, 1987). By contrast, in the pre-election period, the chairman still wants reelection but is not certain who will be reappointing him. The temptation must exist to give the incumbent what he wants, more so when the incumbent’s reelection is more probable, though, paradoxically, the more likely the president is to be

61

Ducking out of the Storm: U.S. Monetary Policy

reelected, the less he needs the chairman’s help. Essentially, the chairman must balance the probability of being seen to have offered a decisive advantage to whomever won with the probability of being seen to have offered a critical disadvantage to whomever won (in spite of it). However, the chairman does not know who will win the election any more than anyone else does. Much depends on the visibility of the chairman’s actions. Since monetary policy is now followed closely in the media, significant actions are sure to be noticed. Hence his strategy should be to take a middle course, and not be seen to have done anything significant, in view of the fact that he gets another chance to maximize his reappointment probability after the election, certain then of whom he has to please. Such a policy of not being seen, of “ducking out of the storm,” has at least two testable implications. Since you get noticed when you shake things up, you keep things quiet: hence, the variability of targets should decline in the pre-election period, as large and frequent changes are avoided. Similarly, you get noticed when you step out of line: so the incentive to accommodate (e.g., fiscal policy) should also be strong. But there are further implications. In almost any version of rational expectations economics, only unanticipated monetary policy has real effects. “Giving the president what he wants” post-election will be the easier, the bigger a surprise the reelection is. Clearly, when the election outcome is uncertain, some surprise is certain. But the surprise can bigger if the financial community expectations of Fed behavior are less clear. They could be made less clear by a deliberate policy of variability before the election. This sort of behavior is a feature of the model in Havrilesky (1987) and (without the electoral context) in Cukierman and Meltzer (1986). However, while this would increase the value of the post-election policy, it conflicts with the chairman’s best preelection strategy of not being noticed. The chairman’s choice will depend on his attitude toward risk. Finally, you cannot “give the president what he wants” if he does not want anything. “What he wants” is what the central economic issues and promises of the campaign indicate policy should be. That may not seem to be much of a restriction, since 1972, 1976, 1980, and even 1984 all featured campaigns with evidently central economic themes that had implications for monetary policy. It is less clear that the 1988 campaign called for a major innovation in monetary policy as an essential ingredient of the president’s program. This general rule, that elected politicians do not rush off to do something they have not bothered to promise to do, was a strong result of earlier work on partisan cycles in unemployment rates (Alt 1985) and should be a source of variability in the monetary policy cycle as well. To sum up, the presidential-principal-political monetary cycle is rooted equally in the (constrained) reelection and reappointment motives of the principal and agent. It is contingent on visibility: the chairman’s decision rule would be different if no one paid attention to monetary policy, since caution

62

James E. Alt

comes from fear of being dragged into the campaign. Presidential suasion or legitimacy of the electoral mandate and the chairman’s reappointment motive coincide in the post-election period, so it is here that presidential influence on monetary policy should be greatest, with form and extent varying as described above. While the shocks could be big, they should be transitory. If there is a pre-election cycle in monetary policy, it should more recently be a by-product of accommodation or caution, as the visibility and attention paid to monetary policy has increased. The next sections show that quantitative evidence, extended where possible to include the 1980s, is reasonably consistent with this model. So are some brief and admittedly journalistic accounts of recent appointment and reappointment highlights: some thoughts on Bums’s quest for reappointment in 1977 and Volcker’s in 1983, and a view of why Carter chose Volcker in spite of the Democratic preference for easier monetary policy. 2.4.3 Ducking out of the Storm There is a little recent evidence of pre-election cycles. Hibbs’s (1987) history of the Carter administration shows it was the opposite of the political business cycle, with the boom first and the contraction before the election. The Reagan incumbency looks just right, but it is a Republican presidency, so the pattern of recession early and boom later is also consistent with partisan theory. From 1984 to 1988, unemployment trends down (like 1964-68), offering no evidence either way. However, growth in real personal income is highest in the two presidential election years in the 1980s, as the literature on economics and election outcomes predicts (see table 2.1). But if there is a political business cycle in real per capita disposable income growth, are its origins in monetary policy? The first part of table 2.3 shows Table 2.3 Year

Money Growth Differences before and after Elections he-election Growth Rate

1960 1964 1968 1972 1976

-2.2*

1980 1984 1988

- 1.8*

1.1

2.7 1.8 .8

- 1.8*

- 8.4*

Year 1961 1965 1969 1973 1977 1981 1985 1989

Post-election Growth Rate +2.1

+ .6

-3.8 - 2.4 -b 2.4

- .7

+ 4.7* -3.2

Party Dem Dem Rep Rep Dem Rep Rep Rep

Nore: Entries are annual growth rates of M1 in percentages. Post-election rate is DecernberDecember, minus the average of the two previous years (source: Economic Report ofrhe Presidenr 1989). Pre-election growth rate is the average of the monthly changes over the election year minus the average of the previous two years (source: Citibase). Dem = Democrat; Rep = Republican. *Sign is opposite of expected value.

63

Ducking out of the Storm: U.S. Monetary Policy

that, while money growth accelerated in the year before a presidential election (relative to the two-year average before) in 1964, 1968, and 1972 (statistically significantly so only in 1968 and 1972, according to Hibbs 1987), the effect in 1976 is smaller and in the 1980s there is never a positive pre-election surge in monetary policy. This implies that the empirical evidence for and against an American political-monetary cycle may rest on the choice of the period examined. Anyone who continues to include pre-1972 data (observation of which led to the theory in the first place) is likely to find a four-year cycle; this is less likely for anyone using the post-1972 period. This does not mean there is no pre-election politics of monetary policy, however. The “credit crunch” of March 1980 is sometimes cited (Beck 1987) as contrary to political business cycle arguments, but there was an even more extraordinary back-off in targets and policy across the summer of 1980. The decision to raise the federal funds target in March was more than reversed over the next two months, leading to wide swings in money growth over the year. While Mayer (1987) quotes Volcker at the time as saying “the sooner the recession begins, the better,” the rapid policy reversal suggests that his carte blanche from the administration to reduce inflation was actually a postdated check, meant to take effect no sooner than late autumn. Greider (1987) alleges that when Volcker was outflanked by a majority favoring further tightness in the summer of 1984, he personally overrode the majority and instructed the open market operations desk at the New York Fed to maintain an unchanged course. Similarly, all the highest rates of money growth in 1988 are observed in the summer, as the actual money growth rate bows upward and stays above the targets all the while the targets are falling. All these episodes are the true pre-election politics of monetary policy, devices to stay out of trouble, keep policy from becoming too visible, hence “ducking out of the storm.” Moreover, Beck (1987) claims that statistical evidence of variations in money growth in election years cannot be found. However, he finds that Fed policy accommodates budget policy. If an electoral cycle in fiscal policy produced increased demand for money before elections, the Fed would accommodate this, introducing pre-election easing apparently without policy changes. The Fed’s operating procedures appear to continue to accommodate money demand at given interest rates. So in election years, if fiscal policy is the source of the cycle, prudent politics plus usual procedures dictate accommodating it. This will hardly show up in statistical analysis, since accommodation is passive and not discontinuous. Finally, consistent with this view of pre-election politics, there is evidence that target variation from month to month declines in election years. This incentive to stay out of trouble follows from the visibility of policy, increased by the more rapid publication of agreed targets in directives after the mid1970s. Table 2.4 shows that, apart from the unnus horribilis of 1980 described above, with its big crunch in March cancelled in April until the election was over, monthly federal funds rate target variability is always lower in

64

JamesE. Alt

lsble 2.4 Years 1976 Rest of 1970s 1980 1984, 1988 Rest of 1980s

lsrget Variability, 1974-88 Observations

Standard Deviations of Targets of Federal Funds

Standard Deviations of Targets of M2 Growth

12 69 12 24 72

.22 .73 1.73 .51 .91

1.62 1.35 1.81 .72 1.02

the 12 months preceding the election (covering 1976,1984, and 1988) than in other years. It is not that the levels of targets are higher or lower, they are just less variable. This is less true for monetary targets, but for politically sensitive interest rates, the point seems to be to convey “no change.” Consequent on less target variability in election years, as above, is that money demand is even more likely to be accommodated if demand changes through the budget. 2.4.4 Post-Election Partisan Cycle Evidence for the post-election partisan cycle in output (Alesina and Sachs 1988) and unemployment (Hibbs 1987) is well known. Hibbs (1990) shows that this unemployment cycle induces (with lags) an inflation cycle that affects rates of return on stocks and bonds as well as corporate profits and disposable incomes. Moreover, the cycle appears to be monetary in origin, at least to some extent. The second column of table 2.3 shows that money growth in the immediate post-election year (measured December-December) follows the partisan cycle neatly, though irregularly in size, save that 1985 is way out of line.’*In every other case, the prediction of partisan theory is fulfilled and the president “got what he wanted” from the Fed. The absence of the cycle in 1985 probably does not mean that the partisan cycle is declining. Figure 2.6 shows forecasts from Hibbs’s (1987) monetary policy partisan model, estimated from 1953 to 1980. These are very similar to results he publishes based on data through part of 1982. These parameters are then used to forecast the model through the 1980s, supplying money growth and inflation exogenously. The model is on course through the recession, which implies that 1981-82 economic policy was not an outlier vis-avis long-run Republican post-election policy. The forecasts do not quite pick up the 1983 boom, which is bigger than expected, but are back on course for the 1984 election. It then misses the 1985-86 monetary expansion, the second policy experiment. However, by 1987-89, the forecasts are back in range. Since the long-run model is still on course at the end of the period, the period offers no support for throwing out the model. Does the chairman run for reappointment? Martin apparently did so in 1967, as revealed in documents from the time (Woolley 1984, 1 16). Greider quotes a Fed official as saying that Bums in 1977 expanded the money supply

65

Ducking out of the Storm: U.S. Monetary Policy

(consistent with partisan theory predictions about what the president would have wanted): “There was a rapid shift in monetary policy and it was designed to ingratiate Burns with Carter so he would be reappointed chairman” (1987, p. 346). Greider also contrasts Bums’s post-election expansion with the fact that he did not “play political games to help re-elect Gerald Ford.” Exactly. If you want to be reappointed and you do not know who will be doing the appointing, caution before and cooperation after is the way to go about it. In 1981, what Volcker wanted was what Reagan wanted (but see below), and Greider extensively describes Volcker’s hopes of reappointment after 1984. Of course, just as the president’s reelection motive is not the only source of policy, the chairman’s reappointment motive is not the only source. But incentives for it and its effect on post-election policy are strong.l9 So why then did Carter choose Volcker, who was known not to be a team player and who had strong anti-inflation preferences? Anyone who looks at the annual data in King and Ragsdale (1988) can estimate a model of president Per cent

15

1

A

5 -

-5

I

1977

I

,

,

1982

,

I

,

,

I

1987

Month and year

......... +t +

-

In-samplefitted values Forecasts Actual M I growth

Fig. 2.6 Rate of growth of money supply M1 (quarterly changes at annual rates, smoothed) and in-sample fitted values forecasts from Hibbs’s (1987) partisan model, updated to 1953-89

66

James E. Alt

approval ratings in the postwar period in which approval is increased by the (longer-term) perception that the president’s party is best able to handle whatever the individual regards as the most important problem facing the country, but reduced by the amount of current inflation (in peacetime; prolonged wars also reduce approval ratings).2oBy 1979 the cost of living was seen as the most important problem facing the country by 60 per cent of the public; it was the seventh straight year it had led the list, beating Watergate, energy, and unemployment in turn. So Carter faced a double problem in terms of his approval rating. He needed to reduce inflation and be seen to be best able to do that; he needed a credible agent and he needed a solution. The reelection motive explains the choice: Carter could not have been reelected without improving his approval rating through improved inflation performance. 2.4.5 Divided Government and Agency Independence

The 1980s saw the longest period of divided partisan control of Congress in nearly a century, and having a president whose party did not control at least one House persisted throughout the decade. Does this have implications for monetary policy? That is, is the Fed’s latitude wider in a period of divided government, when the preferences of the president and Congress are less uniform? Generally, how does the number of principals and the heterogeneity of their preferences affect the Fed’s freedom of action and policy stability? Two principals can coordinate an agent’s behavior the more closely if they share similar preferences. If two principals want different things, and cannot reach a stable bargain, even a closely controlled agent’s behavior could display variability if he alternates between principals’ preferences in choosing a course of action. But if two very different principals can bargain and agree to choose an agent whose own strong preferences reflect their bargaining outcome, even a highly independent agent’s behavior will be stable. Add a third principal and the bargaining space may become more complex, but coordination still works the same way. For example, one principal (say the financial community) might have strong preferences for stability over volatility in markets but is relatively indifferent across levels of aggregate economic activity, while the others (president and Congress) might have strong preferences over economic activity but not volatility. However, they could still coordinate through bargaining and achieve stable policies by appointing someone who would strongly wish to carry out the bargaining outcome, as described above. Hence, a lot seems to depend on the agent’s nature, as well as the principals’ preferences. Volcker was chosen to implement a counterinflation policy in full view of feelings in the administration that he was no team player and might indeed prove difficult to control. His strength of commitment to these outcomes meant that the transitions through a period of divided control of government did not matter much. Earlier in the period of unified party control, Carter had chosen as his Chairman G. William Miller, whom Greider describes as a “team player.” He was not a successful chairman, possibly be-

67

Ducking out of the Storm: U.S. Monetary Policy

cause his desire to be cooperative immobilized him in the face of conflicting goals and signals as inflation worsened. Policy might well have been even more inconsistent under a “team player” in a period of divided government. In fact, splits over policy within the administration could have as much effect as divided party control of Congress. If the principal with reappointment power lacks clear goals, the agent may have more latitude.21The extensive policy disagreement in the Reagan administration between supplysiders-who felt the Fed’s policy during the first experiment was too severe, frustrating the recovery-and monetarists had two related consequences. First, there was a lot of signaling from the administration, as Havrilesky’s (1988) index shows. In fact, his index, a count of published calls for changes in monetary policy by members of the administration, predicts a small part of changes in money growth (about 3 percent of the variance) and is itself a function of economic conditions, like the congressional attention index.22But signaling is what you do when you have no direct control over policy. Greider also recounts the efforts Regan made to get Volcker out in 1982-83 and replace him with someone easier to deal with (presumably like Miller). Regan failed, apparently in the face of strong support for Volcker in the financial community. This makes Ben Friedman’s suggestion (in this volume) that the financial community in the 1980s moved from supporting stability to preferring volatility in markets extremely interesting, since it would mean that by 1987 Volcker’s position was no longer that of any of the principals, and he was not reappointed. 2.5

New Targets, Old Problems

2.5.1 International Origins Those issues surround the domestic politics of monetary policy. The second experiment of the 1980s, reducing the dollar’s exchange value, would require a whole new paper, which would cover many of the same subjects. For example, differences between Carter’s multilateralism and Reagan’s more unilateral approach to foreign policy would figure in the story. The economic policy coordination literature had ups and downs, initially optimistic about improving policy, later foundering on empirical findings of small benefits that evaporated completely once even a little uncertainty or the possibility of error was allowed for. And the Congress would certainly appear, particularly through the linking of international money and trade policy, since the real effects of high real exchange rates are generally felt most directly through the pressure they exert on exporting and import-competing sectors. But was monetary policy, the actions of the Fed, different in this period? There was unprecedented international cooperation over at least the orchestration of policy, from the Plaza Hotel meeting and agreement in September 1985 through further sessions in Tokyo and Paris in 1986 and 1987. There were

68

JamesE. Alt

regular meetings of the “Group of Seven,” and other countries certainly altered their interest rates and fiscal policy to accommodate the dollar. There was a further flurry of activity after the stock market plunge of October 1987, but by 1988 the dollar was no longer so obviously overvalued, and American real interest rates, though still high and positive, were at least no longer higher than GermanykZ3 It might be tempting to conclude from all this that international concerns became a major element of American monetary policy in this period, but I am doubtful. There was some intervention in exchange markets. Fiscal policy gestures were made about reducing the deficit. More important, the dollar’s decline began in the spring (March or May, depending on choice of measurement) of 1985, months before the Plaza meeting. After the peak year for American real interest rates in 1984, there is a sharp reduction by 1985 which does reflect domestic policy: the target federal funds rate (that is, the center of the range that the FOMC now wrote into its directive) was lowered by two percentage points over the last two months of 1984, nine months before Plaza, and shortly before the dollar’s decline began. The target was lowered by a point and a half in the summer of 1985 (before the Plaza meeting) and raised to its previous level again by September (see fig. 2.1). In fact, the federal funds rate itself hardly follows this downward blip: it is allowed to lie at the top of its range for a couple of months. This blip, like the reduction in 1984, may have been aimed at sending the dollar down, but it is quite clear that after 1985, the harmonization of real interest rates owes at least as much to increases in the German and Japanese rates as to reductions in the American rate. In fact, except for the brief blip in summer 1985-before the “coordination” got under way-the federal funds rate target range is constant from December 1984 to March 1986. Nevertheless, speculatively, two political questions seem to be raised by this episode that deserve a systematic answer. First, exchange rates might have become a target of monetary policy at least for a while. Second, while real relative interest rate changes explain a lot about movements in the real dollar exchange rate, politics may have a role here too. 2.5.2 Moving targets

Interest rate stabilization was the goal of monetary policy, subject to periodic surges of partisan change and occasional other politically-motivated shocks. Woolley (1988a) estimates transitory shifts in policy weights on inflation and unemployment, with bursts of anti-inflationary policy in 1969 and 1979.24This (along with his failure to find a surge in the weight on antiinflation policy in 1974-75) verifies systematically the interpretation of documentary evidence by Romer and Romer (1989). Variations of both methods can be used to look at the question of exchange rates and interest rates, to see whether changing international conditions also produce changes in the goals pursued. Furlong (1989) lists the order in which

69

Ducking out of the Storm: U.S. Monetary Policy

major targets were discussed in the FOMC directive as evidence of the priority they received in FOMC discussions. (The interpretation was originally made by former Fed Governor Robert Heller, so it has surface plausibility.) Exchange rates appear in spring 1985 (the first date for which data are provided), listed after the monetary aggregate, strength of expansion, inflation, and credit market conditions. Exchange rates reach center stage briefly in the spring of 1987 and recede in importance after the events of October 1987. Unfortunately, I could not reliably replicate Furlong’s list from my reading of FOMC directives, so I cannot extend the data back before 1985. More systematically, table 2.5 contains reaction function estimates for the federal funds rate, extending Woolley ’s (1988a) moving regressions for overlapping four-year periods from 1977-80 through 1986-89. The dependent variable is the federal funds rate. Independent variables include a lagged dependent variable to stabilize the results, and the possible targets of policy, the inflation rate, unemployment rate, change in industrial production, change in the dollar-deutschmark exchange rate, change in money supply M1, and level of nonborrowed reserves .= These overlapping “moving” regressions imply that the federal funds rate responded only to inflation and money growth in the early years of the policy experiment, and then, after the peak of the recession, there is some systematic evidence of a response to unemployment (rates come down in the face of high unemployment), and rates are also lowered when nonborrowed reserves are highest. After the 1984 election, inflation reappears along with (not quite at conventional significance levels) the dollar-deutschmark exchange rate, at Targets in Moving Regressions in the 1980s

Table 2.5

Lagged Target of Policy and Expected Sign Years 1977-80 1978-81 1979-82 1980-83 1981-84 1982-85 1983-86 1984-87 1985-88 198689

Inflation Rate( + )

Unemployment Rate( - )

ADeutschmark Exchange Rate( + )

Nonborrowed Reserves( - )

AM 1 Growth( + )

* * * *

* * *

* * (*) (*)

Nore: The asterisk, *, indicates coefficient from reaction is statistically significant at .05 level with the expected sign. The asterisk in parentheses, (*), indicates significance at . l level with the expected sign. Moving regressions were estimated over 48 months for the time periods shown. The dependent variable was the federal funds rate; its first lag was included and was always significant.

70

James E. Alt

least intermittently. Thus it is conceivable that the federal funds rate was reacting to the exchange rate, but at most episodically and not very strongly. Any reaction to money growth disappears in the later years. None of this contradicts the Furlong-Heller data. The last two years were probably dominated by the aftermath of the stock market collapse, with the focus on credit market conditions explaining the absence of the usual targets from the reaction function.

2.5.3 Politics and the Demand for Dollars Historically, real exchange rates and relative interest rates are related, but what determines the preference for dollars at a given combination of exchange and interest rates? That is, monetary policy was certainly looser in 1978, tighter in 1982, but would 1982 monetary policy in 1978 have produced the same strength in the dollar? If not, what is the role of politics in the difference? Journalistic accounts of the strong dollar period abound with references to politics, and private conversations with investors suggest the importance to them of political stability. They seem to think about stability as the intersection of two separate components, stability of the world and stability of the United States, so: World U.S. Stable Unsettled

Stable

Unsettled

$ less strong $ very weak

$ very strong $ less weak

That is, when the world seems to be in trouble but the United States does not, the dollar is at its strongest; when conditions are reversed, the dollar is at its weakest. My guess is that a stable United States in a stable world produces a stronger dollar than weakness on both sides. But can this be measured and is there any evidence for it? The United States looks most stable to investors when the approval or popularity of the president is strongest, allowing for the cycles that exist in popularity. Reagan’s early term and short so-far incumbency of Bush have the strongest ratings. Reagan after 1986 is much weaker, and Carter after 1977 weaker still, though the lowest ratings of all are observed in the early 1970s. I have not yet systematically collected world stability data back into the 1970s, so the equation in table 2.6 has omitted variable bias.26The result is clearly interesting enough to merit continuing work. The real exchange rate partially adjusts to the current real interest differential between the dollar and the average of its trading partners, and also to the real level of the current account balance of payments. Independent of this, there is also a significant effect of the annual change in presidential approval. When this goes up, the real exchange rate goes up. I do not set too much store by this result, with all of 16 observations, but the general idea behind it-that is, to look for political

71

Ducking out of the Storm: U.S. Monetary Policy

Table 2.6

Determinants of the Dollar’s Surge and Decline

Independent Variable

Estimated Coefficient

Robust Standard Error

r-Statistic

29.3 .69 3.93

22.7 .23 1.65

1.3 3.0 2.4

10.9

5.06

2.2

.19

2.2

Constant Lagged real exchange rate Interest rate differential Real current balance of payments Change in presidential approval Observations Corrected R2 Durbin-Watson statistic

.47

16 .64 2.22

Note: Dependent variable was the U.S. real relative exchange rate. Data on exchange rates and interest rates is from International Financial Statistics, various issues. Balance of payments data

is from Citibase. Popularity data is from King and Ragsdale (1986).

factors that influence market judgments about the relative values of currencies-seems worth following up.

2.6 Conclusions The congressional choice of regulatory design produced an institution in which the demand for stable financial market conditions would normally be attended to, but in which both Congress and the president could have an influence on monetary policy. This agency framework, and particularly the interaction of the president’s reelection motive and the Fed chairman’s reappointment motive, provided incentives for pre-election monetary cycles to be weak and post-election cycles to be strong. Of course, big parts of the explanation of monetary policy lie in less systematically explored factors like the changing ideas of economists and changes in the structure of world financial markets. But even if each of the systematic political effects we discussed only explain a small part of the variation in outcomes, they cumulate: 5 per cent here, and 5 percent there, and pretty soon it starts to add up to a political model of monetary policy. It is important to note that it is not a question of “Which model is right?’ A focus on multiple principals instead points to several systematic incentives that exist simultaneously. Multiple principals means changing goals, and for the Fed it means partial independence and room for maneuver. Partial independence-that is, limited discretion for the Fed with the possibility of presidential control only at a cost to the president-is a solution to the “rules versus discretion” problem that economists have argued over for several years (Lohmann 1990). At least partial independence in the central bank is worth having, in terms of lower long-run inflation. Moreover, we have seen that Fed

72

James E. Alt

policy has been quite consistent vis-&+is its principals. In spite of the oftalleged incentives it has to exploit its private information to engineer surprises, in fact it seems to follow a simple rule, namely, to smooth interest rates. It does a good job of it since the relevant principal, the financial community, is rarely vocal. Whether the Fed does what it says is harder to say, since money growth makes inconsistent appearances in Fed’s reaction functions. Hence, it may have pursued a consistent policy, even if it has not consistently pursued the goals economists would have liked. Bills are introduced in every Congress to make the Fed more accountable, and economists often suggest institutional reforms. Some suggestions reviewed by Clark (1989) include prompt announcement of FOMC decisions, compulsory publication of Fed budgets, annual publication of targets with presidential removal of the FOMC for failure to achieve them, adding the Secretary of the Treasury to the board and FOMC, and Senate approval of regional bank presidents. Prompt announcement might induce speculation, undercutting credit market stabilization without aiding any other principal, but of each of the others the same thing can be said. Add a power, aid in discovery, and reform strengthens the relevant principal. Hence each reform makes it more likely that the affected principal will get what he wants, rather than what the reformer wants. Let the president dismiss the FOMC and they will not only work to hit targets but choose ones the president wants. This (like putting the Treasury Secretary on the FOMC) will reinforce the existing partisan cycle and promote the “team player” Fed behavior characteristic of Miller’s term as chairman. The empirical evidence on congressional behavior is inconclusive, so the effect of adding a congressional power is uncertain. Probably no amount of power and information would create a geographically redistributive “pork barrel” in credit allocation. If this cannot be done, which seems likely, then Congress will continue not to vote itself powers that have no electoral return, and the blame-shifting attention cycle will continue to characterize congressional involvement. However, the reappointment-reelection cycle discussed above has important implications for the most popular current proposal for institutional reform. Meltzer (1989) concludes his review by suggesting that the president should be able to appoint the chairman upon taking office, in order to ensure cooperation with the president’s program. This proposal featured prominently in the bill introduced in 1989 by Representative Lee Hamilton, chairman of the Joint Economic Committee. But we have seen that the agency relationship that exists produces strong incentives for the Fed to give the newly elected president what he wants, regardless of partisanship. Immediate appointment upon election would not necessarily add to this incentive.*’ Paradoxically, it would change the pre-election incentives. Under the existing system, the chairman, not knowing who will win, has incentives to duck and leave existing policy intact. Under the proposed reform, a chairman can only try to keep his job by throwing everything into the president’s reelection

73

Ducking out of the Storm: U.S. Monetary Policy

effort since, if the incumbent loses, the chairman is out of a j o b no matter what h e has done before the election. Hence incentives to run a pre-election monetary cycle, absent under the present system at least since monetary policy became broadly more visible, would reappear if the system were changed to allow the president immediately to appoint a new chairman. So the positive analysis of policy undertaken here turns out to have relevance for contemporary debates about the design of institutions, as well.

Notes 1. This paper ignores the activities of the Federal Reserve System in coordinating and regulating banking activities, restricting attention to the management of interest rates and money stock carried out by the Federal Open Market Committee, whose membership includes the appointed Board of Governors (including the chairman) of the Federal Reserve, as well as presidents of the regional banks. Good recent reviews of the institutional structure are available in Woolley (1984), Melton (1985), and Kettl (1986). 2. These positive questions complement normative discussions by economists on “What should the goals of monetary policy be?’ and “Should there be discretion or should the Fed follow rules?’ 3. Controversies abound over whether the Fed should have targeted total or nonborrowed reserves and have used lagged or current reserves accounting procedures (they changed to using current reserves in 1984). 4. According to congressional testimony of Treasury Undersecretary Sprinkel in July 1981, the annual growth rate of money MIB was supposed to decline by one percentage point per annum from seven in 1981 to three in 1985 (cited in Hibbs 1987, 287). The experiment was backed by the Administration: the 1982 Report of the Council of Economic Advisers (republished in Tobin and Weidenbaum 1988) gave a central role to reducing inflation, which it described as “essentially a monetary phenomenon” (54).

5. By early 1982, M1 was growing more slowly than the overall level of prices. This “real” contraction is reflected in the severity of the recession induced by the policy experiment. 6. “Implausibly optimistic” is Hibbs’s (1987) phrase; the “economics of joy” is what Stein (1988) calls it. The optimism conjoined the monetarist belief in the possibility of a painless adjustment to lower inflation rates through the change in inflationary expectations upon announcement of a simple and credible anti-inflation policy and the hoped-for “supply-side” effects on investment and output of lower marginal tax rates. 7. The Report projected real personal disposable income; table 2.1 gives the outturn in per capita terms. Population growth slows in the 1980s, making per capita growth performance look better compared to earlier decades, but the change in the projections into per capita terms would have only minimal effects. 8. Net investment would be below trend except for a surge in housing-related investment in the mid-1980s. Extrapolations are polynomial trends fitted to the investment series in unlogged constant-price form. 9. “Real” means the dollar exchange rate v i s - h i s the currencies of the U.S. trading partners, weighted by trade volume, net of their relative price levels. This netting out

74

James E. Alt

of relative inflation removes from the chart that part of changes in nominal exchange rates that offset international inflation differentials. 10. Woolley (1984, chap. 5) relates economists’ academic backgrounds to policy ideas. Gildea (1987) suggests that members of the Board of Governors have backgrounds that affect their propensity to dissent from FOMC decisions. 11. In February 1986, Volcker’s opposition to easing was overruled by a majority, including several other Reagan appointees. This episode emphasizes the division over economic policy that existed in that administration. 12. Belden (1989) shows that this finding continues to hold and also relates dissent to the different personalities of the chairmen and the extent of uncertainty about the economy. 13. See, e.g., the Republican threat to introduce a windfall tax on high interest rates in 1982. 14. Grier (1988) proposes that the changes of chairs in the Congress from Patman to Reuss and Sparkman to Proxmire (both of whom were at the opposite-liberalend of the ideological spectrum from their predecessors) in the mid-1970s and Reuss’s introduction of HR212 in spring 1475 were sufficient to persuade the Fed to lower interest rates. This would be an important observation, since the next big change would be expected when the Republicans gain the Senate in 1980, an event that coincides with the change in presidential control, but the implication that Patman was less in favor of low interest rates than Reuss is troubling. 15. The regression controls for odd and even numbered years, since more bills are introduced in the first year of congressional sessions, more than the difference in volume of business in the two years. 16. Havrilesky (1987) proposes a theoretical unification in which monetary surprises provide real stimuli to raise revenue to pay for other redistributive policies, but he gives little evidence in support. 17. This assumes that the chairman dominates FOMC voting, an assumption for which there is evidence, but oversimplifies in ignoring the role of other advisors in the chairman’s appointment. 18. Output rises and falls in the second year after the election (Alesina and Sachs 1988), so systematic money growth changes in that year (for which there is no evidence) would reflect accommodation. 19. Volcker held off easing to end the recession longer than someone trying to curry favor for reappointment would have, but he enjoyed tremendous support in the financial community. Again, it seems that first actions speak loudest. 20. Presidential approval is increased by a half point for each percentage point of belief that his party is best able to handle the most important problem; each point of inflation takes two points off approval. 21. In fact, the chairman should worry more about the president than Congress, since he is appointed by the president without Senate consent, while amending the Federal Reserve Act without the president’s approval would require a supermajority in Congress. 22. Administration calls for policy could induce changes in private demand for money which the Fed may accomodate by smoothing interest rates. Havrilesky does not test for this. 23. Recently, the G-7 agreement not to coordinate policy to alter the value of the yen was taken as further evidence that while the dollar was everyone’s problem, the yen was Japan’s. 24. He uses Brown, Durbin, and Evans’s (1975) method of moving regressions, in which the same regression is run over successive overlapping time periods in order to investigate shifts in the coefficients.

75

Ducking out of the Storm: U.S. Monetary Policy

25. The targets for the federal funds rate in FOMC directives could appear on the right-hand side here, too, but those targets accommodate changes in the rate, whose behavior I analyze directly. Each new regression drops the first year of the previous one and adds an extra year at the end. Monthly observations are used. 26. To generate a quantitative indicator of sentiment about world stability, one could use a measure of conflict or subjectively turn to the annual responses to the Gallup International survey (“Will the next year be one of peace or of conflict?”) carried out in countries that are financial centers. Responses to this survey trend upward in the later 1980s, so dollar demand, ceteris paribus, should be considerably weaker than in the earlier part of the decade, when the world was widely seen as more unstable. 27. Clark (1989) quotes Jerry Jordan as recommending the appointment take place a year after the president takes office. This would seem to be the most dangerous idea of them all, since a partisan chairman would now have clear incentives to run both the pre-election and post-election cycles in policy.

References Alesina, A., and J. Sachs. 1988. Political Parties and the Business Cycle in the United States, 1948-1984. Journal of Money, Credit, and Banking 20:63-82. Alt, J. 1985. Political Parties, World Demand, and Unemployment. American Political Science Review 79: 1016-40. Banaian, K., L. Laney, and T. Willett. 1986. Central Bank Independence: An International Comparison. In Toma and Toma. Beck, N. 1982. Presidential Influence on the Federal Reserve in the 1970s. American Journal of Political Science 26:415-45. . 1987. Elections and the Fed: Is There a Political Monetary Cycle? American Journal of Political Science 31: 194-216. . 1988. Politics and Monetary Policy. In Willett. Belden, S. 1989. Policy Preferences of FOMC Members as Revealed by Dissenting Votes. Journal of Money, Credit, and Banking 21:432-41. Brown, R., J. Durbin, and J. Evans. 1975. Techniques for Testing the Constancy of Regression Relationships over Time. Journal of the Royal Statistical Socieo. Ser. B, 37:149-63. Chant, J., and K. Acheson. 1986. The Choice of Monetary Instruments and the Theory of Bureaucracy. In Toma and Toma. Chappell, H., and W. Keech. 1988. Choice and Circumstance: The Consequences of Partisan Macroeconomic Policies. Presented to the annual meeting of the APSA, Washington, D. C. Clark, L. 1989. Remaking the Fed: Maybe It’s Time. Wall Street Journal, 18 September, p. 1. Cukierman, A. and A. Meltzer. 1986. A Theory of Ambiguity, Credibility and Inflation under Discretion and Asymmetric Information. Econometrica 54: 1099-1 128. Davidson, J., D. Hendry, F. Srba, and S. Yeo. 1978. Econometric Modelling of the Aggregate Time-Series Relationship between Consumers’ Expenditure and Income in the United Kingdom. Economic Journal 88:661-92. Friedman, B. 1985. Conducting U.S. Monetary Policy: What Changes Do We Need? In Monetarism, ZnJation, and the Federal Reserve. Joint Economic Committee. 99th Congress, 1st sess.

76

JamesE. Alt

Furlong, F. 1989. International Dimensions of U.S. Economic Policy in the 1980s. Economic Review no. 2 (Spring): 3-16. Federal Reserve Bank of San Francisco. Gildea, J. 1987. A Theory of F.O.M.C. Policymaking Behavior. Wheaton College, Norton, Mass. Typescript. Goodfriend, M. 1988. Bureau Analysis and Central Banking. Journal of Monetary Economics 22517-22. Greider, W. 1987. Secrets of the Temple. New York: Simon & Schuster. Grier, K. 1987. Presidential Elections and the Federal Reserve. Southern Economic Journal 54:475-86. . 1988. An Agency Model of Congressional-Fed Interaction. California Institute of Technology and George Mason University. Typescript, March. Havrilesky, T. 1987. A Partisanship Theory of Fiscal and Monetary Regimes. Journal of Money, Credit, and Banking 19:308-25. . 1988. Monetary Policy Signaling from the Administration to the Federal Reserve. Journal of Money, Credit, and Banking 2093-101. Haynes, S . , and J. Stone. 1988. Does the Political Business Cycle Dominate U.S. Unemployment and Inflation? In Willett. Hibbs, D. 1977. Political Parties and Macroeconomic Policy. American Polirical Science Review 7 1 :1467-87. -. 1987. The American Political Economy. Cambridge, Mass.: Harvard University Press. . 1990. The Partisan Model of Macroeconomic Cycles: More Theory and Evidence for the United States. Presented to the NBER Conference on Political Economics, Cambridge, Mass. Kettl, D. 1986. Leadership at the Fed. New Haven, Conn.: Yale University Press. King, G., and L. Ragsdale. 1988. The Elusive Executive. Washington, D.C.: Congressional Quarterly Press. Krehbiel, K. 1991. Information and Legislative Organization. Ann Arbor: University of Michigan Press. Lohmann, S. 1990. Optimal Commitment in Monetary Policy: Credibility versus Flexibility. Carnegie-Mellon University, Pittsburgh. Typescript. McCubbins, M., and T. Schwartz. 1984. Congressional Oversight Overlooked: Police Patrols versus Fire Alarms. American Journal of Political Science 28: 165-79. Mayer, T. 1987. U.S. Monetary Policy, 1973-1987. Research Program in Applied Macroeconomics and Macro Policy Working Paper no. 47. University of California, Davis. Melton, W. 1985. Inside rhe Fed. Homewood, Ill.: Dow Jones-Irwin. Meltzer, A. 1989. The Federal Reserve at Seventy-five. Paper presented to the conference “The Fed at 75 ,” Federal Reserve Bank of St. Louis, October. Myerson, R. 1982. Optimal Coordination Mechanisms in Generalized Principal-Agent Problems. Journal of Mathematical Economics 10:67-8 1 . Nordhaus, W. 1975. The Political Cycle. Review of Economic Studies 42:169-90. Parkin, M., and R. Bade. 1985. Central Bank Laws and Monetary Policy. University of Western Ontario. Qpescript. Peterson, P., and M. Rom. 1988. Macroeconomic Policymaking: Who Is In Control? In New Directions in American Politics, edited by J. Chubb and P. Peterson. Washington, D.C.: Brookings. Poole, W. 1976. Benefits and Costs of Stable Economic Growth. Carnegie-Rochester Conference Series on Public Policy 3: 15-50. Rogoff, K . 1985. The Optimal Degee of Commitment to an Intermediate Monetary Target. Quarterly Journal of Economics 100:1169-89. . 1987. Reputational Constraints of Monetary Policy. Carnegie-Rochester Conference Series on Public Policy 26: 141-82.

77

Ducking out of the Storm: U.S. Monetary Policy

Romer, C., and D. Romer. 1989. Does Monetary Policy Matter? NBER Working Paper no. 2966. Cambridge, Mass. Shepsle, K. 1978. The Giant Jigsaw Puzzle. Chicago: University of Chicago Press. Shughart, W., and R. Tollison. 1986. Preliminary Evidence on the Use of Inputs by the Federal Reserve System. In Toma and Toma. Stein, H. 1988. Presidential Economics, 2nd rev. ed. Washington, D.C.: American Enterprise Institute. Tobin, J., and M. Weidenbaum, eds. 1988. Two Revolutions in Economic Policy. Cambridge, Mass.: MIT Press. Toma, M. 1986. Inflationary Bias of the Federal Reserve System. In Toma and Toma. Toma, E., and M. Toma, eds. 1986. Central Bankers, Bureaucratic Incentives, and Monetary Policy. Dordrecht: Martinus Nijhoff. Willett, T., ed. 1988. Political Business Cycles. Durham, N.C.: Duke University Press. Woolley, J. 1984. Monetary Politics. New York Cambridge University Press. . 1988a. Partisan Manipulation of the Economy: Another Look at Monetary Policy with Moving Regression. Journal of Politics 50:335-60. . 1988b. When Regulators Disagree: Financial Regulatory Conflict in Light of the Congressional Dominance Hypothesis. Presented to the annual meeting of the APSA, Washington, D.C.

COlllmeIlt

Benjamin M. Friedman

The centerpiece of James Alt’s thoughtful analysis of U.S. monetary policy is his representation of the Federal Reserve System as an agent simultaneously serving three principals: the president, Congress, and the private financial markets. As is standard in principal-agent models, the objectives governing the behavior of the agent here are not merely to satisfy the objectives of the principal. Alt posits that Federal Reserve policymakers-including especially the chairman-also act so as to maximize prospects for reappointment when their terms expire. In addition to this specification of the familiar tension between the objectives of the principal and those of the agent, however, Alt’s multiprincipal model opens up another dimension of conflict. No one expects the president and Congress to agree at all times on the proper course of monetary policy, and the introduction of “the market” as yet a third principal only complicates matters further. The resulting framework for studying monetary policy is so rich with possibilities that it is surely no criticism of Alt to say that many of its potential implications remain to be explored. In order to reach a model with so much potential richness, Alt must rely on many simplifying assumptions, some of which warrant closer inspection than he provides in his paper. Perhaps the most obvious of these is that entities like Congress or the financial markets-or even the president-have clearly deBenjamin M. Friedman is the William Joseph Maier Professor of Political Economy at Harvard University and a research associate of the National Bureau of Economic Research.

78

James E. Alt

fined objectives that monetary policymakers can either seek to achieve or not. As Alt is well aware, modeling Congress in this way sidesteps a long-standing and voluminous body of thinking, ranging from the popular to the professional, which analyzes the complexities of an institution made up of representatives of different political parties, different geographical regions, and different economic constituencies. Similarly, although in principal “the president” is a single individual, in practice the administration that the president heads also typically reflects at least some diversity of constituencies. AS the example of the open conflict between “monetarists” and “supply-siders” in the Reagan administrationclearly demonstrated, the resulting differences of opinion within the administration can and do bear quite directly on objectives for monetary policy. Imputing any very unified set of monetary policy objectives to participants in the private financial markets is likewise problematic. The ability to attribute unambiguous objectives to each of the three principals in Alt’s model is not a mere nicety, but a central assumption underlying the model’s capacity to generate testable implications. For example, one half of Alt’s conclusion about how political forces influence monetary policy is that, in the period following a presidential election, the Federal Reserve tends to deliver the monetary policy that the new president wants. (The other half is that in the period leading up to a presidential election, the Federal Reserve tends to steer a neutral course.) In the 1980 election campaign, candidate Ronald Reagan harshly criticized the stagnation of U.S. economic activity and placed major emphasis on the need to create new jobs. But Mr. Reagan was just as critical of the current double-digit price inflation. Nor did the intramural monetarists-versus-supply-sidersdebate clarify the issue once the new administration had taken office. Unless the Federal Reserve was operating under an economic model promising disinflation without recession-something that neither the central bank’s internal discussions at the time nor subsequent experience indicates is likely-how was Paul Volcker to know whether the tight monetary policy that the Federal Reserve imposed in 1981 and the first half of 1982 was what the new president wanted? From the perspective of scholarly inquiry, how is the economist or political scientist to treat this observation in submitting Alt’s model to empirical verification? A second central assumption of Alt’s model is that the private objective governing the behavior of Federal Reserve policymakers (again, especially the chairman) in their capacity as agent is the desire for reappointment. The role of this objective is to motivate the central bankers not only to give a newly elected or reelected president the monetary policy he wants, but also to steer a neutral course in advance of the election so as not to risk alienating whichever candidate wins. This assumption too bears closer inspection. To begin, the institutional facts describing the job tenure of Federal Reserve Board members and Federal Reserve Bank presidents, both of whom serve on the Federal Open Market Committee, render this hypothesized objective implausible except in the case

79

Ducking out of the Storm: U.S. Monetary Policy

of the chairman. The bank presidents are formally appointed by the boards of directors of their respective Federal Reserve Banks. Although input from Washington is often important in the initial appointment of a new bank president, it is typically not a factor in reappointment of an incumbent bank president to successive terms (which last five years). Each Board member is appointed either for a full term of 14 years or for the unexpired portion of a currently running 14-year term. A member who has served for a full term is ineligible for reappointment. A member initially appointed to serve out an unexpired term is eligible for reappointment to one full term, but the very fact that so many Board appointments are to fill out unexpired terms indicates that many Board members not only do not seek reappointment but, in contrast, choose to leave well before they have to. The average tenure of Federal Reserve Board members appointed during the last three decades (not including those still in office as of the time of writing) has been less than six years. More important, it is not clear what implication of this key element in Alt’s principal-agent model provides a handle for empirically testing it. Even if one were able to document beyond doubt that the Federal Reserve adjusts monetary policy after each election to conform to the new president’s wishes, the counterexample of the U. S. Supreme Court-which reportedly “follows the election returns” despite the justices’ life appointments-shows that such a finding would not be strong evidence of a reappointment objective. What alternative objective might motivate the behavior of the Federal Reserve in this principal-agent setting? One alternative that would be familiar to observers of many governmental agencies is that Federal Reserve officials attach great importance to preserving or even enhancing the position of the Federal Reserve as an institution. In the specific case of the Federal Reserve System in recent decades, this institutional objective has usually taken the form of seeking to protect the central bank’s independence: independence from direct intervention by the administration in monetary policy decisions (e.g., by putting the Secretary of the Treasury on the FOMC, as was the case until 1936), independence from “excessive” congressional interference, independence from GAO budget audits, and so on. (To be sure, it is possible to attribute such institutional objectives to private motives too, but presumably those would not include the desire for reappointment.) What kind of principal-agent model might this alternativeobjective for Federal Reserve policymakers generate? I have suggested elsewhere that the effective scope for independent monetary policymaking by the central bank in the United States is bounded on one side by the monetary policy sought by the administration and on the other side by that sought by Congress (again, as if these two entities had unambiguous objectives). When the president and the dominant forces in Congress differ sharply about monetary policy, the Federal Reserve has much room for genuine choice, and U.S. monetary policy becomes independent in reality as well as in theory. By contrast, when the president and the relevant consensus within Congress are in agreement about the

80

JamesE. Alt

proper course for monetary policy, the Federal Reserve has little choice but to deliver the policy that these two principals deem appropriate. As is consistent with Alt’s discussion of “divided government and agency independence,” therefore, this model implies that the available range within which the Federal Reserve makes monetary policy is at times fairly wide, but at other times quite narrow. In this case the effective threat constraining the Federal Reserve to keep monetary policy within this range is that the president and Congressacting togerher-have the power to amend the Federal Reserve Act, and therefore alter the Federal Reserve’s internal structure as well as its relationship to other parts of the government, in whatever way they choose. This discussion is not the place to carry out a comparative evaluation of this alternative model and Alt’s; the point of the example is merely to show that other, very different models are equally plausible a priori. A further important consideration that does not appear explicitly in Alt’s paper, but that is entirely compatible with his model (or with my suggested alternative), is that the desire to avoid financial crises has traditionally been an overriding concern of monetary policymakers, one at least on a par with the macroeconomic objectives that Alt discusses. After all, the Federal Reserve itself grew out of the reaction to the series of banking panics that had so disrupted not only the financial markets but U.S. economic activity more generally in the late nineteenth and early twentieth centuries, and the original Federal Reserve Act was explicit that the new institution’s chief charge was to avoid such episodes. Further, as Hyman Minsky has persuasively argued, the subsequent decentralization of lender-of-last-resort responsibilities in the 1930s (importantly including creation of the Federal Deposit Insurance Corporation) stemmed in large part from the recognition that the Federal Reserve had failed to protect the nation from just such a crisis following the collapse of Credit-Anstalt in May 193 1 . Indeed, the desire of Federal Reserve officials to avoid a financial crisis fits easily into almost any principal-agent representation of the monetary policymaking process. In the context of Alt’s model based on a reappointment objective, a Federal Reserve chairman who had allowed a financial crisis to occur “on his watch” might well anticipate sharply reduced prospects for reappointment. In the context of my suggested alternative based on an institutional-independenceobjective, a financial crisis might easily precipitate another devolution of Federal Reserve responsibilities or even a wholesale restructuring of the institution. Moreover, there is no reason to suppose that avoiding a financial crisis is merely an objective of Federal Reserve policymakers in their role as agents acting for other principals. No doubt the president and Congress care about such matters as well. So do private participants in the financial markets. The role of the financial markets as yet a third principal in this interaction is one of the potentially most interesting aspects of Alt’s model. Unfortunately, however, it is far from obvious what “the market” would like the Fed-

81

Ducking out of the Storm: U.S.Monetary Policy

eral Reserve to do. Part of the problem is the great heterogeneity that has traditionally characterized the US. financial system. Commercial banks, investment banks, brokerage firms, securities dealers, thrift institutions, insurance companies, pension funds, and money management firms need not-and do not-share identical objectives. The challenge is to draw generalizations that are useful in the context of a principal-agent model of monetary policy. Alt assumes that “the market” wants the central bank to provide a financial environment characterized by stable interest rates. But apart from sharing in the common desire to avoid a financial crisis, do market participants really seek interest rate stability? Especially in light of changes in the structure of the U.S. financial markets over the past two decades, interest rate stability is less likely to be a desideratum of professional market participants than may once have been the case. Earlier in the postwar period, the prevailing market structure, in which the yield curve was generally upward sloping and in which highly creditworthy corporate borrowers often relied on commercial banks for a major part of their credit needs, provided both banks and securities firms with ready profit opportunities with relatively modest risk. Securities firms made profits, on average over time, by borrowing at short term and positioning longer-maturity securities. Banks made profits by issuing deposits (or deposit-like instruments) and lending to commercial, industrial, or other high-quality borrowers. Interest rate volatility merely added unwanted risk. More recently, however, the yield curve is less reliably upward sloping, and highly creditworthy corporations turn to the commercial paper market far more than to the banking system when they need to borrow. One result of these changes in the market environment is that both banks and securities firms have had to seek out other-typically more risky-profit opportunities: for example, banks’ lending to developing countries and financing leveraged buy outs and securities firms’ issuing junk bonds and making bridge loans. Another result has been the increasing predominance of a trading mentality, in which interest-rate volatility represents not so much an unwelcome risk as a badly needed profit opportunity. Alt’s idea of representing the Federal Reserve as also responsive to “the markets” in a principal-agent setting is intriguing, but further thought is necessary to establish what objective to attribute to this third principal. A final reservation about Alt’s analysis in this paper arises not fiom the specification of his model but from his approach to testing it empirically, As in much work in this area, Alt represents actual monetary policy outcomes by the growth rate of the money stock (here M2). But there has always been doubt about whether observed money growth in fact reflects deliberate Federal Reserve policy choices, and developments in the 1980s have increased grounds for doubt along these lines. Indeed, since 1980 the relationship between money growth and the growth of either income or prices in the United States has collapsed to such an extent

82

James E. Alt

that there is little reason why the Federal Reserve should have relied on monetary targets for monetary policy. As Kenneth Kuttner and I have shown, data since 1970 provide no evidence that fluctuations in any of the familiar monetary aggregates contain information that is useful for predicting subsequent fluctuations in either nominal or real income. Further, because the mid- 1980s brought both the fastest money growth of the postwar period and the greatest disinflation, the correlation between money growth and price inflation calculated in the way recommended by Milton Friedman (using two-year averages to smooth out short-run irregularities, and a two-year lag between the money growth and the inflation) is now negative for postwar data samples including this decade. In light of these developments, it is hardly surprising that many observers no longer think that the Federal Reserve places much emphasis on moneygrowth targets in planning and carrying out U.S.monetary policy. If money growth no longer occupies a central role in the monetary policy process, then tests that rely on the growth of some monetary aggregate to measure the stance of monetary policy are unlikely to provide a valid test of an underlying model of Federal Reserve behavior.

3

Party Governance and U. S. Budget Deficits: Divided Government and Fiscal Stalemate Mathew D. McCubbins

The president and members of Congress have been grappling with runaway budget deficits for over a decade. The 12-digit budget deficits of the 1980s have been blamed for everything from a decline in private investment and personal savings to trade imbalances with Japan. Collectively, the nation's policymakers seem akin to credit-card junkies, hooked on living beyond their means, borrowing to pay the interest on their debt. Federal budget deficits are of course not new. Every president in the postwar period has run on a platform that included a promise to bring federal spending under control. Indeed, in only 10 of the last 60 years have federal revenues exceeded expenditures. Recently, however, real debt more than doubled from 1981 to 1988. As shown in figure 3.1, the swift growth in the debt as a percentage of the gross national product (GNP) in the 1980s reversed the general trend of the postwar period in which the ratio of debt to GNP had been declining.' The debt as a percentage of GNP in 1981 stood at slightly more than 32 percent. This jumped to 45 percent by 1985 and to over 50 percent in 1987. Whole forests have been felled by analysts sawing through the causes and consequences of the runaway deficits of the 1980s. 'Ikro explanations for the Occurrence of these deficits have received widespread currency. First, though much pulp was pressed over the four-bit deficits of the late 1970%the plunge into the deficit abyss that followed Ronald Reagan's entry into the White House was a clear break from previous policies. While constant-dollar-valued revenues for fiscal years 1982-84 were less Mathew D. McCubbins is professor of political science at the Department of Political Science, University of California, San Diego. The author thanks Neal Beck, Gary Cox, Gary Jacobson, Rod Kiewiet, Sam Kernell, Sam Popkin, Mike Rothschild, and Barry Weingast for their comments and criticism. Also acknowledged is the support of the National Science Foundation, grant SES-8421161,

83

84

Mathew D. McCubbins 1.5 L

1.0

-

0.5

-

4a

68

aa

flscALYEAR

Fig. 3.1 National debt as a percentage of GNP, 1928-88

than the revenues collected under Jimmy Carter’s last budget in 1981, constant-dollar levels of spending increased more than 8 percent. Indeed, over Reagan’s first seven budgets, constant-dollar federal revenues increased by less than 21 percent, while constant-dollar spending increased by 25 percent. The discrepancy between spending and revenue in the 1980s led to constantdollar federal deficits ranging from $50 billion to over $90 billion. This explanation focuses on the central role of the president in budget policy. Presidents, it is argued, through their powers of persuasion, control of budgeting information, and use of the Office of Management and Budget (OMB) manipulate legislative outcomes, controlling the direction, if not always the magnitude, of policy change (Neustadt 1954, 1980; Wayne 1978). Other analysts have seen the deficits of the 1980s as a consequence of a trend begun in the late 1970s. To them the deficits of the last decade arose from a change in the way Congress sets the budget. The Congressional Budget and Impoundment Control Act of 1974 (hereafter, Budget Act) prescribed new controls on presidential impoundments and it is important to note, established a new budget process in Congress. But the new process, instead of

85

U.S. Budget Deficits

balancing the budget, unleashed the spending profligacies of Congress from the restraining hand of the appropriations process. Spending decisions, these scholars said, had been disciplined by the House Appropriations Committee, whose members saw as their role guarding the Treasury from the expensive tastes of their colleagues (Fenno 1966). The 1974 Budget Act, by reducing the role of the Appropriations Committee in determining spending, ruined the incentives for members of that committee to police the spending of their colleagues: members of the House committee, it is argued, no longer seek to cut executive spending requests, but instead seek to secure their own slice of the federal largess (Schick 1980). These explanations represent two fairly common perceptions of American politics and the sources of American economic policy. The first is premised on the widely held belief that the president, in the twentieth century, has come to dominate policy-making. The second is based on a similar belief that congressional committees, in this century, have come to dominate policymaking. The purpose of this essay, by way of examining the deficit crisis, is to examine these two views. I do so by questioning the basic tenets and inferences of each explanation. In so doing I also offer an alternative way to think about American politics, one that examines the relationships between the president and Congress and between Congress and its committees, within the institutional context of decision making. My central thesis is that congressional parties, to a greater extent than commonly thought, govern and determine public policy. I argue that divided partisan control of Congress in the 1980s, and divided control of government, led to rapidly increasing budget deficits, as the Democrat-controlled House and Republican-controlled Senate and presidency were unable to overcome the fiscal stalemate originally created in 1981. 3.1 Of Checks and Balances: The Thesis of Presidential Ascendency in

American Politics

For much of the twentieth century, Congress has been seen to be in decline relative to the president. The greater efficiency of the executive branch and the inability of members of Congress to overcome their diversity of interests led members of Congress to “abdicate” their responsibilities in many key policy areas (Sundquist 1981,28, 35-36). Without presuming to undertake a full development of this thesis or my critique here, in the next few pages I reexamine the thesis of presidential dominance. My goal is to reexamine critically the thesis that the runaway deficits of the 1980s are the result of Reagan’s policies. 3.1.1 Abdication or Delegation The framers of the Constitution did not imagine that the president could attain a dominant position in federal policy-making. But it has become part of

86

Mathew D. McCubbins

the lore of American politics that the president has come to dominate national politics (Binkley 1962; Bryce 1924; Burns 1965; Laski 1940; Milton 1965; Neustadt 1980; Schlesinger 1973; Sundquist 198 1). There is much evidence to support this thesis. The executive branch submits roughly 200 proposals to Congress every year, including budget requests, and legislative proposals relating to fiscal management, executive reorganization, and general policy. Congress frequently accepts these proposals without amendment (Pfiffner 1979; Peterson and Rom 1989) and, moreover, Congress almost never takes action on an issue until it has received a proposal from the president (Edwards 1980). Presidents also nominate several thousand people for federal posts. Rarely are these nominations rejected, and most receive only perfunctory review. So how did this alleged transformation of national politics come about? Most of this increase in authority in the twentieth century has been the result of congressional delegation (Kiewiet and McCubbins 1991, chap. 7,9). Congress, for example, in the 1921 Budget and Accounting Act redelegated the authority to transmit budget estimates from the various departments to the Bureau of the Budget and the President.* The thesis of executive domination argues that in delegating, Congress actually has abdicated its authority to make decisions. The distinction here between delegation and abdication is more than ~emantic.~ Many have argued, given the facts summarized above, that Congress no longer affects decision making on issues, such as the budget, that it has delegated to the executive branch. But what can we actually infer from these stylized facts? They are entirely consistent with an interpretation that Congress, in delegating, actually has retained all of its authority over policy-making. The argument is as follows: Congress delegates to the executive branch its authority to make public policy. Members of Congress then use direct and indirect means (such as appointments, oversight, appropriations hearings, amendments to appropriations bills, etc.) to discipline those charged with carrying out this delegated authority (Kirst 1969; McCubbins and Schwartz 1984; McCubbins, Noll, and Weingast 1987; McCubbins, Noll, and Weingast 1989; Weingast and Moran 1983). Members therefore shape executive decisions as they are being made, thus largely relieving themselves of the need for the type of post hoc intervention that is the object of so many studies. So how can we tell if Congress, in delegating, has retained control over policy-making? Evidence can be found in the ways Congress structures its delegations. Wherever Congress delegates to the executive branch or the president, we see the same attention to the details of structure, for the same reasons. The Budget and Accounting Act of 1921, for example, contains many provisions to control the revision, compilation, and transmittal of agency budget requests by the president (Kiewiet and McCubbins 1991, chap. 7).4These provisions have been amended and expanded many times in the intervening years (including the 1950 Budget and Accounting Procedures Act, the Gov-

87

U.S. Budget Deficits

emment Accounting and Procedures Act of 1956, the 1958 Budget and Accounting Act, the 1970 Legislative Reorganization Act, among others; see Kiewiet and McCubbins 1991, chap. 7, for a complete discussion). Those who have argued that Congress has abdicated its responsibilities over policy-making have not ignored these facts, but they have missed their significance. The question of whether delegation necessitates abdication can be asked in a somewhat different form. Others have admitted that Congress has the ability to exercise control; they ask instead whether its members lack the will to exercise control (Fiorina 1981). At some times, of course, members of Congress do seem more willing to exercise control than at others. The efforts by Congress to restructure presidential discretion in the 1970s are prominent examples. Members of Congress take on the mantle of responsibility most often, however, when the president is from a different party than the one that controls Congress. Delegation is greatest, it seems, only when the president and the majorities in Congress share the same party label. The abdication hypothesis, in its more subtle formulations, contends that the delegation of policy initiation, as in the budget, has tilted the balance between the branches toward the pre~ident.~ And, it was the establishment of legislative clearance at the Budget Bureau that has been heralded as the principal source of presidential dominance in this century (Sundquist 1981). What must be remembered, however, is that “Congress created OMB. Congress can uncreate it-or change it” (quote found in Brand 1985, 1815; see also Kiewiet and McCubbins 1991; Wilmerding 1943). Even with an explicit delegation of authority to the president, as in the Budget and Accounting Act of 1921, Congress need not even consider the president’s proposal. Refemng to his troubled budgets at a press conference on October 22, 1987, Reagan observed that “every year under the law I submitted a budget program early in the year, and as they’ve done every year I’ve been here, they’ve put it on the shelf and have refused to even consider it” (Congressional Quarterly WeeklyReport [24 October 19871, 2626). “When the delegation is not a power to act but only responsibility to recommend,” Sundquist observed (1981, 12), “the executive budget, for instance-the Congress explicitly retains not only its full authority but also its responsibility to act.” Presidents do, nonetheless, gain some influence over outcomes through the recommendations they make. Because their electoral fates are partially linked, the president and his party in Congress will often find it in their best interest to cooperate. How much influence the president holds in this process is open to debate, particularly when his party is in the minority in both houses of Congress. 3.1.2 Presidential Power and the Veto In contrast to the powers delegated the president by Congress, the veto conveys to the president a property right in the legislative process. The president’s

88

Mathew D. McCubbins

proposal may matter on some occasions and not on others, but the veto applies to all acts of Congress. Some people, in analyzing presidential vetoes, have inferred that it is ineffective because it is rarely used. Again, as in the thesis of presidential dominance, this is a mistaken inference. Members of Congress logically anticipate the president’s reaction to their proposals. Their proposals, then, usually are designed to avoid a veto. The influence that the veto gives the president is, however, asymmetrical. The president can use the veto to restrain Congress, to some extent, when he prefers to spend less than its members prefer. But it provides the president with no means of extracting greater appropriations (Kiewiet and McCubbins 1988). This asymmetry derives from inherent limitations in the veto power. The veto provides the president with only the power to reject; it does not provide him with the power to amend. On spending bills, the president’s position is even more precarious: on receiving a bill from Congress, the president can either accept the appropriations contained therein or veto it and let Congress write a continuing resolution.6 Because of the emergency nature of continuing resolutions, they are virtually veto proof. Also, because continuing resolutions almost always contain less spending than is contained in the corresponding appropriations bill, the president is able to reduce spending (to the level contained in the continuing resolution) through the use or threat of the veto, but he cannot get increased spending from a Congress that does not favor it. Again, the details of the situation did not escape Ronald Reagan’s understanding: The President of the United States cannot spend a nickel. Only Congress can authorize the spending of money. And for six years now I have repeatedly asked the Congress for less money and they have turned around and given more-given more to spend, and done it in such a way that I can’t veto it when they put it all together, instead of appropriations, in a continuing resolution-we haven’t had a deficit-or, a budget since I’ve been here. No-the Congress is the one that’s in command. . . . And every budget I’ve sent up there has been put on the shelf and I’ve been told that it’s dead on arrival. (Congressional Quarterly Weekly Report [24 November 19871,2628) The limited and asymmetric influence conveyed to the president by the veto is illustrated by the budget debates in Reagan’s second term. For fiscal year 1985, Reagan proposed a 13 percent real (inflation-adjusted) increase in defense spending for the following year. He coupled this with a proposal to slash Social Security and domestic spending and to eliminate cost of living adjustments (COLAS)for federal pension payments (including Social Security). The political reality on Capitol Hill, however, did not favor a package of defense increases and domestic spending cuts. Reagan buckled to the pressure and reduced his defense request to a 5.9 percent inflation-adjusted increase.

89

U.S. Budget Deficits

On the other hand, he requested steep “cuts” in domestic spending totalling $34 billion in current dollars. The package of domestic spending cuts and defense increases was still unacceptable to Democrats and liberal Republicans. In March 1985, the Senate Budget Committee voted to recommend an inflation-adjusted freeze on defense spending and a freeze on Social Security. Faced with a projected deficit exceeding $200 billion (current dollars) for 1986 and trying to unite a fractured party, the Senate Republican leadership negotiated with Reagan a 3 percent real increase in defense spending. The package also included deep “cuts” in domestic spending. Reagan declared that a 3 percent real increase was the “rock-bottom level” he would accept (Congressional Quarterly Weekly Report [27 April 19851, 771). But, Reagan was holding the wrong end of the veto stick. Despite Reagan’s veto threat, the Senate rejected the Republican leadership’s package. Ultimately, the Senate approved a budget resolution that only reduced, but did not eliminate, COLASand that provided no real increase in defense spending. On the other side of the Capitol, the House, voting along party lines, recommended more spending for domestic programs and less for defense than the budget passed by the Senate. Reagan pronounced the House budget resolution as “unacceptable” (Congressional Quarterly Weekly Report [25 May 19851, 971). In conference the House and Senate compromised on defense, but accepted much of the domestic spending increases advocated by the House. As expected, Reagan accepted the bills passed under this resolution. In the final analysis, presidents do make numerous proposals to Congress. Congress generally takes no independent action on an issue prior to the president’s request. This does not, however, imply congressional abdication or presidential ascendency. If presidential proposals are to succeed, the president must anticipate the reaction of members of the House and Senate to his proposals and accommodate their demands and interests. Presidents know this, so they rarely submit proposals that are likely to fail. Those who ignore this lesson have their proposals ignored. 3.1.3 The Reagan Revolution and the Determinants of Federal

Spending Policy The budget for fiscal year 1982 was heralded (or decried) as a victory by a powerful president over an institutionally weakened Congress. But what sort of victory was this for Reagan? The much-ballyhooed $36.6 billion budget “cuts” were measured not against the spending totals in the fiscal 1981 budget, but against Carter’s proposed budget for fiscal year 1982 (in which he requested a whopping 17 percent nominal increase in spending relative to spending enacted for fiscal year 1981). A vast majority of the programs and agencies that suffered cuts in Reagan’s budget had previously suffered cuts by the Democrats and Jimmy Carter, and almost half had their budgets cut by Democratic Congresses when Gerald

90

Mathew D. McCubbins

Ford was president. Real spending in the commerce, energy, housing and urban development, interior, and labor departments, and in the Postal Service, had declined through most of Carter’s administration. Spending on public works, other than sewage treatment plants and other environmental programs, had been declining since the late 1960s. Further, defense, expanded under Reagan’s budget, also grew in Carter’s last two budgets. Real spending on other federal programs, such as those in the Department of Education and Department of Transportation (DOT), declined in 1982, but then climbed steadily to new highs over the remainder of Reagan’s first term. Thus, though the 1982 budget may have accelerated an existing spending reallocation, it was not a radical change. Where Reagan did try to depart from the budgetary consensus of the previous administration, he was rebuffed. Almost all of the programs scheduled for termination by Reagan in fact survived his tenure in office. What responsibility, then, does Reagan bear for the runaway deficits of the 1980s? Constant-dollar spending grew at a faster rate during Reagan’s administration (25 percent from 1981 to 1988) than during the Eisenhower (16 percent growth), Nixon (13 percent), Ford (3 percent) and Carter (12 percent) administrations. Annual changes in real spending varied from a low of minus eight percent in 1954 to a high of 13 percent in 1967. Reagan averaged a 3.3 percent annual increase (constant dollars) in spending. On the other hand, Reagan, in his first three budgets, requested decreases for most domestic programs, whereas on average, most postwar presidents have requested nominal increases averaging close to 10 percent per year for domestic programs .s Expenditures for many domestic programs declined in the 1980s. Whereas nominal spending had grown, on average, between 8 and 11 percent per year (for the programs and agencies I examined) for each president from Truman to Carter, spending growth under Reagan was held to less than 1 percent per year. Thus, there is some reason to believe that a new direction in American politics was blazed by Ronald Reagan in his first term in office. But, Reagan did not have to contend with large Democratic majorities in both chambers of Congress, as did Eisenhower, Nixon, and Ford. Indeed, the Republicans owned a majority of seats in the Senate for the first time in over two decades during Reagan’s first six years in office. In the 1980s with the Republicans controlling the Senate and the Democrats controlling the House, spending was cut, on average, for only 30 percent of the items in my sample of 69 agencies (which includes 63 domestic and six defense agencies). When the Republicans controlled both houses of Congress (1947-48 and 1953-54), by contrast, they cut over 42 percent of the items in my sample. In fact, the Republican Congress of 1953-54 cut spending for almost two-thirds of the programs in this sample, more than twice the rate of cuts garnered by Reagan. The question becomes, what was the net effect on spending of Reagan’s occupancy of the White House? Certainly, Reagan was no more successful,

91

U.S. Budget Deficits

under the circumstances in Congress, then expected.I0 Reagan was actually less successful than any other postwar president at pushing his spending cuts through Congress. Truman succeeded in getting Congress to enact cuts for all of the relatively few items for which he requested a cut in spending. Eisenhower succeeded 84 percent of the time. Reagan by contrast succeeded less than 60 percent of the time. Also, his batting average at getting spending increases out of Congress was less than any recent president. More generally, Reagan’s success rate in influencing congressional votes for each of his eight years in office was less than that of Dwight Eisenhower in each corresponding year (Stanley and Niemi 1988, 220-21). Indeed, Reagan’s success rate in 1987-88 was less than Richard Nixon’s success rate even during the period of the Watergate scandal (Nixon won roughly 60 percent of his key votes that year). Thus, by this comparison Reagan was a weaker president than his predecessors. 3.1.4 Federal Expenditures in the 1980s

So what accounts for the budgets of the 1980s? In the first place, though there has been much talk of budget “cuts” in the 1980s, constant-dollar spending nearly doubled from fiscal year 1981 to 1989. And, whereas gross national product (GNP), valued in 1972 dollars, grew less than 24 percent from 1981 to 1988, real spending grew over 25 percent. As shown in table 3.1, real spending at the Department of Agriculture grew by more than 50 percent from 1981 to 1986; defense spending in constant 1972 dollars grew from $79 billion in Carter’s last budget to $117 billion by 1986. At the same time, the Department of Health and Human Services grew by 35 percent in real terms, while spending at the Justice Department, the State Department, DOT, and the Treasury also grew quickly. Second, those agencies and programs chosen for spending reductions in the 1980s were largely ones whose budgets declined throughout the latter part of the 1970s. The Great Society programs (especially housing programs) and the regulatory activities of the federal government started declining under Gerald Ford, and their decline accelerated under Jimmy Carter. The reduction of these programs was therefore begun under a Democratic Congress and was accelerated when both branches were controlled by the Democrats. Real spending declined in Reagan’s first budget (relative to the budget for 1981) for eight departments and programs: Commerce, Education, Energy, EPA, Housing and Urban Development, Interior, the Post Office, and DOT. Of these eight declining spending categories, two (Energy and the Post Office) were continuing declines begun under Carter’s 1981 fiscal year budget. Carter had requested spending cuts for three others: Commerce, Housing and Urban Development and DOT. The remaining three departments and programs that declined in Reagan’s first budget-Education, EPA, and Interior-constituted new spending reductions and can rightfully be attributed to Reagan and the Republicans in the House and Senate. For each of these three departments,

92

Mathew D. McCubbins Constant-Dollar Appropriations for Selected U.S. Executive Departments, 1969-88

Table 3.1 ~

Fiscal year

DoA

DOC

DoD

1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986

8.5 10.3 10.3 12.7 12.7 10.2 11.5 11.0 9.5 10.0 15.4 11.8 12.4 12.0 15.6 16.5 15.3 18.6

1.4 1.1 1.2 1.4 1.6 1.3 1.3 1.7 2.7 1.5 1.6 1.3 1.4 .8 .8 .9 .9 .9

92.1 82.5 74.2 72.6 72.6 66.0 66.8 69.9 75.1 73.0 72.4 72.8 79.4 93.3 102.8 106.5 115.2 117.2

Energy

3.3 4.4 4.4 6.4 15.5 5.2 4.2 4.6 4.8 4.9 5.0

HHS

113.9 122.6 125.0 132.6 146.3 153.9

HEW

HUD

DoL

DOT

65.8 70.1 73.6 78.3 85.0 92.7 92.2 96.1 106.5 110.8 111.7 115.5

3.3 8.8 3.6 3.9 4.1 3.8 40.0 20.2 13.6 25.0 19.2 18.1 17.9

6.0 5.7 6.9 9.0 9.1 7.9 20.9 13.9 18.2 14.8 11.9 14.9 15.5 14.8 15.1 21.5 13.9 13.6

9.0 9.0 12.0 8.6 6.7 20.0 14.9 5.7 9.0 8.7 10.1 9.5 12.5 8.6 14.7 11.3 11.6 10.8

10.3 5.2 5.6 5.2 6.7

Note: Abbreviations: DoA = Department of Agriculture; DOC = Department of Commerce; DoD = Department of Defense; Energy = Department of Energy; HHS = Department of Health and Human Services; HEW = Department of Health, Education and Welfare; HUD = Department of Housing and Urban Development; DoL = Department of Labor; DOT = Department of Transportation.

however, real spending rose quickly in fiscal years 1983-86, undoing much of the reallocation of 1982. The budget story of the 1980s, then, is not the fiscal contractions so often advertised by Congress and the president. But several questions remain unanswered. Why should spending have grown during the administration of a fiscally conservative president supported by a Republican Senate? Should not the deficit have receded during Reagan’s two terms? To answer these questions I must first discuss congressional politics and the effect of the budget process on spending decisions.

3.2 Party Governance in Congress To some analysts of American politics, the runaway deficits of the 1980s are not the unintended result of the Reagan revolution, but rather a consequence of a change in congressional procedure. This perception is based on a well-accepted view that Congress is not so much a democratic institution as a “pluralistic leviathan” (for a model of congressional politics based on this

93

U.S. Budget Deficits

view, see Stigler 1971 and Becker 1983). Central to this view is the thesis that congressional politics is committee politics. The “interest group liberalism” that purportedly dominates congressional politics, if true, has profound implications for budgeting. As Shepsle and Weingast (1984, 355) argue, “Legislators invent programs, seek funding, and are especially attentive to policy areas that create or maintain jobs within their electoral constituency. . . . Expenditure programs are, as a consequence, biased away from least-cost methods of production.” As each subgovernment pursues its policies in the way Shepsle and Weingast describe, the end result could be that the government outspends its receipts. To mitigate the effects of interest group liberalism, members of Congress purportedly have relied on members of the House Appropriations Committee to make the hard choices between supporting their colleagues’ programs and the need to economize on spending (Fenno 1966). This system was supported by rules and procedures in Congress that separated authorization from appropriations. And the system appeared to work reasonably well, producing small but manageable deficits through the 1950s and 1960s. This all changed with the Budget Act of 1974. By transfemng authority for establishing overall spending limits to the budget committees, the act so weakened the House Appropriations Committee that it could no longer guard the Treasury. Instead, its members seemingly became claimants on the federal Treasury rather than its protector (Schick 1980). Two assumptions underlie this explanation of the deficit crisis. First, that “power in Congress has rested in the committees or, increasingly, in the subcommittees,” and thus, as a consequence, “throughout most of the postwar years, political parties in Congress have been weak, ineffectual organizations (Dodd and Oppenheimer 1977, 40). Second, that the House Appropriations Committee was once the “guardian of the federal Treasury” and now is only a subdued guardian. I examine these assumptions in turn, arguing that congressional parties and party leaders exercise more control and greater influence in congressional politics, and in budgeting in particular, than commonly has been perceived. I then seek to explain the budgetary decisions of the 1980s in light of this new understanding.

3.2.1 The Institutions of Agency: Parties and Committees The common view of weak parties and autonomous committees in Congress, in its logical form, is identical to the view of presidential dominance presented in section 3.1 above. The membership of each house has delegated to the committees in each house wide-ranging authority to write legislation, hold hearings, and oversee the executive branch. This delegation, as was the case in interpreting delegation to the executive, has been mistaken for abdication. Two factors underpin the importance and autonomy of committees and sub-

94

Mathew D. McCubbins

committees, according to this view: committees have jurisdictional monopolies, and thus an ex ante veto over proposed legislation; and committees honor each others’ jurisdictional autonomy through a system of deference and reciprocation (Shepsle and Weingast 1984, 351, 353). This view of committee and subcommittee power has been developed to explain a set of generalized observations on congressional behavior. It explains, for example, why coalitions within Congress are seemingly universal and nonpartisan-the reason is that all members face the same necessity, to bring home particularistic benefits, and the institutions are geared toward establishing and enforcing vote trades across projects and benefits. It also follows, for the same reason, that party discipline will be very lax-the vote trades cross party lines. Committees also use their powers, particularly their ex post veto, to ensure that amendments rarely get offered to their bills, and, when they are offered, few, if any, are successful. These observations also are consistent with party control of committees. Indeed, none of the things listed above discriminate between the two views.I’ However, we do observe obvious violations of this cozy view of subcommittee autonomy. For example, multiple referrals, where legislation is sent to several subcommittees, are increasingly common in the House. Members of Congress design their institutions to fit their purpose. Students of American politics have tended to focus on those aspects of congressional institutions that enable members to bring home private goods (projects or programs for their own districts). Studies of congressional behavior have focused largely on how members secure water projects, military bases, roads, and post offices for their districts, and the consequences of these activities for their political survival. These studies, of course, assume that voters appreciate projects in their district and that members can build reputations as good providers of federal pork. But, party affiliations are also an important ingredient in voters’ decisions: party labels signal information that is otherwise very expensive for voters to obtain about the policy positions of candidates. As a result, politicians, in seeking office, also establish reputations as partisans. Thus, politicians adopt a mixture of collective, that is, partisan, and individual, that is, districtoriented, activities in seeking reelection. It follows that members will seek to structure Congress in such a way as to facilitate both of these activities. Party organizations, their leadership, and the committees that serve them, provide the institutional means for pursuing the collective goals of party members. But these pursuits are not to the exclusion of individual district-oriented benefits. Thus, the majority-party leadership uses its agenda powers, in concert with the agenda powers assigned to committees, to secure the omnibus pork-barrel logroll so familiar to congressional scholars. The congressional parties, of course, delegate much of the authority to make these kinds of decisions to the leadership and to committees, though the

95

U.S. Budget Deficits

Democratic caucus has at times sat down as a whole and made policy for the Democratic majority. In delegating, the congressional parties encounter the agency problems ubiquitous to human experience: for a variety of reasons, intentional or not, the person to whom authority is delegated may not carry out their authority in the best interests of those doing the delegating. So pernicious are these problems that their existence has led many scholars to conclude that the congressional parties have in fact abdicated their authority to the standing committees and subcommittees of Congress. The abdication conclusion, however, ignores the efforts on the part of the congressional parties to mitigate delegation problems. In essence, like the separation of powers designed into the structure of the federal government, party organizations-in particular the party leadership-and the system of standing committees, form a separation of powers, a system of checks and balances. Party leaders use a hierarchical system of committees (capped by the “control committees,” i.e., Rules, Appropriations, and Ways and Means in the House), control of committee appointments and procedural restrictions to steer policy outcomes in favor of the collective interests of the majority party (Rohde and Shepsle 1973; Shepsle 1988; Cox and McCubbins 1991; McCubbins and Schwartz 1984; McCubbins and Page 1987; McCubbins et al. 1987, 1989). Some standing committees, of course, more closely involve the party label than do others. The actions of these committees affect everyone in the party, and, collectively, the party and its leaders have a greater interest in mitigating the agency problems that arise vis-h-vis these committees. Other standing committees, such as Post Office and Civil Service, Interior, and Merchant Marine and Fisheries, have jurisdiction over issues unrelated, or only minimally related, to the issues that voters identify with the party. Delegation, in this case from the majority party caucus to the standing committees of Congress, in the conventional view of committees has been mistaken for abdication (Kiewiet and McCubbins 1991; Cox and McCubbins 1991). The transfer of authority has been recognized, the actions of the agents (i.e., committees) witnessed. What have not been appreciated, however, are the mechanisms used by the majority party to direct the actions and choices of their committees and the effect these mechanisms have on committee actions. If committees are agents of party caucuses then we expect that most of the decisions in Congress would be made in committee-that is, after all, their function-and that committee members would acquire expertise in the committee’s jurisdiction-that is after all why they were delegated the jurisdiction in the first place. We do not expect these functions to be uncontrolled, however, and we witness many and varied attempts by majority party caucuses and their leadership to control committees. Further, having anticipated the reaction of the majority party to its proposals, the committee can expect that few of its bills will be amended or rejected on the floor.

%

Mathew D. McCubbins

3.2.2 The Role of the House Appropriations Committee In Richard Fenno’s (1966) classic account, the House Appropriations Committee was depicted as a budget-slashing “guardian of the federal Treasury” (353), protecting the House from the budgetary excesses of its own committees and from budget-maximizing bureaucrats. Fenno reported that the House Appropriations Committee cut the amount requested by the president in 73.6 percent of the 575 cases in his data set (Fenno 1966, 353, table 8.1). For a set of 69 agencies and programs, including almost all of Fenno’s 36 bureaus, for a period extending from 1948 to 1985, I found that the committee cut the president’s requests for 70.4 percent of 1,983 cases.I2 But, do these statistics constitute evidence that the “dominant pattern” (Fenno 1966, 353) for the House Appropriations Committee is to guard the Treasury? If the procedural restrictions on the Budget Bureau constrain the bureau’s ability to revise agency budget estimates, and agencies compile estimates in accordance with authorizing legislation, then presidential requests will reflect, to a large extent, the level of funding preferred by the authorizing committees in Congress.l 3 If many committees prefer more spending on items within their jurisdiction than preferred by the House majority party as a whole, and if the House AppropriationsCommittee is relatively representative of the majority party,14 then House Appropriations will often appear to “guard” the Treasury. In Fenno’s account of the House Appropriations Committee, party politics plays essentially no role. I have argued here and elsewhere (Cox and McCubbins 1991; Kiewiet and McCubbins 1991) that the committee functions as an agent of the majority party, pursuing the collective policy goals of the majority party’s membership. These goals may sometimes be to cut the budget, but they are not necessarily so. Consequently, the varying goals of the parties controlling the White House and the House of Representatives determine, for example, the treatment afforded the president’s budget requests by the House Committee on Appropriations, with Democratic majorities favoring higher spending on domestic programs than Republican majorities. How often the committee cuts the president’s requests varies with partisan factors. The House committee is most likely to cut the president’s requests when the president is a Democrat and the House is controlled by the Republicans (cuts amount to 93 percent of actions taken). The committee is somewhat less likely to cut the executive’s requests when the same party controls both bodies-whether controlled by the Democrats or the Republicans, the committee cuts nearly 80 percent of the requests in my sample; but, a Democratic committee is far less likely to cut a Republican president’s requests: only 57 percent of requests were cut by the committee under these circumstances. Under Reagan, the committee cut only 38 percent of the president’s requests, and actually proposed increases for 55 percent of the items. Fenno described such changes as “mood” swings, where the committee would shift from an “econ-

97

U.S.Budget Deficits

omy mood” to a “spending mood.” My data suggests that the mood toward executive budget requests by the House Appropriations Committee is determined by partisan differences between the House, Senate, and the executive. Further, if the guardianship hypothesis is correct, I would never expect the House, in its amendments, to decrease the committee’s recommendations. The committee, after all, is supposedly doing a job the House is incapable of doing, holding back spending. However, I found that more than 58 percent of individual, floor changes were decreases, and only 42 percent were increases. Of the appropriations bills amended by the House that I examined, 106, or 43 percent, reduced the totals recommended by the House Appropriations Committee. Has the 1974 Budget Act changed the Appropriations Committee from the Treasury guardian it once was? There is no evidence that it was a guardian or that its function has in fact changed. It has been, and still is, a check used by the majority-party leadership to ensure that the policies pursued by the other standing committees in the House reflect the collective goals of the membership of the majority party (Kiewiet and McCubbins 1991, chap. 3). 3.2.3 The Congressional Budget Process in the 1980s The budget process created by the 1974 act has three key components. First, it created budget committees responsible for setting guidelines for all aspects of federal spending and revenue. Second, it requires authorizing committees to “reconcile” spending policy within their respective jurisdictions with the guidelines. Third, if a committee fails to recommend legislation providing satisfactory reconciliation, then the budget committees can write the reconciliation legislation for them. To ensure that the budget committees are responsible to the majority-party caucus and its leadership, the act also established a special relationship between the committee and the party leadership. Members are handpicked by the leadership, and tenure and seniority norms observed for other committees do not apply. The majority-party leadership, by these procedures, has relatively greater control, year in and year out, over the composition of the budget committee than it does over the composition of any other committee, allowing the majority party leadership to use the budget process to inject the majority party’s priorities into the decisions of all committees. Indeed, from the perspective of floor majorities in the House and Senate, the budget process has been strikingly effective in the 1980s. In 1981 the Republicans in the House and Senate, together with some conservative Democrats used the budget and reconciliation process to “cut” some $36.6 billion in spending for fiscal year 1982.15The first budget resolution called for the House Committee on Agriculture to write legislation bringing about “cuts” of $2.2 billion in fiscal year 1982 budget authority; required the House Committee on Banking and Urban Affairs to “cut” budget authority in its programs by $12.9 billion; the Committee on Education and Labor to cut $13.5 billion; the

98

Mathew D. McCubbins

Committee on Energy and Commerce, $6.4 billion; and the Committee on Public Works and Transportation, $6.6 billion. In this case a majority consisting of Republicans and conservative Democrats in the House rolled the House Democratic leadership in the budget process. This coalition’s use of the process demonstrated how effective the process could be at submitting congressional committees to the will of floor majorities. It would also mark the only time the House Democratic leadership was rolled. The first budget resolution in 1982, for fiscal year 1983, required “cuts” of $2.2 billion in budget authority. This time, the House Democratic party, using inventive procedures, such as the “King of the Mountain” rule,16managed to reestablish some control over the course of the budget process. The resolution required the revenue committees to recommend legislation to raise $20.9 billion in additional funds. The budget resolution also required “savings” of $7.8 billion in defense spending. The House Democrats used the budget process in 1983 to draft their own budget blueprint as an alternative to the Republican budget submitted by Reagan. The Democrats’ budget added $33 billion in domestic spending to Reagan’s proposal, requested $30 billion in new revenues, and “cut” Reagan’s defense request by $16 billion. The Republicans in the Senate and the Democrats in the House chose alternative strategies in using the budget process to further their programs. The Republican leadership used the budget process to give direction to Senate committees. The Democrats used the budget process as a means to unite the party behind a common program, with the House Budget Committee holding hearings with the entire Democratic caucus. Ultimately, the result of these efforts on the part of party caucuses to control the product and actions of committees is that spending policy reflects the desires of the majority party in each chamber. Indeed, the single best predictor of changes in spending policy for almost the whole range of federal programs and agencies is party control of Congress and the White House (Kiewiet and McCubbins 1991, chap. 8). This analysis suggests that policy is influenced, to a far greater extent than commonly believed, by party politics.

3.3 The Partisan Roots of Deficits Having rejected the two most common explanations of the runaway deficits of the 1980s, how can they otherwise be explained? To explain deficits, I first need to explain federal spending and revenue policies, and how they are affected by divided government. With regard to federal spending decisions, it can be shown that these, at every stage of the spending process, reflect party politics.L7Domestic agencies, for example, do better under Democratic administrations and Democratic Congresses than under Republicans. Defense and high-technology programs do better under Republican administrations and Republican Congresses than under Democrats.

99

U.S.Budget Deficits

When the president is a Republican and both houses of Congress are controlled by Democrats, spending on domestic and social programs is somewhat restrained, but is close to the levels that would have been adopted had the Democrats controlled the White House as well. In such instances, though the Republican presidents act as a restraint on domestic spending, the ability of Congress to package various spending items into an omnibus bill makes it difficultfor even the most ardent Republicans to limit spending. But, what happens when control of Congress is divided, when a Democratic majority controls the House and a Republican majority controls the Senate?’* This has happened several times, mostly in the waning decades of the nineteenth century, though in this century it has occurred twice, once during the second half of Hoover’s administration and most recently between 1981 and 1986. The Constitution established a bilateral veto game between the two chambers, and each chamber holding a check on the actions of the other. The cooperation and coordination necessary to overcome these constitutional checks and balances is frequently inadequate. Budget deficits present members of Congress with a collective dilemma: everyone would be better off if deficits could be reduced, but, individually, members are not willing to reduce spending on their preferred programs or to raise taxes for their constituents. Party discipline is often required to solve such collective dilemmas. Neither congressional party is likely to go along with a solution to a problem such as the deficit for which the other party can claim credit, and each will use its institutional position to defeat the other party’s attempts to solve the problem. What is implied by divided control, then, is that the cooperation to solve collective problems, like the deficit, will largely be nonexistent. What then will be the equilibrium to this bilateral veto game? In an effort to model this game I make a simplifying assumption about the preferences of the members of each party. I assume that there are two types of programs: domestic programs favored by the Democrats and defense programs favored by the Republicans. The preferences of the two parties over budgetary allocations to these two goods is given in table 3.2. The Democrats, I assume, prefer most that spending on their programs be increased, while spending on the Republicans’ programs be decreased (denoted D, r in the table). Democrats next most prefer that spending on all programs be increased (denoted D, R). Democrats are then assumed to prefer decreases in both their own programs and the Republicans’ programs (denoted d, r) to a decrease in their programs with an increase in the Republicans’ programs (d, R). I assume that Republican preferences are similar with respect to their own programs. In bilateral veto games, the reversionary outcome (i.e., if no solution is adopted) determines what if any cooperative solution will be an equilibrium. The spending reversion point for most federal programs is zero. Congress must annually enact legislation appropriating money for most of these activities if they are to continue. Typically, however, Congress will pass a continuing resolution that pegs spending at some low baseline level (often at the

100 Table 3.2

Mathew D. McCubbins Party Ranking of Spending Allocations Democrats

Republicans

Note: D denotes increased spending on Democratic programs; R denotes increased spending on Republican programs; d denotes decreased spending on Democratic programs; and r denotes decreased spending on Republican programs.

spending rate for the previous fiscal year) if no appropriations bill is enacted. Continuing resolutions typically yield little or no growth in spending and may even entail a modest decrease (adjusting for inflation) in spending for the programs covered by the resolution. Thus, the reversion, if no spending policy is agreed to by the two parties, is to decrease appropriations for all programs (i.e., the outcome denoted d, r in the table is the reversion point). The only alternative in the table, then, that is preferred by both parties to the reversionary outcome and thus will not be vetoed by one or the other party, is the outcome in which spending for the programs of both parties is allowed to increase (denoted D, R). Thus, under conditions of divided control, I expect overall spending to increase. In 1983, for fiscal year 1984, for example, reconciliation legislation did not “cut” budget authority in any area. In 1985, for fiscal year 1986, reconciliation legislation passed by Congress and signed by the president, though requiring “cuts” in agriculture, defense, energy, Medicare, and 10 other programs, increased spending for the EPA’s Superfund, income and social security programs, veterans’ affairs, and three other programs. The following year, there again were no significant “cuts” in any area. Throughout the decade, the Democrats in the House and the Republicans in the Senate forged a union that enacted policies confrury to the basic tenets of Reagan’s budget policy. In his second term this “coalition” enacted increases in social and education programs and “cuts” in defense spending, despite Reagan’s strenuous opposition. I have already noted that spending increased during Reagan’s two terms in office-nearly doubling in current-dollar terms during these eight years. The Democratic-Republican spending compromise can be seen in an examination of the budget. Though some areas of domestic spending were reduced, largely those that were under pressure in earlier decades, other areas increased tremendously. Defense spending also increased during Reagan’s first term. As expected, this spending compromise was abrogated once the Democrats regained control of the Senate. With unified party control of Congress, the Democrats could undertake to cut back on those programs favored by the Republicans, namely defense. With regard to federal revenues, the story is much simpler. The logic of the

101

U.S. Budget Deficits

revenue game is presented in figure 3.2. The figure presents the ideal points of congressional Democrats (denoted D in the figure), Southern Democrats (SD), and Republicans (R), in an abstract policy space whose dimensions are the incidence of the tax system (along the horizontal axis) and the tax rate (along the vertical axis). The incidence of the tax system can be thought of as identifying the individuals who will carry the tax burden (from richer to poorer); the tax rate merely runs from zero to 100 percent of income. I assume that the loss of welfare to each player as the policy moves away from that player’s ideal point is proportional to the distance from the ideal point (that is, their indifference contours can be represented as circles). If this is the case, the triangle defined by the line segments connecting the three ideal points contains all the Pareto-optimal outcomes-that is, the policy choices that cannot be changed without making one of the three members worse off. Before the victory by Reagan and the Senate Republicans in 1981, I assume tax policy was set by the Democrats and was at D, their ideal point. Had the Senate remained under Democratic control, or had the Democrats maintained a large enough majority in the House, no change in tax policy would have been expected, as the Democrats would have had the institutional means to

TAX RATE

D

R

HIGH INCOME

LOWINCQME IWDENCE OF TAX BURDEN

Fig. 3.2 Divided government and tax reform

102

Mathew D. McCubbins

veto any changes in the status quo (which was D). But, after the 1980 election, the Republicans were able to strike a deal to change the tax system with conservative, mostly Southern, Democrats in the House. The new tax policy was then chosen somewhere between SD and R in figure 3.2, at a point such as TA. This point was a new equilibrium: once Reagan proposed the tax cut, and the Senate Republicans endorsed it, the Democrats felt they could not oppose it. Once the tax reduction was enacted, however, tax increases could not be passed, even in the face of mounting budget deficits. Reagan promised to veto any tax increase, and with a large Republican minority in the House and Republican control of the Senate, such vetoes were certain to be sustained. The 1981 tax cut reduced current revenues for each succeeding year by over $100 billion. It was not until the 99th Congress that a new deal, this time between the Democrats, who controlled the House, and the Republicans, who controlled the Senate and White House, could be struck. This new tax policy, denoted TR in the figure, is not much different than the previous tax policy, TA. The reforms of 1986, however, were preferred by both the Democrats and the Republicans to the 1981 tax policy. Tax policies that would raise revenue, as proposed by the Democrats, require that one or the other of the parties (presumably the Republicans) be made worse off. Since each party controls a veto in this game, by virtue of their holding one or the other branch of government, no increase in revenue was possible. Divided party control within a setting of a bilateral veto game as established by the Constitution, then, led to a form of stalemate in which the deficit problem was allowed to fester. Spending continued to climb at the same constant rate throughout Reagan’s administration. Revenues were climbing at almost exactly the same rate as expenditures prior to 1982, with a relatively small difference between the two trends. As a result of the 1981 tax cut, revenue declined for three years from 1982 to 1984. After 1984, revenues again paralleled expenditures, but with a dramatic increase in the deficit. Taken together, the effect of divided government on revenue and spending decisions, and the effect of the tax cut of 1981, produced the runaway deficits of the 1980s. The pattern is not new, it has recurred throughout the twentieth century: since 1929, divided government has yielded sizable increases in the national debt. Indeed, the increase in the debt attributable to divided government exceeds the effects of national unemployment and inflation by an order of magnitude (these results are reported in the appendix below). If the Democrats continue to hold majorities in the House and Senate, and the Republicans continue to occupy the White House, then little progress will be made toward reducing the national debt. 3.4

Conclusion

Drawing from the literature on American national government, two explanations for the runaway deficits of the 1980s have received widespread com-

103

U.S.Budget Deficits

ment. The first is that the president did it-that Ronald Reagan, on his way to forging a revolution in American politics, put into place policies that pushed America over the deficit precipice. The second has its roots in Congress-that the Budget Act of 1974 led to the unraveling of fiscal restraints in Congress. Both of these explanations have, at their core, a perception that congressional parties are merely shells within which policy is bartered and to which no control over policy is granted. Congressional majorities, in these models, have abdicated their collective responsibilities over national policy on the one hand to the president and, on the other, to congressional committees and subcommittees. The deficits of the 1980s are the consequence of a structural problem: divided government. Once the deficits of the 1980s were in full bloom, the check Ronald Reagan held over increases in revenue was sufficient to prevent Congress from enacting a tax increase. The compromise required to overcome the mutual checks held by the House Democrats and the Senate Republicans over each other’s spending programs led to increased spending on nearly every function of government. Though spending has been held in check since the Democrats took control of both houses of Congress in 1986, especially spending for Republican programs such as defense, Republican threats to veto tax increases will keep budget deficits in the headlines for some time to come.

Appendix The central thesis of my paper is that parties exert substantial control over the policy-making apparatus in Congress. It follows that divided government will lead to increased budget deficits. In this appendix, I examine the effects of party politics on the level of the deficit for the period from 1929 to 1988.

On Measuring Changes in Spending and Deficits Before describing my test and results, however, it is important to discuss some measurement issues. The first has to do with choosing baselines in comparing fiscal policy from one year to the next. For example, is it possible to compare the defense budget in the 1950s to the defense budget in the 1980s? There are many pitfalls. Several data transformations are commonly used to facilitate this comparison. Budget figures are deflated to account for price increases; budgets are described as a percent of total spending or of GNP. Each transformation serves a purpose, and each introduces errors, distortions, and biases into the comparison. Using data transformation willy-nilly especially in making comparisons across many years, can introduce more confusion than clarity. On a more abstract level, without explicating a specific theory of the demand for government goods and services, we can make use of the general

104

Mathew D. McCubbins

properties of demand functions as an approximation and apply the Slutsky equation. The Slutsky equation states that changes in the demand for a commodity can be decomposed into a substitution, or price, effect and an income effect. Thus, in modeling fiscal policy decisions, most analysts have included GNP as an explanatory variable (Kiewiet and McCubbins 1991; Barro 1979). Dividing your dependent variable-whether it be federal spending or deficit-by an independent variable is highly problematic. To avoid these problems, in the text I compare spending figures over time for federal programs in constant dollars. The second measurement issue has to do with how to measure the deficit. It may seem odd that this is problematic. The deficit, it seems, should just be the difference between how much the government spends and how much it takes in. Something akin to this is the measure commonly reported by the government as the deficit. But, this measure of the deficit minimizes the apparent size of the deficit. The current surplus in the Social Security trust fund is used to offset the deficit in spending on other items. This measure, however, ignores the obligations created by the Social Security system. The surplus in the Social Security trust fund is not treated as tax revenue for general obligations; the Social Security Administration purchases Treasury bonds with the surplus. These bonds must be repaid if the Social Security system is to remain solvent. Thus, while the deficit measure commonly used by the federal government treats Social Security contributions as general revenue, in actuality the surplus in contributions just adds to the debt, which in this case the government owes to the Social Security trust fund. Another measure, which treats the bonds held by the Social Security trust fund-and other federal agencies-as debt instead of revenue, is to examine changes in the national debt from one year to the next. This measure captures all government borrowing from private and public sources and gives a measure of how far government expenditures exceed general revenue. I examine both measures here. There are still yet more complications, however. Analysts, members of Congress, and the administration often seek to subtract various components of federal spending from deficit measures. These, of course, all have the characteristic that they make the deficit smaller. Indeed, as is true at many savings and loans, the book juggling often yields black ink instead of red. As long as people are clear about what they are doing and why, book juggling is not necessarily a problem. Subtracting interest payments, or some fraction thereof, from the deficit might be the right measure to test some hypothesis in which you are unconcerned about wealth transfers between individuals in society. The hypotheses examined here are about policy decisions, made by political actors in the federal government. In choosing how to measure and report deficit figures in, for example, Gramm-Rudman-Hollings, interest payments were included. In fact no measure of spending is excluded from the deficit measures or measures of the debt in legislation.

U.S. Budget Deficits

105

On the Determinants of the Deficit In the 1980s, the annual increments to the national debt (i.e., the budget deficit) became larger than the entire budget of just a few years earlier. To test my hypothesis that divided government leads to increased budgetary imbalances, I estimate a model of the choice of the deficit, derived from models of spending and revenue developed elsewhere (Kiewiet and McCubbins 199 1; McCubbins 1990). The estimation involves income effects (modeled as GNP and unemployment) and substitution effects (various political variables, including dummy variables corresponding to periods of divided government, wars and inflation). The model is represented by the independent variables listed in table 3.A1. The estimation in table 3.A1 was affectedby a few econometric problems, the most serious of which was autocorrelation. In each case I included a set of lagged dependent and independent variables to correct for potential inefficiencies. Other details of the estimation can be found in McCubbins (1990). In table 3.A1, I estimate a model of deficit determination using three different measures of the deficit. In column 1, I estimate the model using the common definition of the deficit that is reported each year by the federal government (measured in constant collars). This measure is simply the difference Table 3.A1

On The Determination of the Federal Budget Deficit, 1929-88 Dependent Variable

Independent Variable Constant WWII

KOREA PRES DEMkONGRESS REP PRES REPkONGRESS DEM PRES REP~ONGRESSDIV

DEFICIT

ADEBT

(1)

(2)

.oo

.29 .13

AGNP

.09

DEFICIT,_,

- .09

DEFICIT,-

ADEBT,-,

- .02 .92*

I

Observations

R2

58

58 .89

Sum of Squared residuals ~

.oo

.12*

POSTWAR GNP

DEBT/GNP,-

.01*

.04* .oo3*

.01*

-.17*

POSTWAR

.oo

.09*

.07

AGNP

.01 .29* .03*

.01

.50

GNP

(3)

- .06 .45* .02 - .08*

28.69 -84.39* - 9.40* 2.81 - 2.68 -26.99*

U

I

DEBT/GNP

58

.68 .22

6,945

.98 .06

~~

Note: For definition of variables, see appendix text. Values in table are estimated coefficients. The coefficient is significant with probability greater than or equal to .95.

106

Mathew D. McCubbins

each year between federal expenditures and federal revenue from all sources (including, e.g., Social Security). In column 2, I estimated the model using as a measure of the deficit the percentage change from one year to the next in the size of the (constant-dollar) federal debt. This measure has the advantage that it includes purchases of Treasury notes by Social Security (which in fact must be repaid by future taxes) as part of the debt, and, therefore, as part of the deficit. In column 3, at the request of the editors, I measure the deficit as the percent of GNP accounted for by the debt (both in current dollars). This latter measure, of course, necessitated the elimination of GNP as an independent variable. The variables in table 3.A1 are defined as follows: DEFICIT is the constantdollar federal budget deficit, calculated from the current-dollar figures reported in Historical Statistics of the United States (U.S. Department of Commerce, Census Bureau 1975) and summary tables in The Budget of the United States Government (U.S. Executive Office of the President, Office of Management and Budget, various years), and the Implicit Price Deflator for federal government purchases of goods and services as reported in Survey of Current Business (U.S. Department of Commerce, Bureau of Economic Analysis, various years); ADEBTis the percentage change in the (1972) constant-dollar federal debt,Igusing the Implicit Price Deflator for federal government purchases of goods and services to deflate the current debt figures, as reported in Historical Statistics (U. S. Department of Commerce, Census Bureau 1975), Historical Tables, Budget of the United States Government (U.S. Executive Office of the President, Office of Management and Budget, 1989) for fiscal 1990, and the Economic Report of the President (U.S. Council of Economic Advisers, 1989); DEBT/GNP is the ratio of current debt to current GNP for each year, drawn from Historical Statistics, Historical Tables, Budget of the United States Government for fiscal 1990, and the Economic Report of the President; W I I is a dummy variable that takes a value of one for the fiscal years of World War 11, 1943-45, and zero otherwise; KOREA is a dummy variable that takes on a value of one for the fiscal years 1952 and 1953, and zero otherwise; PRES R E P ~ O N G R E S SDEM is a dummy variable that takes on a value of one when the presidency was held by a Republican, and the Congress was controlled by the Democrats (fiscal years 1956-61, 1970-77, and 1988), and zero otherwise; PRES DEWCONGRESS REP is a dummy variable that takes on a value of one when the president is a Democrat and the Congress is controlled by the Republicans (1948-49), and zero otherwise; PRES R E P ~ O N G R E S SDIV is a dummy variable that takes on a value of one when the president is a Republican and control of the two houses of Congress is divided (fiscal years 1932-33 and 1982-87); U is the unemployment rate during the fiscal year, as reported by Survey of Current Business; I is the rate of inflation during the fiscal year, calculated as a change in the Consumer Price Index, as reported in Historical Statistics and Survey of Current Business; GNP is the (1972) constant-dollar GNP as reported by Survey of Current Business; POSTWAR GNP is a variable

107

U.S. Budget Deficits

that is equal to GNP for fiscal years after the end of World War I1 (1946-88), and is zero otherwise; AGNP is the percentage change in real GNP;*O POSTWAR AGNP is a variable that equals AGNP for 1946-88, and is zero otherwise. From previous research on spending decisions, I expected increasing unemployment to lead to increased spending. I therefore expect increases in the unemployment rate, all else constant, to lead to increased deficits, and thus, U to have a negative coefficient in column 1 (deficits are measure as negative numbers, surpluses as positive numbers) and positive coefficients in columns 2 and 3. I have no prediction about how changes in I will affect budget deficits. Wars increase defense and related spending and seem to have a mixed effect, but no net decrease, on domestic spending. I thus expect the wartime variables (WII and KOREA), all else constant, to yield negative coefficients in column 1 and positive coefficients in columns 2 and 3.21If spending has increased faster than revenue as the economy has grown, then the coefficients for the two GNP measures will be negative in column 1 and positive in the other two columns. My model of party politics implies that deficits will not increase (Republican majorities might decrease spending and therewith the deficit) when both houses of Congress are controlled by the same party (the majority party in Congress will control the agenda over spending and taxes), even when the president is of a different party, all else constant. Thus, PRES DEM~ONGRESSREP and PRES R E P ~ O N G R E S SDEM should be nonnegative in column 1 and nonpositive in the remaining regressions. It is important to note, however, that divided control of Congress should yield higher deficits: that is, PRES R E P ~ O N G R E S SDIV should be negative in column 1 and positive in column 2 and 3. The results here strongly support the hypothesis that divided control of Congress leads to increased deficits: in all three equations, the coefficient on divided control of Congress was significant and in the predicted direction. We have had two occurrences of divided congressional control in this century, once in the latter half of the Hoover administration and for the first six years of the Reagan administration. If we estimate two dummies, one for each occurrence, both yield significant coefficients in an auxiliary regression. Divided control of government, otherwise, when the effects of divided control of Congress and the tax act of 1981 are held constant, shows either no effect on deficits or confirms that Republican control of Congress produces decreased spending and therewith decreased deficits. The coefficients for unemployment and GNP had the predicted sign, but were not always significant. The war dummies always had the right sign, and the dummy for World War I1 was always significant.

108

Mathew D. McCubbins

Notes 1. The measure of the debt used in fig. 3.1 is the gross federal debt. This “is the broadest generally used measure of the Federal debt. It is composed primarily of the public debt (direct borrowing by the Treasury) but also includes agency debt (such as borrowing by the Tennessee Valley Authority or the Postal Service). About threefourths of the gross debt is held by the public, and about one-fourth is held by Government accounts” (U.S. Department of Commerce, Census Bureau 1975, 1097). This is the measure of the debt most commonly applied by Congress, e.g., in defining debt ceilings. As the debt and deficit have become important political issues, many new and creative measures of each have been prescribed. I address some of the measurement issues involved in measuring the debt and the deficit in the appendix. 2. The 1921 act redelegated the responsibility to compile budget estimates to the president from the Treasury Department, Division of Book Keeping and Warrants. The act specified that the new budget conform to existing law on agency reporting, which included such minutiae as estimates from the Navy Department for printing and stationary; advertising; postage; fuel oil and candles for navy yards; funeral expenses, etc. (see Kiewiet and McCubbins 1989; see also U.S. Congress, House 1921). 3. Sundquist (1981, 12) describes the difference between delegation and abdication as merely semantic, saying that any delegation by Congress is abdication. 4. In granting authority to the president under the 1921 act, Congress required that the estimates submitted conform to existing law. The earlier legislation, e.g., required that estimates be made only for those items authorized by law and that large deviations in estimates from the previous year’s appropriations must be explained in detail (for details on the requirements of the 1921 act, see U.S. Congress, House 1921). 5. Why members of Congress have delegated the authority to propose legislation to the president is explored in Kiewiet and McCubbins (1988). They argue that Congress will always delegate to someone the jobs of fact-finding and the drafting of legislation. The choice is between an executive agency or department, which will be heavily influenced by congressional committees or the president, who is independent of such “iron triangles.” When the members want to establish a check on their own committees, they must secure information from sources independent of those committees. Only one official in the federal government satisfies the requirements-the president. 6. Continuing resolutions are joint resolutions that may provide temporary funding for affected agencies when Congress fails to complete action on one or more regular appropriations bills before the start of a fiscal year (see Oleszek 1989). 7. Reagan, in each of his budget requests, sought to terminate several dozen programs, ranging from the Jobs Corps, to Amtrak, to the Small Business Administration, to urban mass transit subsidies and rural water and waste disposal grants. Congress voted only to terminate the U.S. Travel and Tourism Agency and to sell Conrail. Indeed, of those programs the president tried to kill in, e.g., 1985, only half even had their budgets reduced. 8. Based on an analysis of a sample of 63 domestic programs. Elsewhere (McCubbins 1990, table A-2), I present results that compare the effects of all postwar presidents on budget requests. The regression reported there shows that Reagan, on average, all else constant, requested far deeper spending cuts than any postwar president. Further, Republicans requested greater cuts in spending, all else equal, than did Democrats, with the exception of Carter. See section 3.2.2 below. 9. Restraint in spending growth contrasts with the budget “cuts” most often referred to in Reagan’s fiscal 1982 budget, which were changes in authorizations relative to policies set under Carter (see Muris 1989).

109

U.S. Budget Deficits

10. These expectations reflect projections of the success of a Republican president facing a Democratic majority in the House and a Republican majority in the Senate. 11. For some critical tests, see Kiewiet and McCubbins (1991) and Cox and McCubbins (1989). 12. The obverse, of course, is that the committee increased spending for over 19 percent of the budget items that came before it, and they granted an amount equal to the president’s request in almost 10 percent of the cases. The proportion of times the committee recommended an increase for one of these items over the president’s request ranged from zero percent in fiscal year 1947 to over 13 percent in fiscal years 195961, to a high of 51 percent in 1984. 13. This, of course, is implicitly Mayhew’s (1974) model of agency estimates. If we are to believe that the committee is protecting members from themselves, the estimates the committee deals with must be a reflection of their own desires. If they were not, then the committee would be protecting members from the executive branch, not from themselves. 14. On the unrepresentativeness of many House committees see Cox and McCubbins (1991). On the representativeness of the House Appropriations Committee, see Kiewiet and McCubbins (1991). 15. “Cuts” referred to changes in authorization so that spending was less than it would have been under existing law. For example, “cuts” in Medicare were said to exceed $50 billion for the 1980s. Yet current dollar Medicare outlays actually grew from $32 billion to $94 billion between 1983 and 1988 (Muris 1989). Growth in real terms was only 45 percent. 16. Under this rule, adopted by the House in May 1982, seven budget alternatives were considered (and 68 perfecting amendments), with the House Budget Committee’s recommendation being voted last. The rule requires that the last alternative to win a majority is the plan that prevails. This is just one example of the extraordinary rule changes used by the Democrats to control the budget process and the conservative Boll Weevil faction. In 1982 none of the alternatives won a majority. It was used again in 1983, when the Democratic House Budget Committee’s plan prevailed. 17. See, generally, Kiewiet and McCubbins (1991). McCubbins (1990, tables A-2, A-3) also presents results that model partisan differences in presidential spending requests and congressional appropriations decisions. 18. Alesina and Tabellini (1989) and Roubini and Sachs (1989) discuss crossnationally the impact of party policy and divided government on budget deficits. 19. I use an approximation for the percentage change that is calculated as ADEBT = log(DEBT/DEBT,-,) where DEBT, is the federal debt measured in constant dollars in year t. This transformation is referred to as a partial log. 20. I again use a partial log transformation. 21. A dummy variable for the Vietnam war was dropped from the specification, all tests showed that its effect was insignificant.

References Alesina, Alberto, and Guido Tabellini. 1989. A Positive Theory of Fiscal Deficits and Government Debt. Typescript. Barro, Robert I. 1979. On the Determination of the Public Debt. Journal of Political Economy 87 (October): 940-71. Becker, Gary S . 1983. A Theory of Competition among Pressure Groups for Political Influence. Quarterly Journal of Economics 98:37 1-400.

110

Mathew D. McCubbins

Binkley, Wilfred. 1962. President and Congress. 3d ed. New York: Vintage Books. Brand, Stanley. 1985. Congressional Quarterly Weekly Report (September 14): 1815. Bryce, James. 1924. The American Commonwealth. New York: Macmillan. Bums, James M. 1965. Presidential Government: The Crucible of Leadership. Boston: Houghton Mifflin. Chubb, John E., and Paul E. Peterson. 1989. Can the Government Govern? Washington: Brookings. Cox, Gary W., and Mathew D. McCubbins. 1989. Political Parties and the Appointment of Committees. Paper prepared for the Conference on Congressional Structure and Elections. University of California, San Diego, February 1 1 . . 1991. Parties and Committees in the House of Representatives. Berkeley and Los Angeles: University of California Press, forthcoming. Dodd, Lawrence C., and Bruce I. Oppenheimer. 1977. The House in Transition. In Congress Reconsidered, ed. L. C. Dodd and B. I. Oppenheimer, chap. 2. New York: Praeger. Edwards, George C. 1980. Presidential Influence in Congress. San Francisco: W. H. Freeman. Fenno, Richard J. 1966. The Power of the Purse. Boston: Little, Brown. Fiorina, Moms P. 1981. Control of the Bureaucracy: A Mismatch of Incentives and Capabilities. In Congress Reconsidered, 2d ed., edited by Lawrence C. Dodd and Bruce I. Oppenheimer, 332-48. Washington, D.C.: Congressional Quarterly Press. Kiewiet, D. Roderick, and Mathew D. McCubbins. Presidential Influence on Congressional Appropriations Decisions. American Journal of Political Science 32:7 13-36. . 1991. The Logic of Delegation: Congressional Parties and the Appropriations Process. Chicago: University of Chicago Press. Kirst, Michael W. 1969. Government without Passing Laws. Chapel Hill: University of North Carolina Press. Laski, Harold J. 1940. The American Presidency. New York: Harper & Brothers. McCubbins, Mathew D. 1990. Divided Party Control and Budget Deficits. Paper presented at a conference on the Causes and Consequences of Divided Government. University of California-San Diego, La Jolla, California, June 30. McCubbins, Mathew D., Roger G. Noll, and Barry R. Weingast. 1987. Administrative Procedures as Instruments of Political Control. Journal of Law, Economics and Organiztions 3:243-77. . 1989. Structure and Process, Politics and Policy: Administrative Arrangements and the Political Control of Agencies. Virginia Law Review 75:431-82. McCubbins, Mathew D., and Talbot Page. 1987 A Theory of Congressional Delegation. In Congress: Structure and Policy, edited by Mathew D. McCubbins and Terry Sullivan 409-25. Cambridge: Cambridge University Press. McCubbins, Mathew D., and Thomas Schwartz. 1984. Congressional Oversight Overlooked: Police Patrols versus Fire Alarms. American Journal of Political Science 28 :167-79. Mayhew, David. 1974. The Electoral Connection. New Haven, Conn.: Yale University Press. Milton, George F. 1965. The Use of Presidential Power, 1789-1943. New York: Octagon Books. Muris, Timothy. 1989. The Uses and Abuses of Budget Baselines. Hoover Institution Working Papers in Political Science P-89-3. Stanford University. Neustadt, Richard. 1954. Presidency and Legislation: The Growth of Central Clearance. American Political Science Review 48641-7 1 . . 1980. Presidential Power. New York: Wiley. Oleszek, Walter J. 1989. Congressional Procedures and rhe Policy Process. 3d edition. Washington, D.C.: Congressional Quarterly Press.

111

U.S. Budget Deficits

Peterson, Paul E. and Mark Rom. 1989. Macroeconomic Policymaking: Who Is in Control? in Chubb and Peterson. Piiffner, James. 1979. The President, the Budget, and Congress. Boulder, Colo.: Westview Press. Rohde, David, and Kenneth A. Shepsle. 1973. Democratic Committee Assignments in the House of Representatives: Strategic Aspects of a Social Choice Process. American Political Science Review 67: 889-905. Roubini, Nouriel, and Jeffrey Sachs. 1989. Government Spending and Budget Deficits, Industrial Economies. Working Paper no. 2919. Cambridge, Mass.: National Bureau of Economic Research. Schick, Allen. 1980. Congress andMoney. Washington, D.C.: Urban Institute. Schlesinger, Arthur M., Jr. 1973. The Imperial Presidency. Boston: Houghton Mifflin. Shepsle, Kenneth A. 1988. The Changing Textbook Congress. In Chubb and Peterson 1988. Shepsle, Kenneth A., and Barry R. Weingast. 1984. Legislative Politics and Budget Outcomes, In Federal Budget Policy in the I980s,edited by Gregory B. Mills and John L. Palmer. Washington, D.C.: Urban Institute. Stanley, Harold W., and Richard C. Niemi. 1988. Vital Statistics on American Politics. Washington, D.C.: Congressional Quarterly Press. Stigler, George. 1971. The Theory of Economic Regulation. Bell Journal of Economics and Management Science (Spring): 3-21. Sundquist, James. 1981. The Decline and Resurgence of Congress. Washington, D.C.: Brookings Institution. U.S. Congress, House. 1921. Communicationfrom the President of the United States. Transmitting Laws Relating to the Estimates of Appropriations, the Appropriations, and Reports of Receipts and Expenditures. 67th Cong., 1st sess., H. Doc. 67-129. U.S. Department of Commerce, Bureau of Economic Analysis. Various years. Survey of Current Business. U.S. Department of Commerce, Census Bureau. 1975. Historical Statistics of the United States, Colonial Times to 1970, Bicentennial Edition, 2 vols. U.S. Executive Office of the President, Office of Management and Budget. Various years. Budget of the United States Government. U.S., Executive Office of the President, Office of Management and Budget. 1989. Historical Tables, Budget of the United States Government, Fiscal Year 1990. Wayne, Stephen J. 1978. The Legislative Presidency. New York: Harper & Row. Weingast, Barry R., and Mark Moran. 1983. Bureaucratic Discretion or Congressional Control? Regulatory Policymaking by the Federal Trade Commission. Journal of Political Economy 91 :675-700. Wilmerding, Lucius, Jr. 1943. The Spending Power. New Haven, Conn.: Yale University Press.

Comment

Robert J.

B

~

O

Over the last few years, many costs have been assigned to federal budget deficits. Until now, I have argued that these costs were small. I see now that Robert J. Barro is professor of economics at Harvard University and research associate of the National Bureau of Economic Research.

112

Mathew D. McCubbins

there are also offsetting benefits; after all, if Reagan had not run his deficits I would not have experienced the joy of discussing this paper. One thesis in McCubbins’s paper is that Reagan’s record in holding down federal expenditure was unimpressive in comparison with his presidential predecessors. The data presented to support this proposition are hard to understand because of the random switching between real and nominal magnitudes (“current-dollar spending nearly doubled from fiscal year 1981 to 1989”) and the failure to make historical comparisons on a consistent basis. (“Constantdollar spending grew at a faster rate during Reagan’s administration [25 percent from 1981 to 19881 than during the Eisenhower [16 percent growth], Nixon [ 13 percent], Ford [three percent], and Carter [ 12 percent] administrations .” Even after noticing that the figures represent accumulations over administrations of different length, I could not reproduce some of these figures.) Table 3.C1 shows the annual average growth rates of real federal expenditures during the presidential administrations beginning with Eisenhower’s. I use calendar-year figures that correspond as closely as possible to the actual dates in office (e.g., 1961-63 for Kennedy and 1969-74 for Nixon). These numbers (col. 1 of the table) show that Reagan’s growth rate, 3.1%, exceeded the growth rates for Eisenhower (1 .O%) and Nixon (2.6%), but were below those of the other presidents, including Carter (3.5%) and Ford (4.9%). The figures based on total federal spending are somewhat misleading because they credit Eisenhower and Nixon with the cutbacks in defense spending associated with the terminations of the wars in Korea and Vietnam, respectively. For nondefense expenditures (col . 2 of table 3.C 1) , Reagan’s growth rate of 2.6% was below that of the others. The growth rate for Eisenhower and Nixon are now 5.8% and 6.7%, respectively, and the secondlowest growth rate is 3.7% for Carter. My inference from these figures is that a substantial slowing of the growth rate of real nondefense federal spending has occurred, although this slowing began during the Carter administration. Two types of adjustment for inflation arise in analyses of budget deficits and public debt. The first is the division of nominal magnitudes by a price index or by nominal income or product. As mentioned before, this type of adjustment occurs only occasionally in McCubbins’s paper. The second adjustment involves the inflation rate, rather than the price level, and concerns the distinction between nominal and real interest rates. The natural definition of the government’s real budget deficit is the change over time in the government’s net real obligations. If the government holds no assets and if its liabilities are all in the form of nominal bonds, then the real deficit equals the change in the real public debt: 1. Real federal spending is the ratio of nominal spending to the GNP deflator. The results are similar with the deflator for federal purchases. The federal purchases deflator is problematic, however, because it is based mainly on cost measures, especially wage rates. This deflator is also inappropriatefor transfer payments.

113

U.S. Budget Deficits

Table 3.C1

Growth of Real Federal Spending during Presidential Administrations

President and Period in Office

Growth Rate of Real Federal Spending (1)

Growth Rate of Real, Nondefense Federal Spending (2)

.010

Eisenhower, 1953-60 Kennedy, 1961-63 Johnson, 1964-68 Nixon, 1969-74 Ford, 1975-76 Carter, 1977-80 Reagan, 1981-88

.053 ,062 ,026 ,049 .035 .031

,058 ,075 ,067 ,067 ,071 ,037 .026

Overall, 1953-88

,033

,053

Note: Real federal spending is the ratio of nominal spending (as defined in the national accounts) to the GNP deflator. The figures shown are average annual growth rates from the calendar year preceding each president’s first term to his final year in office.

(1)

d(BIP)/dt = (1/P) * (dB/dt) - IT*(B/P),

where B is the nominal quantity of government bonds, P is the price level, and IT = (UP) (dP/dt) is the inflation rate. The first term on the right side of equation (1) is the conventionally defined nominal deficit, dB/dt, divided by the price level. The second term, IT.(B/P), is the adjustment of the deficit to account for the erosion of the real value of the nominal debt due to inflation. The real budget deficit can also be expressed as the difference between real spending and real revenue. The result corresponds to equation (1) if the term IT.(B/P)is subtracted from the real value of the government’s nominal interest payments, R.(B/P), where R is the nominal interest rate. That is, real interest payments are computed by using the real interest rate, R - IT,rather than the nominal rate, R. From an ex ante standpoint, the adjustment involves the expected inflation rate, IT=,rather than the actual rate, IT.That is, if IT^ appears instead of IT in equation (l), then the relation determines the change in the real debt that would occur in the absence of unanticipated inflation. The adjustment of budget deficits for inflation is quantitatively important. For example, using numbers that correspond roughly to the present situation, if the expected inflation rate is 5% per year and the stock of debt is $2 trillion, then the adjustment, I T ~ * B reduces , the deficit by $100 billion. In the empirical work (table 3.A1), McCubbins uses the growth rate of the real public debt (his variable ADEBT) as one of the dependent variables. This concept corresponds to the real deficit as defined in equation (1) (except that he does not net out the portion of the debt held by government agencies and trust funds or the Federal Reserve). In another specification, he uses the conventional nominal budget deficit divided by a price index as a dependent vari-

114

Mathew D. McCubbins 0.25

0 . 1 6 ! , , , , , , , , , , , , , , , , , , , , , , ~ , ~ , , , . , , 1 , ~ , 1955 1960 1965 1970 1975 1980 1985

Fig. 3.C1 Federal spending as a ratio to GNP (CN: interest payments adjusted for expected inflation)

able (his variable DEFICIT). This measure does not adjust for inflation; in fact, McCubbins criticizes this type of adjustment. Apparently, he does not realize that his ADEBT variable makes the (reasonable) adjustment that he is criticizing. Figures 3.C1-3.C8 provide measures of the history of U.S. government spending, revenue, budget deficit, and public debt. Figure 3.C1 plots the ratio of total federal spending, as customarily defined, to GNP. The dashed line in the figure shows the ratio when federal spending is revised downward to eliminate the portion of nominal interest payments that corresponds to expected inflation, .rr'*(BIP). I use the Livingston survey of expected future values of the consumer price index (provided by the Federal Reserve Bank of Philadelphia) to compute 7~e. During the Reagan administration, the (adjusted) spending-GNP ratio rises from .208 in 1980 to .234 in 1982-83 and then falls to .213 in 1988. The increase in the ratio from 1980 to 1982-83 corresponds mainly to increases in defense spending (fig. 3.C2) and interest payments (fig. 3.C3), as well as to the decline in real GNP because of the 1982-83 recession. After 1983, defense spending and adjusted interest payments are roughly constant as ratios to GNP; therefore, the decline in the overall spending ratio involves cuts in other components of spending in relation to GNP. One thing to notice in figure 3.C2 is that the trough in the defense ratio occurs in 1978-79. The rise in the ratio starts with Carter, not Reagan. Figure 3.C4 plots the ratio of federal revenue to GNP along with the ad-

115

U.S. Budget Deficits 0.14

0.044,,, , , , 1955 1960 ,

I

I

I

1

I

1965

,

I

I

i

I

1970

I

I

1

1

1

1975

I

I

1

1

4

1980

1

8

1

I

8

1

1

1985

Fig. 3.C2 Defense spending as a ratio to GNP

0.035 0.030 0.025 0.020

0.015 0.010

0.005 0.000

Fig. 3.C3 Federal net interest payments as a ratio to GNP (FEDINTN:adjusted for expected inflation)

Mathew D. McCubbins

116 0.24

0 . 1 6 1 , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,, ,, 1955 1960 1965 1970 1975 1980 1985

Fig. 3.C4 Federal spending and revenues as ratios to GNP

justed spending ratio. The gap between these two series equals the ratio of the budget deficit (adjusted for expected inflation) to GNP; this ratio is plotted as the dashed line in figure 3.C5. (The solid line shows the ratio for the conventional budget deficit.) The onset of a substantial deficit in 1982 corresponds to an increase in the spending ratio and a decrease in the revenue ratio. This behavior is customary during a recession; for example, figure 3.C4 and 3.C5 reveal similar patterns for 1958 and 1975. From the perspective of the longterm U.S. history, the data for 1982-83 accord with the usual relation between budget deficits and economic contraction. Starting in 1984, the budget deficit gets out of line with the behavior of the economy. The spending ratio does not decline much until after 1986 and the revenue ratio rises slower than usual during an economic recovery (because of the Reagan tax cuts). The spending and revenue ratios do converge gradually over time, and the ratio of the adjusted deficit to GNP falls from a peak of 4.0% in 1983 to 1.4% in 1988 and 0.9% in 1989. Figure 3.C6 shows the ratio of privately held U.S. public debt to GNP. (The debt figures are at the end of the calendar year since 1916 and at midyear before 1916.) This measure of the debt nets out the amounts held by federal agencies and trust funds and the Federal Reserve. Note that the debt ratio rose from 21.5% in 1979 to 38.6% in 1987. Then, with the cutback in the deficitGNP ratio, the debt ratio fell to 38.0% in 1989. In particular, the reduced deficits meant that the debt-GNP ratio was no longer rising after 1987. Figures 3.C7 and 3.C8 provide information for the total government sector. The overall story is similar to that for the federal government. Because of the

117

U.S. Budget Deficits

o.02h 0.06

0.05 0.04 -

0.03 0.01

0.00 -0.01

1955

1960

1965

1970

1975

1980

1985

Fig. 3.C5 Federal budget deficit as a ratio to GNP (DEFR: adjusted for expected inflation)

1.25

1.oo

0.75

WWll

-->

0.50

0.25

0. Since the St Germain amendment was supported by the administration, Republicans would be more likely to be in favor of it, so we expect P, 0. The committee had, of course, reported the $5 billion package, but this amount had been approved in committee by a one-vote margin. The committee chairman was now sponsoring an amendment to raise the recapitalization level-a switch from the way he voted in committee. This does not imply any clear prediction about P3,but if committee members tend to support the chairman’s position on the floor (as part of an implicit bargain in a continuing relationship), then we should see p, > 0.

198

Thomas Romer and Barry R. Weingast

The predicted sign of p, is negative if WEAKDIS is primarily a measure of constituent pressure from insolvent thrifts. But, as we noted in our earlier discussion, having a larger number of insolvent thrifts in one’s district also heightens depositors’ concerns in the district. This may lead to increasing willingness to vote for a larger recapitali~ation.~’ The net effect on p, is unclear. As to p4, it would be natural to suppose that higher contributions from the thrift PACs would be associated with a lower probability to vote for the amendment. But studies of the relationship of campaign contributions to voting on individual roll calls have shown no clear indication of such direct association. Table 6.4 presents our probit estimates. As expected, Republicans were more likely to vote for the amendment than Democrats-and liberal Democrats (there are no Republicans with high IDEOL values) were more likely to vote yea than their more conservative brethren. Committee members were more likely to favor the amendment, ceteris paribus, than nonmembers (indeed, committee members voted 30 to 18 in favor3*).Money from PACs does not appear to have had a significant, independent association with voting yea. 39 We see that the coefficient on STRONGDIS is negative and quite precisely estimated in all the specifications. Both the estimated coefficient and its standard error are insensitive to which of the political variables are included in the specification. We can say with some confidence, therefore, that the probability of voting against raising the FSLIC borrowing limit was significantly Table 6.4

Voting on the St Germain Amendment to HR 27, Robit Estimates ( N = 411) 2

3

.287 (.199) - 1.225 (.303) ,819 ( ,228) ,728 (.227) ,037

,014 (.182) - ,250 (.130)

-.I30 (. 168)

- ,072 (.029)

- ,078 (.029)

1

Constant PARTY IDEOL COM

PACM 0N EY STRONGDIS WEAKDIS

In likelihood

(.@w

,675 (.224)

,042

,130

,060

(.071)

(.068)

- 251.45

- 258.04

- .011 ( ,099) ,679 (.223) ,038 (. 040) - ,077 ( .029) .060 ( ,069) - 259.87

Note: Estimated asymptotic standard errors are in parentheses.

4 .081

(.142) - 1.164 (.291) ,778 (.217) .756 (. 2 w

- 255.92

5 - ,134 (.I@)

.680 (.222) ,038 (.040) - .077 (.029) .061 (.068)

- 259.88

199

Political Foundations of the Thrift Debacle

Table 6.5

Democrat at median Democrat IDEOL Republican at median Republican IDEOL

Probability of Yea Vote on St Gerrnain Amendment to HR 27 Means (1)

Low

High

Low

High

STRONGDIS

STRONGDIS

WEAKDIS

WEAKDIS

(2)

(3)

(4)

(5)

.305

,414

,192

,258

,397

,394

.510

,265

.341

.492

Note: Computations are based on estimates in column (1) of table 6.4. (a) For all computations, COM = 0. ( b ) Median IDEOL for Democrats = 0.745. (c) Median IDEOL for Republicans = -0.456. ( d ) Maximum value of IDEOL = 1.34;minimum value of IDEOL = - 1.41. (e) Col. 1: Economic variables at their means: PACMONEY = 1.112, STRONGDIS = 5.032, WEAKDIS = 1.087. Col. 2: STRONGDIS = 1 and PACMONEY and WEAKDIS at their means. col. 3: STRONGDIS = 10 and PACMONEY and WEAKDIS at their means. Col. 4: WEAKDIS = 0 and PACMONEY and STRONGDIS at their means. Col. 5: WEAKDIS = 3 and PACMONEY and STRONGDIS at their means.

higher, ceteris paribus, for members from states where the average number of healthy thrifts in a district was higher. The effect of the presence of sick thrifts is somewhat ambiguous; though the estimate of p, is positive, it is much less precisely estimated than p,. To get a sense of the importance of the thrift constituency variables, we used the specification in column 1 of table 6.4to compute the probability of voting yea on the amendment for two types of hypothetical congressmen not on the House Banking committee.40One is a Democrat whose IDEOL has the median value for all Democrats (0.745);the other is a Republican at the median of Republican IDEOL (-0.456).For each of these congressmen, table 6.5shows P,the estimated probability of voting yea, as the value of WEAKDIS or STRONGDIS varies. Not surprisingly, P is generally below 0.5, reflecting the fact that the amendment lost by a 2:l majority among noncommittee members. As noted in column (1) of table 6.5,for our hypothetical Democrat, P = 0.305 when PACMONEY, STRONGDIS, and WEAKDIS are at their sample mean values (for the Republican, P = 0.394).Moving from the mean value of STRONGDIS (= 5) to its lowest value in the sample ( = l), increases the probability of a yea vote by about 33%, to P = 0.414.On the other hand, a change from column (1) to a STRONGDIS = 10 (a change of about two standard deviations from the mean) reduces P by about one-third, to P = 0.192.Changes for the Republican are similar, though less pronounced. The table displays similar calculations for changes in WEAKDIS, which go in the opposite direction. To summarize, the condition of thrifts in the congressman’s state was clearly related to his legislative preferences in voting on the key amendment in the recapitalization debate. Because the number of healthy thrifts exceeded that of weak thrifts in most states (by more than a factor of two in many cases), the effect of STRONGDIS predominated for most congressmen. The net

200

Thomas Romer and Barry R. Weingast

effect, by our estimates, was to reduce the probability of a yea vote in nearly all districts. There were no comparable roll calls in the Senate. Presuming that similar forces operated in that chamber, our findings are consistent with Proxmire’s willingness effectively to veto legislation in 1986, and with his intransigence throughout the FSLIC debate. In a state with only healthy thrifts, any constituency pressure he faced on the issue would be in the direction of delaying recapitalization or reducing its level. With the defeat of the St Germain amendment, HR 27 was adopted by the House, 402 to 6. There were no other controversial amendments to the bill reported by the committee. The Senate had also approved its committee’s bill (S 790) with only minor changes. This lack of controversy on the floor provides additional evidence for our thesis that there was not a significant constituency for confronting the full magnitude of the thrift problem or to curtailrather than to extend-forbearance. Even a small group of dissenters can require roll call votes. In this way, the dissenters can force their opponents to go on record against a position, while the dissenters signify their concern on the issue. The virtual absence of roll call votes on the thrift issue in 1986-87 is like Sherlock Holmes’s dog that did not bark in the night. It is mute testimony to the fact that members of Congress did not believe there was an audience to whom it was worth sending stronger signals about the thrift problem.41 6.4.4 Compromise in Conference: The Competitive Equality Banking Act As the House and Senate conference committees prepared to meet to resolve differences between the two bills, administration spokesmen indicated that the president was seriously considering a veto of any bill that did not raise the recapitalization limit or-more important-relax the restraints on nonbank banks. As to raising the FSLIC limit, Proxmire “told an industry group . . . that the Treasury Department and the bank board are exaggerating the urgency of the FSLIC’s problem and that he would not be pressed into quick action” (New York Times, 5 May 1987). By the end of July, as conferees were meeting, the General Accounting Office reported that FSLIC was $6 billion in the red and was confronting future claims up to $50 billion. In an agreement with the administration, leaders of the conference committee agreed to raise the FSLIC borrowing limit to $10.8 billion, with not more than $3.75 billion to be raised in any one year. The forbearance provisions were left intact. In early August, both chambers passed the conference report (the House by 382 to 12 and the Senate by 96 to 2).42The president signed the Competitive Equality Banking Act on 10 August 1987. The final legislation reinforced continued forbearance by allowing thrifts in farm and oil-patch states to continue to use the lenient regulatory accounting practices adopted in 1982. Thrifts in these areas (and other areas deemed economically depressed) would also be allowed to stay open with a .5% capital-

201

Political Foundations of the Thrift Debacle

asset ratio, instead of the 3% required by existing law. The act also explicitly reaffirmed the commitment that the “full faith and credit” of the U.S. government stands behind FSLIC. Given the limited amount of new FSLIC funding relative to the magnitude of the thrift insolvency problem, the act ensured that many failing thrifts would have considerable more time to gamble for resurrection. The crisis would grow.

6.5 The Thrift Debacle and Beyond The policy of forbearance did not, of course, lead to the resurrection of the zombie thrifts. Instead, as many economists had warned, the losses incurred by thrifts gambling for resurrection continued to escalate. The net income of the thrift industry for 1988 was - $12.0 billion, down from - $7.8 billion in 1987. Barth, Bartholomew, and Bradley (1989) report that the estimated cost of the 205 thrift resolutions begun in 1988 is $31.8 billion. Over half of these thrifts had been GAAP-insolvent for more than three years before they were closed or merged. At year-end 1988, there were 364 thrifts that were insolvent but still open. Many of these thrifts had been insolvent for a long time-some for as long as 10 years (Barth, Bartholomew, and Labich 1990). Congress did not return to the thrift problem in 1988. By the time the woes of FSLIC reappeared on the legislative agenda after the 1988 elections (in which they played little or no role), the costs of delay and forbearance were increasingly evident to nearly everyone. The 1989 “bailout” (Financial Institutions Reform, Recovery, and Enforcement Act, or FIRREA) was enacted in the context of predictions that the liabilities facing FSLIC were expected to exceed $200 billion.43The FIRREA and its legislative history are beyond the scope of this paper. But its enactment was necessitated by the legislative failures of the previous five years. While it is now widely recognized that a policy of forbearance under the existing structure of deposit insurance has been a prime contributor to the escalation of the thrift debacle, the political foundations for this behavior are not as widely appreciated. This paper has pointed to these foundations as being rooted in the logic of the connections between elected officials and regulatory policy. By this view, regulatory agencies are viewed not as autonomous decision makers, but as actors closely tied to the political system. Constituency pressures work on regulators through the (often implicit) connections with politicians. The legislative response to the problems of the thrift industry in the 1980s is an excellent example of how this process works. In summary, we note the following: 1. The political process provides at least tacit support for regulatory policy when the policy is consistent with the preferences of active constituent groups. This characterized policy toward thrifts until the late 1970s. 2 . Signijicant changes in the economic environment typically lead to con-

202

Thomas Romer and Barry R. Weingast

stituency pressures for regulatory change. This change may be initiated by the regulatory agency, but to be sustainable, it must meet congressional and presidential approval. The 1982 legislative response to the erosion of thrift profitability-deregulation coupled with forbearance for sick thrifts-can be seen in this light. 3 . Regulatory change that is resisted by active constituencies and not supported by other active constituencies rarely succeeds. In 1985 the FHLBB appeared to be deviating from its earlier course. The 1986 stalemate in refinancing FSLIC and the 1987 legislation worked-through delay, direct intervention, and explicit limitation on FSLIC resources-to rein in the FHLBB. The president, because he can command national attention, is the primary political actor with the power to counter narrow constituency pressure. In the case of the thrifts, the president opted for a stance that did not oppose the congressional tendency toward forbearance. 4. The emergence of constituencies for whom an issue becomes more salient can alter legislative policy preferences. As the scale of insolvencies grew, depositor concern over FSLIC solvency also mounted. By late 1988, the policy preferences of the insolvent thrifts (in favor of continued forbearance) were more strongly opposed by concerns about the viability of deposit insurance guarantees. As our empirical results on the St Germain amendment suggest, this would lead to an increase in support for larger levels of FSLIC capitalization. Moreover, once the required funding exceeded an amount that could be covered mostly by assessments on the industry (around $15420 billion over five years), healthy thrifts would no longer oppose additional amounts. The financing costs of such increments would come from general revenues; deposit guarantees financed in this way would benefit healthy thrifts.

6.5.1

Information and Policy Choice

It is sobering to note that the political behavior surrounding the thrift debacle is absolutely ordinary. Through 1988, congressmen behaved as they do in ordinary circumstances: paying solicitous attention to active, wellorganized interests, provided that the readily apparent costs to their other constituents are not noticeably high. It is an intriguing question whether congressmen would have acted differently had they “really known” in 1986 that delay, forbearance, and intervention would lead to a $400 billion (or more) debacle. If by “really known,” we mean that there was a widely shared consensus held by broad constituencies, then the answer to the question would probably be yes. But this was not the situation. It is no doubt correct to say that any economist who thought seriously about the situation would have seen that the combination of forbearance and deposit insurance was a recipe for disaster. Some economists actually said so, within the regulatory agency and outside it. But the more widespread belief, reinforced by constituency pressures, was that-as in 1982-84-the problem would abate rather than explode. To a

203

Political Foundations of the Thrift Debacle

large extent, of course, this perception was created by the very process of the forbearance that politically endorsed policies made possible. Accounting procedures and regulatory reporting effectively minimized the magnitude of the losses. This put most congressmen in the position of “not knowing” the consequences of forbearance. The lack of an active constituency against forbearance (together with knowledge that agencies often cry “wolf”) allowed congressmen to act in standard ways; namely, to intervene in the regulatory process on behalf of their constituents. Thus, even in the face of the GAO predictions that FSLIC was facing a $50 billion problem in 1987, nothing near this level of recapitalization was ever on the political agenda in that year. Nor was there any consideration of restructuring the guarantees under deposit insurance.

6.5.2 Beyond the Thrift Debacle Given the nature of incentives facing politicians, it is extremely difficult for Congress to deal at a sufficiently early state with emerging policy problems that may turn into catastrophes if left unattended. Unless they can claim credit for actions to stem the crisis today, congressmen face great difficultiesin bucking the current set of interest group forces. Perversely, even when congressmen know that policy crises are emerging, they may have to wait for the crisis to occur before they can take and be rewarded for remedial actions. Part of the foundation for this tendency is that individual legislators are not seen by their voters as playing a role in the problem. They are not penalized for letting the problem grow. These considerations are particularly relevant because many aspects of the thrift debacle appear now to be replicated in several other financial problems that loom on the congressional horizon-the solvency of commercial banks, farm loan guarantee programs, and various government-guaranteed pension systems.” If the wrong lessons are drawn from the thnft debacle, similar crises may reappear in new settings. It is tempting to explain the thrift debacle as a story about greed, fraud, and criminal behavior. Surely these are part of the story. It is also tempting to argue that this was a case of regulators in bed with the industry they were supposed to oversee. This certainly also played a role. But underlying these factors is the essential component: Congress sanctioned regulatory forbearance and actively intervened when regulators sought a new, more restrictive path. In order to argue that the next potential crisis would unfold differently, one would need to demonstrate that structure or incentives have changed. It is possible that, in commercial banking, for example, healthy firms will behave differently than such firms did in the thrift industry. They may perceive that early action to deal effectively with insolvencies is in the long-run interest of the rest of the industry. But it is also possible that, in each case, the regulated interests will remain major constituents of politicians and will argue against forceful r e g ~ l a t o r sWill . ~ ~ politicians be more wary? The answer to this de-

204

Thomas Romer and Barry R. Weingast

pends on the degree to which currently diffuse interests-generally, taxpayers as a group-see their representatives as having played a major role in the crisis. By moving the focus to shifty operators in the industry or by blaming incompetent or biased regulators, the political source of the crisis will be obscured.

Notes 1. We will sacrifice some precision and use the terms “savings and loan” and “thrift” interchangeably to mean federally insured thrift institutions (which include both S&Ls and some mutual savings banks). 2. The Wall Street Journal reported on 6 April 1990 that Congressional Budget Office and General Accounting Office projections of spending through the 1990s will be between $300 billion and $350 billion, “before factoring in increased net losses from an unexpectedly greater number of [thrift] insolvencies” (emphasis in original). The article quotes estimates by close observers of the industry that the likely 10-year cost will exceed $400 billion. 3. By 1981, market-value net worth of federally insured thrifts had fallen to - 17.3% of total assets (Brumbaugh 1988, 50, table 2-7). 4. In 1982 nearly 10 percent of FSLIC-insured thrifts were insolvent by GAAP standards, more than in any prior year (Brumbaugh 1988, 37, fig. 2-1). 5 . Prior to the Garn-St Germain Act, the Federal Home Loan Bank Board (FHLBB) had eased some accounting rules to help ailing thrifts, by lengthening the list of items that could be excluded as liabilities and those included as assets, in computing net worth. Under Garn-St Germain, “well-managed” thrifts with net worth between 0.5% and 3% of assets could issue “net worth certificates” that could be exchanged for FSLIC promissory notes and counted as assets. In this way a thrift institution could convert what was, in effect, a liability into an asset. Moreover, once FSLIC agreed to buy such certificates from an institution, it was committed to continue buying them as long as the institution was deemed to be “well-managed.’’ 6. A typical report, carried in the 13 August 1983 issue of National Journal, proclaimed that Garn-St Germain had “rescued” the thrift industry and that thrift executives were optimistic about their new options. 7. Kane (1989b) provides a good overview of the first and third of these hypotheses. Explanations that lean heavily on fraud and skulduggery are presented in a spate of “inside story” books: e.g., Adams (1989), Pizzo, Fricker, and Muolo (1989), Pilzer and Dietz (1989). Movie versions cannot be far behind. 8. Our discussion is based on a large and growing literature on this topic: Fiorina (1981), Fiorina and No11 (1978), Ferejohn and Shipan (1989), McCubbins and Schwartz (1984), Kiewiet and McCubbins (1991), McCubbins, Noll, and Weingast (1989), Moe (1985a, 1989), Weingast (1984), and Weingast and Moran (1983). For a survey of some of this work, see Romer and Rosenthal(l987). 9. On recognition, see Jacobson (1987). 10. This line of work has provided insights into equilibrium and comparative statics, applied in a variety of settings. See, e.g., Romer and Rosenthal(1978), Weingast and Moran (1983), Ferejohn (1986), Ferejohn and Shipan (1989), and Kiewiet and McCubbins (1991). 11. This is emphasized in Gilligan, Marshall, and Weingast (1989).

205

Political Foundations of the Thrift Debacle

12. On the role of the House leadership in the 1980s, see Sinclair (1989). 13. This argument is based on McCubbins and Schwartz (1984) and Weingast (1984). On oversight, see also Aberbach (1990) and Ogul and Rockman (1990). 14. See Kiewiet and McCubbins’s (1991) discussion of delegation and Ferejohn and Shipan (1989) on the legislative threats of intervention in agency decisions. 15. See Weingast (1984) for a review of the literature and evidence for this claim. 16. Other examples abound. See esp. Ferejohn and Shipan’s (1989) study of Federal Communications Commission policies after the AT&T divestiture and Weingast and Moran’s (1983) study of the rise and fall of consumer activism by the Federal Trade Commission. 17. Thrifts had objected to the federal mortgage insurance provisions of the Housing Act, fearing that they would increase the ability of commercial banks and insurance companies to compete in the mortgage market. On this, and for other details of the history of the thrift industry, see Woerheide (1984). 18. This arrangement differed from the case of the FDIC, which was made an agency independent of the Federal Reserve Board. 19. The economic argument for subsidies to mortgage lenders as a way to encourage demand for housing is a weak one and had been frequently challenged-by economists, if not by developers and thrift institutions; see e.g., Meltzer (1981) and Weicher (1988). 20. Robert Litan’s comments (in this volume) provide some details on regulatory developments before 1982. 21. Since the early days of deposit insurance, economists had pointed to the perverse incentives for bank risk taking that it created. Barth and Bradley (1988) quote observers from 1931 and 1936 on this point. By the early 1980s, some FHLBB economists were warning about the riskiness of allowing zombie thrifts to operate, and those outside the industry were calling attention to this problem as well; see Kane (1985, 1989b). 22. In the 1986 Economic Report of the President, for example, the president’s report itself (as distinct from that of the CEA) carried the message that continued deregulation of financial institutions should proceed apace. The thrifts were not mentioned. 23. This quote is taken from an Associated Press dispatch filed by Martin Crutsinger, 26 July 1985; see the analysis in Barth et al. (1985a, 1985b) and Brumbaugh and Hemmel(l984). 24. Figs. 6.1 and 6.2 are based on year-end data, using GAAP definitions. 25. Their position, as described by banking consultant Bert Ely, was, “After all, Ford and GM didn’t bail out Chrysler. Boeing didn’t bail out Lockheed (Wull Street Journal, 22 July 1987). 26. Given the economic status of the thrifts, even a $15 billion recapitalization would have required assessments of over $1 billion per year and would have cut severely into the income of the segment of the industry that was healthy in 1986. By 1987, tangible net worth (GAAP net worth minus certain items such as goodwill) of the industry had shrunk to $9 billion. Thrifts with GAAP capital-to-asset ratio above 3% had tangible net worth of $34 billion. Tangible net worth of the rest of the industry was negative $25 billion (Barth et al. 1990, table 1). 27. Congressional deadlocks of this sort can sometimes be broken by presidential intervention. This did not occur, as the administration was willing to acquiesce in the existing policy of forbearance. 28. Kane (1989b) provides a good discussion of the extent to which accounting numbers systematically provided rosy pictures of the thrift situation. 29. The lobbying and campaign financing activities of thrift industry groups and individual thrifts on this issue have been well documented; see, e.g., Jackson (1988).

206

Thomas Romer and Barry R. Weingast

30. Although we refer to FSLIC borrowing, under all of the plans discussed here the actual borrowing was to be done by a newly created financing corporation (FICO). The principal on the borrowed amounts would be secured by zero-coupon bonds issued by the Treasury. Interest on FICO borrowing would be paid by FSLIC-insured thrifts. 31. The day before Wright’s announcement of his support for the $15 billion limit, the FHLBB “filed a $350 million lawsuit against seven former officers of the Vernon Savings and Loan Association of Dallas charging them with looting the organization of hundreds of millions of dollars. Vernon was one of the thrift institutions for which Mr. Wright had sought leniency” (New York Times, 29 April 1987). 32. The St Germain amendment is CQ Roll Call 83. Democrats voted 81 yea and 160 nay; Republicans voted 72 yea and 98 nay. 33. The categories are: (GAAP net worth/GAAP assets) less than 0; between 0 and 1.5%,between 1.5%and 3%;between 3% and 6%;and over 6%. 34. Using year-end 1987 data instead did not af€ect our qualitative results, nor did defining WEAK as thrifts with GAAP net worth less than 1.5% of GAAP assets or STRONG as those with GAAP net worth above 6%. 35. The PACs are those affiliated with the U.S. Savings League and with the National Council of Savings Institutions (Thriftpac). These were the only two thrift PACs among the 500 largest (in terms of campaign contributions) PACs in 1985-86. The totals do not include “soft money” or honoraria received by congressmen from thrift industry groups. 36. We thank Keith Poole for the data. The IDEOL variable for each congressman equals the coordinate of the first dimension estimated from running two-dimensional NOMINATE on the 100th House. The IDEOL variable correlates highly (over .9) with the more familiar ADA score, but is based on a much wider set of roll calls. For details, see Poole and Rosenthal(l991). 37. This could be more directly tested with a variable that measured the number of depositors with accounts at insolvent thrifts in the district. We did not have such a variable at our disposal. 38. That this may have involved strategic voting by at least some committee members has not escaped our attention. All members who, in committee, had voted to keep the borrowing limit at $15 billion, voted for the St Germain amendment on the floor. Five members (including St Germain) who had voted to reduce the limit from $15 billion to $5 billion in committee, switched their position and voted for the St Germain amendment on the floor. 39. It should be recalled that PACMONEY is an imprecise and almost certainly understated measure of actual contributions (see n. 35 above). It is not clear how the “true” measure and our measure would be correlated. 40. Column 1 of table 6.4 is the specification that follows directly from our discussion. Likelihood-ratio tests of comparisons with the results of the other columns also argue in favor of the col. 1 specification. 41. In 1986 there was one roll call vote in the House dealing with thrifts; there were none in the Senate. In 1987, aside from final passage, there were two roll call votes on HR 27 in the House and two on S 790 in the Senate. 42. At first, the U.S. Savings League announced that it would oppose the compromise, saying that the borrowing limit was excessive. Given the GAO report, however, it was unlikely that many congressmen would have been willing to reopen the issue for a major floor fight. Final passage of the bill in the House did require a parliamentary manoeuvre. “Because the conference report broke new ground, beyond the scope of either the House or Senate’s original bills, it was vulnerable to a point of order on the House floor. St Germain appealed to the Rules Committee for a waiver of this and other procedural points that might have deterred final passage. In the end, the Rules Committee granted the waivers on a voice vote” (Congressional Quarterly 1988,636).

207

Political Foundations of the Thrift Debacle

43. Some observers have noted that FIRREA does not address many of the fundamental problems facing the industry (see, e.g., Barth and Brumbaugh 1990; Scott 1989b). 44. Commercial banks are the most directly related concern. See Starobin (1989) and Brumbaugh, Carron, and Litan (1989). 45. As of this writing, congressmen from New England have called bank regulators to task for being too vigorous in dealing with commercial banks with weak balance sheets (Wall Street Journal, 12 April 1990). While gratifyingly consistent with our thesis, such a development hardly bodes well for taxpayers in the 1990s.

References Adams, James Ring. 1989. The Big Fir: Inside the S&L Scandal. New York: Wiley. Aberbach, Joel D. 1990. Keeping a Watchful Eye: The Politics of Congressional Oversight. Washington, D.C.: Brookings. Barth, James R., Philip F. Bartholomew, and Michael G.Bradley. 1989. The Determinants of Thrift Institution Resolution Costs. Research Paper no. 89-03. Washington, D.C.: Office of Thrift Supervision, Department of the Treasury. Barth, James R., Philip F. Bartholomew, and Carol J. Labich. 1990. Moral Hazard and the Thrift Crisis: An Empirical Analysis. Consumer Finance Law Quarterly Report 44(1):22-34. Barth, James R., and Michael G.Bradley. 1988. Thrift Deregulation and Federal Deposit Insurance. Research Paper no. 150. Washington, D.C.: Office of Policy and Economic Research, Federal Home Loan Bank Board. Barth, James R., and R. Dan Brumbaugh. 1990. The Rough Road from FIRREA to Deposit Insurance Reform. Stanford Law & Policy Review 2 (Spring), 58-67. Barth, James R., R. Dan Brumbaugh, Daniel Sauerhaft, and George H. K. Wang. 1985a. Insolvency and Risk-Taking in the Thrift Industry: Implications for the Future. Contemporary Policy Issues (Fall): 1-32. . 1985b. Thrift Institution Failures: Causes and Policy Issues. In Proceedings of a Conference on Bank Structure and Competition, 184-216. Chicago: Federal Reserve Bank. Brumbaugh, R. Dan. 1988. Thrzfts under Siege: Restoring Order to American Banking. Cambridge, Mass.: Ballinger. Brumbaugh, R. Dan, and Andrew S. Carron. 1987. The Thrift Industry Crisis: Causes and Solutions. Brookings Papers on Economic Activity, no. 2 , 349-88. Brumbaugh, R. Dan, Andrew S. Carron, and Robert E. Litan. 1989. Cleaning Up the Depository Institutions Mess. Brookings Papers on Economic Activity, no. 1, 24395. Brumbaugh, R. Dan, and Eric I. Hemmel. 1984. Federal Deposit Insurance as a Call Option: Implications for Depository Institutions. Washington, D.C.: Office of Policy and Economic Research, Federal Home Loan Bank Board (October). Congressional Quarterly. 1988. 1987 CQ Almanac. Washington: CQ Press. Denzau, Arthur T., and Michael C. Munger. 1986. Legislators and Interest Groups: How Unorganized Interests Get Represented. American Political Science Review 80:89-106. Fenno, Richard. 1978. Home Style: House Members in Their Districts. Boston: Little, Brown. Ferejohn, John A. 1986. Logrolling in an Institutional Context: A Case Study of Food

208

Thomas Romer and Barry R. Weingast

Stamp Legislation. In Congress and Policy Change, edited by G. C. Wright, L. N. Rieselbach, and L. C. Dodd, 223-53. New York: Agathon Press. Ferejohn, John A., and Charles R. Shipan. 1989. Congressional Influence on Administrative Agencies: A Case Study of Telecommunications Policy. In Congress Reconsidered, 4th ed. Edited by Lawrence C. Dodd and Bruce I. Oppenheimer. Washington, D .C .: Congressional Quarterly Press. Fiorina, Moms P. 1981. Congressional Control of the Bureaucracy: A Mismatch of Incentives and Capabilities. In Congress Reconsidered, 2d ed. Edited by Lawrence C. Dodd and Bruce I. Oppenheimer. Washington, D.C.: Congressional Quarterly Press. . 1989. Congress: Keystone of the Washington Establishment, 2d ed. New Haven, Conn.: Yale University Press. Fiorina, Morris P., and Roger G. Noll. 1978. Voters, Bureaucrats and Legislators: A Rational Choice Perspective on the Growth of Bureaucracy. Journal of Public Economics 9:239-54. Gilligan, Thomas W., William J. Marshall, and Barry R. Weingast. 1989. Regulation and the Theory of Legislative Choice: The Interstate Commerce Act of 1887. Journal of Law and Economics 3235-62. Jackson, Brooks. 1988. Honest Grqfr. New York: Knopf. Jacobson, Gary C. 1987. Running Scared: Elections and Congressional Politics in the 1980s. In Congress: Structure and Policy, edited by Mathew D. McCubbins and Terry Sullivan. Cambridge: Cambridge University Press. Kane, Edward J. 1985. The Gathering Crisis in Federal Deposit Insurance. Cambridge, Mass: MIT Press. . 1989a. The High Cost of Incompletely Funding the FSLIC’s Shortage of Explicit Capital. Journal of Economic Perspectives 3:3 1-47. . 1989b. The S&L Insurance Mess: How Did It Happen? Washington, D.C.: Urban Institute Press. . 1989c. The Unending Deposit Insurance Mess. Science 27:451-56. Kiewiet, D. Roderick, and Mathew D. McCubbins. 1991. The Logic of Delegation: Congressional Parties and the Appropriations Process. Chicago: University of Chicago Press. McCubbins, Mathew D., Roger G. Noll, and Barry R. Weingast. 1989. Structure and Process, Politics and Policy: Administrative Arrangements and the Political Control of Agencies. Virginia Law Review 7 5 4 3 1-82. McCubbins, Mathew D., and Thomas Schwartz. 1984. Congressional Oversight Overlooked: Police Patrols vs. Fire Alarms. American Journal of Political Science 28: 165-79. Mayhew, David. 1974. Congress: The Electoral Connection. New Haven, Conn.: Yale University Press. Meltzer, Allan. 1981. The Thrift Industry in the Reagan Era. In Managing Interest Rate Risk in the Thrift Industry: Proceedings of the Seventh Annual Conference, 513. San Francisco: Federal Home Loan Banks. Moe, Terry M. 1985a. Control and Feedback in Economic Regulation: The Case of the NLRB. American Political Science Review 79:1094-1116. . 1985b. The Politicized Presidency. In New Directions in American Politics, edited by John E. Chubb and Paul E. Peterson. Washington, D.C.: Brookings. . 1989. The Politics of Bureaucratic Structure. In Can Government Govern?, edited by John E. Chubb and Paul E. Peterson. Washington, D.C.: Brookings. Ogul, Moms S., and Bert A. Rockman. 1990. Overseeing Oversight: New Departures and Old Problems. Legislative Studies Quarterly 155-24. Pilzer, Paul Zane, with Robert Dietz. 1989. Other People’s Money: The Inside Story of the S&L Mess. New York: Simon & Schuster.

209

Political Foundations of the Thrift Debacle

Pizzo, Stephen P., Mary Fricker, and Paul Muolo. 1989. Inside Job: The Looting of America’s Savings and Loans. New York: McGraw-Hill. Poole, Keith T., and Howard Rosenthal. 1991. Patterns of Congressional Voting. American Journal of Political Science 35 (February): 228-78. Romer, Thomas, and Howard Rosenthal. 1978. Political Resource Allocation, Controlled Agendas, and the Status Quo. Public Choice 33(4):27-43. . 1987. Modem Political Economy and the Study of Regulation. In Public Regulation: New Perspectives on Institutions and Policies, edited by E. E. Bailey, 73116. Cambridge, Mass.: MIT Press. Scott, Kenneth E. 1989a. Deposit Insurance and Bank Regulation. Business Lawyer 44:907. . 1989b. Never Again: The S&L Bailout Bill. Stanford University School of Law. Typescript. Sinclair, Barbara. 1989. House Majority Leadership in the Late 1980s. In Congress Reconsidered, 4th ed. Edited by Lawrence C. Dodd and Bruce I. Oppenheimer. Washington, D.C.: Congressional Quarterly Press. Starobin, Paul. 1989. Will the Banks Be Next? National Journal. December 30. Weicher, John C. 1988. The Future of the Housing Finance System. In Restructuring Banking & Financial Services in America. Edited by William S. Haraf and Rose Marie Kushmeider, 296-342. Washington, D.C.: American Enterprise Institute. Weingast, Barry R. 1984. The Congressional-Bureaucratic System: A Principal-Agent Perspective (with Applications to the SEC). Public Choice 44:147-91. Weingast, Barry R., and Mark J. Moran. 1983. Bureaucratic Discretion or Congressional Control? Regulatory Policymaking by the Federal Trade Commission. Journal of Political Economy 91:765-800. Woerheide, Walter J. 1984. The Savings andLoan Industry. Westport, Conn.: Quorum Books

Comment

Robert E. Litan

Thomas Romer and Barry R. Weingast join the growing industry of economists, journalists, and now political scientists who have been attempting to explain the worst financial crisis since the 1930s: the thrift disaster of the 1980s. The entry of political scientists into this fray is welcome. For Romer and Weingast are correct when they argue that the fundamental causes of the thrift mess are political rather than economic or criminal. Not that economics or flagrant abuse of the law have not mattered, because they have. The lifting of the deposit insurance ceiling from $40,000 to $100,000 in 1980 contributed to the massive risk-taking in the thrift industry that occurred thereafter. The flat-rate feature of deposit insurance pricing, which allowed the drunk drivers of the system to pay no more for their insurance then the safe drivers, also played a part. And the number of books and articles on the rampant insider abuses and fraud among thrift owners and managers clearly demonstrate that criminal activity played a role as well. Robert E. Litan is a senior fellow and the Director of Economic Progress and Employment at the Brookings Institution, Washington, D.C.

210

Thomas Romer and Barry R. Weingast

But each of these explanations raises still additional questions. Take deposit insurance, for example. While I have been among those who have argued that the banking industry is in poorer health than reported either by the banks themselves or by the Federal Deposit Insurance Corporation (FDIC) (Brumbaugh and Litan 1990), no responsible observer claims that the cost of actual or hidden bank failures has been even close in magnitude to the costs of resolving the thrift crisis. Yet the deposit insurance system for banks and thrifts has been identical (but for a somewhat higher differential in the flat-rate insurance premium for thrifts since 1985). How then can deposit insurance be the sole cause of the thrift problem, as certain would-be reformers of deposit insurance implicitly, if not explicitly, claim.? Meanwhile, although fraud and insider abuse appears to have contributed to many, if not most, thrift insolvencies, there are no reliable estimates to indicate what portion of the cleanup cost is due to criminal activity. My own guess-and it is just that-is that the figure is 20% or below. Others certainly will have different intuitions. But the exact figure does not matter because, in my opinion, there is a more fundamental reason why so much criminal or near-criminal activity apparently took place, as well as why deposit insurance for thrifts in particular turned out to be so disastrous. That reason was the virtual abandonment of capital standards by thrift regulators who not only formally lowered required capital-to-asset ratios in the early 1980s, but who also introduced new regulatory accounting principles (RAP) that effectively allowed failed thrifts to hide their insolvency. Romer and Weingast, as well as others, are also correct to point to the utterly senseless refusal by the Office of Management and Budget in the mid-1980s to increase the number of thrift supervisors precisely when they were needed most: that is, after the congressional decision in 1982 to broaden thrift asset powers. This action, too, had the effect of abolishing capital regulation for many thrifts. It is now well recognized that the decision to let thrift operators play with federally insured deposits, but with little of their own money, was like throwing a lighted match into a pool of gasoline.' But it is less well understood that, by abandoning meaningful capital regulation, thrift regulators virtually invited high rollers and crooks into the industry. Had capital standards been enforced, few of the Donald Dixons (Vernon Savings and Loan of Texas), Charles Keatings (Lincoln Savings and Loan of Arizona), and David Pauls (Centrust Savings and Loan of Florida) would have ever bought thrifts. In short, capital deregulation in my view lies at the bottom of the thrift disaster. But, then, this explanation too simply leads to another question. Why did thrift regulators, but not their bank counterparts, effectively gut 1. This point has now been so heavily discussed that one hesitates to single out any particular authors who have advanced it. Nevertheless, a small sample of the literature includes: Benston and Kaufman (1990), Brumbaugh, Carron, and Litan (1989), and Barth et al. (1985).

211

Political Foundations of the Thrift Debacle

preexisting capital standards? Or, to put the question in somewhat more political terms, why did regulators and Congress wait for so long to put insolvent institutions out of business? Romer and Weingast tells us that the answer is simple: Congress was behaving in a business-as-usual mode by encouraging regulatory forbearance, not only in now-celebrated particular cases (the Vernon and Lincoln S&Ls, for example) but in a systemic fashion by denying FSLIC sufficient funds to clean the insolvent institutions out of the industry. Stripped to its essentials, the Romer/Weingast story is one we find over and over again in public policy circles. A narrow constituency, in this case the thrift industry, strongly wants a policy outcome, the costs of which are widely diffused throughout the economy. Congress then adopts the policy, in this case forbearance, largely by ignoring the problem and entrusting its resolution to its own narrow specialists, members of the banking committees. Of course, Romer and Weingast tell a fuller story, embellishing it with an interesting econometric demonstration of the important role played not only by insolvent thrifts, who of course wanted forbearance in their own particular cases, but also by healthy thrifts, who feared that they would have to pay for the cleanup and thus encouraged their legislators to deny the thrift insurance fund of all the resources it required. My only significant quarrel with Romer and Weingast is that by focusing so heavily on congressional and regulatory forbearance in the mid- 1980s they provide an incomplete analysis of their topic-the “political foundations of the thrift debacle.” In fact, there were two stages to the thrift debacle of the 1980s, which, as Edward Kane (1989) has demonstrated, can be usefully analogized to a massive oil spill. Like the Exxon spill that dumped millions of barrels of oil onto the Alaska coastline, the first stage of the thrift crisis occurred when doubledigit interest rates in the early 1980s caused thrifts to spill billions of dollars of red ink onto the financial landscape-by several estimates, over $100 billion in present value. But unlike the relatively rapid cleanup of the oil spill, the administration and the regulators did not ask for the funds, nor did the Congress voluntarily supply the funds, for cleaning up the initial thrift spill. Why? Romer and Weingast do not directly answer this question-prefemng instead to concentrate on later stages of the crisis in the mid-eighties-but the answers are straightforward. To have “cleaned up” the initial thrift spill would not only have cost far more money than anyone at the time was willing to spend, but it would have required the liquidation or assisted merger of most thrifts. At that time, few believed that effective substitutes for financing home ownership existed. The mortgage-backed securities market was growing, but it was not then as well developed as it eventually has become. In any event, even if all new mortgages could then have been securitized, it is highly doubtful that without some institutions dedicated primarily to buying those mort-

212

Thomas Romer and Barry R. Weingast

gages that they would have been fully absorbed by other buyers (pension funds, insurance companies and banks). All of these reasons help explain why Congress and regulators essentially chose to gamble on interest rates coming down, which eventually they did, rather than either shrinking the industry or, less radically, not permitting it to grow. Nor do Romer and Weingast discuss in detail the political foundations of why the first thrift spill happened in the first place. Part of the answer, of course, lies in the Regulation Q deposit interest ceilings, whose origins could have been usefully explored. But Regulation Q applied to banks first, and, as I have already noted, banks did not get into nearly as deep a mess as thrifts in the 1980s. The reason, of course, is that banks have been free to invest in assets of varying maturities as well as to lend at floating rates. In contrast, thrifts have been locked into long-term mortgages, which, until the early 1980s, had to be at fixed rates (except for certain state-chartered thrifts, such as those in California, that were permitted to extend adjustable rate mortgages in the 1970s). One of the great policy mistakes of the 1970s that Romer and Weingast should have given more emphasis was that Congress essentially refused to consider thrift industry proposals (as well as those of the industry’s regulator, the Federal Home Loan Bank Board) to permit the extension of adjustable-rate mortgages: this would have dramatically reduced the magnitude of the initial thrift crisis of the early 1980s. By the logic advanced in the RomedWeingast paper-wherein thrifts always get their way-this should not have happened. But it did because of heavy opposition from consumer groups. The strength of the Romer and Weingast paper, as I have said, is that it focuses on what happened after the first thrift spill, or, using the oil spill analogy, why Congress and the regulators delayed cleaning up the initial thrift spill even after the “oil” turned toxic. Not only did thrifts have a singleminded interest in forbearance, but politicians could reasonably assume that there would be no immediate or near-term (two years for a Representative, or even six years for a Senator) costs to a forbearance policy. After all, depositors were protected against all losses, both by the formal deposit insurance system and, in the case of the larger thrift institutions, by the “too big to fail” doctrine implicitly developed by federal regulators when they protected uninsured depositors of Continental Illinois Bank in 1984. During the 1984-88 period, given the administration’s adamant opposition to higher taxes, there also was little immediate or even intermediate-run prospect that taxpayers would suffer from forbearance. Finally, politicians could ignore warnings by the few economists, and eventually by FHLBB Chairman Ed Gray, that the FSLIC fund was effectively bankrupt by arguing that the problems thrifts were experiencing would be temporary. After all, had not the doomsayers who had pointed to the $100 billion-plus market value insolvency of the industry in .the early 1980s been proved wrong by the subsequent drop in interest rates, which seemingly restored many thrifts to financial health?

213

Political Foundations of the Thrift Debacle

Eventually, however, the toxic effects of the thrift spill became too large for even healthy thrifts and banks to ignore. Of course, by then the cleanup costs were also too large for them to pay, so with a free conscience they could presumably advocate a massive cleanup effort. But it is far from clear that this was why the Bank Board and then the new Bush administration finally began to attack the thrift problem with vigor. My simple-minded explanation is that eventually the press became convinced that the economists who had been warning about the immense magnitude of the problem had been right all along, which helped force the new administration into doing something major, at least after the 1988 Presidential election was over. In addition, by extending roughly $40 billion in government guarantees to buyers of failed thrifts in the 1988, Danny Wall (the chairman of the FNLBB at the time) also helped force Congress and the new administration confront the mess. The size of the expenditure, coupled with the off-budget way in which it was made, literally embarrassed both branches of government into going about the cleanup in a more straightforward way. As a reader I would have liked to have seen Romer and Weingast spend a little more time explaining why they think Congress and the administration backed off forbearance when they did. Finally, I would have liked to see Romer and Weingast attempt to apply their logic to both predicting how, if at all, Congress will attempt to reform the system to prevent future “thrift spills.” At a minimum, their paper persuasively suggests the broad outlines of what should be done. We now know what happens when business-as-usual politics confronts a major crisis: defenders of the status quo will prevent an immediate solution to the problem as long as no one feels the costs of not addressing it (or the costs of not doing so are diffused broadly throughout the population). Armed with this knowledge, it surely makes sense to develop mechanisms for forcing much earlier action to be taken when depository institutions get into trouble, but before they become insolvent.* One possible approach is to require regulators to intervene early and to base their judgment on more realistic marketbased measures of the financial condition of institutions (Benston et al., 1989). A supplement, or alternative, is to rely more heavily on market mechanisms, whether discipline by depositors, holders of subordinated debt, or private deposit insurers. I realize this is not the forum in which to debate the relative strengths and weakness of these alternative proposals. The concluding question that I would like to raise here is, What do Romer and Weingast believe their political analysis has to say about which, if any, are likely to be adopted? My own forecast, 2. Earlier action would surely mitigate losses to the insurance funds. In the 1980s the FDIC lost approximately 12 cents per dollar of recorded assets on failed banks. Since 1986, the thrift insurance fund (first the FSLIC and then later the Savings Association Insurance Fund of the FDIC) has lost more than 30 cents per dollar. These loss figures demonstrate that, by the time the capital of a depository institution, measured at book value, falls below zero, the market value of the institution is far less than that.

214

Thomas Romer and Barry R. Weingast

based on my inside-the-Beltway knowledge and intuition, is that, first, given the huge dimensions of the thrift problem and the certainty that more money will have to be authorized to solve it, Treasury will recommend and in 1991 or 1992 that Congress adopt significant reforms in deposit insurance. These will include: (1) authority for the FDIC to introduce risk-sensitive pricing of deposit insurance premiums; (2) some variation of the American Bankers Association proposal requiring mandatory “haircuts” on uninsured deposits; and (3) authorization for regulators to assume conservatorship of troubled depositories before they become insolvent. Without discussing the merits of any specific proposal, I believe that, in combination, the set of reforms I have just described would force regulators to act much earlier than they do now to assume conservatorship or to force the merger of troubled depositories. I hope that Romer and Weingast in their future work lend their considerable talents to forecasting whether the political system that brought us the thrift crisis can ever help prevent another, and, if so, how.

References Barth, James R., R. Dan Brumbaugh, Daniel Sauerhaft, and George K. Wang. 1985. Thrift Institution Failures: Causes and Policy Issues. In Proceedings of a Conference on Bank Structure and Competition, 184-216. Chicago: Federal Reserve Bank. Benston, George J., and George G. Kaufman. 1990. Understanding the Savings-andLoan Debacle. The Public Interest no. 99 (Spring), 79-95. Benston, George J., et al. 1989. Blueprint for Restructuring America’s Financial Institutions. Washington, D.C.: Brookings. Brumbaugh, R. Dan, Andrew S. Carron, and Robert E. Litan. 1989. Cleaning up the Depository Institutions Mess. Brookings Papers on Economic Activity, no. 1, 34988. Brumbaugh, R. Dan, and Robert E. Litan. 1990. The Banks Are Worse Off Than You Think. Challenge (January-February), 4-12. Kane, Edward J. 1989. The High Cost of Incompletely Funding the FSLIC’s Shortage of Explicit Capital. Journal of Economic Perspectives (Fall) 3 1-48.

7

The Spatial Mapping of Minimum Wage Legislation Keith T. Poole and Howard Rosenthal

7.1 Introduction Federally legislated minimum wages, first enacted in the Fair Labor Standards Act of 1938, are an enduring legacy of the New Deal.' While economists may argue the merits of minimum wages-along the lines that they impede efficiency and redistribute from poor and unskilled outsiders to the insiders of organized labor, minimum wages are determined by a political process. This process is highly partisan. The original minimum wage law was enacted by a Democratic Congress. Periodically, the (nominal) level of the wage has been increased. But none of the increases occurred during the 10 years since 1938 when the Republican party controlled at least one of the houses of Congress. Increases in the nominal wage have also been blocked or moderated by Republican presidents. Figure 7.1 shows that the real minimum wage rose steadily from the end of World War I1 until 1968, the end of the Kennedy-Johnson administration. When the Democrats both defeated Dewey and regained control of Congress in the 1948 elections, the minimum wage, in 1988 dollars, was only $1.96. Four years later, when Eisenhower was elected, the wage stood at $3.34. After the eight years of the Eisenhower administration, the wage again increased, Keith T. Poole is professor of politics and political economy at the Graduate School of Industrial Administration, Carnegie Mellon University. Howard Rosenthal is professor of political science and industrial administration at the Graduate School of Industrial Administration, Carnegie MelIon University. This paper was written while Rosenthal was visiting professor of economics at the Massachusetts Institute of Technology. The authors thank participants at the NBER Conference on Politics and Economics, especially our discussant, Charles Brown, for many constructive comments that, unfortunately, must be pursued in future work rather than in this paper. The extent to which this paper relates to the theme of the volume directly reflects the persistence of Albert0 Alesina and Geoffrey Carliner.

215

Keith T. Poole and Howard Rosenthal

216 56.00

,

JSH

Control Congress

1939

1944

1949

1954

1959

1964

1969

1974

1979

1984

1989

Fig. 7.1 The minimum wage

but far less rapidly, to $4.00. After increases were passed in 1961 and 1966, the real minimum wage reached its all time peak, $5.44 in 1968.2 The demise of the Great Society initiated a steep decline in the real minimum wage. Under Nixon the minimum wage fell-and, without Watergate would probably have eroded further-to $4.78. The 1977 increases only moderated the decline, since double-digit inflation in the late seventies eroded the value of the increments. When Ronald Reagan took office, the real wage stood at $4.36. When he left, the wage had declined to $3.20. The increases, passed at the end of 1989, will recoup only a small portion of the substantial decline in the real minimum wage that took place in the eighties. (The 1990 and 1991 entries in fig. 7.1 are based on the wage set by the 1989 law and an assumed inflation rate of 5 % . ) The purpose of this essay is to explore the history of congressional roll call voting on minimum wages, with an emphasis upon the recent increase enacted in 1989. We begin, in section 7.2, with a more detailed look at the time series on minimum wages in order to more firmly document the partisan character of increases in the wage. In section 7.3, we briefly present a dynamic spatial model that we use to analyze roll call voting on minimum wage votes. The spatial model is based on the hypothesis that voting on nearly all issues, not just minimum wage but also foreign policy issues, regulatory issues, and so on, reflects a legislator’s general ideological orientation on a few “liberalconservative” dimensions. In contrast, “economic” models of congressional voting are motivated by the premise that members of Congress seek reelection and therefore seek to match their voting decisions to issue-specific economic

217

The Spatial Mapping of Minimum Wage Legislation

preferences of their constituents. After, in section 7.4, critiquing the economic models, we, in section 7.5, directly address three previous “economic” studies of roll call voting on minimum wages and show how the roll calls analyzed in those studies fit into the spatial model. A systematic spatial analysis of all (71) House roll calls on minimum wages between 1937 and 1985 is conducted in section 7.6. We find that minimum wage voting is related not only to a “prolabor” or “economic” dimension but also to a “pro-civil rights/ anti-civil rights” or “social” dimension. In section 7.7, we present a more detailed analysis of events in 1989. Once again, the roll calls are highly spatial. Finally, section 7.8 contains some rough calculations that suggest that, for individual senators, support for minimum wages has waned somewhat over time. A brief conclusion then follows.

7.2 Congress, the Presidency, and Minimum Wage In the introduction, we noted a most striking indication of the partisan character of the minimum wage-no increase has been voted when the Republicans controlled at least one House of Congress. We would now like to anchor this observation more firmly by contrasting it with some simple, exploratory regressions of the process in those years when the Democrats had full control of Congress. We eliminate the war years and begin our sample with 1945; we end with 1990. For the years of full or partial Republican control of Congress (1947-48, 1953-54, and 1982-873), the time path of the nominal minimum wage, m,, can be described exactly as m, = m,-,, Similarly the real minimum wage, w,, is exactly

where P, is the price level. In other words, given that the reversion of the nominal wage cannot be altered with partial Republican control, the real wage atrophies with inflation. As it makes no sense to include these “error free” years in a regression, we delete them from the sample, but we use the two equations as null hypotheses to test whether years of Democrat control represent a different regime. Figure 7.1 shows both the real wage and, for periods of Democratic control of Congress, the nominal wage. When the Democrats have been in control of Congress since the war, the nominal wage (in spite of the fact that revisions to the law have always been separated by at least four years) increases quite regularly at an average annual rate of 7.1%. (The null hypothesis-zero-is rejected at p = .002.) Much of the fall in the real minimum wage in recent years, is, from this perspective, a combination of higher inflation rates in the seventies and a prolonged period of Republican control of the Senate in the eighties. When the Democrats control the White House as well, the rate averages

218

Keith T. Poole and Howard Rosenthal

9.3%. Divided government, with a Republican president and fully Democratic Congress, averages 4.6%. Sample sizes are small, however, for the comparison of Republican and Democratic presidents. On the one hand, the Republican average is not significantly greater than zero. On the other, the difference between the Democratic and Republican averages is not significant, either. A similar pattern of small differences appeared in other regress i o n ~The . ~ remainder of this section does not distinguish among presidents. We also reject the null hypothesis for the real wage. Table 7.1 presents two regressions where the natural logarithm of the real wage is the dependent variable. In the first column, the regressions include a constant, the log of the lagged real wage, and the log of the price ratio. The null hypothesis is a zero constant and unit slopes. The null hypothesis is rejected for the constant, the wage slope, and for the joint hypothesis (F-test). In the second column, lags of the independent variable are added. Under the null hypothesis, their coefficients should be zero, but they are significantly nonzero. The regression suggests that the political process of adjustment to the minimum wage has a lagged response to changes in the price level. For both regressions, one can compute a steady-state real wage assuming a constant rate of inflation. This, in contrast to the equation for Republican control of Congress, always shows a nonzero wage, but, consistent with our earlier observation that the nominal wage grows at 7%, a real wage that is decreasing in the inflation rate. These regressions suffice to show that the real-wage time series is quite different when the Democrats hold sway on Capitol Hill than when they must at least share power with the Republicans. Fully modeling the dynamics of the real minimum wage is beyond the scope of this paper. A major concern is that a nonmarket mechanism like Congress does not make a series of daily or even annual adjustments. IdiosynTable 7.1

Logarithmic Regressions for the Real Minimum Wage, w, Regression Model Variable

(1)

Constant

,394 (.Ill) .750 (.076) .516 (.579)

ln(w,-J

(2) ,423 (.093) ,366 (.170) - 1.016 (.754) .388 (.160) 2.046 (.775)

R’ p level, Null hypothesis Note: Standard errors are in parentheses.

,744 .0027

,786 .oO07

219

The Spatial Mapping of Minimum Wage Legislation

cratic matters are likely to affect the exact timing of increases. Nixon’s vulnerability during Watergate may have been instrumental to enactment of the 1974 bill that he signed after successfully vetoing a nearly identical bill in 1973; an exceptionally conservative Republican president may have delayed any increases when the Democrats regained control of the Senate in 1987. Once a package of nominal increases is eventually enacted (the 1977 bill mandated increases in 1978, 1979, 1980, and 1981, and the 1989bill in 1990 and 1991), further legislation has always followed the initiation of the last mandated increase. As a result, the real wage appears to be at the mercy of unanticipated inflation at least for several years after any bill has been signed into law. In this section, we have found that how the wage changes depends on who controls Congress. We now present a model that enables us to understand the coalitions that form to pass legislation during periods of Democratic control.

7.3 The Spatial Model of Congressional Voting In this paper, we present, in figures 7.2-7.4 and also in 7.7-7.1 1, a number of “spatial maps” of each house of Congress. The horizontal dimension on each plot can be thought of as extending from the economic left, or prolabor side to the economic right, or promanagement. The top of the vertical dimension represents opposition to civil rights, the bottom, support of civil rights legislation. Members of Congress are represented as points on the map. Legislator positions are shown by d (Northern Democrat), s (Southern Democrat), and r (Rep~blican).~ Roll calls are represented by lines, known as cutting lines. Members on one side of the line tend to vote “yea” on the roll call, those on the other side tend to vote “nay.” Members “very close” to the cutting line tend to split about 50-50. Those very far from the line almost always vote as predicted. The spatial model (Enelow and Hinich 1984) of Congressional voting (Poole and Rosenthal 1991) asserts that any bill, regardless of content, can be represented in a low (one- or two- ) dimensional space. This model reflects the fact that legislators’ positions on a wide variety of public policies are interrelated. For example, a legislator who opposes raising the minimum wage is very likely to have favored aiding the Nicaraguan Contras, to have supported Bork’s nomination to the Supreme Court, opposes cutting defense spending, supports a constitutional amendment to outlaw flag burning, and so on. In the language of contemporary American politics, this collection of policy positions would be held by a “conservative” and the opposite collection by a “liberal .” The consistency of policy positions held by legislators allows us to develop a spatial map on which we can place cutting lines not just for minimum wage votes but also votes on foreign policy, regulation, affirmative action, and so on. In essence, the two dimensions in the maps represent abstractions-the labor and civil rights labels are provided as simple heuristics.

220

Keith T. Poole and Howard Rosenthal

How are the members of Congress and cutting lines placed on the maps? Many readers will be familiar with the related topics of principle-components factor analysis and eigenvector extraction. What we do is similar in spirit to these methods, but the actual techniques are quite different since they are recovery techniques based on a maintained hypothesis of probabilistic spatial voting. The techniques are explained in detail in Poole and Rosenthal (1985a, 1991). The maps in Figures 7.2, 7.3, and 7.4 are based on applying the techniques in a simultaneous, dynamic estimation involving every roll call between 1789 and 1985.6 We found (Poole and Rosenthal 1991) that throughout American history at most two dimensions suffice to capture about 85% of the individual decisions, even on close votes. In fact, over 80% classification can be obtained with a one-dimensional model, essentially the horizontal projections of the points in the figures in this paper. We also found that relative spatial positions exhibit remarkable intertemporal stability, particularly since the Great Depression. As a result, we would get virtually identical spatial maps if we excluded minimum wage votes from the calculation of legislator positions or used different time periods in the estimation. One important source of change in the maps (cf. figs. 7.2-7.4 with 7.8 below) is the position of Southern Democrats. As the civil rights conflict evolved, they moved, as a group, to a distinct position at the top of the maps. But, then, following the passage of the Voting Rights Act in 1966, they became more similar to Northern Democrats. This movement, largely resulting from the replacement of old blood with new legislators with different locations (Bullock 1981) will, as we shall see in section 7.6, have a major impact on minimum wage votes. Quite typical of minimum wage votes captured by the spatial model are the three shown in figures 7.2, 7.3, and 7.4.’ The top panels of figures 7.2 and 7.3 show those who voted yes and were paired or announced yes, and the bottom panels show those who voted no or were paired and announced no. The top panel of figure 7.4 shows those individuals who took the “liberal” position of voting against the Erlenborn amendment but for final passage on the (successfully vetoed) 1973 bill. The middle panel shows those who voted “moderate” by supporting both the amendment and the bill, while the bottom panel shows the “conservative” position of voting for the amendment and against the bill. Classification errors for figures 7.2 and 7.3 are those individuals on the “wrong” side of the cutting line. (Figure 7.4 is discussed in section 7.5.) The two figures represent correct classification of 84% and 93% respectively. Even higher classifications could be obtained from cutting lines chosen to minimize classification error rather than to maximize a likelihood function. In a nutshell, votes can be correctly classified on the basis of the legislator’s general ideological orientation, without recourse to information about the specifics of minimum wage.

-

rAmerican Labar d4emocrat f-Farmeriabar n=Nm-Parliaan r-Republican adouthemDemocrat

Yes Paired or AMomced Yes

-

Cutting Lima

I

~

i

i

i

~

~

i

~

~

i

i

Pared or Announced No No

-

-

-

-

r

-

Culling Lime

I

I

I

I

I

i

I

I

I

I

I

I

l

l

I

l

l

i

i

i

i

I I

t rr

d d

r r rr

d

Yes Paidor Amaneed Yes

\

r

r

Fig. 7.3 Kitchin (D-North Carolina) amendment, 30 June 1960

223

The Spatial Mapping of Minimum Wage Legislation

Inspection of the figures reveals that while legislators tend to vote along party lines, there is considerable diversity within each party.8Most important, minimum wage votes tend to split both parties, with some liberal Republicans joining Northern Democrats and with Southern Democrats historically voting with the Republicans. Political scientists describe the alliance of the Southern Democrats and the Republicans as the “conservative coalition.” Minimum wage is a conservative coalition issue. While a three-party (Northern Democrats, Southern Democrats, and Republicans) model will classify much better than a two-party model, figures 7.2-7.4 show that a three-party model is still inferior to a spatial model. The three-party model fails to tell us which Southern Democrats are likely to oppose minimum wage and which Republicans are likely to break party ranks and support minimum wage. In contrast, the spatial model can capture the diversity within party and regional blocs. 7.4

Economic Models

Economic models link roll call voting to demographic measures and measures of costs and benefits. Peltzman (1984) investigated a large set of roll calls in an analysis that included a bevy of sociodemographic and economic variables, party, and an interest group rating. He found that after controlling for constituency interest measures, the marginal effects of party and interest group rating were small. Peltzman’s methodology has two weaknesses. First, interest group ratings lose some information, since they are folded (Poole and Daniels 1985) and not fine grained (Cox and McCubbins 1991, chap. 6). Second, there is no investigation of the marginal explanatory power of the constituency interest variables once an ideology measure has been used. We carried out this exercise (Poole and Rosenthal 1985b) and found that broad-brush measures of economic interests made little contribution to classification after controlling for spatial position. Unlike Peltzman’s analysis of large numbers of roll calls on a variety of issues, Kalt and Zupan (1984) focused on strip-mining voting. Although they were careful in accounting for constituency interests, they found that longterm ideology was an important determinant of roll call voting behavior. Reanalyzing their data, we found (Poole and Rosenthal 1985b) that the economic variables made only a small, very marginal contribution to classification once our spatial positions were In Gilligan, Marshall, and Weingast (1989), the enactment of the Interstate Commerce Act was modeled as a trade-off, made within the institutional structure of Congress, of interests on the short-haul pricing constraint issue and the issue of regulating by statute rather than by regulatory commission.I0 Their empirical work is a standard “economic” analysis. Since no interest group ratings were available for 1887, they could not run the type of tests performed by Kalt and Zupan.L1But party is significant, even with the “economic” controls.

Keith T. Poole and Howard Rosenthal

224

I

I

d Cutting h e . Passage I

I

I

I

I

I

I

a

8 S

8

I I , ,

I

I

I

225

The Spatial Mapping of Minimum Wage Legislation

,4

CuHi l i e , Erlenbom

For Amendment, Against Passage

I

rr r r

cutting Line. Pastage I

I

I

I

I

I

I

I

~

I

I

I

I

I

Fig. 7.4 Erlenborn (R-Illinois) amendment and passage, June 1973, House of Representatives

Similarly, previous research on minimum wage voting by Bloch (1980), Silberman and Durden (1976), and Krehbiel and Rivers (1988) has shown that measures of constituency interest, in terms of wage levels, unemployment levels, and union membership, are far less important to voting decisions than party membership. For example, Krehbiel and Rivers (1988; hereafter K-R), in an ordered probit analysis of 1977 votes on minimum wage amendments, found that Democrats, ceteris paribus, would prefer a minimum wage 17$22$ higher than Republicans, whereas a 10% increase in the percentage of the labor force belonging to unions in the state would induce a preferred increase in the minimum wage of under 8$. Similar conclusions can be drawn from Bloch’s probits on 1966 and 1974 Senate voting. In disaggregated results for each party, Bloch found that neither the wage nor the union variable was significant at the conventional 0.05 level in 1966. To put matters simply, two senators from the same state will tend to oppose each other on minimum wage if they are from different parties. The partisan effect overwhelms aggregate measures of constituency interest.

226

Keith T. Poole and Howard Rosenthal

7.5 The Spatial Model and the Economic Model Compared: The Senate in 1977 “Ideology,” as measured by spatial position, is, however, a more powerful explanatory variable than political party. In fact, a spatial model clearly outperforms the economic models in the literature, even when party is included as an economic variable. To demonstrate this point, we begin with the K-R study, in which the authors attempted to integrate the “economic” approach with a one-dimensional spatial model. The dimension is simply the nominal value of the minimum wage. When an increase to $2.90 an hour is being voted against a status quo of $2.70, the cutting line should be to the right of the cutting line for a higher proposal of $3.05 versus $2.70. That is, a more moderate proposal should attract broader support. The K-R study used constituency characteristics (party, union membership, wages, unemployment, South, and percentage black) and observed votes on the Bartlett and Tower amendments to estimate ideal points on minimum wages. The strength of the K-R approach is that it incorporates the quantitative information about the wage levels directly into the analysis. The weakness is that, for the ideal points to be estimated accurately, the economic constituency variables must be highly correlated with the true ideal points. Unfortunately, the economic variables (including party as an economic variable) bear little relationship to the roll call votes they analyzed. As we go on to show, the spatial model provides a much better accounting of the data and does so at a smaller cost in terms of estimated parameters. Our alternative procedure, to recapitulate, is to view ideal points on specific issues as arising from more general liberal-conservative or “ideological” preferences onto which specific issues, such as minimum wage, are mapped. We estimated our model using three alternative data sets. First, we used all roll calls from 1789 to 1985. We investigated both one- and two-dimensional models. This data set includes, except for those senators serving after 1985, all roll calls in a senator’s career. Second, we used only roll calls in 1977 prior to the votes on the Bartlett and Tower amendments. Third, we used all roll calls in the previous Congress (1975-76). For this last data set, we restrict classification computations to senators that had served in 1975-76. For the second and third data sets, we report results only for one-dimensional models where we chose cutting lines to optimize classification, using the previously estimated ideal points as exogenous information. Thus, in the second and third data sets, we are estimating one parameter per roll call. The K-R study estimates six to eight parameters in joint estimations covering both roll calls. The results appear in table 7.2. For those members serving in 1976, onedimensional liberal-conservative coordinates correctly classify 93.3% of the 134 votes cast on the two minimum wage amendments. For all members vot-

227

The Spatial Mapping of Minimum Wage Legislation

Table 7.2

ClassificationAccuracy Percentage of Votes Correct

Predictor

Bartlett

Tower

Combined

N

A. Yes, No, Each Roll Call Spatial Models: 1975-76 roll calls Earlier 1977 roll calls Dynamic NOMINATE, one-dimensional Dynamic NOMINATE, two-dimensional Marginals

95.8 95.5 91.0 91 .O 80.9

90.5 88.0 88.0 91.3 65.2

93.3 91.7 89.7 91.2 73 .O

134‘ 181b 181 181 181

B. Yes-Yes, No-Yes, No-No Both Roll Calls Dynamic NOMINATE, two-dimensional Khrebiel-Rivers MarEinals

n.a. n.a. n.a.

n.a. n.a. n.a.

85.7 73.9-76.1 68.1

91‘ 91 91

‘N reflects those serving in 1975-76 and voting in 1977. bNreflects those actually voting in 1977. ‘N reflects those voting or announced on both Bartlett and Tower roll calls.

ing, using coordinates estimated from all 1977 roll calls preceding the minimum wage votes, we correctly classify 91.7% of the 181 votes cast. We also classified the votes using the coordinates estimated from our dynamic model. In this case, the cutting line is a maximum-likelihood, rather than optimal classification, estimate. However, we again do quite well, correctly classifying 89.7%. There is little payoff, on these roll calls, from using a two-dimensional model. For the Bartlett amendment, we make eight classification errors with both one- and two-dimensional models. With the Tower amendment, a second dimension reduces the errors from eleven to eight. The overall percentage correctly classified becomes 91.2%. The little gain is not surprising since minimum wage had become a liberal-conservative “economic” issue by 1977 (see below). To have an appropriate baseline for comparison with K-R, we used our estimated cutting lines to classify senators into three classes: predicted yes on both amendments, predicted no on Bartlett but yes on Tower, and predicted no on both. (No senator was predicted yes-no, and no senator actually voted yes-no.) We then compared actual votes to predicted votes and found that we had correctly classified 86% of the individual decision pairs. Our 13 classification errors seem unlikely to be “residuals” that can be reconciled by appeal to standard economic interests considerations. To illustrate this point, we consider the 13 errors in terms of union membership, a key

228

Keith T. Poole and Howard Rosenthal

independent variable in K-R, and the presence of a right-to-work law.I2 Our two most serious errors were Danforth (R-Missouri) who voted yes-yes when predicted to vote no-no and Gam (R-Utah) who voted no-no when predicted yes-yes. Neither deviation would seem to have an “economic” interpretation. Missouri was above the national average in union membership and did not have a right-to-work law, while Utah had a low degree of unionization and a right-to-work law. l 3 Moreover, any economic considerations that would “explain” Garn’s vote would “unexplain” the vote of Hatch, the other Republican senator from Utah. A similar inspection of the 11 less serious errors (no-yes predictions where the actual votes fell into the other two categories or vice versa) also failed to disclose any consistent pattern in terms of either unionization or right-to-work laws. The “errors” of the spatial model are likely to be linked as much to internal horse-trading within Congress as they are to underlying economic interests in the constituencies. However, K-R’s results are much poorer; they are able to classify correctly only 74%-76% of the individual voting decisions. (The percentage correct varies over alternative specifications.) Thus, as table 7.2 indicates, their model only modestly betters the classification success that could be obtained by simply using the marginal distribution of votes to predict that everyone would vote as the majority voted (no-no).I4(Of the 91 senators in the sample, 68% fall into the modal class, no-no). We also looked at the 1966 and 1970 Senate roll calls studied by Bloch (1980). We achieved classification success of 90% and 86%, respectively, for these two roll calls. Bloch does not report classification success, but it is unlikely his results would better ours since his independent variables are a subset of those used by K-R.15 The final earlier study we compare to our “ideological” approach is Silberman and Durden (1976). They applied ordered probit to two 1973 House roll calls; the Erlenborn substitute and its final passage by the House.I6 The independent variables were a South dummy, campaign contributions to the 1972 Congressional winner by labor, contributions by small business organizations, and measures of low-wage workers and teen-age workers. It is difficult to compare this study to the spatial model as the authors report only the estimated coefficients. Two points can be made. First, the most statistically significant coefficients are the region dummy and the two campaign contribution coefficients. In this respect, we note that campaign contributions and region cannot be specifically linked to minimum wages but are relevant to a whole set of interests that are captured by our spatial coordinates. Indeed, as shown by Poole and Romer (1985) and Poole, Romer, and Rosenthal (1988), campaign contributions, particularly by labor, are highly related to spatial position. In other words, there is an identification problem. Region and campaign contributions have a logical relationship to minimum wage interests. On the other hand, they relate to a whole set of other interests as well. Since Southerners, for example, tend to vote as a bloc on a

229

The Spatial Mapping of Minimum Wage Legislation

whole set of issues, it is difficult to distinguish the economic impact specific to minimum wage from general regional interests. Second, campaign contributions do not measure within-constituency interests. Since the contribution variables are contributions to the winning candidate, variable values for individual districts will be highly sensitive to the outcomes of House races. Moreover, many contributors, such as the United Auto Workers or the National Association of Manufacturers, are not local groups. So if the Silberman-Durden specification is valid (as against a Bloch or K-R), we will have to believe that, while economic interests may matter, they are not median voter interests. In figure 7.4, we have plotted the cutting lines and shown the crosstabulation of the two roll calls. Table 7.3 contains the classification analysis. Excluding against-against types and individuals voting on only one of the roll calls (as did Silberman and Durden), we correctly classify 81% of the joint decisions. The most serious errors occur in the upper-right and lowerleft comers of the table. These represent only 1% of the representatives. (Classification of the two roll calls separately shows 89% for Erlenborn and 87% for passage.) More research would be needed to ask if the Silberman-Durden variables would account for the errors of the spatial model. One of them clearly will not, at least on its own. An inspection of figure 7.4 shows that Southern Democrats are spread out over the three categories. There is not a strong pattern to the classification errors for southern representatives. To summarize the results of this section, there appears to be little interest in using economic models once the spatial nature of voting has been recognized. Why do the economic variables fail to have an impact? We hardly deny that economic interests are important to congressional decision making. But at this time, little progress has been made in either modeling or measuring those interests. As political scientists (Fiorina 1974; Fenno 1978; Weingast, Shepsle, and Johnsen 1981) have long pointed out, a member of Congress is an agent who faces a multitude of principals. Whose preferTable 7.3

Classification Analysis for the 1973 House Votes Actual Vote

Predicted Vote Against Erlenborn, For passage For Erlenborn, For passage For Erlenborn, Against passage

Against Erlenbom, For Passage

For Erlenborn, For Passage

188

16

3

11

35

27

2

17

102

For Erlenbom, Against Passage

230

Keith T. Poole and Howard Rosenthal

ences-typical constituents, registered voters in his or her party, party militants, or campaign contributors not even resident in the district-are operant is a complicated problem. And even if we knew whose preferences counted, measuring those preferences is likely to be at least as complicated as the cost benefit calculations of Kalt and Zupan. Simply using explanatory variables culled from government documents is unlikely to do the job. On the other hand, most of the time most of the relevant information gets picked up by the legislator's Euclidean coordinates.

7.6 Minimum Wage Roll Calls in the House of Representatives, 1937-83 The highly spatial nature of the handful of minimum wage roll calls analyzed in the literature is indeed generally true for the entire history of minimum wage roll calls. We show this by considering the entire set of minimum wage roll calls that occurred in the House of Representatives during the period, 1789-1985, spanned by our dynamic estimation. The first roll call occurred in the 75th House in 1937, the last in the 98th House in 1983. In figure 7.5, we show the classification success for each roll call. The roll calls are inversely ordered by the size of the majority. The bottom curve shows the classification success, equal to the size of the majority, achieved by the majority (marginals) model. The intermediate curve shows classification success for a one-dimensional estimation. The top curve is for two dimensions." The Congress number for each roll call appears above the top curve. It can be seen that the classification success is generally independent of the size of the majority. On roll calls where no one is close to being pivotal, gratuitous expressions of opinion can occur. For example, on the final passage of a bill, extreme liberals and extreme conservatives can both vote against and express their dissatisfaction with an inevitable compromise. But when the game is on the line, spatial considerations predominate. The spatial model provides a better accounting of minimum wage voting after World War I1 than before. Voting on minimum wage occurred before the war in the 75th Congress, when the initial legislation was passed, and in the 76th, when revisions were considered. Subsequently, new legislation was made moot by the command economy of the war. Divided government occurred in 1947-48, with a Republican Congress and a Democratic president. In the labor area, the Republicans devoted their energies to overriding Truman's veto of the Taft-Hartley Act. Minimum wages did not get considered until the 81st Congress. Consequently, we can divide the roll calls neatly into pre-World War I1 and post-World War I1 samples. Classification of minimum wage roll calls using only the one-dimensional dynamic model is very high after World War 11, averaging 88.2%, but is much lower, at 71.2%,before World War II.18 Moving to a two-dimensional model improves matters considerably before World

231

The Spatial Mapping of Minimum Wage Legislation

70

-

tW

U

65

n 60

17

, i

One Dimension

- Two Dimensions

o

Majority

Fig. 7.5 Minimum wage roll calls in the House, classification and model dimensionality

Note: The top line in the figure is broken by the Congress number for each roll call. The line shows the classification accuracy for the two dimensional model. The intermediate line shows the classification accuracy for the one dimensional model. The bottom line shows the percentage voting on the majority side. This line descends from left to right, since the roll calls are ordered by the size of the majority.

War 11, with classificationsjumping to 82.0% but still below the 88.6% obtained for the postwar period.I9 Since the second dimension was needed to obtain good classifications in the prewar period, we know that initially minimum wage was an unusual, nonstandard issue for its time. Conceivably, since our dynamic model is constrained to using the same abstract dimensions for an entire 200 years period, the second, vertical dimension, could capture the main lines of debate in certain periods of history. But when we look across all roll calls, and not just at minimum wage roll calls, we find that since the 1850s the first, horizontal dimension has always dominated the second dimension in terms of classification (Poole and Rosenthal 1991, fig. 4).20In particular, for the 75th and 76th Congresses, computing optimal classification cut points using the estimated legislator coordinates from the first dimension classifies over 80% of the individual decisions across all roll calls. Optimal classification for the second dimension results only in about 70% correct, barely better than the marginals. Comparable results hold, by and large, for the postwar period. The finding that the second dimension is the key to classification of minimum wage prior to World War I1 shows that, initially, minimum wage was not

232

Keith T. Poole and Howard Rosenthal

part of the main line of liberal-conservative conflict. Related to this result, figure 7.2 shows that, in terms of projections onto the horizontal dimension, Southern and Northern Democrats were not differentiated. As this differentiation increased, as shown in figure 7.3, Southern Democrats had become more similar to Republicans than to Northern Democrats on the horizontal dimension, and the minimum wage conflict turned into a quintessential liberalconservative battle. Another important point is made by the comparison of the prewar and postwar periods. The fact that, even in two dimensions, minimum wage voting is significantly less structured before World War I1 is indicative of the potential multidimensionality of most economic legislation. Since, for example, the level of the wage can be traded off against the definition of which jobs will be covered, a vote between two alternative bills may not fit readily into the preexisting spatial pattern of voting. The complex nature of such trade-offs should, we suggest, be most apparent in the initial legislative handling of an issue. Eventually, however, the multidimensionality is packaged and shoehorned into the spatial structure. The packaging indeed results, at any one point in time, in minimum wage voting being nearly unidimensional. In figure 7.6, we plot the cutting line angles in chronological order. If the votes were unidimensional at a point in time, all the angles for a year would be identical. While there is some variation within years, it is quite small, particularly for roll calls with close (less

loo 80

I

I

,,,,,,

7 0 j , : 1 , , , 1 60 75

76

81

, / ,

84 87

/ / ,

89

, , , , , , ,

-,,

9092

,,,, 93

, , , , , , , , , , , 1, 1: 1

95

1 1 1 1 1 , 1 ( 1 1 /

98

CONGRESS

Fig. 7.6 Angles in the House of Representatives, Congress labeled for margins < 65-35

The Spatial Mapping of Minimum Wage Legislation

233 Table 7.4

Cutting Line Angles, IIme, and Margin Coefficient

Independent Variable Constant Calendar year (Year x Margina)/lOO

R= Standard error of estimate

(1)

(2)

3,385.37 (268.1 I ) - 1.667 (.136)

3,291.22 (260.56) - 1.622 (.132) ,8059

.679 15.037

(.336) .703 14.467 _ _ _ _ _ _ _ ~ ~

Note: Dependent variable is cutting line angle in degrees. Standard errors are in parentheses. 'Margin = (%Yes - %No(.

than 65-35) margins. (We have not yet carried out an estimation where all minimum wage roll calls in a Congress were constrained to have an identical angle.) In contrast, the angles vary strongly and linearly across time. In table 7.4, we present the results of a regression of the angle against a constant, calendar time, and roll call margin. Angle declines sharply with time, especially for close roll calls.21The standard error of the estimate of 14.5 degrees indicates the small variability in angle within a give year. Results in Poole and Rosenthal(l991) help to understand why the angle has gradually shifted. At the beginning of the New Deal, positions of Northern and Southern Democrats did not have significant differences. The Roosevelt coalition represented a reasonably coherent voting bloc. The civil rights issues introduced significant and increasing polarization within the party. The passage of the Voting Rights Act in 1966, however, marked the beginning of a period in which Southern Democrats have drifted back toward the party mainstream. This pattern is evidenced in figures 7.2-7.4. By 1940, some of the separation of the Southern delegation had occurred. They had moved above the Northern Democrats on the vertical dimension but had not moved to the right on the horizontal dimension. By 1960, a time when civil rights began to dominate American domestic politics, Southern Democrats had made the full transition. By 1977, there was less separation, and by the 1980s the regional differences among Democrats had dampened considerably (fig. 7.7-7.11). The changing positions of Southern Democrats are tracked by the minimum wage cutting lines. The opposition of Southern Democrats as well as Republicans to the initial minimum wage legislation led to a nearly horizontal cutting line before the war. As Southern Democrats, flush with the manna of the New Deal, switched from seeking to alleviate their economic situation vis-avis the North to seeking to protect the internal status quo vis-h-vis blacks, they

234

Keith T. Poole and Howard Rosenthal

became more conservative on the economic dimension. The cutting line angle echoed this movement. While a standard economic explanation for the opposition of the South is that minimum wages cause low-wage regions to lose their comparative advantage in attracting investment, an alternative view is that enfranchised whites in the South sought to maintain disenfranchised blacks in a low-wage situation. Congressman Dies (&Texas) was unambiguous on this point in stating: “There is a racial question involved here. Under this measure whatever is prescribed for one race must be prescribed for the others, and you cannot prescribe the same wages for the white man as the black man” (New York Times, 14 December 1937). Later developments support the view that the race issue played a key role. First, as the debate on extending coverage unfolded after the war, Southern Democrats fought hard to prevent extension to sectors such as tobacco where competition with the North was not an issue. Then, in the seventies, as Southern Democrats acquired a black constituency, more of them became favorably disposed to increased minimum wages, even though Southern states had low levels of unionization and nearly all had right-to-work laws. Concurrent with the new liberalism among Southem Democrats, liberal Republicans became a vanishing breed. Echoing this movement, the cutting line became nearly vertical. 7.7

Minimum Wage in the 1980s

The most recent developments in minimum wage legislation are a microcosm of the long-term historical analysis of the preceding section. With the election of Ronald Reagan and a Republican Senate in 1980, the debate shifted from increases in the nominal wage to retrenchment of coverage in the form of administration proposals for a subminimum wage for teenagers. The subminimum wage, however, was not successful in the Democratic controlled House. It is interesting that one factor that contributed to the impasse on the subminimum wage was the failure of McDonald’s and other large fast-food chains to support the proposal. These large-market-share firms preferred not to risk their consumer image.22 In contrast, the National Restaurant Association, which represented many small firms, supported the proposal. It is also possible that large firms saw wage regulation as enforcing their competitive advantage with respect to smaller firms. In any event, it appears that oligopoly coupled with an active media may result in outcomes that anticipate legislation and vitiate much of the agenda of Republican conservatives. As soon as control of the Senate switched back to the Democrats after the 1986 elections, proposals for an increase in the wage began to make their way through committees. An anticipated presidential veto removed much of the

235

The Spatial Mapping of Minimum Wage Legislation

impetus from these proposals until a “kinder, gentler” George Bush endorsed an increase in the minimum wage during his campaign. The new Bush administration proposed an increase to $4.25 an hour over three years and a training wage, equal to 80% of the minimum, to be paid to new workers in the first six months on the job. Democratic bills were introduced in the House and Senate. The relevant committees were chaired by two ultraliberals, Augustus Hawkins in the House and Edward Kennedy in the Senate. Hawkins’s committee reported out a bill calling for an increase to $4.65 without a training wage. A GOP substitute bill, containing the administration’s proposals, was rejected. However, the Hawkins bill was known to be in disfavor at the White House. It was amended via a “compromise,” offered by Murphy (D-Pennsylvania) that subsequently passed as the House bill. This called for an increase to $4.55 and a training wage set at 85% of the minimum wage with sharp restrictions on the conditions and time for which the training wage would apply. The president indicated he would veto this bill, and the passage vote of 248 to 171 clearly indicated a veto would be sustained. Nonetheless, the Democratic leadership chose to invite the veto by proceeding to pass a nearly identical bill in the Senate and to gain acceptance of the conference report in both Houses. After the president vetoed the bill and the veto override failed, the White House and the Democratic congressional leadership negotiated. The supposedly powerful committee chairs were reportedly frozen out of the negotiations. Bush got his way on the level-the final bill went no higher than $4.25. But he agreed to a two-year rather than three-year phase-in, to restrict the training wages to teenagers, and to make the limit 85% rather than 80%. There are three key lessons in this scenario. First, politics has symbolic, as well as real, victories. The Republicans finally got the teenage training wage but at a higher level than the $2.50 an hour (75% of the $3.35 minimum wage then in effect) proposed by the Reagan administration. The Democrats got an increase in the nominal wage but at a level that barely dents the erosion of the seventies and eighties. Second, committee chairs do not always carry the clout found in other discussions (e.g., Weingast and Moran 1983). Third, attempts to model the interaction between committees, the two Houses, and the president, as a game with a small number of stages may be misplaced. The actual process appears to be pure bilateral bargaining between a Democratic legislature and a Republican executive in which both parties have veto powers. While Bush could successfully veto any bill that did not match his proposal, the Democrats could prevent any bill from passing. Since both parties were committed to a bill, negotiated compromise occurred. The voting process continued to fit into our spatial model. The votes occurred in the lOlst Congress. Our dynamic estimation ended with the 99th Congress and full data on the lOlst Congress is not available. Consequently, we did a separate spatial estimation based on roll calls in the 100th Congress

236

Keith T. Poole and Howard Rosenthal

and then analyzed the 1989 votes of those members who had served in both Congresses. Thus, the remaining analysis is based on spatial positions estimated from votes that occurred prior to the minimum wage votes. The very first House roll call, a procedural motion to accept the rule on the bill (fig. 7.7) was nearly a straight party-line vote. The few defectors tended, as called for in the spatial model, to be in the moderate wings of their parties. Searching for optimal classification found a cutting line that had only 15 errors of 382 voting, paired, or announced. The first substantive vote in the House concerned the GOP substitute (fig. 7.8). Many Southern Democrats were attracted by this proposal. On the other hand, some liberal Republicans defected. Again, spatial positions from 1987-88 successfully picked out the likely supporters and opponents of the substitute. The substitute lost by only 20 votes; the margin is partly represented by Republican losses since the early Reagan years, partly by the increasingly liberal character of the Southern Democrats. On this vote, we had 31 classification errors of 389 voting, paired, or announced. Most other votes were intermediate between these two, with fewer Southern Democrat defections than occurred on the GOP substitute bill. One vote that was not was the final passage vote on November 1, shown in figure 7.9. The spatial model presumes a choice between two alternatives. But final passage is often an opportunity for extremists to voice their displeasure. Negative votes were cast not only by the right wing of the Republican party but also by two literal Democrats, Miller (California) and Perkins (Kentucky). For Republicans, the bill went too far, for the two liberals, not far enough. Since many final passage votes involve mainly symbolic protest, they should typically not be treated in either spatial or economic analyses. The Senate exhibited patterns similar to the House. A cutting line can be found for the GOP substitute vote that results in only five errors (fig. 7.10). All Democrats are predicted to vote against the substitute. All but one of the Republicans voting against the substitute were on the liberal end of the party. The one exception was Jesse Helms, so opposed to any form of minimum wage that he voted against the administration bill. Economic models like those discussed earlier would also fail to capture this form of protest voting. Finally, as seen in figure 7.1 1, the vote to kill the Gramm (R-Texas) amendment shows that strategic considerations do little to upset the spatial pattern of voting. Gramm proposed to strike from the final bill a provision that kept teenagers in agriculture from being paid the subminimum rather than the minimum wage. Voting to kill the amendment introduced by this conservative Republican was in fact supporting the president’s position. Republicans Waffled between their support for the president’s negotiated compromise and their preferences for lower minimum wages. The motion was made by Strom Thurmond (R-South Carolina) who then voted against his own motion. The waffling also resulted in fairly noisy voting by the more liberal Republicans. Nonetheless, the spatial model correctly classifies all but 9 of the 87 senators

d d d Yes Paired or Annomced Yes

-I

Pated or Announced No No

Fig. 7.7 HR2, procedural motion, 22 March 1989

Yes Paired or Announced YI

-

,

I

I

I

I

I

r

r r r r

d d

Paired or Announced No No

,I

I

I

I

I

I

I

I

I

I

Fig. 7.8 HR2, GOP substitute, 23 March 1989

I

I

i

I

l

-

I

Yes Paired or Announced Yes

1

Perkins r

r

Miller

r

Pared M Announced No No

I

I

I

I

1

7

I

1

Fig. 7.9 HR2710, passage, 1 November 1989

I

I

I

1

7

S

S

S

R Yes Paied or AnnouncedYes

i

S

S

Helm8

R R

R

1 Fig. 7.10 S4, GOP substitute, 11 April 1989

Paired or Announced No

S R

R Ye8

Paired or Amomed Ye8 I

I

I

I

I

I

ss

I

I

I

1-

R

k

S

R

Pared or Announced No

No I

I

I

I

I

I

I

I

I

I

I

Fig. 7.11 HR2710, Gramm amendment, Senate, 8 November 1989

I

-

242

Keith T. Poole and Howard Rosenthal

voting. Because Southern Democrats are now increasingly like other Democrats, the motion passed easily. (Only two Democrats were predicted to vote against the motion.) Roll call voting on minimum wage during the Bush administration remains, in summary, a highly partisan matter. 7.8

Is Support for Minimum Wage Waning?

Until this point, we have indicated that minimum wage voting obeys a spatial mapping. While the mapping has always been present, outcomes have differed dramatically, with the real minimum wage first rising sharply and then, since 1968, falling sharply. Indeed, the minimum wage has lagged, in real terms, minimum wages in other countries where the standard of living is not above that of the United States. For example, in the mid-l980s,the minimum wage was about $4.65 an hour in Belgium (Emerson 1988) and France (Rotbart 1989) when it was $3.35 in the United States. Part of the falling minimum wage in the United States may be attributed to Republican resurgence. On the other hand, while minimum wage voting may always be spatially patterned, the induced mapping of the real wage may well have changed. The “ideal” real minimum wage of a legislator with the same relative degree of “liberalness” may have fallen over the past two decades. We need to compute the spatial mapping of the real wage. Making comparisons is difficult because each bill is in fact a multiattribute item. As a result, we will only work through some rough, back-of-theenvelope calculations. We find that the cutting lines of the Tower amendment in 1977 and the GOP substitute bill in 1989 similarly partition senators serving in both years.23In particular, Senator Lugar, who took the conservative position both in 1977 and 1989 and Senators Hatfield, Packwood, and Heinz, who voted with the liberals both times, were reasonably close to both cutting lines. Therefore, if we can map a wage to a cutting line in both cases, we can get a rough estimate of how the preferences of moderate Republicans have changed over time. If those senators now have lower support for a real minimum wage, the assumptions underlying the spatial model force the conclusion that the ideal points of all senators shifted in the direction of lower support. As K-R point out, the Tower amendment was in fact a vote between the proposals of Tower and Williams. Tower proposed increases to $2.65 in 1978, $2.85 in 1979, and $3.05 in 1980. Williams proposed $2.65 in 1978, $2.90 in 1979, $3.15 in 1980, and $3.40 in 1981. Here K-R do not deal with the problems posed by the differences in the schedules. Rather than attempt to deal with discounted streams and anticipated inflation rates, we simply compare the 1980 wages and impute a value of $3.10 to the cutting lines. In 1988 dollars, this amounts to $4.45. The 1989 vote on the GOP substitute is also complicated by multiattribute considerations, since the GOP substitute had a lower and more extended train-

243

The Spatial Mapping of Minimum Wage Legislation

ing wage as well as a lower minimum wage. Nonetheless, we compare the 1991 wages of $4.55 in the Democratic substitute and $4.25 in the GOP substitute and impute a cutting line value of $4.40 or $3.81 in 1988 dollars. Comparing the $4.45 figure from 1977 to $3.81 for 1989 suggests an ebbing support for minimum wage. The differential would be greater if one argued that the inflation of the late seventies was largely unanticipated or that inflation in 1990 and 1991 will exceed our assumed 5% rate. On the other hand, the differential would be less if one argued that senators were conditioning their votes with a view to adopting a bill that would win presidential approval. Carter’s acquiescence to labor demands and Bush’s firmness may have influenced the spatial mapping in the legislative branch.

7.9 Conclusion We have traced out how minimum wage voting fits into the spatial structure of congressional voting and indicated that the spatial structure better accounts for the data than statistical analyses based on constituency characteristics. Voting on minimum wage bills is a highly partisan, liberal-conservative matter. Over time, the mapping of wages into the spatial structure has probably changed. Individual legislators who supported a given real wage in the seventies probably prefer a somewhat lower real wage today. Much work remains to be done on the multiattribute nature of the bills. Specifically, we need to investigate the interaction between the level of the wage and coverage. The major policy implication of our research for those interested in affecting the level of the minimum wage is that they should direct their attention not to forming realigning coalitions on the issue but rather to moving the location of the “cutting” line that separates liberals from conservatives along the dimension that represents the stable mapping. Still, reasoned argument is likely to be far less important than changing control. Given the Democratic lock on the House, Republicans can affect change only by allowing inflation to erode the value of the nominal wage. On the other hand, any presidential landslide for the Democrats would be likely to lead rapidly to a minimum wage indexed to about 50% of the average wage in manufacturing. If we think it is likely that Democrats will pursue their traditional platform, it is partly because we think that the eighties were mostly business as usual in American politics. The parallels between Nixon and Reagan are as striking as the differences. Both were elected after overseas debacles by the Democrats. Both were reelected by landslides. Both began their administrations by advocating a subminimum training wage. Both saw no increase in the minimum wage during their first terms. Nixon finally conceded on minimum wage when the Watergate scandal hit early in his second term. More speculatively, Bush’s campaign promises on the minimum wage may have been one of several moves to moderation induced by a fall in Republican popularity from the IranContra scandal, which hit late in Reagan’s second term. In any event, the

244

Keith T. Poole and Howard Rosenthal

minimum wage continues to be hostage to the larger ebb and flow of liberal and conservative fortunes.

Notes 1. In our descriptive account of minimum wage legislation, we have relied on a number of secondary sources, mainly various volumes of Congress and the Nation and Congressional Quarterly Weekly Report, both published by Congressional Quarterly Press. 2. Similar conclusions pertain to other evaluations. The minimum wage as a percentage of the poverty line gives similar results. The minimum wage, as a percentage of the average wage in manufacturing, rose to 50% in 1950 and remained at that level through 1968. Subsequently, it fell gradually below 50% until 1980, after which it declined sharply. 3. Most increases in the wage have become effectiveafter the first year but before the end of a party’s reign in the White House. There are two exceptions. In 1961, a raise became effective in Kennedy’s first year. In 1981, a raise, voted in 1977, took effect in Reagan’s first year. To give Kennedy the raise he supported and to avoid giving the Republicans a raise engendered by a unified Democratic government, we coded Carter’s presidency as having lasted through 1981 and Republican control of the Senate as lasting from 1982 to 1987. Very similar results were produced by moving all terms of office forward one year. 4. These results are based on logarithmic regressions. The regression slopes were sufficiently close to 1.O that we chose to base the discussion in terms of means. Adding terms for trend, trend squared, and a post-sixties dummy did not improve the fit. 5. We follow Congressional Quarterly in defining the South as the 11 Confederacy states plus Kentucky and Oklahoma. 6. The estimation was carried out on the Cyber 205 and ETA-10 supercomputers at the John Von Neumann National SupercomputingCenter. 7. The space in the figures is 1.8 units square. Bootstrap results presented in Poole and Rosenthal(l991) suggest that about 75% of the first dimension coordinates of the representatives have standard errors less than 0.03 and 99% have standard errors less than 0.05. (Although the bootstrap results are for a one-dimensional estimation, the first dimension coordinates from the two-dimensional model correlate over 0.99 with the one-dimensional coordinates). Bootstrap results are not available for twodimensional problems, but it is highly likely that the points in the figure are very precisely estimated relative to the range of the space. 8. In this context, the argument made by “distributive politics” advocates that party is not a constraint on politicians appears to be overstated (e.g., Marshall and Weingast 1988). If voting were independent of party, we would not see the party clusters in the figures. In fact, party discipline and leadership, even in the modern era, may well be the key to explaining why politics, potentially explosively multidimensional (Riker 1982), looks as if nearly all issues can be placed in a low-dimensional space. For example, Ferejohn (1986) details how the Democratic leadership bundled agricultural subsidies and food stamps into a single piece of omnibus legislation that represents an institutionalized logroll. Our analysis of roll call votes on this issue suggests that the logroll is not a coalition of fann interests and urban interests but of Democratic representatives from farming areas and urban representatives. 9. We thank Tom Romer for drawing our attention to the economics literature on roll call voting and for participating in the reanalysis of the Kalt and Zupan data.

245

The Spatial Mapping of Minimum Wage Legislation

10. It is interesting to note that a similar debate took place when the initial minimum wage legislation was drafted in 1937-38. Some members of Congress pressed for setting the minimum wage by independent commission. Similarly, during the 1988-89 debate, liberals advocated a commission that would propose future adjustments to the wage. 11. The votes on the Interstate Commerce Act in the House are in fact much like minimum wage. The spatial model works well. 12. Our data is for 1975. Our source is the Statistical Abstract ofthe United States (1988,415). 13. Gam’s votes in 1977 are even more puzzling when one observes that he was a stalwart opponent of minimum wage increases in 1988. 14. To be fair, we point out that the intent of K-R was to present a methodology rather than to contribute to our understanding of minimum wages. They were in part motivated by what they saw (p. 1158) as a consistency problem with our scaling procedures. In Poole and Rosenthal (1991), we present extensive Monte Car10 evidence that shows that consistency need not be a concern. K-R (1988) also has problems as a methodological piece. Equations (6),(7), and (1 1) contain errors that lead to an incorrect result for the covariance matrix. 15. Another problem with the Bloch study is that the votes analyzed were final passage votes. These generally tend to be “Hurrah” votes with large majorities. The key votes generally tend to be earlier votes on amendments or substitute bills. See our later discussion of the 1989 legislation. 16. The bill passed by the House went to the Senate and to conference. After the House accepted the conference report, the bill was vetoed by President Nixon. 17. Both estimations are for models where legislator positions were constrained to be a linear function of time. 18. The t-statistic for the null hypothesis of equality is 8.600. 19. The r-statistic for equality is 2.072. Since we hypothesize that classifications improve as new issues result in permanent legislation, a one-tailed test is appropriate, withp = .021. 20. We get very similar results when we evaluate the estimation in terms of geometric mean probability of observed choices, a measure based directly on the likelihood function. For simplicity, we focus on classificationin this paper. 21. This regression provides a better description than the use of a pre-World War I1 dummy variable. Examination of the residuals did not suggest that margin was a source of heteroscedasticity. 22. An analogous situation appears to be the recent ban, initiated by H. J. Heinz (Starkist) on canning tuna caught in nets that trap dolphins. Heinz is a highly visible firm with over one-third of the tuna market. A similar ban would, it seems, be unlikely to arise in a market with only very small canners. 23. With comprehensivedata for all roll calls from 1989, we could include the 1977 and the 1989 votes in a dynamic estimation. This would be a preferable method for checking the comparability of the two votes.

References Bloch, Farrell. 1980. Political Support for Minimum Wage Legislation. Journal of Labor Research 1~245-53. Bullock, Charles S., 111. 1981. Congressional Voting and Mobilization of a Black Electorate in the South. Journal of Politics 43663-82.

246

Keith T. Poole and Howard Rosenthal

Cox, Gary, and Mathew McCubbins. 1991, forthcoming. Parties and Committees in the U.S. House of Representatives. Berkeley and Los Angeles: University of California Press. Emerson, M. 1988. What Model for Europe? Cambridge, Mass.: MIT Press. Enelow, James, and Melvin Hinich. 1984. The Spatial Theory of Voting, Cambridge: Cambridge University Press. Fenno, Richard, 1978. Home Style. Boston: Little, Brown. Ferejohn, John. 1986. Logrolling in an Institutional Context: The Case of Food Stamps. In Congress and Policy Change, edited by Leroy Reiselbach et al., 22353. Agathon Press. Fiorina, Morris. 1974. Representatives, Roll Calls, and Constituencies. Lexington, Mass.: Heath. Gilligan, Thomas, William Marshall, and Barry Weingast. 1989. Regulation and the Theory of Legislative Choice: The Interstate Commerce Act of 1887. Journal of Law and Economics 32:35-62. Kalt, Joseph, and Mark Zupan. 1984. Capture and Ideology in the Economic Theory of Politics. American Economic Review 74:301-22. Krehbiel, Keith, and Douglas Rivers. 1988. The Analysis of Committee Power: An Application to Senate Voting on the Minimum Wage. American Journal of Political Science 32:1151-74. Marshall, William J., and Barry R. Weingast. 1988. The Industrial Organization of Congress; or, Why Legislatures, Like Firms, Are Not Organized as Markets. Journal of Political Economy 96: 132-63. Peltzman, Sam. 1984. Constituent Interest and Congressional Voting. Journal of Law and Economics 27: 181-200. Poole, Keith T., and R. Steven Daniels. 1985. Ideology, Party and Voting in the U.S. Congress, 1959-80. American Political Science Review 79:373-99. Poole, Keith T., and Thomas Romer. 1985. Patterns of Political Action Committee Campaign Contributions to the 1980 Campaigns for the U.S. House of Representatives. Public Choice 47, no. 1,63-111. Poole, Keith T., Thomas Romer, and Howard Rosenthal. 1987. The Revealed Preferences of Political Action Committees. American Economic Review 77:298-302. Poole, Keith T., and Howard Rosenthal. 1985a. A Spatial Model for Legislative Roll Call Analysis. American Journal of Political Science 29:357-84. . 1985b. The Political Economy of Roll-Call Voting in the “Multi-Party” Congress of the United States. European Journal of Political Economy 1:45-58. . 1991. Patterns of Congressional Voting. American Journal of Political Science 35228-78. Riker, William H. 1982. Implications from the Disequilibrium of Majority Rule for the Study of Institutions. American Political Science Review 74:432-47. Rotbart, Gilbert. 1989. Au voisinage du SMIC. Economie et Statistique 221:15-21. Silberman, Jonathon, and Gary Durden. 1976. Determining Legislative Preferences on the Minimum Wage: An Econometric Approach. Journal of Political Economy 841317-29. Weingast, Barry R., and Mark Moran. 1983. Bureaucratic Discretion or Legislative Control? Regulatory Policy Making by the Federal Trade Commission. Journal of Political Economy 9 1:765-800. Weingast, Barry R., Kenneth A. Shepsle, and Christopher Johnsen. 1981. The Political Economy of Benefits and Costs: A Neoclassical Approach to Distributive Politics. Journal of Political Economy 89542-63.

247

The Spatial Mapping of Minimum Wage Legislation

Comment

Charles Brown

Much of the analysis of the politics of minimum wage legislation by economists emphasizes both economic (rational choice) models of legislator behavior and economic differences as the ultimate explanatory variables in predicting legislative votes. Poole and Rosenthal, on the other hand, emphasize ideological factors-in effect, positions on apparently unrelated roll call votes-which turn out to permit quite accurate discrimination between those favoring higher and lower minimum wage increases. Based on similarities and differences on virtually all roll call votes in each Congress, they assign each legislator a “position” in a hypothetical twodimensional space. Much like factor analysis, the computer finds the dimensions and the analyst brings the labels. Poole and Rosenthal identify their more important dimension as general liberalkonservative leanings, and the other as attitudes on racial questions; this second dimension serves to separate Northern and Southern Democrats. The key point is that these “positions”-in effect, scores on two hypothetical variables-are based on votes in general, not votes on minimum wage questions in particular. To relate these general positions to minimum wage votes, they find the “line” relating the two variables that distinguishes those who vote “yea” on a particular minimum-wage vote from those who vote “nay.” Finding such a line in effect estimates two parameters; the resulting line lets one “predict” more than 90 percent of the votes on a minimum wagerelated question, versus more than 80% for roll call votes in general. I find the result that a one- or two-dimensional characterization of general voting stance predicts minimum wage votes well an interesting but not a surprising one-the minimum wage controversy in Congress is a liberals-versusconservatives affair. (It would be interesting to see how Poole and Rosenthal’s “spatial” variables would do in predicting legislators’ choice between minimum wage increases and the Earned Income Tax Credit in raising the after-tax earnings of the low-paid. My guess is that positions would be harder to predict.) On the other hand, the later finding that North-South splits among the Democrats have been reduced (as the Southerners respond to enfranchisement of blacks) was both interesting and unexpected. Poole and Rosenthal then argue that their “spatial” model predicts better than “economic” models. They are able to compare their predictions against those based constituent characteristics (e.g., wage levels and union membership) as reported by Krehbiel and Rivers (1988). They achieve much more accurate discrimination between minimum wage supporters and opponents than do Krehbiel and Rivers, and they use considerably fewer parameters (two vs. six, including one reflecting the influence of party membership). Those attached to constituent-characteristic models will detect a tilt in the Charles Brown is a professor of economics at the University of Michigan.

248

Keith T. Poole and Howard Rosenthal

track on which this horse race is run. The race between “political” and “economic” models here is very much like that between “time-series” and “structural” models in macroeconomics. While the former often provide more accurate predictions, the latter (arguably) give deeper insights into why things are changing. Similarly, one could argue that “economic” models are atrempring to explain why some legislative districts are represented by those who rank high in Poole and Rosenthal’s conservative dimension, while their model “merely” confirms that whatever explains the general pattern of other votes will explain minimum wage votes too. Indeed, if one did nor see a high degree of consistency across votes, this would be puzzling for a constituentcharacteristics model.’ I do, however, believe that Poole and Rosenthal have a point, that positions on the minimum wage are part of a larger mosaic of positions, and that larger mosaic (which they would call ideology) is important in understanding the politics of the minimum wage. One does not get very far in understanding the Autoworkers’ or Steelworkers’ support for minimum wages by focusing on their desire to exclude lower-cost labor: if their deepest fear was that Pintos or pig iron would be manufactured by nonunion U.S. workers making less than $3.90 an hour, they would sleep very well indeed. Poole and Rosenthal note that constituent-characteristic models will have a hard time explaining why Senators Gam and Hatch (same party, same constituents) would ever disagree. This suggests to me an empirical strategy to supplement the results reported in the paper: What sort of predictive accuracy would one get (in Senate votes) using state dummy variables (and perhaps party)? Since senators from the same state have the same constituents, the dummy variables give an upper bound of how well we might ever hope to do with an “ideal” set of constituent variables, at least for Senate votes. Poole and Rosenthal also argue their model’s predictions compare with those of Silberman and Durden, whose “economic” model includes amount received from unions and from small business as well as constituent characteristics. Poole and Rosenthal note that adding these variables moves things in the direction of a model like theirs (contributions depend on general orientation more than on an individual vote like the minimum wage). Unfortunately, Silberman and Durden did not report how many votes their model predicted correctly, so we really cannot tell whether Poole and Rosenthal’s approach is stronger or weaker than theirs. Poole and Rosenthal then argue that the real value of the minimum wage preferred by Congress has been falling. This conclusion rests on a comparison of votes on the Tower amendment in 1977 and the GOP substitute bill in 1989. These produced very similar partitions of the Senators voting on both, but the alternatives at stake on the former vote were higher than those on the latterhence the inference that a lower real minimum is preferred. In addition to I

This point was made by Herschel Grossman.

249

The Spatial Mapping of Minimum Wage Legislation

problems that come from other provisions of the bills, which the paper notes, there is also the fact that the 1989 discussion concerned a larger increase (at a time of lower expected inflation) than the 1977 vote.* Consequently, agnosticism on the question of Congress’s preferences over time seems the safer verdict to me. A fine paper ends on a terribly speculative note: “any presidential landslide for the Democrats would be likely to lead rapidly to a minimum wage indexed at roughly 50 percent of the average wage in manufacturing.” My own prediction is that a Democratic landslide would require a Democratic focus that demonstrates some independence of organized labor. If labor-led Democratic victories of the past could not produce an indexed minimum wage, why should the next (less labor-led) one do so?

References Krehbiel, Keith, and Douglas Rivers. 1988. The Analysis of Committee Power: An Application to Senate Voting on the Minimum Wage. American Journal of Political Science 32: 1 15 1-74. Silberman, Jonathan, and Gary Durden. 1976. Determining Legislative Preferences on the Minimum Wage: An Econometric Approach. Journal of Political Economy 84:3 17-29. Anne Krueger noted that the number of individuals affected by the new minimum may be at least as important as either Poole and Rosenthal’s “real level” or my “relative increase” standards.

This Page Intentionally Left Blank

8

U. S . Trade Policy-making in the Eighties I. M. Destler

8.1 Introduction As the 1970s wound to a close, Robert Strauss, President Jimmy Carter’s Special Trade Representative, won overwhelming congressional approval of the Tokyo Round agreements. This was a triumph of both substance and political process: an important set of trade liberalizing agreements, endorsed through an innovative “fast-track” process that balanced the executive need for negotiating leeway with congressional determination to act explicitly on the results. And it was accompanied by substantial improvement in the nonoil merchandise trade balance. Both the substance of trade policy and the process of executivecongressional collaboration would be sorely tested in the 1980s. During the first Reagan administration, a mix of tight money and loose budgets drove the dollar skyward and sent international balances awry. The merchandise trade deficit rose above $100 billion in 1984, there to remain through the decade (see table 8.1). The ratio of U.S. imports to exports peaked at 1.64 in 1986, a disproportion not seen since the War between the States. The constant-dollar ratio of imports to domestic goods production-the best single measure of change in the import pressure faced by U.S. firms-shot up from 18.9 percent in 1982 to 26.0 percent in 1986 (see table 8.2), and the figure for manufuctured goods jumped from 23 to 34 percent (Destler and Henning 1989, 121). Such increases had no precedent in modern U.S. history. Political reaction was substantial. Had the U.S. Congress been generally I. M. Destler is professor at the School of Public Affairs, University of Maryland, where he directs the graduate program on Public Policy and Private Enterprise. He is also a visiting fellow at the Institute for International Economics. His publications include American Trade Politics: System Under Stress. The author wishes to thank Paul Baker for research assistance, and Anne Krueger, Dani Rodrik, Robert Pastor, and the editors for their helpful critical comments.

251

252

I. M. Destler

Bble 8.1

U.S. Merchandise Wade, 1970-89 Billions of Current Dollars

Ratios

Year

Exports

Imports

Balance

MIX

1970 1971 1972 1973 1974 1975 I976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989

44.5 45.6 51.7 13.9 101.o 109.6 117.5 123.1 144.7 183.3 225.1 238.3 214.0 206.1 224.1 220.8 224.4 225.1 322.0 368.9

40.9 46.6 56.9 71.8 104.5 99.0 124.3 151.9 176.5 211.9 247.5 266.5 249.5 271.3 334.3 340.9 367.8 412.8 449.0 480.2

3.6 -1.0 -5.2 2.1 -3.5 10.6 -6.8 -28.8 -31.8 - 28.6 - 22.4 - 28.2 -35.5 -65.2 - 110.2 - 120.1 - 143.3 - 157.7 - 127.0 -111.3

.91 1.02 1.10 .97 1.03

.90 1.06 1.23 1.22 1.16 1.10 1.12 1.16 1.32 1.49 1.54 1.64

I .62 1.39 1.30

Defi citlGN P

... ,001

,004

...

,002

...

,004 ,014 .014 .011 .008 ,009 ,011 .019 .029 ,029 ,034 ,035 .026 .021

Source: Department of Commerce, Bureau of Economic Analysis. Various years. Survey ofcurrent Business. Nore: M = imports, X = exports.

disposed toward trade protection-as often assumed in press commentaryand had its members given priority to actual impact on policy-as assumed in journalistic accounts and academic models alike-then there surely would have been a 180-degree turn in postwar U.S. trade policy. In fact, there was not. The dikes constructed over the decades to protect liberal U.S. trade policy did spring some leaks, but in general they held. Congress did insist on enacting a new, omnibus trade law, but this tinkered at the margins rather than making fundamental changes. The overall policy erosion was far less than analysts and practitioners would have forecast in 1980 had they known the 12digit trade deficits the decade would bring. There was, over the decade, a significant shift in policy emphasis. Both branches became much more aggressive in pressing for the opening of foreign markets, and, in the decade’s final year, the incoming Bush administration’s trade representative, Carla Hills, was driven to designate entire nations as unfair traders: “priority foreign countries” singled out for the “number and pervasiveness” of their “acts, policies, or practices” that impede U.S. exports (Omnibus Trade and Competitiveness Act 1988). And the “super-301’’ law that forced this action was a congressional creation. But even this was in the

253 Table 8.2

U.S. Trade Policy-making

U.S. Merchandise Trade and Output, by Volume, 1970-89 Billions of Constant (1982) Dollars

Percentage

Year

Output of Goods

Imports

Exports

MIOutput

X/Output

1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989

1030.0 1037.6 1093.8 1175.0 1159.2 1125.0 1194.7 1256.2 1329.1 1354.6 1344.2 1386.0 1319.1 1367.0 1509.2 1553.6 1592.6 1669.0 1771.6 1837.1

150.9 166.2 190.7 218.2 211.8 187.9 229.3 259.4 274.1 277.9 253.6 258.7 249.5 282.2 351.1 367.9 413.7 440.5 467.1 494.4

120.6 119.3 131.3 160.6 175.8 171.5 177.5 178.1 196.2 218.2 241.8 238.5 214.0 207.6 223.8 231.6 245.9 285.7 344.3 386.8

14.6 16.0 17.4 18.6 18.3 16.7 19.2 20.6 20.6 20.5 18.8 18.6 18.9 20.6 23.3 23.7 26.0 26.4 26.3 26.9

11.7 11.5 12.0 13.6 15.1 15.2 14.8 14.2 14.7 16.1 18.0 17.2 16.2 15.2 14.8 14.9 15.4 17.1 19.4 21.0

Source: Department of Commerce, Bureau of Economic Analysis. Various years. Survey ofcurrent Business. Nore: M = imports, X = exports.

long-standing tradition of choosing toughness on (trade-expanding) export issues as an alternative to toughness in imposing (trade-contracting)import controls.* Reinforcing support for devices such as Super-301 was a growing concern about the relative competitive position of the United States. By late in the decade, Paul Kennedy (1987) had published his unlikely best-seller, Clyde Prestowitz (1988) had brought forth a widely noticed work, entitled Trading Places: How We Allowed Japan to Take the Lead, and Robert Gilpin had concluded his comprehensive work of contemporary political economy by associating “many of the troubles of the world economy in the 1980s” with “the relative decline of American hegemony.” It was “not likely,” he declared, “that a liberal world economy could survive without a liberal hegemon committed to its preservation” (1987, 365). Well before this, concern about the relative position of the United States had contributed, along with the trade deficits, to an atmosphere of dissatisfaction with liberal trade policies and interest in “strategic” alternatives. Politicians were reinforced in their propensity to believe that other nations were taking advantage of us, either through unfair trade practices or through clev-

254

I. M.Destler

erer and more coherent pursuit of their self-interest. Political scientists and economists (Cohen and Zysman 1987; Krugman 1986; Richardson 1990) were exploring matters of “dynamic comparative advantage” and “strategic trade policy.” Mercantilism became semilegitimate. Dissatisfaction with liberal policies (and trade outcomes) did not lead to a strong alternative consensus: in fact, the prescriptive harvest from the strategic trade (or dynamic comparative advantage) school of economists, political economists, and business leaders was strikingly thin. But it did increase the pressures on and from Congress and the demands on executive-branch trade policy leaders, weakening the still-persistent U. S. bias toward open-market trade policies. For an analyst of U.S. trade policy-making, therefore, one major task in reviewing the 1980s must be to explain why the system held as well as it did. To this end, this paper will challenge (in sec. 8.2) the widely held notions that members of Congress are “protectionist” and that they seek to control trade policy. The body of the paper (secs. 8.3 and 8.4) will be a recounting of the major policy events of the decade. The evidence supporting the assertions of section 8.2 will be contained therein, and summarized in section 8.5. The events to be recounted will include: The rise of traditional protectionism early in the decade. The Reagan administration ended up, in the words of then-Treasury Secretary James Baker 111, granting “more import relief to U.S. industry than any of his predecessors in more than half a century” (1987). The response of Congress. Despite the unprecedented pressure on importimpacted industries, Congress did not “go protectionist.” But there was a sharp rise in legislative entrepreneurship, as members both felt increased constituency pressure and saw a growing “issue opportunity.” The role of partisan politics. In the House, much of the Democratic leadership saw trade as an issue promising partisan advantage; in the Senate, however, dissatisfaction was bipartisan, and the Omnibus Trade and Competitiveness Act of 1988 reflected a bipartisan consensus for a tougher U.S. approach to international trade. The administration’s shift to international economic activism. In 1985, led by Baker, the second Reagan administration shifted from neglect to activism on exchange rates, and to aggressiveness in pushing U.S. exports. The latter, it was hoped, would appease Congressional critics, win such foreign concessions as proved available, and buy time for the J-curve to play itself out.3 The continuing, but somewhat diluted, U.S . commitment to multilateralism. The single largest executive branch priority remained, at decade’s end, the Uruguay Round: because of what might be achieved substantively and because it offered a vehicle for managing domestic trade policy pressures. But it was less central to U.S. trade policy-making in the eighties than the Tokyo Round had been in the seventies. And bilateral trade deals-notably the freetrade pact with Canada-assumed greater prominence.

255

U.S. Trade Policy-making

As these points illustrate, this paper will offer an analytic description of the main policy developments of the decade, as seen through one analyst’s lens. It will draw substantially on my research over the decade and recent published work (Destler 1986; Destler and Ode11 1987; Destler and Henning 1989), while seeking to avoid tracking that work too closely. It will emphasize throughout how the trade imbalance, even though it was not caused by trade policy-and most key trade policy players knew that the causes were elsewhere-had enormous impact on the balance of trade politics: enlarging the population of the “trade-afflicted”; diminishing those gaining on the export side; weakening the legitimacy for existing trade policy approaches; driving Congress and the administration to take action. With the 1985-87 decline of the dollar came a halt to the rise in import pressure and strong gains on the export side. As a result, trade-political players felt some relief as the decade drew to a close. But the macroeconomic adjustment was incomplete, and the trade imbalance was threatening-by most projections-to grow larger in the early 1990s.

8.2 Congressional-Executive Relations and Bade: Eight General Propositions This paper seeks to make explicit a view of congressionalbehavior on trade that is implicit in my broader work.4 It can be summarized in eight propositions: PROPOSITION 1. On the surface, both Congress and the executive seem to be struggling over policy outcomes, but there is an asymmetry in stakes: executive players give greater priority to controlling trade policy outcomes than do congressional players. PROPOSITION 2 . For individual members of Congress, direct control over trade policy is not a necessary means to their broader goal, which is to maintain and enhance political standing, at home and in Washington. PROPOSITION 3 . For the great majority of members for whom trade is but an occasional concern, it is sufficient to advocate the cause of interests important in one’s district and to strike general trade policy postures that appeal to one’s support coalition. PROPOSITION 4.For legislators on the key trade committees (Senate Finance and House Ways and Means) and the (growing) minority that has singled out trade policy for special attention, controlling policy on core trade matters is neither the only, nor in most cases the best means, of enhancing tradepolitical standing. PROPOSITION 5 . When individual legislators do seek influence over policy, legislation is often not the most effective means to that end. PROPOSITION 6 . This overall pattern of behavior tends to diminish the impact of party divisions on policy outcomes, though trade issues will frequently be employed for partisan point scoring.

256

I. M. Destler

PROPOSITION 7 . Hence, members of Congress find their interests wellserved by a system of power-sharing that gives them ample opportunity for initiative and visibility but allows the buck to stop elsewhere. PROPOSITION 8. Administration leaders, who give greater priority to policy impact, alsofind their interests better served than in a system where spec@ trade barriers were legislatively determined. Such propositions, this paper will suggest, continued to have explanatory power even in a decade of record trade pain for U.S. producers, and hence pressure on Congress that was unprecedented in the postwar period. These assertions diverge sharply from much of the “new institutionalist” literature on Congress, which is built on the assumption that what members are seeking is impact on policy. They need backing from constituents, which they purchase through provision of policy payoffs. Legislators are thus, in practice, conduits of pressure from interest groups. Within Congress, the committees and subcommittees most responsive to the dominant constituencies in an issue area come to play the pivotal role, and they use their power to control p01icy.~ Such a model fits U.S. trade policy-making quite well, up through 1930. But the Smoot-Hawley Act of that year proved to be a blame-generating law if ever there was one, with the Great Depression and World War I1 among the events to which it has been causally linked. In its wake, therefore, Congressional leaders cooperated with executive officials in constructing a system that minimizes the direct impact of legislation on U.S. trade barriers. This system is characterized by: delegation to the president of de facto authority to set the level of trade and nontariff barriers, which are arrived at through negotiations with foreign the use of quasi-judicial procedures for industries claiming injury from imports and/or unfair foreign trade practices; negotiation by the executive of special, “voluntary” export restraint (VER) arrangements in cases where a major industry (textiles, steel, autos) is demanding trade relief; creation of a trade-brokering agency, the Office of the United States Trade Representative (USTR), to be the target of congressional and industry pressure, and to take the heat as it works for compromise on contentious trade matters in the United States and in foreign markets. Notable is the fact that this system deals with industry pressure by diversion as well as accommodation, through buffering institutions that give legislators ample opportunity either to duck or to assert themselves on trade issues. It is ideal for “blame avoidance.”’ What it reliably provides is not protection for industry, which often finds its claims rejected or deferred, but “protection for Congress” (Destler 1986, chap. 2). Why is this system consistent with legislators’ interest in maintaining and enhancing their political standing? The simple answer is that striving for direct personal influence over policy outcomes is not, for most members in most instances, a cost-effective means to that end. Recall Mayhew’s (1974) classic

257

U.S. Trade Policy-making

formulation, centered on the goal of securing reelection, that cites three activities of House members as crucial: advertising, position taking, and credit claiming. None of these requires actual impact on policy. For the majority with limited interest in trade, it will usually be sufficient to engage in position-taking responsive to key trade-affected interests in one’s constituency, and/or to evoke sentiment in broader ideological support groups. More interesting, however, is the growing minority of members who take special interest in trade issues or whose committees give them special power. Their interest in legislating on trade will certainly be greater, for they are “trade policy politicians” who build their broader political standing, in part, on trade policy engagement. But even they need not give priority to actual impact on outcomes. They have open to them time-tested activities such as making speeches, issuing press releases, holding hearings, traveling to the capitals of U.S. trading partners, crafting marginal legislation for claiming credit, or backing seemingly major legislative proposals (ideal for advertising and position taking) that no one expects will become law. And even when the activists do want to influence policy, legislation is only one of several routes. Others include lobbying the executive, using legislation as a threat-as did Senator John Danforth on auto quotas in 1981-and pressuring foreign governments-as illustrated by the same example. Indeed, the executive branch (or international bargaining) game, where “downtown” officials take the issue-specific heat, may offer well-connected members of Congress greater opportunity for impact than the procedurally cumbersome legislative game. Or they may collaborate with the executive in crafting more efficient legislative procedures: the “fast-track” rules established by the Trade Act of 1974 give members of the key committees the inside track in influencing the president’s proposed legislation to implement a major trade agreement, legislation that cannot be amended once it is formally submitted. In a variety of ways, therefore, Senate and House trade specialists can be visible policy players, responding to their constituents, bashing the Japanese, able to claim credit for diplomatic or administrative actions taken in their behalf, while retaining their ability to duck blame for outcomes unfavorable to them. There will be occasions, of course, when a critical mass of congressional players becomes committed to the enactment of major trade legislation. This happened in a minor way in 1984 and in a major way in 1985-88. Even then, however, Congress tended to refine at the margins the mechanisms for indirect influence rather than taking direct control of specific product issues or specific bilateral trade relationships. What role does party competition play in these politics? Politicians are regularly looking for issues to use in partisan conflict. In the context of the 1980s, therefore, Democrats would naturally seek to pin on a Republican White House the blame for a 12-digit trade deficit or the clear pain of tradeimpacted firms and workers. But they would give priority not to overturning

258

I. M. Destler

present policy but to establishing a politically rewarding posture: blaming the administration, arguing for “tougher,” pro-American policies. Nor would the Republicans allow them to capture the issue: they would respond with their own trade aggressiveness, particularly in the Senate, seeking to close the window of political opportunity. An overall system in which Congress eschews direct responsibility can be attractive to legislators because it gives them greater flexibility and initiative in preserving and advancing their careers as politicians while avoiding blame for mistakes. It can be attractive to executive players because they give greater priority to policy outcomes, and congressional delegation gives the executive greater leeway in influencing policy outcomes. In any case, such a system of American trade policy-making was well established by 1980. Initiated in the wake of Smoot-Hawley, it matured while trade policy receded from public prominence and lost its partisan character in the early postwar years. It proved rather effective in combating “normal” protectionism; that is, from industries losing comparative advantage. Elements of this would continue into the eighties for steel, autos, and textiles. But the Reagan decade would bring new challenges, specifically arising from the incredible rise of the dollar and enormous trade imbalance and from the decline in the relative U.S. position in the world. New trade pressures emerged in consequence. To how the American system responded, we now turn.

8.3 1980-84: The First Five Years Defining the decade as “the eighties” offers an awkward fit with the political calendar. We get the last and first years of two administrations and eight years of the one in between. We get no clear beginning and no definitive ending. Dividing the eighties at their mid-point, however, has strong empirical as well as numerical appeal. Not only does it separate the first from the second Reagan administration (Reagan I and Reagan 11), but the two periods exhibited rather different patterns of trade politics, both in the executive branch and on Capitol Hill. The first featured a range of product issues and the building of a trade imbalance with large political consequences, but it contained no central trade policy event, either at home or abroad. The second featured one main event at home-the trade bill-and the beginnings of one abroad-the Uruguay Round. It was further defined by the sharp administration policy departures, as Reagan I1 sought to contain the whirlwind sown by Reagan I. The first five years began quietly: the Tokyo Round had been ratified, the trade bureaucracy was being modestly reorganized, and the trade balance was improving, fueled by a boom in exports. The economic policy scene was dominated by the oil price surge and its domestic consequences. One casualty-the U.S. auto industry-generated the main trade issue of the early decade.

259

U.S. Trade Policy-making

8.3.1 Autos 1980-81 Even as the overall U.S. trade balance was improving, the largest of American industries was in crisis. The oil shock of 1979 brought a doubling in the price of motor fuel. This resulted, simultaneously, in a sharp drop in total demand for autos and a shift of demand into energy-efficient small cars, the bulk of which came from Japan. Between 1978 and 1980, car imports rose from 17.7 to 26.7 percent of the U.S. market, as sales of American-made vehicles plunged from 9.3 to 6.6 million, the lowest since 1961. Auto industry unemployment rose above 300,000 out of a total of almost one million directly employed there (Winham and Kabashima 1982,76; U S . Department of Transportation 1981, 83-85; Cohen and Metzger 1982, chap. 3). Auto companies suffered record losses, and Chrysler needed a federal bailout to avoid bankruptcy. In terms of its economic impact, this was surely the greatest trade-related disruption any U.S. industry had experienced since the Great Depression. The trade-political response was remarkable for its moderation. An “escape clause” petition seeking temporary import relief was submitted by Ford and the United Auto Workers (UAW), who were not joined by their industry brethren. House Trade Subcommittee Chairman Charles L. Vanik (D-Ohio) held hearings in 1980, signaling an interest in some trade-restrictive action, but he also reiterated that year his hope to “depoliticize” such issues, to “run trade on economic law” (U.S. House Ways and Means Committee 1980, 140). In this case, “economic law” failed to deliver, on a technicality. The U.S. International Trade Commission (USITC) ruled, by a 3-2 margin, that it could not recommend relief because auto imports failed the “substantial cause of injury” test-other causes on the industry’s troubles, in particular the oil price rise, were more important. Since this ruling came after the November election, the buck was now passed to the new Congress and the Reagan administration. Reagan had promised in his campaign to “try to convince the Japanese” to slow down imports. John Danforth (R-Missouri), the new Trade Subcommittee chairman in the Republican-controlled Senate, was promising hearings by December, and he introduced an auto import quota bill in early 1981. But hearings and markup were scheduled not to pass a law, but to pressure the Japanese, and to help ensure that Reagan’s campaign commitment prevailed over the free-market leanings of his economic advisers. Senate Majority Leader Robert Dole (R-Kansas) put out the word that he could count twothirds of the Senate in support of the Danforth bill, enough to override a Reagan veto. But when Tokyo announced on May l that Japan would restrain exports, reducing the number of cars sold in 1981-82 by 7.7 percent below the previous year, U.S. Trade Representative William E. Brock immediately assured the Japanese that, in light of their action, the bill had no chance of enactment. Danforth cancelled a scheduled markup and shelved the bill, even

260

I. M. Destler

though the Japanese plan was less restrictive than he had wanted. In particular, it did not provide for a reduction of the Japanese quota in the event that the overall U.S. market shrank further-as in fact it would in 1982.* 8.3.2

Autos 1982-83

Japan’s auto export restraints remained in place through the eighties. Their trade impact rose as the Reagan recovery gained force in 1983-85, then diminished as the dollar declined and the U.S. output of Japanese firms increased. Economically, the main long-term beneficiaries were probably those same Japanese firms, since they captured billions of dollars in rents from the price rises the quotas permitted. Politically, however, the quotas put a cap on the auto-trade issue. And had the U.S. economy not plunged into severe recession in the year after their enactment, nothing much more would have occurred on the matter. But with that recession, the plight of automakers worsened, and the Japanese producers’ share of the U.S. market grew further even with their lower absolute number of sales. In response, the UAW pressed for a bill that would have imposed a “domestic content” requirement on cars sold in the United States. The larger a company’s total sales, the more onerous the requirement: for Toyota (and General Motors) it would have been 90 percent. If enacted, the bill would have cut imports of Japanese autos to a fraction of current levels. Because the bill ostensibly regulated production rather than trade, its proponents were able to circumvent the House Ways and Means Committee. Hence the bill was twice considered, and twice reported favorably, by the House Energy and Commerce Committee. And it twice passed the House of Representatives: by 215 to 188 in December 1982 and by 219 to 199 in November 1983. Unlike the situation in 1980-81, division on rhis auto issue was very much along partisan lines. Northeastern and midwestern Democrats almost uniformly voted “Aye” in the House, and the bill never got out of committee in the Republican-controlledSenate. But unlike with the Danforth bill of 1981, no one could even pretend that “domestic content” was a serious legislative threat-the fact that it was not, in fact, allowed many of its supporters to set aside their policy convictions and give their votes to the UAW, a long-time supporter of liberal Democratic causes. Because it was not a credible threat, the bill had little impact on either the administration or the Japanese. It was, however, one of several events suggesting that party might become the main definer of politicians’ positions on major trade legislation, as it had been in the 1930s. 8.3.3

Executive Branch and Congressional Rivalries

As the domestic content bill showed, trade politics in the early eighties were punctuated also by conflicts and rivalries built into the structure of leg-

261

U.S.Trade Policy-making

islative and executive politics. In the House of Representatives, John Dingell (&Michigan), chair of the Energy and Commerce Committee, was challenging the long-time leadership of Ways and Means. In so doing, his committee was challenging also the tradition of delegation and deference to the executive on trade. With its middle ranks dominated by entrepreneurial Democrats from the Watergate Class of 1975, the Dingell committee was entrepreneurial and policy active, and Ways and Means had been weakened by the procedural reforms of the 1970s. In the end, Energy and Commerce did not make major inroads, but the threat it posed drove Ways and Means counterparts to greater activism. Trade Subcommittee Chairman Sam Gibbons, for example, began pushing very hard for tightening of countervailing duty and antidumping laws (Destler 1986,59-62,73-74). More visible was the conflict between U.S.Trade Representative Brock and Secretary of Commerce Malcolm Baldrige. Before the former was even named to his position, it became public knowledge that presidential assistant Edwin Meese favored abolishing USTR and had promised Baldrige that he, not Brock, would exercise trade policy leadership. Two years later, Meese and Baldrige persuaded Reagan to propose legislation to implement this proposal by creating a Department of International Trade and Industry which would subsume USTR. Brock fought back, using his policy and political skills and the relationships developed as a former Senator and U.S. Representative. So while Baldrige won battles on several trade issues, USTR ended up winning the war: the department legislation never reached the floor of either chamber, and it was Brock who spoke for the administration as general trade legislation was hammered out in 1984. In particular, USTR’srelationships with Ways and Means and Finance proved ~ r i t i c a las , ~they had a decade earlier when Richard Nixon had proposed to eliminate the office (Destler 1980, 154-55). But the BrockBaldrige rivalry underscored a broader structural anomaly built into executive branch policy-making: the stronger the Secretary of Commerce, the more fractious an administration’s trade policy-making. There was no policy sphere, save trade, that was worthy of a Commerce incumbent seeking Cabinet status in fact as well as form. Yet trade already had a leader, at Congressional insistence: the USTR. Moreover, that agency’s established role as broker of congressional as well as executive (and foreign) interests was a key to the role that legislators wished to play. They could delegate authority to that office, in the expectation that it would be responsive-not just (and often not mainly) to their policy views, but also to their wish to be credited with important roles in the policy process. 8.3.4 Steel (and Copper) 1984

One of the ways Congress delegated trade power was through regulatorytype procedures, through which industries with specific complaints could seek import relief. Most important were the “unfair trade” statutes-the counter-

262

I. M. Destter

vailing duty law through which producers could gain offsets for officially subsidized products; the antidumping law providing similar relief for imports sold at “less than fair value”-and the “escape clause” though which an industry could seek temporary protection from imports causing “serious injury,” fairly or unfairly traded. The early and mid-eighties brought a sharp rise in the number of traderemedy petitions seeking import relief, particularly those alleging unfair foreign practices (see table 8.3). And the prime user of these procedures was the American steel industry. In 1982, exploiting changes in the unfair trade laws, which its congressional supporters had gotten included in the Tokyo Round implementing legislation of 1979, U.S. steelmakers jointly delivered to the Commerce Department, on a single day, 494 boxes containing 3 million pages of documentation for 132 countervailing duty and antidumping petitions, mainly against European Community (EC) exporters. Loud European protests forced a reluctant Commerce Secretary into brokering a deal: the EC agreed to restrain carbon steel exports. And that industry opened 1984 with a new onslaught of petitions, targeted this time at the newly industrializing countries (NIC), whose share of the U.S. market was rising rapidly. At the same time, the industry got 210 representatives to cosponsor a steel import quota bill, and Bethlehem Steel and the United Steelworkers union submitted an escape-clause petition timed to force presidential action before the 1984 election. It was this last factor that led to action. In September 1984, Ronald Reagan was faced with two politically timed escape-clause petitions in response to which the USITC had found injury and had recommended relief. He rejected one from U.S. copper producers, perhaps fortified by the strong resistance of Table 8.3

’kade Cases By Qp,1980-89

Number of Cases Year 1975-79, average 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989

Sec. 201

Sec. 301

Countervailing Duty

Antidumping

8.8 2 1 3 0 7 4 1 0

4.2 0 5 6 7 2 5(4)’ 6(4P 5(1P 71)’ IO(7)”

2.6 11 14 124 31 53 41 29 8 13 7

16.6 29 19 71 45 73 62 71

1

0

Source: USTR (1990) and USITC. ‘Numbers in parentheses denote number of USTR self-initiated cases.

15

42 23

263

U.S. Trade Policy-making

U.S. copper users, who would pay the price of protection. On steel, however, he acceded. Though the president formally rejected the USITC recommendation, he ordered Brock to negotiate export restraint agreements with all major suppliers of steel to the U.S. market (Destler and Ode11 1987, 15-18,43-49). 8.3.5 The Sudden Enactment of Legislation Reagan’s decisions on copper and steel took place during the last major trade policy event of the eighties’ first five years, the enactment of the Trade and Tariff Act of 1984. Unlike other recent comprehensive trade laws, this one neither authorized a new multilateral trade negotiation nor acted on the results of one. The action began when the Senate Finance Committee, at Brock’s encouragement, reported out a bill that stitched together several loosely-connected proposals: an extension of United States law providing trade preferences (GSP)to developing countries, a necessary if unpopular mecsure because existing authority expired at year’s end; an authorization for the administration to negotiate a bilateral free-trade agreement with Israe1,’O a popular measure quite marginal to broader trade concerns; and several other measures, including a Danforth “reciprocity” bill (forerunner of Super-301) seeking access to foreign markets equal to that which foreign producers of like-products had to U.S. markets. Danforth brought this package to the Senate floor in September, without time- or amendment-limiting agreements, and it survived a chaotic week. It emerged encumbered with protectionist amendments for producers of a range of products, including wine, copper, shoes, and ferroalloys. The House followed quickly with its own legislation, which included a bill sponsored by Ways and Means Chairman Dan Rostenkowski (D-Illinois) to support and enforce President Reagan’s newly established steel protection program. Brock then managed, working in conference with Danforth and Rostenkowski in particular, to get the bill transformed, in the words of the Washington Post (12 October 1984), into “pretty respectable legislation. . . . Most of the bad stuff got thrown out and all of the good stuff stayed.” In an atypically chaotic way, this legislation confirmed the old pattern: Congress would threaten product-specific restrictions, perhaps even pass them in one house (particularly the more open Senate), but it would seldom enact them into law. In the end, members would content themselves with measures that delegated authority to the executive, albeit with strong general signals about how that authority should be used. Even the steel title of the bill was consistent with this practice: what it did was to provide enforcement authority, and rhetorical support, for action the president had already taken.

8.3.6 Macro Policy: Sowing the Wind But through the first Reagan term, the trade balance was rapidly growing worse. The causes were (and are) well known, and they had nothing to do

264

I. M. Destler

with trade policy. There was the Latin debt crisis that curtailed exports. There was the economic recovery-earlier and stronger than that in major trading partners-that pulled in imports. There was, above all, the remarkable rise of the dollar with impact on both sides of the balance. As calculated by the IMF’s Multilateral Exchange Rate Measure ( M E W ) , the dollar’s overall, tradeweighted international value rose 63 percent between its average for 1980 and its peak in March 1985, And while no explanation of this rise is fully satisfactory, most experts saw an important cause in the unusual U.S.macroeconomic policy mix of this period-tight money to fight inflation, large fiscal deficits to stimulate growth. The effect was reinforced because policy in Europe and Japan was moving in the opposite direction (Marris 1985, chaps. 1,2). The new U.S.policy mix-widely labeled “Reaganomics”-began with the tax cuts and defense increases of 1981, their impact persisting as the decade reached midpoint. A “conspiracy” between certain congressional and administration leaders had brought some remedial action (Stockman 1987), but as the Reagan recovery gathered force, the president’s resistance to such countermeasures stiffened. This increased the reluctance of his close advisers to press such measures, particular on the revenue front. When 1984 brought Reagan both a 49-state election landslide and a real GNP growth rate of 6.8 percent, the highest since 1955, his macroeconomic formula became particularly hard to change. It was natural for him to conclude that he had been right all along. But as a by-product of his policies, 1984 also brought the first 12-digit trade deficit in any nation’s experience: exports at $219.9 billion, imports at $332.4 billion. And these were nominal figures: the strong dollar meant cheap imports, and hence pressure on domestic industry even fiercer than the value statistics suggested. Measured in constant (1982) dollars, merchandise imports shot up in proportion to domestic output: from 18.8 percent in 1980 to 23.3 percent in 1984 (see table 8.2 above). The comparable ratio for exports declined from 18.0 to 14.8 percent. For manufactured goods, the changes were greater-from 19.7 to 29.2 percent for imports, and from 25.5 to 18.4 percent for exports (Destler and Henning 1989, 120-21). The causes lay outside the trade policy arena, but the effects were felt within it. Exporters were frustrated; those competing with imports were up in arms. The second Reagan administration would reap the whirlwind. Trade issues had been rising in visibility through Reagan I, yet at its end Congress could pass a relatively unaggressive trade bill hardly mentioning Japan. Things changed rather quickly thereafter. 8.4

1985-89: The Second Five Years

8.4.1 Macro: Reaping the Whirlwind Before moving to the events of 1985-89, let us recall our initial characterization of executive-congressional politics on trade. In essence, the picture

265

U.S. Trade Policy-making

painted was one of a positive-sum political game: legislators got leeway to play issues so as to enhance their status as policy politicians; executive leaders, through adroitness in dealing with these legislators and in helping them manage trade-political pressures, retained the final word, the leeway on specific product decisions. And since they cared more about the final outcomes, this was a good deal for them too. It was also a good deal for the United States, at least from the liberal-trade perspective. Executive leaders typically used their leeway to tilt policy toward trade expansion, toward barrier reduction, toward limiting concessions to protection-seeking claimants and balancing these with progress toward barrier reduction. But to make such a system work, administration leaders had to be sensitive to the domestic forces at play. That is why Congress created, and regularly protected and strengthened, the office of the presidential trade representative, (U)STR. Imagine, then, a situation where the pressures on Capitol Hill multiply but the White House no longer seems interested in playing the game, no longer responds credibly to these pressures. This was what Congress confronted in early 1985. The result was a political explosion, a quick build-up of the drive that led ultimately to the first congressionally initiated piece of major trade legislation since Smoot-Hawley, the Omnibus Trade and Competitiveness Act of 1988. When the 98th Congress adjourned in October 1984, the sense among trade specialists was that the legislative decks had been cleared, and the 99th Congress would turn its attention to other issues. But several things happened, in quick succession. First, the trade statistics for 1984 were released, and they showed a deficit of $107.6 billion-or $123.3 billion with imports measured cif (cost/insurance/freight),as Congress had mandated. The 12-digit outcome had long been foreshadowed by the monthly statistics, of course, but the final confirmation was still a jolt for those members who had not been watching closely and a spur and rationale for action for those who had been. Second, the Reagan administration announced in early March that it would not ask Japan to renew restraints on auto exports. Third, with congressional tradeoutrage growing, particularly in the Senate, and with a key April 1 deadline approaching in telecommunicationsnegotiations with Japan, the White House announced on March 20 that Trade Representative Brock, who seemed finally to have won intra-administrationtrade primacy, was being made Secretary of Labor. Nothing was said about his replacement. Finally, a few days later, Japan’s Ministry of International Trade and Industry announced that it would continue voluntary auto restraints, but at a level 25 percent above that of 1984. This was seen as a rebuff by both the administration-which wanted no restraints-and most members of Congress, who saw the quota increase as egregious. All this precipitated a virulent-if largely rhetorical-Capitol Hill reaction. Amidst bipartisan denunciation of the Japanese, the Senate passed by 92 to 0 a Danforth resolution calling for trade retaliation against Japan unless

266

I. M. Destler

U.S. exports expanded by enough to offset the expected increase in imports of Japanese cars. The House followed, 394-19, with a similar measure. None of this was binding: the Senate Finance Committee did endorse on April 4 a mandatory version of the Danforth resolution, but it did not move this to the floor. The atmosphere had clearly changed. In July, three respected DemocratsChairman Rostenkowski of House Ways and Means, future Chairman Lloyd Bentsen of Senate Finance, and rising Democrat Richard Gephardt-introduced the forerunner of the notorious “Gephardt amendment,” a measure that would have imposed an import surcharge on countries running large trade surpluses with the United States. Like the March Senate resolution, this was intended to send a message, not make new law.” The dollar had peaked in late February, but it remained very strong. Tradedgoods producers saw a bad situation getting worse and got no solace from a White House spouting euphoria about how “America was back.” Members of Congress responded to their pressure, blended with outrage about administration obliviousness signaled by the abandonment of auto quotas and the transfer of Brock.I2 Democrats, particularly in the House, saw potential for partisan advantage. Republicans got tough partly to limit this potential, partly because Reagan’s election victory had freed them-and their business allies-from the need to mute attacks on the president’s trade policies. 8.4.2 The Administration’s Two-Track Response Through most of the summer, the words from the administration remained the same. But beneath the surface there was movement. Reagan had a new Treasury Secretary, James Baker 111, who had traded jobs with Donald Regan, now White House Chief of Staff. Baker and his deputy, Richard Darman, were quietly laying the groundwork for a major change in the administration’s posture on the exchange rate. Nonintervention and benign neglect had been the rule under Regan and his Under Secretary for Monetary Affairs, Beryl Sprinkel. Consulting carefully with Fed Chairman Paul Volcker and Secretary of State George Shultz, Baker moved to make dollar decline a major, and visible, international economic policy goal of the United States.l3 At the Plaza Hotel on September 22, Baker and his G-5 colleagues declared themselves in favor of “further orderly appreciation of the major nondollar currencies against the dollar,” and the central banks backed up this declaration by selling dollars in foreign exchange markets. Baker moved on the dollar at his own initiative, never raising the issue explicitly for interagency or presidential review. During that same summer, however, he was presiding-in the administration’s Economic Policy Council-over development of a more aggressive approach to trade policy. This was dramatized by a “fair trade” speech delivered by Ronald Reagan at the White House on September 23, announcing determination to pursue a range of unfair trade practice cases against Japan, Korea, Brazil, and the EC, and

267

US.Trade Policy-making

creation of a “strike force” under Commerce Secretary Baldrige to uncover and-attackpractices that barred U.S. products from foreign markets. The two-track strategy combined economic and political logic. Decline of the dollar would bring first the hope, then the reality, of improvement in the trade balance. Because of the J-curve effect, the impact on the nominal balance would be delayed, but the effect on the real balance-the volumes of exports and imports-would be greater than the dollar statistics showed. Since this real balance was so important for trade politics, the administration could expect better days-if it could hold off pressures in the meantime. It would do so through its new aggressiveness, including-to use the common, barbaric phrase-official “self-initiation” of unfair trade practice cases. It would also accelerate its campaign for a new, multilateral trade negotiating round. But it would resist the congressional push for trade legislation-at least until dollar decline could do its work. It was hardly new for an administration to employ export bargaining as a means of diverting pressure to restrict imports, nor was it new for Congress to support such an effort. Indeed, the post- 1985 administration approach had its roots in a long political tradition, dating from Cordell Hull and the Reciprocal Trade Agreements Act of 1934. But with the growing pressure and the deterioration in trade balances, it assumed new forms. 8.4.3

The Semiconductor Agreement

Among the “unfair trade” issues given new priority, none proved more important-or more complicated-than semiconductors. This product had been independently invented by two Americans in the late 1950s and had received a major early boost from official defense and space programs. U.S. firmsbased particularly in California’s Silicon Valley-had led through the 1970s as hundreds, then thousands of electronic functions were crowded onto the tiny chips that became indispensable for computers, telecommunications,and many other advanced industrial products. But with government help and encouragement, Japan’s integrated electronics firms began making major inroads, particularly with the mass-produced, standardized DRAMS. By 1983, the California-based Semiconductor Industry Association (SIA) was publishing a report entitled The Efect of Government Targeting on World Semiconductor Competition: A Case History of Japanese Industrial Strategy and Its Costs for America. This contrasted the rise in Japanese exports with the low and static share (around 11 percent) which U.S. producers had in the Japanese market. In July 1985, its members hit by growing Japanese competition and a severe slump in overall demand, SIA filed a Section 301 case claiming that this import resistance, encouraged by past government action, was an “unreasonable” barrier to U.S. trade. In the months thereafter, antidumping cases were submitted by individual U.S. firms and by the Secretary of Commerce. There were conflicting interests within the U.S. industry, of course-major chip

268

I. M. Destler

users, like IBM, benefited from high-quality, low-cost Japanese inputs. Nonetheless, the industry case as a whole won atypically broad political support. Anxieties about U. S. technological leadership, from private industry to the Pentagon (U.S. Department of Defense 1987), made support of the industry’s case far broader than would have been received by a “low-tech” industry of comparable size. There followed a year of complex-and fractious-negotiations. These culminated in a unique “three-market” trade agreement. To halt dumping in the United States, the two governments adopted a system of minimum prices and reporting of sales by Japanese firms. There was also a somewhat looser system of price-monitoring in third-country markets that was aimed at protecting U.S. exports against dumping there. Last but certainly not least was a commitment to increase U.S.-based producers’ share of the Japanese market, featuring a side letter from Japanese officials declaring, in the words of one U.S. negotiator, that “they understood, welcomed, and would make efforts to assist the U.S. companies in reaching their goal of a 20-percent market share within five years” (Prestowitz 1988,65). The arrangement had some political and economic logic. To act only against dumping in the U.S. market would mean higher-priced chips there than elsewhere, undercutting IBM and other producers of computers and other downstream products. Moreover, if the Japanese firms-NEC, Hitachi, Fujitsu, and so on-were able to restrict foreign access to their very large home market, it would be hard for U.S. firms to hold their own. Nor did these firms wish for a two-market agreement that would encourage low-priced Japanese sales in the rest of the world. The deal completed was responsive to all of these concerns. But it was also in contradiction to itself. The minimum (nondumping)prices in the United States were based on each Japanese firm’s cost of production, and they were periodically updated. This encouraged large production runs which brought down per-unit costs. But these led to oversupply and falling prices within Japan (the only market where the agreement did not set minimum prices). This undercut U.S. firms’ sales there. It led also to a “gray market” in third countries, as chips were flown out of Japan in suitcases for low-price resale there. The obvious way to combat this was for Japan’s Ministry of International Trade and Industry (MITI) to encourage the Japanese firms to limit production, but when the Japanese had suggested this during the negotiations, both SIA and the U.S. government had opposed it-SIA because of concerns of its user members about supply shortages, U.S. officials because it was contrary to free-market principles. In any case, by the end of 1986 SIA was complaining that both the Japanese and third-country market provisions of the agreement were being violated. U.S. officials expressed growing concern to Tokyo, warning of retaliation. By early spring, MITI was pressing the Japanese firms to cut output-and this would soon drive the price up in Japan and take most of the profit out of the gray market. But before this could be effective, U.S. patience ran out. Pressed

269

U.S. Trade Policy-making

by a Congress considering omnibus trade legislation, President Reagan announced in March 1987 that he would impose sanctions because of “Japan’s inability to enforce” the agreement. In April, he imposed punitive (100%) tariffs against Japanese imports equivalent to the estimated sales lost to U.S. firms in the Japanese market and in third countries. These produced shock and headlines in Tokyo: “sanctions against an ally,” the first such U.S. action against Japan since World War 11. Later that year, with the third-country dumping resolved, Reagan removed a portion of the penalties. But the major share, the sanctions aimed at loss of anticipated U.S. sales in Japan, continued through the decade. So did Japanese dominance in the DRAM market. 8.4.4 Bilateralism: The Free-Trade Agreement (FTA) with Canada

Over this same period, the United States was also negotiating a different sort of departure from trade multilateralism: a bilateral FTA eliminating most remaining barriers to trade with Canada. Members of Congress, frustrated with the slow-moving General Agreement on Tariffs and Trade (GATT) process, had shown growing interest in “alternative strategies” for opening foreign markets (Schott 1988, 11). Reagan’s first trade representative, William Brock, was also interested in bilateral approaches as both an alternative and a prod-a threat to other U.S. trading partners that could help get multilateral talks moving. Hence the 1984 trade law included, in its title authorizing a free-trade pact with Israel, a rather cumbersome procedure by which the administration could negotiate an agreement “with any country other than Israel,” subject to a veto by either of the major trade committees and ultimate approval under the fast-track procedures. The main impetus, however, came from north of the border, where the new Conservative government reversed long-standing Canadian policy in 1985 and sought to negotiate a bilateral free-trade agreement with the United States. President Reagan agreed and gave Congress, in December, the required notification of his intention to proceed. And after a sharp debate in the Senate Committee on Finance, fueled by concern over softwood lumber trade and broader frustration with administration trade policy, a resolution of &upprovul, which would have blocked the effort, failed by a 10-10 vote. After some appeasement of affected interests, negotiations proceeded-somewhat lethargically in 1986, then frantically in the summer of 1987 with the approach of the October 3, 1987, statutory deadline for notifying Congress of the basic substance of the agreement.14 The actual agreement was signed on 2 January 1988, but the President’s implementing bill was not formally submitted until July. The reason was extensive Congressional participation in its drafting, following the precedent of the Trade Agreements Act of 1979, which implemented the Tokyo Round accords (Destler 1986, 62-67). In any case, Congress ended up approving the Canada agreement overwhelmingly-by 366-40 in the House and 83-9 in

270

I. M. Destler

the Senate. In Canada, however, final approval awaited the results of a national election fought mainly over whether that nation should take what was regarded as a historic step (Leyton-Brown 1990,28-31). The agreement went into effect on 1 January 1989. The U.S.-Canada FTA “eliminat[es] tariffs and most import and export measures . . . by the end of the 1990s” (Morici 1990, l), after a 10-year phase-in period. Since Canada is the largest single U.S. trading partner, the agreement represents a significant new step in U.S. trade policy, one inconsistent-on its face-with the GATT most-favored nation principle. In practice, the inconsistency is modest and manageable for two reasons. First, a large amount of the trade between the two nations had flowed freely prior to the agreement. Second, some of its provisions-on dispute settlement and trade in services, for example-were “useful models for new GATT rules . . . in the Uruguay Round” (Schott 1988,34). 8.4.5 Omnibus Legislation: Congress Seizes the Initiative

As both the semiconductor dispute and the Canada agreement were proceeding in early 1986, the House was acting on broader trade legislation. Speaker Thomas P. (‘‘lip’’) O’Neill had already, with some exaggeration, declared it “the number one issue . . . based on what I hear from members in the cloak room” (Washington Post, 19 September 1985). Among his colleagues were a number who saw an ideal partisan issue (toughness and jobs). Ways and Means leaders Rostenkowski and Gibbons were not among them, but given general party sentiment they needed to get moving lest jurisdiction be shifted elsewhere. By May the House had voted 295-1 15 in favor of an “omnibus” trade measure, which curbed presidential discretion in trade remedy cases, required retaliation when countries did not open their markets, and imposed quotas on countries running large bilateral surpluses with the United States (the Gephardt amendment). The White House denounced it as “ominous,” “rankly political,” “pure protectionism.” The last it was not-in fact, it avoided direct imposition of trade barriers. But it revised trade relief statutes in ways that would have led indirectly to the same result. The Republican-controlled Senate did not follow that year, mainly because the Finance Committee was preoccupied with tax reform. It was September before they could markup a trade bill, and with time short and the administration opposed, consensus proved impossible to achieve. That same month, U.S. Trade Representative Clayton Yeutter successfully headed a U.S. delegation that succeeded in winning agreement to begin the Uruguay Round. Hence administration strategy seemed successful: the dollar was still declining, and legislative action had been deferred while exchange rate change had time to do its work. But the 1986 election brought Democrats back into control of the Senate. This meant that both Houses would move on trade in the 100th Congress. The reason was not party differences on substance: in the Senate, trade activism

271

U.S. Trade Policy-making

was bipartisan. But the Republican Senate leadership of 1981-86 saw its job as helping the White House-other things being equal. Their Democratic successors were interested, other things again equal, in scoring points against the White House. Trade was an obvious issue opportunity, the most developed major issue available. So even as the election returns were coming in, Senator Robert Byrd (D-West Virginia), soon to be recrowned as Majority Leader, declared that omnibus trade legislation would have top priority.

8.4.6 The Omnibus Trade and Competitiveness Act of 1988 The House had to go first, however. Trade bills remained revenue measures, at least marginally, which the Constitution ordained that the House must originate. Tip O’Neill’s successor, Jim Wright, was more than willing, and he organized and led a multicommittee hearing and markup process. The bill passed in 1986 was reintroduced as HR 3, and then parceled out for reworking to 1 1 House committees, with Ways and Means first among equals. Seeing the handwriting on the wall, the administration now agreed that omnibus legislation would be useful. In particular, it needed, for the Uruguay Round, extension of fast-track negotiating authority, due to expire in January 1988. But though the administration presented, in late February, its own “competitiveness” legislation, the House moved ahead on the basis of HR 3. By April the leadership was merging the 1 1 separate committee proposals. Wright insisted on excluding product-specific measures: textiles would get a separate vote later, and an Energy and Commerce-reported measure to limit imports of high-quality digital tape recorders was excised from the omnibus legislation. The revised HR 3 passed the House, 290-137, on April 30, with 43 Republicans joining virtually all Democrats. Included along with standard trade policy measures-authorizing for the Uruguay Round, numerous toughenings of trade-remedy laws, strengthening of the authority of the U.S. Trade Representative (USTR), and so on-were provisions on exchange rates and Third World debt, worker retraining, relaxation of national security export controls, agricultural and broader export promotion, and math, science, and foreign language education. (The “omnibus” in the title was not without meaning. ) Included also was the Gephardt amendment requiring import barriers against countries with large bilateral trade surpluses if they did not reduce them. Unlike in 1986, Ways and Means had not included this in the provisions it had reported-part of a broader effort to produce a more moderate bill. This time, members knew they might actually be writing a law, and they tempered their actions accordingly. But Gephardt won inclusion of his amendment by a floor vote of 218-214. The victory helped his presidential campaign, while the narrow margin signaled the unlikelihood of the amendment’s inclusion in final legislation. Meanwhile, the Senate began its work, with new Finance Committee Chairman Lloyd Bentsen in the lead. By early May, his committee had re-

272

I. M.Destler

ported a bill with broad (19-1) bipartisan support, one which excluded the Gephardt amendment, but, among its many provisions, added new procedural conditions on “fast-track’’ authorization for the Uruguay Round, mandated retaliation against unfair foreign trade practices, and curbed presidential discretion in escape-clause cases. Majority Leader Byrd combined this with bills reported by eight other committees and brought it to the Senate floor in June. After a month of debate, it passed by 71 to 27. It included, in a floor amendment, what would, after conference reworking, become known as “Super301”-a provision requiring the U.S. trade representative to name and target countries maintaining patterns of import barriers and unfair, market-distorting practices. It also included a requirement that companies with more than 100 employees give at least 60 days’ notice before plant closings, and a ban on all imports from Toshiba Corporation and a Norwegian defense company, a provision also added on the floor after reports surfaced of their sales of important defense-related equipment to the Soviet Union. In general, this bill-like its House counterpart-was considerably less restrictive in its likely effects than the omnibus bill of 1986. Moreover, the House bill was more restrictive on some matters, the Senate on others, suggesting that the final result could prove more liberal than either. Administration trade officials were considerably more engaged in the legislative process in 1987 than they had been in 1986-though they were certainly less influential at this stage than their counterparts of 1973-74, when the last major negotiations-authorizing bill was under consideration. And White House comments as the bills made their ways forward were certainly more moderate than the denunciations of 1986. Still, there was administration criticism of many specific provisions, and U.S. Trade Representative Yeutter and Treasury Secretary Baker urged Senate Republicans to vote against final passage in order to strengthen the prospects for change in conference. Each house had passed a bill roughly 1,000 pages in length, with the details above giving just a flavor of some of the more important provisions. To reconcile the two, a 199-member conference committee was appointed-44 Senators and 155 Representatives. They were subdivided into 17 subconferences responsible for separated sections. Before any of them met, however, Black Monday (19 October 1987) precipitated a financial (and economic policy) crisis, and in the weeks that followed, leaders of the key committees were preoccupied with the White House-Congressional summit on the budget deficit. (The stock market plunge also made members wary of taking any quick trade action, fearing that it would be labeled “protectionist” and would contribute to deepening of the crisis.) As 1988 began, Rostenkowski and Bentsen-chairs of the two key committees-signaled that they were seeking a bill that the president would sign. This meant compromise with the administration. The conference subgroup they led grew seriously engaged in February and March, with administration officials actively involved. Danforth said with only some exaggeration: “This is not a conference between the two houses. It’s a conference between Con-

273

U.S. Trade Policy-making

gress and the administration” (Congressional Quurrerly Almanac 1988, 2 16). Virtually all wade policy issues were resolved by the end of March, with a rewritten version of Super-301 supplanting the Gephardt amendment, numerous specific authorities transferred from the president to the U.S. trade representative, but with executive flexibility maintained on the escape clause and the imposition of trade sanctions. The final drama was supplied by the plant-closing provision. Following frenetic April negotiations with the administration-and with organized labor-the conferees retained the mandatory notification provision. This was strongly opposed by organized business and thought to ensure a presidential veto. Reagan did in fact veto the omnibus bill on May 24, after both houses had approved the conference report. He cited in particular both the plantclosing provision and a formerly obscure prohibition of certain oil exports from Alaska. However, the veto message did not even mention the major trade provisions of the bill. The House voted to override the veto; the Senate fell five votes short. A new bill was prepared, identical except for the excision of plant-closings and Alaskan oil. To appease organized labor, a separate plant-closing bill was also introduced and scheduled for consideration in advance of omnibus trade. To Democrats’ delight, the issue and the bill caught fire. The administration was put on the (antiworker) defensive, as both houses passed it by lopsided margins. To put an end to that issue, Reagan let that bill become law. In the meantime, the House and then the Senate were passing the slightly slimmer trade bill by large bipartisan majorities. On August 23, with Reagan’s signature, it became Public Law 100-418, the Omnibus Trade and Competitiveness Act of 1988. By this time, the trade imbalance had finally begun to improve. This fact made it easier to remove most of the bindingly restrictive provisions from the bill, though the fact the improvement was relatively modest and so long in coming reduced the improvement’s positive political effect. By this time also, the Gephardt presidential campaign had come and gone: using a TV spot on Korean auto import restrictions to win the Iowa caucuses, but losing after this precipitated a backlash and-what is more important-his campaign ran short of money in the run-up to the “Super Tbesday” primaries in the south. The trade bill was-finally-law. But the enormous legislative effort had brought only a modest reshaping of American trade law. It tilted the U.S. posture toward greater aggressiveness on exports, with Super-301 the most visible manifestation. It gave the executive-the USTR in particular-more deadlines to meet. It extended fast-track authority to cover the Uruguay Round negotiations. It contained countless other specific provisions. But even where it seemed to bind the executive, as on mandatory retaliation for “unjustifiable” foreign trade practices, a closer reading showed that the president and the U.S. trade representative had been left with some flexibility, some window of escape.

274

I. M. Destler

8.4.7 Textile Success-and

Failure

Also under consideration throughout this period was the legislative proposal of that venerable protection claimant, the US textile industry. Throughout the postwar era, industry leaders had shown skill in pressuring both ends of Pennsylvania Avenue. Mill executives worked directly with Congress and the White House and through their American Textile Manufacturers Institute. They would win action commitments from presidential candidates of both political parties: John F. Kennedy in 1960, Richard Nixon and Hubert Humphrey in 1968, Ronald Reagan in 1980. They would show, on Capitol Hill, a capacity to block or threaten trade liberalizing legislation-from the Eisenhower administration in the late fifties to the Carter administration in the late seventies. Sometimes they would spring a “Hollings amendment” on the Senate floor-as in 1968 and 1978-and win on a lopsided record vote. But they seemed to understand that Congress would not, in the end, enact a bill restricting textile imports. Product-specific legislation was outside the post-1934 rules of the game, and there were other ways that members could gain political credit for championing the industry’s cause. So, working with the White House, legislative leaders would find ways to sidetrack quota legislation (Destler 1986, 27-28, 61-62). But if the industry could establish a de fact0 veto power over general trade legislation, especially measures to authorize international trade negotiations, leaders of both branches would respond by strengthening negotiated textile protection. Hence the Long-Term Arrangement on Cotton Textiles initiated by George Ball in 1961 cleared the way for the Trade Expansion Act of 1962, and the Multi-Fiber Arrangement (MFA) of 1973 secured industry acquiescence in the Trade Act of 1974. By blocking legislation needed to complete the Tokyo Round in 1978, moreover, industry leaders won a tightening of bilateral agreements with major East Asian exporters. In the mid-l980s, the industry began by employing a similar strategyusing the Congress not for final action but to win concessions in the executive arena. Unlike in the 1970s, imports were now a really serious problem for the industry, particularly the apparel producers. After slow growth in the 1960s and 1970s, the importkonsumption ratio for textiles and apparel rose from 12.1 percent in 1980 to 22 percent in 1986, and from 18.4 to 31.1 percent for apparel alone (Cline 1987, 49). The vehicle to mobilize pressure for a response was the Jenkins bill, whose formal objective was to impose statutory quotas on textile imports but whose aim was to toughen the U S . position in negotiations for renewal of the MFA. In the broader climate of trade discontent, the bill gained nearly 300 House cosponsors, though on the actual House record vote about 30 backed off, so it passed by “only” 262 to 159. After favorable Senate action and the expected Reagan veto in December 1985, the industry’s House backers developed a new twist. Rather than following the standard practice of going for an override vote as soon as possible

275

U.S. Trade Policy-making

after the bill was returned, they arranged for such a vote to be scheduled eight months later, when the MFA talks were to be concluded. This had the obvious benefit of keeping the heat on USTR negotiators, and once that did achieve a marginal stiffening of the U.S. position, the override effort failed, though this time 276 members went on record in support of the industry’s cause. Moreover, since the textile legislation was considered ahead of the omnibus legislation, the industry retained, through 1986, the possibility of once again using its power in the Congress to extract a price for enactment of the latter. But in 1987-88 the industry overreached, and seemed to forget the formula for its success. Their appetites perhaps whetted by coming so close (at least formally) to statutory quotas, textile executives first moved to punish the Reagan administration by backing Democrats in the textile states in the 1986 midterm elections. In 1987, they continued to push their bill separately. In the House they had no choice: Speaker Jim Wright insisted on keeping productspecific measures out of the omnibus trade bill, offering the consolation prize of a clear House vote after the general bill had been passed. In the Senate, however, where industry leaders could have exploited the chamber’s more open rules by seeking to attach their bill as a rider to the omnibus legislation, they did not do so, settling instead for a similar procedural deal. The result for the industry was a series of hollow legislative victories. As promised, in July 1987 both Ways and Means and Finance reported the industry bill for floor action, without recommendation. The House passed it in September 1987; the Senate did not act until the following September, when it voted in favor of a version that Hollings had recast to add some appeal to farm-state senators. Rostenkowski, who opposed the bill throughout, wanted to go to conference, to delay, and perhaps to kill it. But he reluctantly deferred to the wishes of Democratic colleagues who wanted to force a Reagan veto so they could use the issue in the election campaign. So the House accepted the Senate bill, Reagan vetoed it, and the House failed by 11 votes to override. Again, strong (slightly under two-thirds) majorities in both houses were able to go on record for textile protection, secure in the expectation the bill would fall. But because they held nothing else hostage, textile interests ended with the worst of both worlds: their bill died and the omnibus bill became law without a textile quid pro quo. In contrast to earlier periods, textile leaders made the mistake, apparently, of thinking they could actually get Congress to legislate specific barriers to trade. As a result, they got no additional protection in 1987-88, the first time in three decades that the industry had failed to exploit an omnibus trade bill effectively. 8.4.8 The Bush Epilogue The Bush administration entered office with the task of implementing the new Omnibus Trade and Competitiveness Act of 1988. U.S. Trade Representative Carla Hills found she not only had to energize the Uruguay Roundscheduled to conclude in 1990-but to decide, by the end of May 1989,

276

I. M.Destler

whether to name specific countries as general trade offenders under Super301. Her response was, for the most part, an adroit balancing. Taiwan and Korea were encouraged to make substantial trade concessions to avoid Super301 designation. Japan was named a “priority foreign country”-along with Brazil and India-but in a way that mainly targeted export issues the United States was pushing anyway. Broader Japanese import resistance was considered in a separate negotiation-the Structural Impediments Initiative-which focused also, at least in form, on the trade-related structural problems of the U.S. economy. On steel, Bush had to decide whether to negotiate an extension of export restraint agreements, a decision foreshadowed by a Pennsylvania campaign promise that he would do so. But in the execution of that promise he tilted to the liberal side, easing the bite of the restrictions and promising to phase them out by 1992. But if he was proving antiprotectioniston micro issues, his macroeconomic approach seemed destined to make things worse. For his “no new taxes” pledge precipitated a further retreat from budgetary responsibility in both branches. That, plus the 1988-89 resurgence of the dollar, made it likely that the trade deficit would persist, and perhaps rise again, in the early nineties.

8.5

Conclusions

How did United States trade policy fare during the 1980s? This paper comes to three basic conclusions. Congressional trade activism multiplied. U.S. trade protection increased marginally. And trade protection imposed directly by Congress increased hardly at all. The first of these is beyond dispute. The other two points deserve some elaboration. The increase in American trade protection came mainly in the first five years and largely in the form of “voluntary” export restraints (VERs); this chapter treats the central episodes, involving automobiles and steel, in section 8.3. As shown in table 8.3, the use of the trade remedy procedures for import relief shows a similar pattern-up very sharply through 1985, beginning to recede thereafter. And finally, aggregate assessments of the coverage of U.S. nontariff barriers (NTBs) show a rise during this period. Such assessments face daunting methodological problems: the standard measure, the proportion of a nation’s trade affected by NTBs, has the perverse effect of weighting moderate barriers more heavily than severe ones: when Japan relaxes (increases) her beef import quotas, for example, she increases the share of her total imports subject to NTBs. Nonetheless, several sophisticated attempts have been made both at cross-national comparisons and for the United States alone. Bela Balassa and Carol Balassa (1984, 187) found just 6.2 percent of U.S. manufactured imports subject to visible quantitative restrictions in 1980; in

277

U.S. Trade Policy-making

1981-83, an additional 6.52 percent of U.S. imports came under such restraints.15Gary Clyde Hufbauer and his colleagues (1986, 21) calculated that “U.S. imports covered by special protection,” including high tariffs as well as quantitative restraints, rose from 12 percent of total imports in 1980 to 21 percent in 1984. Other recent studies (Nogues, Olechowski, and Winters 1986; Laird and Yeats 1988, Stoeckel, Pearce, and Banks 1990) come to similar conclusions. The Balassa and Hufbauer estimates seem, at first glance, to indicate very substantial increases in U.S. protection: more than a doubling by the first measure. But on closer examination, the increase is due almost entirely to a single-albeit important-trade policy action, the Japanese VER imposed in 1981 on auto exports to the U.S. market. Without the $29 billion in imports this covered in 1984, the Hufbauer measure would have remained at 12 percent, and the Balassas’ measure of the increase in U.S. protection would have been under 1 percent. The auto VER remained technically in force through the decade, but it had lost its bite by the late eighties. It had been enlarged by 24 percent, and the combination of the weaker dollar and increased production by Japanese firms on American soil made it increasingly difficult for the quotas to be filled. All things considered, therefore, it seems fair to characterize as “marginal” the increase in U.S. protection during the eighties. More than marginal, perhaps, was the rise in U.S. aggressiveness in pressing for opening of others’ export markets. But that was still within the long-standing U.S. tradition of pushing trade expansion as a political counterweight to the forces of import limitation. Finally, what was the role of Congress during this period. Though it is sometimes labeled “protectionist,” the thrust of the Omnibus Trade and Competitiveness Act of 1988 is to open foreign markets, not close American ones. In the development of that legislation, and on trade policy more generally, Congressional behavior seems to have been broadly consistent with the pattern depicted in section 8.2 of this paper. Legislators became publicly engaged. They went on record in all sorts of ways; they forced the administration into a more aggressive international bargaining posture. But they did not-in the end-change American trade policy more than very slightly along its most important dimension, the openness of the U.S. market to imports. And they did not themselves impose specific trade restraints on behalf of specific clients. Had members sought above all to influence specific trade barriers through legislation, they would have responded differently, for the eighties offered an ideal opportunity. The trade imbalance grew to unheard-of proportions; executive leaders were discredited; anxiety about the Japanese-and other foreign economic rivals-was waxing, as was concern about American decline. They did not do so. As policy politicians, they increased their activity and visibility, and their identification with broadly popular general postures of

278

I. M. Destler

trade toughness. They forced the administration’sposture to change in that direction. But when they legislated, they did so in ways that muted their direct responsibility for consequences. They retained flexibility and avoided blame. And, as the decade ended, the trade policy buck continued to stop not on Capitol Hill, but at the White House and the USTR.

Notes 1. In 1980, U.S. trade in manufactured goods was $19 billion in surplus, having rebounded from a then-record deficit of $6 billion in 1978 (US. Council of Economic Advisers 1983, 280). 2. For an elaboration of this argument, see my comments elsewhere (Lawrenceand Schultze 1990,207-14). 3. One result was the US.-Japan Semiconductor Trade Agreement of 1986. This became a textbook illustration of how what seemed logical, even necessary, politically proved contradictory economically, as certain provisions of the agreement undercut others and contributed importantly to Japan’s “failure to implement” that agreement and the U.S. imposition of sanctions in the spring of 1987. 4. This view resembles, in some respects, that set forth by Pastor (1983). 5. For examples, see Shepsle and Weingast (1984), Shepsle and Weingast (1987), and McCubbins and Schwartz (1984). All are sophisticated in their treatment of congressional motivation, and the “fire alarm’’ concept of legislative oversight is certainly applicable to trade policymaking. But all assume that the goal of legislator’s policy-related activity is impact on policy outcomes, on what government actually does. 6. Since the Tokyo Round negotiations of the 1970s, the key element in such delegation is that Congress is committed to act expeditiously,without amendment, on legislation proposed by the president to implement an agreement specifically authorized by Congress. This led to expeditious approval of the Tokyo Round agreements in 1979 and of the bilateral free-trade agreements with Israel and Canada in the 1980s. “Fasttrack” authority was also included for the Uruguay Round in the 1988 legislation. 7. Weaver (1988, chap. 2) finds this a powerful explanation of a set of congressional decisions which, by definition, reduce legislators’ direct power over policy outcomes. See also the work of Moms Fiorina. 8. In this case, the administration emulated the Congress in avoiding direct responsibility-unlike on textile or steel voluntary export restraints, there was no bilateral auto agreement. In form, the Japanese acted on their own. 9. For example, when the Meese-created Cabinet Council on Commerce and Trade held its first meeting under Baldrige’s chairmanship, the chairmen of both Ways and Means and Finance reportedly telephoned President Reagan and told him that this was illegal, since Congress had established, by statute, a Cabinet-level Trade Policy Committee chaired by USTR and assigned it overall coordination authority. 10. The key provision, as with all authorizations of trade negotiations since the Tokyo Round, was that Congress would act on the results under the expedited (“fasttrack”) approval procedures noted earlier in this paper. As discussed later, this title also provided the statutory basis for later negotiation of a free-trade agreement with Canada. 11. When I objected in a telephone conversation to the virulence of the Danforth

279

U.S. Trade Policy-making

resolution that March, an aide to a senior Senate Republican responded: “You don’t understand. The target isn’t the Japanese; it’s the White House!” By November, however, Danforth was introducing an omnibus, bipartisan bill with 33 Senate cosponsors, which was seen as a basis for actual legislation. 12. It took more than three months for Brock’s successor, Clayton Yeutter, to be nominated by the president and confirmed by the Senate. And in a final blow to Danforth, Reagan rejected that August a USITC recommendation that the shoe industryimportant in Missouri-receive import relief under the escape clause. The Senator nonetheless refused that fall to vote for statutory shoe import quotas when they were added to the textile bill. 13. On how Baker and Darman prepared the way, see Funabashi (1988, chap. 3). On why they stressed the exchange rate rather than the underlying problem of the budget deficit, see Destler and Henning (1989,4344). 14. The fast-track authority expired on 3 January 1988, and the president had to notify Congress 90 days before signing an agreement. The authority was renewed for the Uruguay Round-and possible new bilateral accords-in the omnibus legislation of 1988. 15. The Balassas avoided the above methodological problem for the change in NTB coverage by measuring the share of 1980 trade for the products brought under NTB coverage in 1981-83.

References Baker, James, 111. 1987. Remarks at the Institute for International Economics. 14 September 1987. Balassa, Bela, and Carol Balassa. 1984. Industrial Protection in the Developed Countries. WorldEconomy 7 (June 1984): 179-96. Cline, William R. 1987. The Future of World Trade in Textiles andAppare1. Washington, D.C. : Institute for International Economics. Cohen, Stephen D., and Ronald I. Metzger. 1982. United States International Economic Policy in Action. New York: Praeger. Cohen, Stephen S ., and John Zysman . 1987. Manufacturing Matters: The Myth of the Post-Industrial Economy. New York: Basic Books. Destler, I. M. 1980. Making Foreign Economic Policy. Washington, D.C.: Brookings. . 1986. American Trade Politics: System Under Stress. Washington, D.C.: Institute for International Economics, and New York: Twentieth Century Fund. Destler, I. M., and C. Randall Henning. 1989. Dollar Politics. Washington, D.C.: Institute for International Economics. Destler, I. M., and John S . Odell. 1987. Anti-Protection: Changing Forces in United States Trade Politics. IIE Policy Analysis no. 21. Washington, D.C.: Institute for International Economics. Funabashi, Yoichi. 1988. Managing the Dollar: From the Plaza to the Louvre. Washington, D.C.: Institute for International Economics. Gilpin, Robert. 1987. The Political Economy of International Relations. Princeton, N.J.: Princeton University Press. Hufbauer, Gary Clyde, Diane T. Berliner, and Kimberly Ann Elliott. 1986. Trade Protection in the United States: 31 Case Studies. Washington, D.C.: Institute for International Economics.

280

I. M. Destler

Kennedy, Paul. 1987. The Rise and Fall of the Great Powers: Economic Change and Military ConJictfiom 1500 to 2000. New York: Random House. Krugman, Paul R., ed. 1986. Strategic Trade Policy and the New International Economics. Cambridge, Mass.: MIT Press. Laird, Sam, and Alexander Yeats. 1988. Trends in Nontarir Barriers of Developed Countries, 1966-1986. Washington, D.C.: World Bank. Lawrence, Robert Z., and Charles L. Shultze, eds. 1990. An American Trade Strategy: Options for the 1990s. Washington, D.C.: Brookings. Leyton-Brown, David. 1990. Implementing the Agreement. In Making Free Trade Work, edited by Peter Morici, 27-59. New York: Council on Foreign Relations Press. McCubbins, Mathew D., and Thomas Schwartz. 1984. Congressional Oversight Overlooked: Police Patrols versus Fire Alarms. American Journal of Political Science 28, no. 1 (February): 165-79. Manis, Stephen. 1985. Deficits and the Dollar: The World Economy at Risk. Washington, D.C.: Institute for International Economics, Policy Analysis no. 14. Mayhew, David. 1974. Congress: The Electoral Connection. New Haven: Yale University Press. Morici, Peter, ed. 1990. Making Free Trade Work: The Canada-U.S. Agreement. New York Council on Foreign Relations Press. Nogues, Julio J., Andrzej Olechowski, and L. Alan Winters. 1986. The Extent of Nontarif Barriers to Imports of Industrial Countries. Washington, D.C.: The World Bank. World Bank Staff Working Papers No. 789. Pastor, Robert. 1983. The Cry-and-Sigh Syndrome: Congress and Trade Policy. In Making Economic Policy in Congress, edited by Allen Schick. Washington, D.C.: American Enterprise Institute for Public Policy Research. Prestowitz, Clyde. 1988. Trading Places: How We Allowed Japan to Take the Lead. New York: Basic Books. Richardson, J. David. 1990. The Political Economy of Strategic Trade Policy. International Organization 44 (Winter): 107-35. Schott, Jeffrey J. 1988. United States-Canada Free Trade: An Evaluation of the Agreement. Washington, D.C.: Institute for International Economics, Policy Analysis No. 24. Shepsle, Kenneth A., and Barry R. Weingast. 1984. “Legislative Politics and Budget Outcomes.” In Federal Budget Policy in the 1980s. G. Mills and J. Palmer, eds. Washington, D.C.: Urban Institute. . 1987. The Institutional Foundations of Committee Power. American Political Science Review 81, no. 1 (March): 85-104. Stockman, David A. 1987. The Triumph of Politics: The Inside Story of the Reagan Revolution. New York: Avon Books. Stoeckel, Andrew, David Pearce, and Gary Banks. 1990. Western Trade Blocks: Game, Set, or Match for Asia-Pacific and the World Economy? Canberra, Australia: Centre for International Economics. Weaver, R. Kent. 1988. Automatic Government: The Politics of Indexation. Washington, D.C.: Brookings Institution. Winham, Gilbert R., and Ikuo Kabashima. 1982. The Politics of U.S.-Japanese Auto Trade. In Coping With US.-Japanese Economic Conjicts, edited by I . M. Destler and Hideo Sato. Lexington, Mass.: Lexington Books. U.S. Council of Economic Advisers. 1983. Economic Report of the President, February 1983. Washington, D.C.: GPO. U.S. Department of Commerce. Bureau of the Census. Historical Statistics of the United States.

281

U.S. Trade Policy-making

U.S. Department of Defense. Office of the Under Secretary of Defense for Acquisition. 1987. Report of the Defense Science Board Task Force on Defense Seminconductor Dependency. Washington, D.C.: Department of Defense. U.S. Department of Transportation. 1981. The US.Automobile Industry, 1980: Report to the President from the Secretary of Transportation, 83-85. Washington, D.C.: Government Printing Office. U.S. House Ways and Means Committee. 1980. Trade with Japan: Hearings before the House Ways and Means Trade Subcommittee. 96th Cong., 2d sess., 18 September. United States Trade Representative. 1990. 1990 Trade Policy Agenda and 1989 Annual Report of the President of the United States on the Trade Agreements Program. Washington, D.C.: GovernmentPrinting Office.

COm~ent

Anne 0. Krueger

I. M. Destler has provided an interesting and penetrating account of U.S. trade policy-making in the 1980s. There are many insights both into the politics of trade (e.g., the USTR’s position vis-his Congress and the administration, and the relationships of the two Houses of Congress to the administration) that will further economists’ understanding of the trade policy process and the economics of trade (e.g., the importance of the fiscal deficit and the behavior of the real exchange rate in influencing trade policy), which are constraints on political behavior. Destler’s essential hypothesis has three parts. First, he maintains that members of Congress are more concerned with appearing to be against protection than they are with outcomes. Second, he believes that there was an enormous increase in political pressures on Congress for activism in the 1980s. Third, he concludes that despite those pressures, because of the first proposition there was little departure from the traditional American free-trade position. The actual formation of trade policy has long been a subject of puzzlement to economists and political scientists alike. To economists, at least historically, the question was, Why, when free trade was obviously so superior on welfare grounds, did governments adopt such protectionist policies? For political scientists, the question has been, Why, when there are such powerful pressures for protection, is there not more of it? Destler’s answer to these questions, given in his three propositions enumerated above, provides one possible answer. There are, however, alternatives. In these comments, I wish to outline one such alternative, in the process commenting on a few of Destler’s observations and subthemes. As a starting point, however, I must note that Destler’s explanation does not presume full Anne 0. Krueger is the A r t s and Sciences Professor of Economics at Duke University and a research associate of the National Bureau of Economic Research.

282

I. M. Destler

rationality of all actors: in particular he believes that those pressuring for protection (including lobbyists who lobby Congress) are content with measures seen to be protectionist, rather than with securing protectionist outcomes. In that regard, he views Congress as appealing to the attitudes of voters (or of others to whom Congressmen are responsive) who in turn are somewhat irrational. If it is accepted that there is some irrationality in the political-economic system, then one set of hypotheses about the determinants of protection would center on the loci of such irrationalities. An alternative to Destler’s hypothesis would be that individuals (including politicians, lobbyists, and voters) do not fully understand the economic considerations affecting individual industries and that protectionist measures that are adopted fail (to some degree) in their purpose because market pressures mitigate them. In this light, one can adopt a demand-and-supply framework. There is a “demand for protection,” which is presumably a function of voter concerns, lobbyist pressures, and other considerations. Simultaneously, there is a “supply of protection,” which is a function of the economic and political costs of providing protection. The demand for protection is higher the less well informed voters are about economic policy, the less economic knowledge supports a free-trade stance, and the larger the negative net trade balance in individual economic activities. The supply of protection is greater the smaller the negative impact on other producers (and hence greater for consumer goods than producer goods and greater for activities in which domestic firms do not have overseas operations from which they supply part of the domestic market) and the smaller the negative side effects (such as foreign policy considerations) of protection. In this framework, the 1980s experience could be interpreted differently. Clearly, the demand for protection increased. It increased because negative net trade balances emerged and/or increased for a large number of economic activities. It increased because the economics profession was willing to give more credence to considerations of “strategic protection” than it had earlier, thus undermining the professional consensus that had stood as a partial barrier to protectionist measures. However, the supply of protection shifted leftward: a given protectionist measure had higher economic and political costs in the 1980s, because of lowered transport and communications costs, and increased interdependence which in turn implied an increasing share of trade in producers’ goods and an increasing fraction of supplies from overseas affiliates. This same reduction in the “natural barriers” to trade would, in the absence of offsetting increases in “artificial barriers,” have resulted in an even greater expansion in trade flows than in fact occurred. The demand-supply framework as outlined here is consistent with any outcome. I would argue that pressures for protection did increase-based largely

283

U.S. Trade Policy-making

on the economic discomfort resulting from other factors, and that pressures did result in a change in the system. Destler argues that measures such as automobile VERs, steel protection, the semiconductor agreement, and the U.S.-Canada Free Trade Area were only “marginal” changes. However, one can readily argue that there was a qualitative change in the 1980s that tended to undermine the open multilateral trading system.’ First, there is a question as to how much protection is required before the system is no longer regarded as open. To the extent that American industries can now conclude that they will be eligible for protection when their output and employment diminishes sharply, that is a major change in perception which will affect much more than those few industries thus far protected. Incidentally, Destler’s own numbers indicate that, although sales of American-made automobiles fell by 2.7 million in the 1978-80 period, the decline in the total sales volume was 2.3 million. It is interesting that he nonetheless regards the difficulties of the American automobile industry as trade impacted. Second, the U.S.-Canadian Free Trade Agreement is a clear departure from the commitment to multilateralism. To be sure, there are a number of reasons, including a common border and the preexisting volume of trade, while that agreement alone might be consistent with an open-trading stance. However, given policy statements by the USTR and the administration regarding the possibility of other free-trade agreements (even before Mexico), there was certainly an erosion of the American commitment to an open, multilateral system, both because of FTAs and because of bilateral negotiations concerning Super-301 status, VERs, and other issues. Third, and perhaps most important, the momentum toward an increasingly open international trading system that characterized world trade at least from 1945 until the mid-1970s was certainly lost during the 1980s. The earlier momentum had derived, at least in good measure, from the U.S. hegemonic commitment to free trade as part of its foreign policy interests. As Destler’s own analysis indicates, American trade policy by the 1980s had shifted from being an instrument of foreign policy to being an instrument of domestic economic policy. As such, the American adoption of bilateral bargaining and other measures eroded the international system more than similar behavior by a smaller trading country would have done. Why then did increased protection not reduce trade flows? The answer I would provide is that reduced transport costs, lowered costs of communications, and the nature of technical progress more generally (especially in shifting the composition of output toward lighter weight, more easily transportable I Destler argues that there were no major “trade events” during the first Reagan administration. An alternative view would be that the failure of the G A l T Ministerial Meeting in the fall of 1982 was a major event shaping the Reagan administration’s attitude toward the multilateral system, and that failure to launch a new round of trade negotiations at that meeting was the disaster of the 1980s.

284

1. M. Destler

commodities and services) would have increased the share of trade in goods and services in world GNP even more than in fact happened had protectionist measures not increased. It may be noted that this tentative answer does not assume-as Destler’s analysis does implicitly-that lobbyists are irrational in seeking protection (although it would be difficult to argue with the proposition that they may overrate its potential effects, especially in light of lower-cost transport, etc.). It is quite possible that American imports (and, of course, exports) would have been greater, with the same size of the trade and current account deficits, had there been no additional protectionist measures. However, it is also probable that the increased costs of protection that result from increased interdependence has increased resistance to protectionist measures. To the extent that is the case, there is certainly hope that political pressures for protection may have peaked, and that the 1990s may witness moves toward a more open trading system.

Contributors

Albert0 Alesina Department of Economics and Government Harvard University Cambridge, MA 02138 James E. Alt Department of Government Harvard University Cambridge, MA 02138

I. M. Destler School of Public Affairs University of Maryland College Park, MD 20742 John A. Ferejohn Hoover Institution Stanford University Stanford, CA 94305

Robert J. Barro Department of Economics Harvard University Cambridge, MA 02138

Moms P. Fiorina Department of Government Harvard University cambridge, MA 02138

David F. Bradford Woodrow Wilson School Princeton University Princeton, NJ 08544

Benjamin M. Friedman Department of Economics Harvard University Cambridge, MA 02138

Charles Brown Department of Economics University of Michigan Ann Arbor, MI 48 109

Anne 0. Krueger Department of Economics Duke University Durham. NC 27706

Geoffrey Carliner Executive Director National Bureau of Economic Research 1050 Massachusetts Avenue Cambridge, MA 02138

Robert E. Litan The Brookings Institution 1775 Massachusetts Avenue, NW Washington, DC 30036

285

286

Contributors

Mathew D. McCubbins Department of Political Science University of California, San Diego La Jolla, CA 92037

Howard Rosenthal Graduate School of Industrial Administration Carnegie Mellon University Pittsburgh, PA 15213

William D. Nordhaus Cowles Foundation for Research in Economics Yale University New Haven, CT 06520

Charles H. Stewart 111 Department of Political Science Massachusetts Institute of Technology Cambridge, MA 02139

Keith T. Poole Graduate School of Industrial Administration Carnegie Mellon University Pittsburgh, PA 15213 Thomas Romer Graduate School of Industrial Administration Carnegie Mellon University Pittsburgh, PA 15213

Barry R. Weingast Hoover Institution Stanford University Stanford, CA 94305

Name Index

Aberbach, Joel D., 2051113 Abramson, Paul R., 21 Acheson, K., 53 Adams, Gordon S., 20 Adams, James Ring, 204117 Advisory Commission on Intergovernmental Relations, 151 Aldrich, John H., 21 Alesina, Alberto, 13nnl,3, 14nn4,5,9, 23, 33n2, 59, 64,74n18, 1091118, 119 A ~ ~ of MAmerican c Politics, 154 Alt, James, 59, 61 Ansolabehere, Stephen, 22 Bach, Stanley, 167n22 Bade, R., 44 Baker, James A., 111, 254, 266,272, 279n13 Balassa, Bela, 276 Balassa, Carol, 276 Baldridge, Malcolm, 261, 267, 278n9 Ball, George, 274 Banaian, K., 44 Banks, Gary,277 Barro, Robert, 14n9, 104, 120, 122 Barth, James R., 176, 201, 205nn21,23, 207n43,210nl Bartholomew, Philip F., 201 Bauer, Raymond I., 167n17 Beam, David R., 157n, 166n5 Beck, N., 53,59, 60,63 Becker, Gary S., 93 Belden, S., 56,74n12 Benston, George, 210n1, 213 287

Bentsen, Lloyd, 266, 271-72 Binkley, Wilfred, 86 Bimbaum, Jeffrey H., 157, 166115 Bloch, Farrell, 225, 228 Bradford, David F., 165113 Bradley, Michael G., 176, 201, 2051121 Brady, David, 22 Brand, Stanley, 87 Brock, William E., 259, 261-63, 265-66, 269, 279n12 Brown, R.,74n24 Browning, Robert, 132-33, 141nll Brumbaugh, R. Dan, 176, 204nn3,4, 2051123,207nn43.44, 210 Bryce, James, 86 Bullock, Charles S., 111, 220 Bums, Arthur, 45, 62, 64-65 Bums, James M., 86 Burtless, Gary, 126, 132 Bush, George, 2, 17, 20, 21, 33,70, 235, 243, 276 Byrd, Robert, 271-72 Calvert, Randall L., 140117 Campbell, Angus, 24 Campbell, James E.,21, 34n9 Carron, Andrew S., 176, 207n44,21Onl Carter, Jimmy, 1-2,4, 14n6,21,26, 33n6, 62.65-67, 70, 89-91, 131-33, 172, 243, 25 1 Chant, J., 53 Chappell, H., 59 Citrin, Jack, 150

288

NameIndex

Clark, L., 72, 73n1, 74n10,75n27 Cline, William, 274 Cohen, Stephen D., 254, 259 Conable, Barber, 167n20 Congressional Quarterly, 2061142; Almanac, 155; Weekly Report, 88, 89 Conlan, Timothy J., 157, 166115 Converse, Philip E., 24 Cook, Rhodes, 22 Cox, Gary W., 95,96, 109nnl1, 14, 223 Cukierman, A,, 61, 119 Danforth, John C., 228, 257, 259-60, 263, 265-66, 272-73,278n11, 279n12 Daniels, R. Steven, 223 Darman, Richard, 266, 279n13 Davidson, J., 57 Denzau, Arthur T., 181 Destler, I. M., 251, 255-56, 261, 263-64, 269, 274, 279n13 Dewey, Thomas E., 215 Dexter, Lewis A., 167n17 Dies, Martin, 234 Dietz, Robert, 204117 Dingell, John, 261 Dodd, Lawrence C., 93 Dole, Robert, 157, 259 Dougan, William R., 128 Durbin, J., 74n24 Durden, Gary, 225, 228, 229 Edwards, George C., 86 Eisenhower, Dwight D., 19,90-91, 131, 133, 215 Elving, Ronald D., 166n7 Ely, Bert, 2051125 Emerson, M., 242 Enelow, James, 169n26, 219 Erikson, Robert S . , 23, 26, 32 Evans, J., 74n24 Fair, Ray, 13nn1,2, 18, 19-20, 33nn5,6, 34n7 Fenno, Richard J., 85,93,96, 129, 167n22, 181, 229 Ferejohn, John A,, 56, 129, 140nn7,9, 141n15, 1671118,182, 204nn8,10, 205nn14,16, 244118 Fiorina, Moms, 13nn2,3, 20, 22, 24, 28, 32, 33111, 87, 167n18, 179, 184, 204x18, 229, 278n7 Ford, Gerald, 65, 89-91, 131-33

Frankovic, Kathleen A., 167n18 Frey, Bruno S . , 18 Fricker, Mary, 204n7 Friedman, Benjamin, 14nn8,9,44, 67 Friedman, Milton, 82 Funabashi, Yoichi, 279n13 Furlong, F., 68-69 GAO. See General Accounting Office Garn, Jake, 56, 187, 193, 228,245n13, 248

General Accounting Office (GAO), 55, 200, 203 Gephardt, Richard, 266, 271, 273 Gibbons, Sam, 261,270 Gildea, J., 60, 74n10 Gilder, George, 167n12 Gilligan, Thomas W., 204111 1, 223 Gilpin, Robert, 253 Ginsberg, Benjamin, 125 Golden, David G., 18, 19 Goldwater, Barry, 28, 31 Goodfriend, M., 42, 56 Gramm, Phil, 135, 195, 236 Gray, Edwin, 177, 187-88, 193, 196, 212 Greider, W., 63-67, 167nn12,13 Grier, K., 54-56,59,74n14 Gronke, Paul W., 20 Grossman, Herschel, 248111 Gudgin, Graham, 22 Hamilton, Lee, 72 Hansen, Susan B., 167n17 Hatch, Omn, 228, 248 Hatfield, Mark, 242 Havrilesky, T., 61,67, 74n16 Hawkins, Augustus, 235 Haynes, S . , 59 Heinz, John, 242 Heller, Robert, 56, 69 Helms, Jesse, 236 Hemmel, Eric I., 205123 Hendry, D., 57 Henning, C. Randall, 251,255, 264, 279n13 Hibbs, D., 14n14, 19,59,62-63,64, 73nn4,6 Hills, Carla, 252, 275-76 Hinich, Melvin, 1681126, 219 Hollings, Ernest F., 275 Hoover, Herbert, 99 Hufbauer, Gary C., 277 Hull, Cordell, 267 Humphrey, Hubert, 274

289

NameIndex

Jackson, Brooks, 205129 Jackson, John E., 24-25 Jacobson, Gary C., 21, 23, 34nn9.10, 204119 Johnsen, Christopher, 229 Johnson, Lyndon, 26, 131 Kabashima, Ikuo, 259 Kagay, Michael, 25 Kalt, Joseph, 223, 230 Kane, E. J., 14nl1, 176-77, 184,204117, 205nn21,28,211 Kaufman, George G., 210111 Keech, William R., 351121, 59 Keith, Bruce E., 341114 Kemp, Jack, 172 Kennedy, Edward, 235 Kennedy, JohnF., 131, 133, 244,274 Kennedy, Paul, 253 Kernell, Samuel, 18,21, 23, 34nn9,lO Kettl, D., 54, 55, 56, 73111 Kiewiet, D. Roerick, 18-20, 22, 33n2,8688, 95-98, 104-5, 108nn2,5, 109nnl1,14,17, 159, 167n18, 204nn8,10,14 Kinder, Donald R., 20, 22 King, G., 65 Kirst, Michael W., 86 Klein, Rudolf, 125 Krarner, Gerald H., 13nn1,2, 18, 20-23 Krehbiel, Keith, 55, 1411115, 165112, 168n27, 225, 247 Krueger, Anne O . , 249n2 Krugman, Paul, 254 Kuttner, Kenneth, 82 Labich, Carol J., 201 Laird, Sam, 277 Laney, L., 44 Laski, Harold, 86 Lawrence, Robert Z., 278112 Lekachman, Robert, 1671113 Lewis-Beck, Michael S., 21, 34n9 Leyton-Brown, David, 270 Linsey, Lawrence, 167n16 Litan, Robert E., 176, 2051120,207n44, 210 Lohmann, S., 71 Londregan, J., 13nl Lugar, Richard, 242 McCubbins, Mathew, 14n7, 54, 86-88, 9598, 104-5, 108nn2,5,8, 109nnll,14,17,

149, 159, 162, 179, 204nn8,10, 205nn13,14,223, 278n5 McDaniel, Paul R., 165113 McKelvey, Richard D., 128, 14Qn6 MacKuen, Michael B., 26, 32 Manley, John, 167n22 Markus, Gregory B., 21,23 Marra, Robin F., 22 Marris, Stephen, 264 Marshall, William J., 204111 1, 223, 244118 Martin, William McChesney, 64 Mashaw, Jerry L., 140115 Mayer, T., 63 Mayhew, David, 1091113, 179,256-57 Meese, Edwin, 261 Melton, W., 46, 49, 73nl Meltzer, Allan, 44,46,61,72, 119,205n19 Metzger, Ronald I., 259 Miller, Arthur H.,2 1 Miller, G. William, 66-67, 72 Miller, George, 236 Miller, Warren E., 21, 24, 34n16 Mills, Wilbur D., 156 Milton, George F., 86 Minsky, Hyman, 80 Moe, Terry M., 181, 183,204118 Mondale, Walter, 21 Monroe, Kristen R., 18, 33n2 Moran, Mark, 86, 204nn8,10, 2051116, 235 Morici, Peter, 270 Mueller, John E., 18 Munger, Michael, 128, 181 Muolo, Paul, 204117 Muris, Timoth, 108119, 1091115 Murphy, Austin J., 235 Murray, Alan S., 157, 166115 Myerson, R., 43 National Opinion Research Center (NORC), 140113 Neal, Stephen, 196 Neustadt, Richard, 84, 86 New York Times,146, 195, 200, 2061131, 234 Niemi, Richard C., 91 Niskanen, William, 19-20, 152 Nixon, Richard M., 90-91, 131-33,216, 219, 243,261, 274 Nogues, Julio J., 277 NOH, Roger G., 86,95, 179, 204n8 Nordhaus, William, 14nn5,6, 18, 34n8,58 Norpoth, Helmut, 19,25, 34x116

290

Name Index

Odell, John S., 255, 263 Ogul, Morris S., 2051113 O’Higgins, Michael, 125 Olechowski, Andrzej, 277 Oleszek, Walter J., 108n6 OMB. See U.S. Executive Office of the President, Office of Management and Budget O’Neill, Thomas P. (Tip), 193, 270, 271 Oppenheimer, Bruce I., 22,93 Ordeshook, Peter C., 168n26 Orfield, Gary, 141n12 Ostrom, Charles W., Jr., 19, 33n4, 22 Packwood, Robert, 173, 242 Page, Benjamin I., 144 Page, Talbot, 95 Palmer, John L., 123 Panning, William H., 165n2 Parkin, M., 44 Pastor, Robert, 278n4 Patman, Wright, 74n14 Patterson, Kelly D., 140113 Pearce, David, 277 Pechman, Joseph, 165n3 Peltzman, Sam, 223 Perkins, Carl, 236 Persson, T., 14n9 Peterson, Paul E., 55, 86, 132 Petrocik, John R., 25, 34n16 Pfiffner, James, 86 Pigou, A. C., 120 Pilzer, Paul Zane, 204x17 Pizzo, Stephen P., 204117 Pool, Ithiel de Sola, 1671117 Poole, Keith T., 197, 2061136, 219-20, 223, 228,231, 233, 244117, 2451114 Poole, William, 57 Poterba, James M., 18, 19 Prestowitz, Clyde, 253,268 Proxmire, William, 56, 74n14, 187, 193-95, 200 Ragsdale, L., 65 Rapp, David, 168n30 Reagan, Ronald, 1-2, 17, 19-21, 26, 33nn3,6, 34n10,67,70,78, 83-84, 8892, 103, 121-23, 131, 133, 135-36, 138-39, 144, 154-55, 163-65, 171-73, 216,234,243,262-63, 266,269, 27375 Regan, Donald, 67, 266 Reuss, Henry, 741114 Rice, Tom W., 21,34n9

Richardson, J. David, 254 Riker, William H., 244118 Rivers, Douglas, 18-20.23, 33n2, 1671118, 225-26,241 Roberts, Paul Craig, 1671112 Rockman, Bert A., 205n13 Rogoff, K., 60 Rohde, David, 21,95 Rom, Mark, 55, 86, 132 Romer, Christine, 68 Romer, David, 68 Romer, Thomas, 204nn8,10, 228 Roosevelt, Franklin D., 184, 233 Rosenthal, Howard, 13nn1,3, 23, 197, 204nn8,10,206n36, 219-20, 223,228, 231, 233,244117. 2451114 Rostenkowski, Dan, 157, 263, 266, 270, 272, 275 Rotbart, Gilbert, 242 Roubini, Nouriel, 109n18 Rudder, Catherine, 167n22 Russell, Judith, 140n3 Sachs, Jefiey, 59,64,74n18, 1091118 St Germain, Fernand, 193-97 Sauerhaft, Daniel, 205n23, 210nl Sawhill, Isabel, 123 Schick, Allen, 85, 93 Schlesinger, Arthur M., Jr., 86 Schneider, Friedrich, 18 Schneider, William, 17 Schott, Jeffrey J., 269, 270 Schultze, Charles, 278n2 Schwartz, Thomas, 54, 86, 95, 179, 204n8, 205n13, 278n5 Scott, Kenneth E., 176, 207n43 Sears, David O., 150 Selko, Daniel, 167n17 SemiconductorIndustry Association (SIA), 267 Shanahan, Eileen, 149, 1681130 Shanks, J. Memll, 21 Shapiro, Robert Y.,140n3, 144 Shefter, Martin, 125 Shepsle, K., 55,93-95,229, 278n5 Shipan, Charles R., 56, 182, 204nn8,10, 205nn14,16 Shughart, W., 53 Shultz, George, 266 SIA. See SemiconductorIndustry Association Silberman, Jonathan, 225,228,229 Simon, Dennis, M., 19, 33n4 Simon, William, 172

291

NameIndex

Sinclair, Barbara, 205n12 Smith, Steven S., 134, 167n22 Sparkman, John, 56, 741114 Sprinkel, Beryl, 49,73n4, 266 Srba, F.,57 Stanley, Harold W., 91 Stanwood, Edward, 1671117 Starobin, Paul, 2071144 Stein, Herbert, 73116 Stewart, Charles, 167n23 Stigler, George J., 93 Stimson, James A,, 19,22,26,32 Stockman, D., 14n8, 135, 141n13, 167nn12.13, 264 Stoeckel, Andrew, 277 Stokes, Donald, 24 Stone, J., 59 Strahan, Randal, 158, 166n5, 167n22, 168n24 Strauss, Robert, 251 Summers, Lawrence, 127 Sundquist, James, 85-87, 108n3 Surrey, Stanley S., 165n3 Svensson, L., 14n9 Swain, Carol M . , 351121 Tabellini, Guido, 14n9, 109n18, 119 Taussig, F. W., 1671117 Taylor, Peter, 22 Thrumond, Strom, 236 Tobin, James, 47, 73n4 Tollison, R., 53 Toma, M., 53 Tower, John, 242 Truman, Harry,90-91,230 Tufte, E. R., 14n5, 18, 21, 22 Udall, Moms, 167n20 U.S. Congress. House of Representatives, 108nn2,4; Ways and Means Committee, 259 U.S. Council of Economic Advisers, 106, 186, 278111 U.S. Department of Commerce: Bureau of Economic Analysis, 106; Bureau of the Census, 106, 108nl U.S. Department of Defense, 268

U.S. Department of Transportation, 259 U.S. Executive Office of the President: Office of Management and Budget (OMB), 84, 106,144,146 University of Michigan Survey of Consumer Attitudes, 26 Vanik, Charles L., 259 Viscusi, W. K., 14n10 Volcker, Paul, 4,45,49, 54,56,59,62-63, 65-67,74nn11,19,78,266 Wall, M.Danny, 213 Wall Street Journal, 196, 205n25, 207n45 Wang, George H. K., 205n23,210nl Washington Post, 196, 263, 270 Waterman, Richard W., 22 Wattenberg, Martin P., 21 Wayne, Stephen, J., 84 Weatherford, M. Stephen, 33n2 Weaver, R. Kent, 278n7 Weicher, John C., 2051119 Weidenbaum, Murray, 47,73n4 Weingast, Barry R., 86.93-95, 179, 183, 204nn8,10,11, 205nn13,15,16,223, 229, 235,244118, 278n5 Weisberg, Herbert F., 21 Willett, T., 44 Williams, 242 Wilmerding, Lucius, Jr., 87 Winham, Gilbert R., 259 Winters, L. Alan, 277 Witte, John F., 1671117 Woerheide, Walter J., 205n17 Woodworth, Lawrence, 172 Woolley, J., 53-57, 60,64, 68-69 Wright, Jim, 193-96, 206n31,271, 275 Wrightson, Margaret, 157, 166115 Yeats, Alexander, 277 Yeo, S., 57 Yeutter, Clayton, 270, 272, 2791112 Young, John T., 14Gn3 Zupan, Mark, 223, 230 Zysman, John, 254

Subject Index

Approval ratings, presidential, 18-19 Auto industry, 259-60 Balance of payments. See Trade balance Budget Act (1974).See Congressional Budget and Impoundment Control Act (1974) Budget and Accounting Act: (1921). 86-87,

108~2,4; (1958), 87

Budget and Accounting Procedures Act

(1950).86 Budget deficit: adjustment for analysis, 11219;effect of divided government on, 99103, 105-7;effect of party politics on level of, 103-7; explanations for, 5; measure of, 83-84, 108nl;trust-fund surpluses to reduce, 148.See also Taxsmoothing model Budget process: components under Budget Act (1974)of, 97;role of majority party in, 97-98 Canada, 254,268-70, 283 Carter administration: changes in welfare system under, 126;deregulation policy of, 9;fiscal policy of, 4; military spending policy of, 6-7; spending policy of, 8990.See also under Name Index Committees, congressional: alliances with distributional support by, 129-30, 132; gatekeeping role of, 180 hierarchical system of, 95;individual importance and autonomy of, 93-94;party politics of, 94-95;political interdependence among, 192;role of veto points in, 180-81

292

Competitive Equality Banking Act (1987),

200-201 Congress: decline relative to president, 85; delegation of trade power by, 261-63; effect of actions on thrift institution crisis, 10; forbearance policy for thrifts, 178, 185-86, 192-201, 203,211;influence applied to monetary policy, 43;influence of committee system of, 155-58; influence on Federal Reserve of, 43-45, 54-56; institutional reforms in, 144, 155;regulatory policy-making process of, 179-84; tax increase measure (TEFRA,1982), 144;trade-related behavior of, ll-12,255-58,270-78; veto game between House and Senate, 99-

102 Congress. House of Representatives: Appropriations Committee role in, 85,93,9698, 156;Committee on Banking, Finance, and Urban Affairs, 97, 192, 19596;Committee on Education and Labor, 97;Committee on Energy and Commerce, 98;committee system tax policy, 155-58;Ways and Means Committee,

155-58, 163,260-61 Congress. Senate: Committee on Banking, Housing, and Urban Affairs, 55-56, 192;Energy and Commerce Committee, 260-61; Finance Committee, 271-

72 Congressional authority: delegation to committees of, 93-95;delegation to executive branch of, 86

293

Subject Index

Congressional Budget and Impoundment Control Act (1974), 84-85,93,97-98, 126, 135 Constituent interests: role in regulatory policy-making of, 179-84, 202-3; in thrift institution crisis, 184-92 Copper industry (producers and users), 26 163 Cumulative Data File (CDF), 27 Danforth bill: auto import quota, 259-60; trade reciprocity, 263 Democrats: effect of resurgence in Congress of (1982), 143, 155; gains in control of Congress (1982-86), 1-3, 7, 12-13; response to 1980s tax policy, 154-55; rollcall voting on minimum wage, 219-43; spending decisions of congressional, 98102; trade-related behavior of, 12, 25758, 260-61,265-67, 270-78 Deposit insurance: effect on thrifts, 175, 210; lifting of ceiling of, 209; political motivation for. 176 Economic conditions, correlated with: approval ratings, 18-20; party identification, 24-33; voting behavior, 21-23 Economic Recovery Tax Act (ERTA, 1981), 88, 148-49, 162; effect of, 8-9; party response to, 154-55; provisions of, 14344, 170-71 Election outcomes: long-term factors in, 2432; short-term factors in, 21-23 Electoral systems: proportional representation (PR) system, 22; single-member simple plurality (SMSP) electoral system, 22 Exchange rate policy, 67-71 Fair Labor Standards Act (1938), 215 Fair model predictions, 20 Federal Home Loan Bank Board (FHLBB), 177-78, 184; proposal to tighten thrift industry regulations, 193; regulatory powers of, 185; report on insolvent thrifts by (1985). 1987 Federal Insurance Contributions Act (FICA), 145 Federal Open Market Committee (FOMC), 44;announced targets and actual federal funds rate of, 45-48; composition of, 45,56; influences on, 60; management practices of, 72, 73111;policy in 1970s

and 1980%45-48; voting behavior, 5354,72 Federal Reserve System: congressional oversight of, 54, 72; effect of principals on, 3,43-44,53-67, 71-72, 77-80; evidence of presidential influence on, 5962; relation to Congress of, 54-56,72. See also Federal Open Market Committee Federal Savings and Loan Insurance Corporation (FSLIC): Congress prevents recapitalization of, 178; establishment of, 184; legislation reaffirming government commitment to, 200-201; plan to recapitalize, 193-95, 199; problems of (early 1980s), 176-77; shortfall of insurance funds (1985), 187-88 FHLBB. See Federal Home Loan Bank Board FICO. See Financing Corporation Financial community: influence on Fed chairman reappointment, 60; influence on Federal Reserve, 80-82; monitoring of Federal Reserve system by, 56-58 Financial Institutions Reform, Recovery, and Enforcement Act (FIRREA, 1989). 201 Financing Corporation (FICO), 194,2061130 Fiscal policy, 4-9. See also Budget deficit; Economic Recovery Tax Act; Spending, federal; Tax policy; Tax Reform Act; Tax-smoothingmodel FOMC. See Federal Open Market Committee Ford Motor Company, 259 Free-Trade Agreement (FTA), US.Canadian, 254, 269-70,283 FSLIC. See Federal Savings and Loan Insurance Corporation Gallup Poll, 24, 166n10 Gam-St Germain Depository Institutions Act (1982): conversion of liability into asset under, 204nn5.6; deregulatory provisions of, 177, 185 General Agreement on Tariffs and Trade (GATT): failure of ministerial meeting (1982), 283111;getting around, 269-70; Tokyo Round, 251, 254, 258, 262, 269, 274, 278116;Uruguay Round, 254,258, 270-72,278n6, 2791114 Gephardt amendment, 266, 270-73 Government, divided: effect of, 2-3, 99-103; effect on fiscal policy of, 98-102, 119; effect on large budget deficit of, 5 ; pattern of, 2-3; test of hypothesis of, 105-7

294

Subject Index

Government Accounting and Procedures Act (1956). 87 Group of Seven (G-7) meetings, 68

National Election Studies (NES), 24, 26-32, 34n18. See also Cumulative Data File Nontariff barriers (NTBs), 276-77

Inflation: adjustment in budget deficit analyses for, 112-19; effect on election outcomes of, 20; measurement of, 121; Reagan administration action on, 19; relation to money of, 51 Institutional systems, congressional: committees as veto points in, 180-81; influence on regulatory agency policy of, 182-83, 201-2 Institutional systems, governmental, 155-63 Interest groups: influence on congressional politics of, 93; influence on Federal Reserve of, 43-45,60, 80-82. See also Constituent interests

Omnibus Trade and Competitiveness Act (1988), 252, 254, 265, 271-73, 275-77

Japan, 259-60, 265, 267-69 King of the Mountain rule, 98 Legislative Reorganization Act (1970). 87 Long-Term Arrangement on Cotton Textiles (1961), 274 Macroeconomic policy: cyclical theories of, 58-59; in Reagan administration, 264 Majority-rule instability, 128-29 Mama-Ostrom model, 22 MFA. See Multi-Fiber Agreement Minimum wage: partisan political process of, 215-44; role of government in setting, 10-1 1, 215-44; roll-call voting in Congress (19809, 234-42; roll-call voting in House (1937-83), 230-34 Ministry of International Trade and Industry (MITI), Japan, 265,268 Monetary policy: with changing money relations, 50-52; congressional influence on, 43, 72; to control money stock (197088), 45-48; international, 48-50, 67-71; political influence over, 3-4, 41-42, 7172; post-election partisan cycle for, 42, 58-59, 64-66,71-72, 78-79; preelection political influence on, 42-43, 58-64, 71-73, 78; in 1980s. 41. See also Exchange rate policy; Federal Reserve System Money, 50-51 Multi-Fiber Agreement (MFA, 1973), 274-75

Parties, congressional, 92-98. See also Committees, congressional Partisan cycle. See Partisan theory Partisan theory: applied to monetary policy, 42; post-election prediction of partisan cycle, 64-66; predictions of, 58-59 Party identification: analysis of, 27-32, 3840; associated with consumer confidence, 26; factors influencing, 2, 23-32; as long-term factor in elections, 24-27; voting on basis of, 24. See also Democrats; Republicans; Voting behavior Party organizations, congressional, 94-98. See also King of the Mountain rule Plaza Accord (1985), 67-68, 266 Policy of forbearance. See Congress; Thrift institutions Political business cycle: applied to monetary theory, 42-43; predictions of, 58-59; pre-election, 62-64 Political changes: effect on welfare programs of, 133-35; influence on tax debate of (1980s), 150-52 Political support, 128-29 President: effect of popularity of, 70-71; influence on Federal Reserve of, 43-45, 58-67; policy dominance of, 85-87; regulatory policy-making of, 179-84. See also Veto game; Veto power Presidential dominance thesis, 85-87. See also Congress. House of Representatives; Congressional authority; Congress. Senate Public opinion: approval or disapproval of president, 18-23, 70-71; changing preferences on tax policy, 150-52. See also Gallup Poll; Voters’ attitudes; Voting behavior Reagan administration: deregulation agenda, 9-12; effect on welfare system, 135-39; fiscal policy, 4-8, 13, 89-92, 135-38; military spending policy, 6-7; stops inflation, 19, 26; strategy for welfare program change, 125-26; tax-reduction

295

Subject Index

measure of (ERTA, 1981), 144. See also under Name Index Reciprocal Trade Agreements Act (1934), 267 Regulation Q. 212 Regulatory agencies: role in policy-making, 179-84; role in thrift institution, 18492, 203 Republicans: control in Congress (198 1-82), 1-2, 12-13, 17; minimum wage levels under, 215-17; response to 1980s tax policy by, 153-54; roll-call voting on minimum wage, 219-43; spending decisions of congressional, 98-102; traderelated behavior, 12, 257-58, 260-61, 265-67, 270-78 Responsibility hypothesis, 39 Roll-call voting, congressional: economic models for, 223-24; on minimum wage, 219-42; spatial and economic models compared, 226-30; spatial model of, 2 19-23 Savings and loan industry. See Thrift institutions Semiconductorindustry, 267-69 Separation of powers, 155, 158-63. See also Veto game Slutsky equation, 104 Smoot-Hawley Tariff Act (1930), 256, 258 Social Security program: effect of 1983 amendments to legislation for, 171; if separate from deficit calculations, 121; revenues from (1983-89), 143 Spending, federal: changes in growth rates of welfare, 124-25; decisions reflecting party politics, 98; decline in domestic, 6-7; effect of insufficient cuts in, 149; increase in, compared to increase in revenues, 84; increases in (1981-88), 9192; measurement for analysis, 103-4; military, 6-7 Steel industry, 261-63 Stock market crash (October 1987). 68 Super-301. See Trade legislation Tax Equity and Fiscal Responsibility Act (TEFRA, 1982), 149, 157; party response to, 155; provisions and effect of, 144-46 Tax policy: House Ways and Means Commit-

tee role in, 155-56; influences on 1980s, 149-65 Tax Reform Act (TRA, 1986), 145-46, 16162; provisions and effect of, 8-9, 171 Tax-smoothing model, 6, 119-22 Tax system: laws changing structure of, 146, 148-49; reform efforts for, 143; reform in 1980s of, 144-49. See also Social Security TEFRA. See Tax Equity and Fiscal Responsibility Act Textile industry, 274 Thrift institutions: bailout legislation for (1989), 201; congressionaljurisdiction over, 192; congressional policy of forbearance for, 178, 185-86, 192-201, 212; effect of legislation and deregulation on, 9-1 1, 177-78; gambling for resurrection policy of, 175, 177-78, 186, 201; insolvency data (1988), 201; own role in crisis of, 184-92; political foundations of crisis in, 211-14; political motivation for establishmentof, 176; unprofitability of (1980s), 176 Trade Act (1974), 257, 274 Trade actions: rise in 1980s of, 262; sanctions against Japanese semiconductor industry, 269; Section 301 (dumping) case of semiconductor industry, 267-69 Trade Agreements Act (1979), 269 Trade agreement (semiconductortrade complaint), 268-69 Trade and Tariff Act (1984), 263, 269 Trade balance: effect of imbalance in, 255, 263-67; improvement in, 258-59; political reaction to deficit in, 251-52 Trade Expansion Act (1962), 274 Trade legislation: House initiates omnibus bill (1986), 257, 265, 270-72; related to textile industry, 274-75; Super-301 bill and law, 252-53, 272-73. See also Omnibus Trade and CompetitivenessAct Trade policy: response to auto industry crisis, 259-60; in 198Os, 11-12, 252-54, 27678, 281-84. See also Trade legislation Trade protection: hypothesis of demand for and supply of, 282-83; nontariff barriers (NTBs) as, 276-77; rise in early 1980s of, 254; trade flows with, 283-84; voluntary export restraints (VERs) as, 276 Trust funds (in federal budget), 148

296

Subject Index

United Auto Workers (UAW), 259-60 U.S. International Trade Commission (USTIC), 259, 262-63 US.-Japan Semiconductor Trade Agreement (1986), 278113 U.S. Savings League, 195-96 U.S. Trade Representative (USTR): competition with Department of Commerce (1981), 261; creation of office of, 256; strengthening of office of, 265 Veto game: among House, Senate, and presidency, 159-63; between House and Senate, 99-102 Veto points. See Committees, congressional Veto power: limitations to presidential, 8789. See also Veto game Voluntary export restraints (VERs), 259-60, 265,276-17 Voters’ attitudes: legislators’ perception of,

202-3; shift of, 2-3. See also Party identification Voting behavior: in congressional elections, 12-13, 21-23; factors influencing, 20, 94; in Federal Open Market Committee, 53-54; long-term factors influencing, 23-27; short-term influences on, 20-23 Voting Rights Act (1965). 133 Welfare programs: basis of political support for, 128-30; factors influencing growth rates of, 124-25; growth of means-tested and social insurance, 131-32; Reagan administration impact on, 138-39; trends in spending for, 130-33 Welfare system: changes during Reagan administration in, 125-26, 135-36; differences between cash and in-kind transfer, 131; effect of political changes (1960s) on, 133-35; evolution in 1980s of, 131; political foundation in 1970s for, 124-33

This Page Intentionally Left Blank