The Accounts of Nations 9051991568, 9789051991567

National income and products estimates are used extensively worldwide. During the 1950's P. Studenski's work `

176 110 7MB

English Pages 260 [258] Year 1994

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

The Accounts of Nations
 9051991568, 9789051991567

Citation preview

Edited by Z. Kenessey

IOS Press Amsterdam • Oxford •Washington •Tokyo

THE ACCOUNTS OF NATIONS

edited by Zoltan Kenessey

IOS Press 1994 Amsterdam • Oxford • Washington, DC • Tokyo

©Authors mentioned in the table of contents. All rights reserved. No part o f this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior written permission from the publisher. ISBN 90 5199 156 8 Library of Congress Catalog Card Number: 94-075424

Publisher: IOS Press Van Diemenstraat 94 1013 CN Amsterdam Netherlands

Sole distributor in the UK and Ireland: IOS Press/Lavis Marketing 73 Lime Walk Headington Oxford 0 X 3 7 AD England

Distributor in the USA and Canada: IOS Press, Inc. P.O. Box 10558 Burke, VA 22009-0558 U.S.A.

Distributor in Japan: Kaigai Publications, Ltd. 21 Kanda Tsukasa - Cho 2 - Chome Chiyoda-Ku Tokyo 101 Japan

LEGAL NOTICE The publisher is not responsible for the use which might be made of the following information.

V

PREFACE National income and product estimates are used extensively over the world. Data pertaining to concepts such as GDP (or GNP) are quoted daily by the media. Yet the genesis of these concepts is not well known, not even among professional economists or statisticians. On the history of national income concepts and estimates the most significant book appeared in the 1950s. The book, THE INCOME OF NATIONS (1958, New York University Press), became a classic source on the topic. It earned its author, Professor Paul Studenski from New York University wide international acclaim. Since the 1950s the relatively simple ways of estimating national income and product became increasingly complex. The methodology got much more involved, the compilations much more extensive, and the application of the results much more sophisticated. Also, far reaching international agreements have been worked out to ensure the international comparability of the national efforts. The UN System of National Accounts (SNA) became a milestone on the path of this development. It reflected much of the thought of Sir Richard Stone, who received a Nobel Prize for his contributions to this field. Studenski’s magisterial work could not yet reflect the wide sweep of national and international effort that characterized the postwar decades. Therefore the time has come to supplement the research on the earlier times reflected in his INCOME OF NATIONS with a historical view of THE ACCOUNTS OF NATIONS, that can show the development of the present day national income and product accounts. In order to meet this need, the 1992 Conference of the International Association for Research in Income and Weatlh (IARIW), at the editor’s suggestion and under his organization, held a special session dedicated to the topic. First rate international experts participated in the program, most of whom were personally involved in the work on national accounts during the postwar era. The special session generated very much interest among the economists and statisticians attending the IARIW conference and the papers were generally very well received. Now this volume makes the latest research results about the evolution of postwar national accounting accessible to a growing international audience interested in the subject. As also indicated in the overview by the editor below, the assistance of IARIW and of the Professor Wilhelm Keilhau Memorial Fund (Oslo) was essential for this project and is most thankfully acknowledged. Most of the information in these studies about Scandinavia, the United Kingdom, the Netherlands, the United States, Japan, India and Germany is either first time analyzed or pulled together in a comprehensive manner covering typically five or more decades. The extensive bibliographies shown at the end of most studies provide valuable additional reference material. It is expected that these in-depth studies of national and international national accounting (mainly in respect of the UN SNA, but even covering the Material Product System of the former socialist states) will stimulate further research about the past, the present and the future of national accounting systems.

VII

Contents

Preface

V

The Genesis of National Accounts: An Overview, Z. Kenessey

1

The Scandinavian Contribution to National Accounting, O. A ukrust

16

Dutch National Accounts: A History, G.P. den Bakker

66

Keynes’ How to Pay fo r the War and Its Influence on Postwar National Accounting, Y. Kurabayashi

93

American Contributions to the Development of National Accounts, Z. Kenessey

109

Development of National Accounts in India, S.G. Tiwari

124

German National Accounts Between Politics and Academics, U.-P. Reich

144

The SNA: 1968 - 1993 and Beyond, A . Harrison

169

Constancy and Change in the United Nations Manuals on National Accounting (1947, 1953, 1968 and 1993), F. Bos 198 The Material Product System (MPS): A Retrospective, J. Arvay

218

A Historical Note on Quadruple-Entry Bookkeeping, H .H . Postner

237

Intellectual Foundations for the 1968 SNA, G. Pyatt

246

Author Index

251

•»

The Accounts o f Nations Z. Kenessey (Ed.) IOS Press, 1994

1

THE GENESIS OF NATIONAL ACCOUNTS: AN OVERVIEW Zoltan Kenessey* Voorburg, The Netherlands Abstract. These introductory thoughts about the genesis of modern national accounts also contain references to the papers included in this volume, which were originally presented in 1992 at the Conference of IARIW in Flims, Switzerland. Second, they highlight six distinguishing features of contemporary national accounts in comparison with earlier national income estimates. Third, they mention four key national efforts leading to the establishment of the modern System of National Accounts as well as some related matters.

Introduction National accounts have been, arguably, the most important new tools of economic analysis and policy making in the second half of the 20th century. At least four Nobel prizes in economics were awarded to persons who contributed to their development: Simon Kuznets, Sir Richard Stone, Wassily Leontief and Jan Tinbergen. Reports about the Gross Domestic Product of the industrialized countries are quarterly events widely covered by the press, television, and radio. Even relatively small revisions of such data are much followed by the financial markets, and indeed most governments’ monetary and fiscal policies are closely attuned to them. The International Association for Income and Wealth (IARIW) in 1992 dedicated a session during its 1992 Conference (held in Flims, Switzerland) to the evolution of national accounts. Earlier, the more than 300 years of preceding developments in developing the concepts and introducing the practice of estimating national income had been described by Paul Studenski in his widely acclaimed book ’The Income of Nations’ (1958). The IARIW meeting focused on the qualitatively new developments which have occurred in the postwar era, and which had their origins in the Great Depression of the 1930’s and in the challenges of the Second World War. The title of this volume, ’The Accounts of Nations’ reflects the desire to follow up from where Studenski’s account stopped; yet being a collection of individual albeit related studies, it cannot claim the depth of coverage achieved by him. Indeed, the field of national accounts has grown in depths and width tremendously over the last 50 years. There is hardly any country in the world that has not engaged its statisticians and economists in the subject. Even a multi-volume series could not do full justice to all the national and international efforts in this field. Nevertheless, we trust, this volume provides a useful overview of some major developments regarding the establishment of the accounts of nations over the last 50 years or so. * The author is Director of the International Statistical Institute in Voorburg, The Netherlands. The views expressed are his own.

2

Z. Kenessey/The Genesis o f National Accounts

Before providing an overview of the genesis of national accounts, with particular regard to the studies presented in this volume, it is appropriate to express sincere thanks to the IARIW Council as well as to Professor Edward Wolf and Jane Wolf of the IARIW for their interest and crucial help with this project. Similarly, warm acknowledgments are due to Mr. Odd Aukrust, a doyen of the field of national accounts and to the Professor Wilhelm Keilhau Memorial Fund (Oslo, Norway) for their generous support. Having taken care of this most pleasant obligation, let us turn our attention to the evolution of national accounts over the last 50 years. Focusing on the evolution of contemporary national accounting is as important for assessing its future possibilities as for chronicling its performance over the recent decades. One hopes, that some of the accumulated experience, dealt with in this volume, provides certain perspectives for the further development of such accounts. There is wide agreement, both among users and producers of national accounts, that a good number of improvements are needed to sharpen this tool for use in the years and decades to come. The latest revision of the System of National Accounts, adopted by the Statistical Commission of the United Nations in February 1993, incorporates many of the improvements required. Yet still further efforts will be needed and hopefully they will be informed by the manifold historical experiences chronicled in this volume.

1. The Intellectual Stimuli for the New System New intellectual achievements, such as the system of national accounts introduced around the middle of the 20th century, can be always traced to some earlier efforts which had direct or indirect influence upon them. Historians of the sciences and the arts often find evidence from ages gone and even from geographically distant societies and cultures establishing, or at least indicating, intellectual links between recent and past achievements. 20th century national accounting is no exception in this regard. Several strains of earlier thought can be cited as pertinent to the evolution of national accounting as practised today. Some of these appear to be direct antecedents; others seem to have exerted their influence in a more circuitous route. Placing some earlier efforts in the ’direct’ and others into the ’indirect’ category of influence is, of course, subject to challenge by both facts to be unearthed later and by alternative interpretations of the available facts. With such caveats in mind, one can highlight perhaps the following six major strains of intellectual influence upon the establishment of modern national accounts: (a) The thoughts and influence of John Maynard Keynes, (b) The efforts of Ragnar Frisch, (c) The influence of American scholars, such as Simon Kuznets and Irving Fisher, (d) The input-output work of Wassily Leontief, (e) The Dutch contributions connected to the name of Jan Tinbergen, and (f) The approach of ’money-flows’ developed by Morris Copeland. As perusers of this volume will see, there were many other contributors. Yet, those named earned special respect for their efforts.

Z. Kenessey/The Genesis o f National Accounts

3

These intellectual influences were important, directly or indirectly, for the evolution of national accounts in various ways. Therefore, in this section of our overview, a few words will be accorded to each of these main strains. Yet, there are other ways to recognize the intellectual novelties involved in national accounts. These are summarized in the second section of the present overview. They pertain to the dividing line, as postulated here, between traditional national income estimating which was practised over a three hundred year period starting with the work of Sir William Petty and Gregory King in England, on the one hand, and today’s national accounts, on the other. In the third section, relevant developments in four countries are emphasized: the United Kingdom, Norway, the Netherlands, and the United States. In these countries the statistical services paid particular attention to the establishment of modern national accounts. These national efforts require special attention inasmuch national accounting is certainly not just an academic exercise (in either the good or in the bad sense of the word) but also involves a large scale statistical operation, that requires much organizational acumen, considerable technical skills, and the outlay of sizable budgetary funds.1 Finally, in the fourth section of our overview certain related developments are referred to briefly. The intellectual influences and country efforts mentioned were, naturally, not the sole factors in establishing modern national accounts. Other important circumstances must be taken into account and many of them are handled in the various studies of this volume. The brief references in our introduction to the studies of this volume mention them typically in more than one context. However, these references are by no means summaries of the ideas contained in each. In order to get more information about the scope of the individual contributions the readers are referred to the Abstracts provided at the beginning of all articles. Returning to the six main intellectual strains, holding particular importance for the development of modern national accounts, the following summary thoughts are offered at this stage: 1.1 The Influence o f Keynes The most outstanding as well as dramatic influence on modern national accounts is the impetus provided by John Maynard Keynes, both through his General Theory and via his efforts to base financing the Second World War in the United Kingdom on realistic economic intelligence. In this volume, the article of Professor Yoshimasa Kurabayashi focuses on Keynes efforts in this context; his article also contains very interesting references to the attention paid in Japan during the war to Keynes’ work in this regard. Not surprisingly, Keynes’ name comes up in several other articles in this book, including those of Gert den Bakker, Frits Bos, Utz-Peter Reich, Graham Pyatt, and Zoltan Kenessey. Keynes’ effect on the genesis of modern national accounts was not solely academic. Indeed, without his influence at the UK Treasury such work probably would not have materialized during the difficult years of the war. His personal support to the work was quite essential to the establishment of the accounts at the time. The work of Meade and

4

Z. Kenessey/The Genesis o f National Accounts

Stone at the time, and the grand structure of the national accounting system as designed by its main architect, Sir Richard Stone, clearly bear the imprints of the thoughts and concerns of Lord Keynes. 1.2 The Impact o f Frisch Ragnar Frisch, one of the founder of econometrics, also played an outstanding role in developing 20th century approaches to the quantification of macroeconomic processes. His ideas about measurement are not widely known, and even national accountants can be unaware of his efforts in this area. In part, this is related to the limited circulation of his early work pertinent to national accounting (issued in Norwegian, or only in mimeographed format in English). Also, Norway was occupied during the crucial years when national accounting evolved into the important tool of war management in Britain. Nevertheless, influence emanating from Frisch, especially through the major efforts of Odd Aukrust and Petter Jakob Bjerve, became an organic part of postwar national accounting not only in Norway, but internationally as well. In this volume, a major study by Odd Aukrust provides much insight into the Norwegian (and Scandinavian) work in the field, and into Frisch’s influence. 1.3 Kuznets and Irving Fisher In discussing the new system of national accounts, the justifiable tendency is to highlight its qualitatively new features. However, among the various stimuli leading to the evolution of the accounts the intellectual heritage coming from traditional national income studies also played an important role. After all, national accounts are really ’national income and product accounts’ (abbreviated NIPA in American parlance), and the basic concepts of national income and product - inherited from the traditional estimates but refined further for the accounts - played key roles in the new system. The resolution of countless conceptual, estimating, and data problems regarding national income totals and their main components by Simon Kuznets in the United States and the elucidation of certain important international ramifications regarding national income studies by Colin Clark were among the important components of this heritage as it came about in the 1930’s. Other writers in these two countries as well as economists and statisticians from other countries, for example Erick Lindahl in Sweden, were also important sources of this stream of intellectual influence. Regarding the impact of Lindahl, in particular Odd Aukrust’s thoughts in this volume should be cited. Probably the least appreciated intellectual debt modern national accounts owe is the one due to Irving Fisher. Perhaps he was really ahead of his time when at the beginning of the 20th century pointed out the potentials of double entry bookkeeping for use in developing national estimates of income and other macro aggregates. The article in this volume by Zoltan Kenessey, which deals with the American contributions to the establishment of national accounting, contains references to his ideas in this regard.

Z. Kenessey/The Genesis o f National Accounts

5

1.4 The Input-Output Work o f Leontief Input-output statistics and analysis was a crucial invention of the 1920’s and 1930’s for the presentation and interpretation of macroeconomic relationships in general, and the structure of modern economies, in particular. Initially the system of national accounts developed separately from input-output. However in the subsequent revision of the UN System of National Accounts (SNA) it was incorporated as well. Even today, as Graham Pyatt’s work in this volume mentions, the relationships between input-output, commodity balances, and national accounting are insufficiently understood in respect of Stone’s pioneering work. Input-output, as masterly constructed by Wassily Leontief and tirelessly implanted by him in a hundred or so countries, ought to be considered as an important constituent of the intellectual heritage pertinent to the evolution of national accounting. Such a viewpoint is justified conceptually and statistically, notwithstanding the semi­ independent development of input-output and the SNA for some years. Indeed, the early efforts regarding the establishment of the national economic balances of Russia for 1923/24 (which included a kind of interindustry table) should be recognized when speaking of the wide, sometimes rather indirect, intellectual heritage of contemporary national accounting.2 1.5 Tinbergen and the Dutch Contributions Several efforts by Dutch economists and statisticians, including by the best known of them, Jan Tinbergen, were also very important among the intellectual sources of modern national accounting. These works impacted the genesis and the elaboration of the accounts not only in the Netherlands but also internationally. In our volume, the study by Gert den Bakker highlights Tinbergen’s work and its influence. The pioneering econometric work of Tinbergen for the League of Nations stemmed from similar concerns about the Great Depression which lead to the formulation of the General Theory by Keynes. As did the national accountants, Tinbergen’s econometric work also relied on the Keynesian conceptual apparatus, even though he and Keynes differed about the potential of econometric methods in economic inquiries. 1.6 Copeland’s Moneyflows The studies that made Copeland famous in the postwar era pertained to the ’moneyflows’ or the ’flow of funds’ as they are still called. These studies themselves were important milestones in the evolution of postwar macroeconomic accounting. However, in referring to the prewar intellectual heritage of modern national accounts special reference ought to be made to Copeland’s emphasis on the usefulness of accounting concepts for macroeconomic purposes in general, and to the application of double-entry bookkeeping, in particular. References to this matter are found in this volume in the article by this writer about the American contributions to national accounting.

6

Z. Kenessey/The Genesis o f National Accounts

2. The Distinguishing Features of the New System The postwar era of national accounting - in contrast with the preceding centuries of national income estimating since Sir William Petty and Gregory King - has been characterized by major new developments, especially the following six: (a) The systematic application of macroeconomic theory (b) The introduction of double-entry accounting (c) Increased policy orientation (d) The portrayal of the whole economic process (e) The internationalization of methodological development (f) The mathematization and computerization of the work The present section of this overview attempts to show that these characteristics, while all existed to some degree earlier, indeed provide for a divide between the postwar period and the times preceding it. That is to say, national accounts - while an outgrowth of national income estimating - should be considered as a qualitatively new stage in the history of the subject. 2.1 The Systematic Application o f Macroeconomic Theory Theoretical considerations emanating from economic theory, of course, always played a role in working on national income estimates. However, this relationship became much closer with the establishment of national accounts. Historically, the macro emphasis found in Quesnay, Smith, Ricardo or Marx almost vanished by the end of the 19th century. Once the marginalist thinking became dominant, the earlier emphasis got nearly lost. However, the Great Depression and the Second World War rekindled interest in macroeconomic theorizing. Most importantly, the work of Keynes (especially his General Theory) was accepted as guiding by most of those working on the new systems of national accounting. 2.2 The Introduction o f Double Entry Bookkeeping The acceptance of double entry bookkeeping techniques into national accounting should not be considered as a mere technicality. It is, of course, a technical development -but a very important one, that goes to the heart of the new system. Indeed, probably it is a key element of the divide between earlier national income estimating and contemporary national accounting. Indeed, the systematic application of modern tenets of macroeconomic theory to national income seemed to require the introduction of this technical innovation into the work. One can say that the discipline of this bookkeeping technique, and the way it permits relating flows of expenditures to incomes, production to consumption, or savings to investments was indispensable for the systematic integration of statistical macro aggregates. The decisive step for accepting this technique for national accounting was made in the United Kingdom, and Sir Richard Stone, who was involved in that work, later

Z. Kenessey/The Genesis o f National Accounts

7

became instrumental for its international acceptance as well. However, as it is discussed in the present writers article in this volume during the discussion of American stimuli to national accounting, the intellectual impetus apparently originated with Irving Fisher and Morris Copeland in the U.S., even though the immediate influence on the British work seems to have come from Hicks. Also, later Stone emphasized - for example in the 1968 revision of the SNA - the matrix presentation of the account, as underlined in Graham Pyatt’s note in this volume. 2.3 Increased Policy Orientation From the earliest times on, the results produced by estimators of national income were presented and considered for policy purposes. There is no need to provide the particulars - from Sir William Petty and Gregory King on to the 20th century - which attest to such an orientation. Yet, the circumstances of the middle third of the 20th century very much accentuated policy needs for aggregate economic measurement. The two outstanding demands, of course, came from the calamities of the Great Depression and the Second World War both of which engulfed most countries in a profound way. Policy makers, as well as economists connected to them, struggled with seeking answers to the contraction of production and employment first, and to the stresses of the war on their economies thereafter. The significance of this turn in attention, which was already cited in connection with the return to macroeconomic concerns by economic theorists, cannot be overemphasized when reviewing the evolution of the new accounts. 2.4 The Statistical Portrayal o f the Whole Economic Process Another new feature of the contemporary national accounting work has been its comprehensiveness, and the detailed nature of the estimates. Of course, in the interwar years, for example in the United States, there was considerable expansion in national income estimating, with a substantial increase in the detail provided by the estimators. Nevertheless at that time the attempts at interindustry statistics (input-output) and money flow statistics (flow of funds) were essentially separate developments. National accounting work today, internationally as well as nationally, provides for the relating of all such macroeconomic measurement efforts to each other. This permits a much more comprehensive treatment of the interrelated economic phenomena than was the case before the war. Also, there are great advances in the detail provided about all macroeconomic aggregates. 2.5 The Internationalization o f Methodological Development Before the war national income work was essentially a national affair. There was, of course, interaction between the scholars of different countries who were involved in such studies. Yet, actual estimating work (even if its one intention was to compare a

8

Z. Kenessey/The Genesis o f National Accounts

certain country’s conditions with those of another) was pursued separately within each country. The decades since the Second World War, in contrast, witnessed unprecedented international interaction and cooperation in respect of developing national accounts. Indeed, the development of national accounting in the last fifty years would be incomprehensible if this key characteristic was neglected. The internationalization of methodological work came about via the cooperation of national statistical authorities, mainly within the framework of international organizations such as the United Nations and OECD (and earlier: OEEC). Several articles in this volume provide analysis about this aspect; indeed almost all of them touch upon (or are related) to this feature. However, those written by Anne Harrison and Frits Bos present their discussion in this context. 2.6 The Mathematization and Computerization o f the Work As it is well known, many important prewar research results already pointed towards mathematization as a major trend in economics and having a relevance to economic statistics proper. In particular, the input-output system of Wassily Leontief and the econometric work of Jan Tinbergen can be cited in this context. In contrast, computerization, of course, played no role yet in the immediate postwar work on national accounts, neither figures computing as a major factor in the history in prewar calculations of national income (Babbage’s pioneering computing machines from the 19th century were all but forgotten). However, in due course computers became all pervasive and an integral part of statistical work (for and on) national accounting practically everywhere. This is a very important and distinguishing characteristic of postwar national accounting work (especially after the 1960’s).

3. Key National Contributions to the New System Among the major sources of influence upon postwar national accounting it seems particularly useful to highlight the work undertaken in the United Kingdom, in Norway, and in The Netherlands as far as European countries are concerned; and regarding the Western Hemisphere to focus on the efforts in the United States. In both Europe and in the Americas some other works may be cited; in Europe Swedish and in America Canadian efforts come perhaps first to the mind. It seems, however, that research in the selected four countries exercised particularly important affects on postwar national accounting development, especially in the international sphere.

Z. Kenessey/The Genesis o f National Accounts

9

3.1 The United Kingdom The story of the establishment of national accounts in the United Kingdom is a truly fascinating web of efforts by brilliant researchers, administrators and courageous leaders all involved in the organization of the deadly struggle of their nation during the Second World War. An account of this story by the present writer is in the Festschrift issued by the Central Bureau of Statistics of The Netherlands in 1993 on the 50th anniversary of national accounts work in that country.3 Besides Bowley and Stamp, who were involved in the interwar national income work and in some related efforts of significance to the establishment of national accounts later - the figure of Colin Clark looms large in the intellectual preparations for the extended national accounts which were later established in the UK. In 1932 and 1937 he published improved estimates of UK national income. He started working on his most important work, ’The Conditions of Economic Progress’ in 1935 and it first appeared in 1940. Soon after the outbreak of the Second World War John Maynard Keynes published in ’The Times’ three articles in November 1939, which were revised and in 1940 reprinted as his ’How to Pay for the War’ (London, Macmillan). The appendix to this work contained accounts on production, income and capital formation as well as on the income and outlay of the government and the private sector regarding 1938/1939. As already mentioned, Yoshimasa Kurabayashi' $ work in our volume focuses on this work. The role of Meade was very important at this stage. Meade, before the CSO of the UK was established and before Stone arrived to join the work, prepared in 1940 a secret paper involving national accounts estimates that utilized double entry bookkeeping methods for the UK data (published later in ’The collected works of James Meade’, Volume I, Employment and Inflation, Chapter 7, London, Unwin Hyman, 1988). Meade continued to work on national accounts in the CSO for a while with Richard Stone. Keynes recommended that national accounts should be appended to the UK Budget presentation in April 1941. As time went on, Meade got involved in various other matters, while Stone’s focus remained on national accounts throughout the war. His interest in the matter was inspired by his Cambridge tutor, Colin Clark. With a view to their entire life and oeuvre, today we can say that Meade’s initial work on national accounts was crucial for its establishment in the UK, as was his work with Stone 1941 for the White Paper and as was their joint article the same year, (Meade and Stone, 1941). However, Stone’s work over the period of the next decades became seminal for the international establishment of national accounts. Of great importance was Stone’s chairmanship of the 1945 expert international expert meeting on the subject and his authorship of the basic study that was reviewed at that ground-breaking event on the international plane. He also played key roles in other international efforts, including those at the OECD, and was the main architect of the 1968 revision of the UN SNA.

10

Z. Kenessey/The Genesis o f National Accounts

3.2 Norway There are certain similarities to the UK in the story of the Norwegian accounts. Interest in national income was already present before the war; in the person of Ragnar Frisch an outstanding academician provided key stimulus at the University of Oslo to work in this direction; the war years and the expected needs of postwar reconstruction focused thinking on the capacity of the economy to achieve the difficult tasks ahead; and in the governmental statistical apparatus a place existed for the national accounts work needed. Moreover, from the immediate postwar years on, Norway joined the international efforts with great enthusiasm and much dedication. Naturally, the parallelisms should not be carried too far: the size of the population and of the economy of Norway cannot be compared to that of the UK - which at the time of our story was a leading world power - and, of course, Norway was occupied during the war years. Therefore, the situation in Norway, despite some overall similarities, was also unique. The establishment of national accounting in Norway, one might say, was the implementation of a vision of Ragnar Frisch regarding the portrayal of the whole economic circulation process. This vision ultimately aimed at the quantification of his famous Okosirk-systemet. The development of the Norwegian work in this regard, and indeed the scope of the Scandinavian contributions in general, are described in the major study of Odd Aukrust in this volume. The results of a wartime national accounting study of the CBS were published in 1946, which also marked the time Odd Aukrust took over national accounting in the CBS. In the meanwhile Petter Jakob Bjerve joined the Ministry of Finance where national budgeting was started (and until 1950, current national income estimating was also with the ministry). Aukrust, starting in 1946 (and still going on strong today) relied in particular: on ideas from Frisch’s circulation system; on Stone’s report at the League of Nations 1945 meeting (the mimeo version of the report, which was printed by the UN in 1947, were available already in the previous year); and on the commodity flow approach as applied in Denmark earlier by Kampmann. Aukrust’s work preceded the international codification of national accounting by the UN. 3.3 The Netherlands The first and greatest name among the Dutch inspirations to national economic accounting is that of Jan Tinbergen. One of the lesser known facts about Tinbergen (outside Holland) is his prewar service beginning 1927 in the business cycle research unit at the Central Bureau of Statistics, which should be considered a significant matter in the context of our topic. Tinbergen, of course, earned world fame with his pioneering econometric model building which goes back to the 1930’s. Generally, the outside world became attentive to his efforts in connection with his work for the League of Nations, which under its research director, Alexander Loveday explored the various theories and methods known at the time for examining the business cycles.

Z. Kenessey/The Genesis o f National Accounts

11

Actual national accounts compilations, pertaining to 1938, were available first in 1944 in The Netherlands. These were undertaken at Tinbergen’s and Derksen’s direction and apparently reflected much of Derksen’s contribution. In this volume, the study of Gert den Bakker provides a description and analysis of the establishment of national accounting in the Netherlands. Without going into the details, it seems worth mentioning that the Dutch work, just as the efforts in Scandinavia, benefitted from the availability of good industrial statistical data and could make use of the commodity flow approach. Also, similarly to the UK and Norway, the Dutch experts on national accounting from the beginning participated in the international efforts as well. For example, Derksen was a member of the 1945 international meeting under Stone’s Chairmanship, and later served as chief of national accounting for the UN Statistical Office. 3.4 The United States As already mentioned, during the 1930’s a significant expansion of national income work took place in the United States. The most outstanding author of the era was Simon Kuznets. However, many others were also involved in the work, much of which was related to the interest of the U.S. Congress in the problems of the Great Depression. The account about American contributions to national accounting in this volume, by the present writer, complements earlier accounts on the subject by Carol Carson and John Kendrick. As regards data work, the American efforts were apparently more diverse and comprehensive than most other studies at the time. Interestingly, the first major governmental report on the subject sold in more copies than the statistical yearbook of the same year. From the point of view of conceptual development, an interesting circumstance is the early discussion of the usefulness of double-entry bookkeeping techniques. One can find very early and quite explicit references to this already in 1906 by Irving Fisher; and later in the 1930’s by other writers, most notably Morris Copeland. Nevertheless, in the end, the acceptance of double-entry techniques in American national accounting apparently came about more via the influence of Stone, Hicks and the international climate of the time, than from the direct influence of the American writers on the subject.

4, Some Related Matters 4.1 Developments in Other Countries The four countries highlighted above, of course, were not the only ones which contributed to the evolution of national accounts. Yet, for one reason or another, circumstances regarding national accounts were apparently less favorable in many other countries. The political constellations in the 1930’s, and the Second World War itself

12

Z. Kenessey/The Genesis o f National Accounts

played a great role. In those years the situation, for example, in Germany, Russia, Italy or France was certainly not conducive to this type of work, especially in an internationally cooperative fashion. The dangers of leaving out great cultures, such as the French, German, Italian, Indian, or Chinese from thinking about national accounts is best illustrated by Professor Yoshimasa Kurabayashi’s paper in this volume which contains valuable references to Japanese studies during the war. Francois Furquet’s study published in 1980 contains extensive information about French developments in the field. This and other studies done in France are unfortunately not much reflected in this volume because a hoped for contribution could not materialize; even though in the postwar years French national accountants maintained distinct positions on certain methodological issues of national accounting. However, fortunately regarding Germany the reader can be referred to UtzPeter Reich’s article in this volume, which describes the dilemmas faced in that country in respect of national accounts. The review by S.G. Tiwari, also in this volume, portrays developments in India, and Janos Arvay’s account about the Material Product System, that was used for decades in the former socialist countries, introduces the reader to the experience obtained in those countries with a more limited (and different) system of national accounts. 4.2 Developments Regarding Related Subjects Similarly to the story of French national accounts, some other developments, related to the evolution of modern national accounting, also need further research and/or are not adequately explored or represented on the pages of this volume. For example, the evolution of the International Comparison Project, still awaits a more exhaustive description. Similarly, the wide ranging work undertaken under the aegis of IARIW itself requires more analysis. So does the extensive program that has been carried out in OECD. It should be noted that a National Accounts Research Unit was set up in Cambridge in 1949 by the OEEC under Stone’s direction with the purpose of training statisticians from member countries in national accounting methods. Two reports resulted: A Simplified System o f National Accounts (OEEC, 1951) and A Standardised System o f National Accounts (OEEC, 1952). In addition, a number of National Accounts Studies were published dealing with individual countries, including one for France which was prepared by Marczewski (OEEC, 1952). The Unit closed down in 1951.4 Also, in OECD (and OEEC) pioneering work was done by Gilbert, Kravis and others about the international comparisons of the purchasing power of currencies and of national products. The important topic of national wealth accounting needs more coverage as well, including the key studies of Raymond Goldsmith. Overviews about the proposed extension of national accounts, by the Ruggless and others are also yet to come. One of the conceptual issues is, however covered in this volume by Harry Postner who deals with the matters of quadruple entry bookkeeping. Also, as mentioned, Graham Pyatt provides analysis about the relationships of input-output, commodity balances, and national accounts.

Z. Kenessey/The Genesis o f National Accounts

13

Clearly, a field in which four writers: Sir Richard Stone, Ragnar Frisch, Simon Kuznets and Jan Tinbergen saw their efforts crowned with the Nobel Prize, and one that yielded results of considerable interest to business, government, and research must look to the lessons of past experience, some of which are chronicled in this volume. At the same time, one hopes, the look into the evolution of the field will enhance its future development, as the challenges of our days, and the days to come are hardly smaller than those faced in the past. As already mentioned, this overview referred to the articles in this volume in the context of the main topics discussed in it; regarding the contents of each study readers were referred to the abstracts placed at the front end of each author’s piece. However, it seems useful in conclusion of this introductory overview to highlight one or two features of the individual essays. They concern ideas, or emphases, which I found particularly interesting or enlightening for myself. As these selections are unavoidably subjective, readers probably will find other items as equally or more interesting. However, I am confident that all persons considering the evolution of national accounting will find numerous points in the various studies which will provide additions to their knowledge about the subject. The first - and most extensive - study in this volume, written by Odd Aukrust provides an in-depth review of the Scandinavian work on national accounting. Aukrust is, of course, thoroughly familiar with the development of Norwegian accounts, as one of their masterful progenitors. But he gives ample credit to Swedish and Danish efforts in the field - and in a thoroughly convincing manner. It is fitting, that this broad study serves as the frontispiece of the present volume. Also, his tribute to Ragnar Frisch - and his eco-circ system - shows some less well known sides of Frisch, whom economists often consider only in the context of his work for econometrics. Gert den Bakker’s contribution about Dutch national accounts, while not quite as detailed as the review of Scandinavian developments, also provides a wealth of material regarding the evolution of the Dutch accounts and their relationship to Tinbergen’s work. I believe, that both of these studies, about Scandinavia and the Netherlands, respectively, will remain for years to come essential sources regarding their subjects, including their rich bibliographical references. It seems to be poignant, that the praise of Keynes’ contribution to national accounting, and of the early work of Stone and Meade, come in this volume from a Japanese author, Professor Yoshimasa Kurabayashi. His work is made particularly interesting by his references to Professor Yuzo Morita, who during the Second World War (!) got engaged in a deep-going analysis of Keynes’ work on the subject. Among the contributions dealing with the four key country sources of inspiration to modern national accounting (Norway, Netherlands, the United Kingdom, and the United States) the piece by Zoltan Kenessey highlights certain more or less overlooked aspects of early American national accounting work, in particular regarding Irving Fisher and Morris Copeland. The review by S.G. Tiwari about postwar Indian national accounting developments is a veritable store of information about the very wide-ranging, yet relatively little known

14

Z. Kenessey/The Genesis o f National Accounts

(outside India) efforts undertaken in his country. This is the only country study in the volume, that focuses entirely on the evolution of national accounting in a developing economy, albeit others (especially the article by Harrison) do touch upon relevant subjects. The work by Anne Harrison is an important general study about the revision of the 1968 SNA. Persons interested in this process will find many insights into the discussions leading to the 1993 approval of the latest version of the SNA. Also, Harrison does not neglect to show the impact of the procedural processes adopted in the revision work for the outcome of the efforts. The study of Frits Bos, while it can be seen as a companion to the Harrison article, takes a thorough inventory-taking approach to the successive "generations" of SNA. Thus he compares the changes introduced, as well shows the recommendations that remained constant in national accounts over the decades. Two pieces - by Reich and Arvay - touch upon the intricate connections between national accounts and politics. The work by Utz-Peter Reich provides many insights in the context of national accounting developments in Germany, in particular by exploring certain aspects of the postwar resistance to the introduction of such accounts in his country. The study by Janos Arvay, in a way, provides a requiem for the MPS. Arvay has unique qualifications for chronicling the evolution of the MPS in the former socialist countries and the steps toward introducing the SNA in this group of countries as well. He participated in numerous international meetings organized within the former CMEA; yet he has been also an active participant of SNA related work in UN meetings in Geneva, at the IARIW, etc. His rich material also will remain an indispensable source on the topic for many years to come. Finally, Harry Postner’s note on quadruple-entry bookkeeping illuminates a lessdiscussed topic, and provides insight into the evolution of his theme over the last decades, and Graham Pyatt’s piece deals with the very well known work of Stone as related to somewhat less well known aspects of his research, including the Social Accounting Matrix. Both of these contributions provide food for thought in respect of national accounts otherwise rarely discussed.

Endnotes 1. I have elsewhere provided a m ore detailed analysis of the m atters referred to in the second and third sections of this overview. Kenessey (1993). 2.

See Kenessey (1992).

3.

See Kenessey (1993).

4.

Lady Giovanna Stone commented "...I know that my husband attached quite a lot of importance to this initiative of the O EEC, which preceded that of the UN by two years and was so to speak the prelude to the SNA." (Communication to the author, December 1, 1993). I am grateful to Lady Stone for her references and suggestions about the importance of the O ECD work.

Z. Kenessey/The Genesis o f National Accounts

15

References [1]

Andvig, Jens C. (1986) Ragnar Frisch and the G reat Depression. A study in the interwar history of

[2]

A rndt, H.W. (1988) Colin Clark. In the volume National Income and Economic Progress. Essays in

macroeconomic theory and policy. Norsk Utenrikspolitisk Institute. H onour of Colin Clark. Edited by D uncan Ironmonger, J.O.N. Perkins, Trann Van Hoa, St. M artin’s Press, New York. [3]

Bjerve, Petter Jakob (1989) Okonomisk Planlegging og Politikk. Det Norske Samlaget. Oslo.

[4]

Clark, Colin (1940; second ed. 1951) The Conditions of Economic Progress. Macmillan. London.

[5]

Copeland, M A . (1932) Some Problems in the Theory of National Income. Journal of Political

[6]

Fisher, Irving (1919; first ed. 1906) The N ature of Capital and Income. Macmillan. New YorkLondon.

[7]

Furquet, Francois (1980) Les Com ptes de la Puissance. Histoire de la comptabilitc nationale et du

Economy, Vol. XL, February 1932.

plan. Encres Recherches. [8]

Kenessey, Zoltan (1993) Postwar Trends in National Accounts in the Perspective of Earlier Development, in the Festschrift: The Value Added of National Accounting, Netherlands Central

[9]

Bureau of Statistics. Kenessey, Zoltan (1992) The 1923-4 National Accounts of the Soviet Union. In S. Todd Lowry (ed.) Perspectives on the Adminstrative Tradition: From Antiquity to the Twentieth Century. Selected Papers from the History of Economics Conference 1990. Edward Elgar.

[10]

Kravis, I., Kenessey, Z., Heston, A. and Summers, R. (1975) A System of International Comparisons of Gross Product and Purchasing Power. John Hopkins University Press, Baltimore.

[11]

Kravis, Irving B. (1984) Comparative Studies of National Incomes and Prices. Journal of Economic Literature, Vol. XXII. March 1984.

[12]

Meade, J.A. and Stone, R. (1941) The Construction of Tables of National Income, Expenditure, Savings and Investment. Economic Journal, pp. 216-231.

[13] [14]

Ohlsson, Ingvar (1953) On National Accounting. Konjunkturinstitutet, Stockholm. Oomens, C.A. (1985) De Ontwikkeling van de Nationale Rekeningen in Nederland. Statistisch Magazine, Vol. 5, nr. 2, ’s-Gravenhage.

[15]

Pyatt, Graham (1992) In Memoriam: Sir Richard Stone (1913-1991). The Review of Income and W ealth, Series 38, Num ber 2, June 1992.

[16]

Stone, Richard (1947) Definition and M easurem ent of the National Income and Related Totals. Appendix to the R eport of the Sub-Committee on National Income Statistics of the League of Nations Com mittee of Statistical Experts, entitled: Measurement of National Income and the Construction of Social Accounts. Studies and Reports on Statistical Methods, No. 7. United Nations, Geneva, 1947. (Stone’s study - a revised version of his 1945 paper for the famous Princeton meeting is the largest part of the UN publication).

[17]

Studenski, Paul (1958) The Income of Nations, New York University Press, New York.

16

The Accounts o f Nations Z. Kenessey (Ed.) IOS Press, 1994

THE SCANDINAVIAN CONTRIBUTION TO NATIONAL ACCOUNTING Odd Aukrust Statistics Norway, P.B. 8131 Dep, 0033 Oslo, Norway

Abstract. This paper surveys developments in national accounting theory and method in Scandinavia (Denmark, Norway, Sweden). The paper provides a chronology by countries and authors, with particular attention paid to the pioneering efforts of Ragnar Frisch and Erik Lindahl during the 1930s and the theoretical work by Ingvar Ohlsson and Odd Aukrust shortly after World W ar II. The paper summarizes by subject m atter the ideas contributed to national accounting by Scandinavian economists, stressing differences between Scandinavian and main stream Anglo-American thinking.

Introduction Plan o f the paper National accounts, as we know them today, are essentially a creation of the late 1940s and early 1950s. It was during this period that the ideas were brought together and clarified which a decade later enabled the United Nations to achieve almost universal acceptance for the 1968 version of its System of National Accounts ("SNA Mark II"). However, important work had been done prior to 1940 in different corners of the world. In addition to an influential Anglo-American tradition, much original thinking had been done elsewhere, for example in Scandinavia. It is the aim of this paper (1) to review work in Scandinavia (Denmark, Norway and Sweden) from around 1930 to 1955, by which time Scandinavian ideas had taken their final shape, (2) in so doing, to emphasize points where Scandinavian thinking showed originality and/or deviated from mainstream ideas elsewhere, and (3) to consider the extent to which Scandinavian ideas may have exerted an influence on the present SNA. With three countries to be covered, the paper is rather lengthy. It is organized as follows: Sections 2-4 contain a chronology of Scandinavian national accounting work for 1930-1955, by country and author; in particular, the works by Ragnar Frisch and Erik Lindahl are described in some detail. The reader, if not particularly interested, may pass quickly over these sections. Sections 5 and 6, drawing on the above, summarize Scandinavian efforts (1930-1955) by subject matters. Section 5 deals with issues for which it may be claimed that Scandinavian economists contributed significantly to the pool of ideas out of which grew the 1968 revision of the SNA. Section 6 records, as a matter of historical interest, how a selected number of conventional issues, much dis­ cussed in the standard literature, were treated by Denmark, Norway and Sweden in the

O. Aukrust/The Scandinavian Contribution

17

early days when international recommendations had not yet been agreed. Section 7, finally, rounds off the paper with a brief account of developments in Scandinavia after 1955. Chronology o f Scandinavian work It may be helpful, as an introduction, to sketch the main trends, from the early beginning until today, in the development of national income accounting in Scandinavia. For the purpose of this paper five, slightly overlapping, phases may be distinguished. The first phase was characterized by studies aiming at estimating the value of "the national income", typically based on tax assesment statistics. This phase has limited methodological interest and will not be treated here1. The second phase witnessed the work of two first rate economists: Ragnar Frisch in Norway and Erik Lindahl in Sweden. During the 1930s they spent much of their time on conceptual problems, considered by both as a necessary starting-point for fruitful macro-economic research2. Similarly, for long periods, both of them also engaged in empirical national accounting work. Frisch worked intensively for many years around 1940 on the design of a general national accounting system; empirical work on implementing the system was started on an experimental basis but never completed due to events during the war. Lindahl headed the team which during the early 1930s produced the monumental study "The national income in Sweden 1861-1930" [39]. This work by Lindahl et.al, while conceptually having its roots in phase two, is for our purpose more conveniently consi­ dered the start of phase three. Phase three lasted from the first half of the 1930’s to around 1945. It was characterized by empirical work in all the Nordic countries, largely applying (and occasionally improving upon) Lindahl’s methodology. A new epoch began in 1945/46 when Scandinavian economists learnt about the important developments in the field of national accounting in the English speaking countries during World War II. This marked the beginning of phase four, characterized by the introduction in the Scandinavian countries of national accounting systems as we know them today. The challenge facing those responsible for the work (Ingvar Ohlsson in Sweden, Kjeld Bjerke and Paul Milhoj in Denmark, and Odd Aukrust in Norway) was obviously to combine the best of Scandinavian and Anglo-American thinking. This task was completed well before 1955 by which time all Scandinavian countries had established new national accounting systems. These systems were retained largely unchanged until, just before 1970, a switch was made to "SNA revised". This switch marked the start of the last of our five phases during which new statistical practices led to steadily improving quality of data. However, phase five will be considered only briefly in what follows.

18

O. Aukrust/The Scandinavian Contribution

1. The Scandinavian pioneers: Ragnar Frisch and Erik Lindahl 1.1. Frisch’s system o f economic concepts: The Eco-circ System Economic theory in the 1920s, when Frisch started his professional career, was constrained by the fact that no unified system of concepts and terminology existed for use by economists. This created problems for an efficient exchange of views: It made precise presentation difficult, was a source of misunderstandings, and sometimes resulted in disagreements where no real differences existed. Frisch felt strongly that this confusion had to be cleared up if advances in economic theory were to be possible3. During a period of more than 20 years, starting in 1928, Frisch devoted an astounding amount of his time (and of that of his assistants) to conceptual issues, culminating with his "Eco-circ System", a first version of which was ready in 1942. One additional reason why Frisch considered the work important was his vision of a national accounting system which, based on Eco-circ concepts, at some future date would provide the data needed for turning macro-economics into an empirical science and serve as a base for macroeconomic policy. The latter, obviously, was the ultimate aim of it all4. Frisch approached the problem of standardizing concepts for use in economics in the following spirit: "The standardization must aim at various things: The logic of the concepts and of the relations between the concepts, the terminology, the notation in the form of letters, the graphical representation, and the accounting system. All these various forms of expression must be made as conform to each other as possible so that from one of them any of the others may be read off. ([26] p. 106, also in [2] p.2). In other words, the system should allow the concepts and the relations between them to be described, alternatively, in terms of definitional equations, graphs, or sets of accounts. The Eco-circ System in its original (1942) version was general in the sense that it was applicable to any sector. The system allowed for a detailed description of what went on inside each individual sector, each sector being considered in turn, and of that sector’s transactions with all other sectors combined; when aggregated hierarchically over all national sectors a picture of the total economy would result. What the system did not allow for was a description of external flows between pairs of sectors, i.e. there was no place for "from whom to whom" analysis. In contrast, a revised version of the system from 1948 distinguished between a private and a government sector and showed the flows between the two explicitly while in other respects retaining the system unchanged5. The graphic representation of the original Eco-circ System is reproduced in Appendix A. More relevant for our purpose than the graphic representation are the economic ideas of the Eco-circ System. A selection of them, including in particular ideas which have subsequently influenced work on national accounting in Scandinavia, are noted in the following. One favourite idea by Frisch was his insistence that the concepts used in national

O. Aukrust/The Scandinavian Contribution

19

accounting must be established in an axiomatic manner. For instance, in [2] Frisch writes: "It is through these [definitional] relations that the logical content of the various variables is established. The meaning of any given variable can indeed never be exhaustively defined through an enumeration, however extensive, of concrete elements which the variable is assumed to embrace. To be logically complete the definitions must therefore be established by an axiomatic procedure, and the essential part of this axiomatic procedure is the setting up of the definitional relations between the variables. ....when this logical structure is established it becomes a matter of practical convenience and convention to decide how much of the concrete data shall be thrown into one of the variables and how much into another" (p. 18). In Frisch [28] and [29] he reasons along similar lines in an attempt to come to grips with the market price - factor cost issue. To me, there are two ideas which more than anything else give Frisch’s system its distinctive character. One is his insistence that "real" phenomena must be clearly distinguished from "financial" (monetary) phenomena. The other is a recognition that economic life (and therefore national accounting) is about economic objects. The two ideas are interrelated (all quotations in this and the following three paragraphs are from [2]): "The distinction between the real circulation and the financial circulation of the economy is a distinction according to the nature of the economic objects that make up the flows in question, the real circulation being the flows and stocks of real objects (goods and services) and the financial circulation being the flows and stocks of financial objects (economic claims and counterclaims, in a wide sense)".(p.6). The nature of real and financial objects respectively is described in the next paragraphs as follows: "One of the main characteristics of the real objects is that they would be of economic importance even if no property rights existed. In such a case there would be no financial objects at all. A real object may be defined without taking any regard to ownership, while a financial object can only be defined in relation to a certain creditor and a certain debtor. Even if it has an owner (a creditor), a real object has as such no debtor, a financial object has both creditor and debtor .... (p. 6) ... It should be noted that the property right to a real capital is something different from the real capital itself ... the property right may be sold ... without moving the real capital o b ject.... If this is done the property right should be considered as a financial object" (p. 7). In general (though Frisch does not say so) it is possible to consider a real capital (a farm) as being an asset of one sector (agriculture) while at the same time treating the corresponding property right as a financial object with some other sector (the farmer) as creditor and the first sector (agriculture) as debtor. This construct may be used for example for classifying real capital by industries. Having defined economic objects it is possible to give precision to the definitions of transactions and flows:

20

O. Aukrust/The Scandinavian Contribution

"A transaction ("micro-phenomena", "micro-flow") is something which happens to an economic object, e.g. "... the fact that say 1.3 Kg. of a specific commodity is being handed over from the store N.N. to the housewife P .P ...." (a real transaction) o r "... the fact that the housewife P.P. hands over to the store N.N. a certain amount of money ..." (a financial transaction). Flows ("macro-flows") are aggregates of transactions ("micro-flows")" (p.4). Frisch’s picture of the real circulation (any sector) was as follows (symbols refer to the Eco-circ graph, Appendix A starting from the bottom): Services rendered by labour, real capital and organization belonging to a sector (R",Rk,R°) combine to create the net product, or real income, (R) of that sector. The net services of the factors of production (R) together with depreciation (D) defines gross value added (E). Looked at from anot­ her angle gross value added (E) is defined alternatively as gross output (A) minus intermediate input (H). Gross output (A) is used either for gross investment (J), or con­ sumption (C), or intermediate input (H), and possibly for (real) "taxes" (T) (When the sector under consideration is the national economy, T may be thought of as government consumption (Tin) and/or contributions to other countries). Net investment (I) results in an increase in the sector’s stock of real capital (K) and an equally large negative change in its "real funds" corresponding to its savings (S). The real capital may change not only as a result of investment but also because of occasional gains/losses (F and G). In an open sector most of the above flows must be split in order to describe the real flows to and from other sectors separately. Note, in particular, that real factors of production "belonging" to ("resident" in) the sector may be active in production outside the sector. This issue is taken up in 5.7 below. In the orginal (1942) version of the Eco-circ System Frisch made the financial circu­ lation an exact mirror picture of the real circulation6. The correspondance related to concepts as well as their graphic and symbolic representation: Corresponding to net real investment ("real" saving) there was net financial investment. Corresponding to real capital there was financial capital. While real capital generates real income (i.e. services of real capital, or rent), financial capital generates financial income (interest and divi­ dends). Real and financial categories may be added: The sum of the sector’s real capital and net financial capital is its total capital. The sum of the sector’s real income (net value added) and its net financial income is its total income7. In the 1948 version of the Eco-circ System [2] only three financial concepts were retained: financial income, financial saving, and financial investment. Before we leave Frisch’s work on conceptual issues a few word must be said about his views on the market price-factor cost issue. Frisch did not accept the view that national income or product at factor cost, and national income (or product) at market price, as usually defined, were two measures of the same total, only valued differently. Rather, Frisch maintained that the former should be considered a part of the latter, and therefore itself a market price concept. He wrote at least two papers on the issue [28] [29] in which he reasoned roughly as follows: Assume that the value of the national product is estimated directly by valuing its components (e.g. consumption, investment, net exports) at market prices. Then any deduction from this total with the idea of

O. Aukrust/The Scandinavian Contribution

21

arriving at the "true" costs (the "factor costs") of producing the total must imply either that some part of the national product was not created by the factors accounted for, but by something else; or that someone who has not himself contributed to the creation of the national product (e.g. general government) exploits those who have (e.g. through indirect taxation). Frisch clearly preferred the latter way of looking at things. Very little about the Eco-sirc System, or about Frisch’s work on conceptual issues as a whole, was ever published in print, and even less in an international language. By the end of World War II, therefore, this part of Frisch’s work was largely unknown outside Scandinavia and it had no influence on work done internationally at the time. To future generations of Scandinavian economists, however, the Ecocirc System has, as we shall see, provided a solid conceptual foundation for empirical work. Through their efforts Frisch’s ideas may, indirectly, have come to influence the international thinking in the field. 1.2 Frisch’s empirical work on national accounting Frisch seems to have toyed with the idea of constructing a "nasjonalregnskap" already in the early 1930s and the term was used by him in print in 1933 (Bjerve, [15])8. The plans materialized when, in 1937, funds were made available to the University Institute of Economics for a project aiming at "a structural analysis of the Norwegian economy" (0konomisk strukturunders0kelse for Norge, 1937). The project was to include a detailed empirical description of Norwegian economic life, industry by industry, in nati­ onal accounting terms. The national accounting part was to be Frisch’s contribution to the project. For some years, until the University was closed by Nazi authorities in 1943, Frisch had a small group of assistants working numerically on the project. At the theoretical level work proceeded in parallel with work on the Eco-circ System, aiming at an accounting system in conformity with the latter, but accomodating much more de­ tail than could be conveniently shown in a graph. At the empirical level the system was tested for one year (1935) by numerical estimates for selected sectors. No published document exists which describes in full detail Frisch’s vision of a numerical national accounting system. However, his main ideas may be judged from his project proposal in (0konomisk Strukturunders0kelse, 1937) and, in particular, from his prepared statement [24] to a conference of Nordic Statisticians in 1939 at a time when the project was well under way. Some points from the latter document are noted in the following paragraphs. Frisch’s definition of national accounting sounds modern even today: "By national accounts we mean not only a picture of the national income in a given year or the national wealth at a given point of time, but a reasonably complete survey of the total economic activity o f a nation in a year, presented in a way which allows the interrelationships between different data to be clearly demonstrated. The presentation must be such that relationships which are particularly interesting show up by accounting necessity. For instance, change in wealth should by accounting necessity correspond with data on income, consumption and saving. And the correspondance between the real and

22

O. Aukrust/The Scandinavian Contribution

the "personal" method for estimating national income should show up automatically ..." ([24] p. 141-142. My translation). Note, however, that Frisch when talking about"... interrelationships between different data ..." did not have in mind inter-sectoral flows; as explained above, the 1942 Eco-circ System related to only one sector and its transactions with all other sectors combined, and this characteristic was retained in the accounting system. Consequently, the weight was on defining a standard system of accounts for this sector which, when aggregated hierarchically over all sectors, would comprise the national economy. The sector division envisioned treated the economy without geographic specifications, but was in other respects extremely detailed: "We apply a classification of sectors in two directions. First, according to the organizational form of the individual units, this leads to the organizational sectors shown horizontally ... second, according to an industrial classification shown vertically. A cell in this two-dimensional grouping - e.g. the collection of all one-person enterprises in the wood industry - will be called a structural sector. And it is for such structural sectors that a description in accounting terms will be attem pted....... Within each structural sector there may be individual sectors ... we may also try a sub-grouping according to the size of the enterprisse ..." (p. 146-147). The set of accounts to be used for each sector was no less ambitious9. Obviously there would have to be a standard system for use everywhere since " ... a hopeless confusion would result if we were to use different systems for different sectors ..." (p. 147). The accounts should distinguish clearly between real activities and financial activities, and it would be useful to organize the accounting structure in such a way that each production account showed value added as its balancing entry. The problem of valuation was discussed at considerable lenght. Ideally, Frisch said, one would like to value goods and services by valuation coefficents reflecting their utility. At some future date this might be possible. (Frisch was a believer in cardinal utility.) For the time being one would have to fall back, in practise, on valuing things by their prices. According to Frisch three types of prices were of interest: (1) "self cost", (2) current market prices, (3) "anticipated values" (by which Frisch meant the value of things - especially capital goods - when a firm was valued as a going concern). Other possible concepts of value, e.g. reproduction cost, insurance value, scrap value, were less interesting and did not have to be considered. A final point worth noting is Frisch’s observation ([26], p. 121) that a set of national accounts may be drawn up either according to "the principle of centralized description" or, alternatively, according to the principle of "decentralized description". In the former case the estimates are prepared by one single observer applying the same principles of evaluation consistently to all sectors, and a numerical balancing of the system is guaranteed. In other cases the estimates may be prepared by different observers for different objects and sectors according to different principles; this is a case of de-centralized description, and numerical discrepancies will arise. Differing economic valuations by different transactors have a role to play in the business cycle, and there may well be a case for adopting decentralized description in a system of national

O. Aukrust/The Scandinavian Contribution

23

accounts intended primarily for business cycle analyses. (This foreshadows a point made by Ingvar Ohlsson in his discussion of the purposes of national accounting, see later.) However, in the actual work of Frisch and his assistants, the principle of centralized description was applied throughout. The empirical work on national accounting started by Frisch in 1937, as already mentioned, was not completed. It was, perhaps, too ambitious for its time. Lindahl, at the conference of Nordic Statisticians referred to above, commented: "Frisch, as usual, has a more imposing program than the rest of us, and we are greatful to him for show­ ing us at this early stage the goal which statistics on national income and wealth should ultimately aim at. I believe, however, that for practical reasons we are better advised to start with something on a lesser scale". ([9], p. 160, my translation). Lindahl’s scepticism may not have been completely misplaced. Yet, in retrospect, Frisch’s pioneering efforts remain an expression of the remarkable creativity of a great mind10. 1.3 Lindahl’s conceptual work It is useful to distinguish two lines of development of conceptual systems in Sweden during the 1930s, both of them connected with the name of Erik Lindahl. One related directly to empirical national income work, culminating with the publication of [39]. This line is considered in the next section as auguring phase III of the Scandinavian national accounting tradition. The other line related to the design of theoretical systems of concepts intended primarily for economic analysis. Lindahl’s immediate purpose seems to have been that of presenting the ex ante and ex post analysis of the Stockholm school in systematic form. The standard reference here is [41] which we shall consider in the following paragraphs11 . Like Frisch, Lindahl strived to achieve a system of concepts as general as possible. In Lindahl’s case generality meant that the system had to be applicable both to ex ante and ex post values, and equally suitable for the description of micro-economic and macro-economic phenomena. By suitable interpretation of the terms involved, the description at the micro level should be equally applicable to firms, households, or any other economic subject. At the macro level the terms might be thought of as primarily refering to the national economy but other interpretations were also possible. Lindahl starts his exposition by setting out the relationships valid at the micro-economic level. He distinguishes two chief categories of economic subjects: (1) "firms", which engage in production and "possess" all capital, and (2) individuals, or private households, who "own" the firms; households get their income partly from selling labour services to the firms and partly in the form of "financial income" resulting from their ownership of the firms (their "financial capital"). Identical sets of equations are used both for firms and for households. Some relations are valid both for ex ante and ex post values of the variables; they are the results of estimates made at the same point of time. Other relations describe the differences which may exist between estimates made ex ante and estimates made ex post. There are also relations concerned with "sub­ jective valuation" of capital and income, either ex ante or ex post, and the whole system

24

O. Aukrust/The Scandinavian Contribution

of micro-economic relationships and symbols ultimately becomes rather complicated. We shall not consider the micro relationships any further. By aggregating the micro-economic equations for all subjects belonging to a group, we get macro-economic equations valid for that group. If the group is "the nation", the resulting equations will be definitional equations relating to the national economy. A selection of such equations, valid ex post, with Lindahl’s symbols and terms ([41], pp. 114-115), is set out in the diagram on the opposite page. A "receipts-expenditure" equation, which at the micro level was a cash equation, at the macro level becomes the balance of payments. The "earnings equation" shows national income (E) to equal value added (A-B + I) plus net income from capital abroad (R*) Other equations relate to the use of the national income, or show saving to equal net (real) investment plus net financial investments abroad, or the national wealth to equal (real) capital plus net claims on other nations. 1.4. Kindred souls As originally set out in [41] Lindahl’s system of correlative definitions did not immediately address national accounting work, the value and feasibility of which he rather doubted, as seen in his comments on Frisch quoted above. Yet it had obvious implications in that direction. That this was the case was made clear by Lindahl himself 14 years later when, commenting on Ingvar Ohlsson’s dissertation, he returned to his system in an article called "Nationalbokf0ringens grundbegrepp" (The basic concepts of national accounting) [42], This work by Lindahl shows more explicitly than [41] how Lindahl’s views on important points coincided with those of Frisch and deviated from thinking outside Scandinavia. Indeed, Lindahl’s work in Sweden had striking similarities with Frisch’s work in Norway. We have noted how both scholars were led to work on conceptual issues as a preliminary to analytical work. Both aimed at generality. Both obtained generality by deriving their concepts from systems of definitional equations, general enough to be applicable to any economic subject, or group of subjects. The definitional equations used by Lindahl [41] were identical in important aspects with those used earlier by Frisch [23], and so were - whether by intention or accidentally - some of the symbols and notations used. Lindahl and Frisch of course knew of each others work and, as far as is known, thought highly of each other12. Lindahl’s article from 1954 [42] made clear that the kinship between himself and Frisch extended beyond the formal properties of their systems to substance as well. Lindahl, like Frisch, considered a distinction between "real" and "financial" to be important. Real factors of production are real capital and labour; the services of real capital and labour generate the (homeproduced) income (which Lindahl in 1954 calls national product in accordance with what by then had become accepted Scandinavian terminology); ownership of real capital is financial capital; financial capital generates financial income (interest, dividends, etc.). As I read Lindahl, financial capital includes "direct investments" and financial income includes income from such investments, at least at the national level:

O. Aukrust/The Scandinavian Contribution

25

Lindahl’s system of macro-economic equations All equations are valid ex post for the national economy. All variables are in value terms. The "equation of international payments " corresponds to what at the microlevel was termed "the cash equation". The "earnings equation" shows national income (E) to equal value added (A - B + I) pluss (net) income from capital abroad. The remaining equations relate to the use of the national income, or show saving to equal net (real) investment pluss (net) financial investments abroad, or the national wealth to equal real capital pluss (net) claims on other countries.

All equations are valid ex post for the national economy. (All variables are in value terms) Equation of international payments: A' (B‘ + C ') Exports Imports •. Exportsurplus

+

R' Income from capital abroad

+

B

+

F' Net lending to abroad and net import o f securities

(-T ) Gifts, etc. from abroad

M* Net import o f gold and cash

The earnings equation E

=

A >

+

»*'

Income from factors (value added)

The use of income equation

E

The saving investment equation: The equation of wealth:

=

C

+

T"

+

R’ Income from financial capital (neglecting accrued but not yet realized interest on investments abroad)

S

S = I + J + M W = K +

//'

+

M"

List of terms and symbols (* used as superscript indicate relationships with other economies) A B C E F

= = = = =

sales purchases for production purchases for consumption net income net lending

I J S K R T W

= = = = = = =

net investment financial investment saving value of real capital income from capital taxes and transfers net value o f wealth

A* BC‘

= Export, (sales to abroad) } = import (purchases from abroad)

F* H’

= =

net lending to abroad net value of financial capital

M' R’ T-

= = -

holdings of international money income from capital abroad taxes and transfers to abroad

26

O. Aukrust/The Scandinavian Contribution

"... If a foreign firm has a branch within the country we must imagine that the branch is a national firm, even though it is controlled by foreign interests..... This construction, which is based on the distinction between real and financial capital, makes it possible for the national product [domestic product in SNA terms, my remark] to embrace the total product value generated by factors of production (labour and real capital) within the country...... It follows, moreover, that all the capital investment which one country makes in another is considered as financial investment ..." ([42], p. 88; here and elsewhere in what follows quotations and references are from the English version of the paper.) These were ideas which were consistent with those of Frisch. They were at the same time in sharp opposition to contemporary Anglo-American thinking, a fact about which Lindahl was very explicit13. A related point, admittedly more a convention than a principle, where Lindahl seems to side with Frisch, concerns the question of residence and in that connection the delimitation of the national economy. Finally, Lindahl is in complete agreement with Frisch, and in opposition to Anglo-American thinking, on the issue of market price-versus-factor cost valuation, and, in this connection, the meaning of the concept factor incomes itself. Lindahl’s views are neatly expressed by him in one single sentence: "Factor income is thus the part of product value accruing to the factors of production after the government has taken its share in the form of indirect taxes" (p. 82). As a final point it is worth noting Lindahl’s choise of product and income totals for use at the national level. Here, again, he is in accordance with Frisch and his Norwegian followers but in strong opposition to mainstream Anglo-American writing: The national product, net or gross, is defined by Lindahl as the sum of all the product values, at market prices, that relate to resident factors of production. If indirect taxes (the product value seized by the government) are deducted we arrive at total factor income. To the national product may be added (net) financial income from abroad, this gives national income14 which should be used only as a net concept. Finally, there is total disposable income which exceeds national income by the amount of any (net) gifts or other transfer payments received from abroad; this is a measure of the income at the disposal of the nation for consumption and (real or financial) investment purposes.

2. Empirical work in the Lindahl tradition (1930-1945) 2.1. The birth o f commodity flow accounting While Frisch’s work on national accounting during the 1930s had a distinct theoretical bias, Lindahl during much of the same period was engaged in purposeful empirical work. As early as in 1926 Lindahl was given responsibility for a project which, more than a decade later, resulted in "The national income in Sweden 1861-1930", a monumental study in two large volumes [39]15. This study, in which Lindahl was assisted

O. Aukrust/The Scandinavian Contribution

27

by Einar Dahlgren and Karin Kock, marked the beginning of phase three of the history of national income calculations in Scandinavia. It was succeeded, with differences in timing, by work by official institutions in all the three Scandinavian countries in the following years16. A common characteristic of the statistical systems of the Scandinavian countries at the time was that, as compared with other countries, they all had well developed industrial statistics while statistics on incomes were weaker. This favoured estimating national income from "the production side", i.e. building the national totals from estimates by industries using industrial statistics when available, the industry classification being adapted in each case to the statistics available. As initiated in Sweden and developed further in Denmark, this approach led to a fairly advanced form of "commodity flow" accounting which, methodologically, may have been the greatest Scandinavian contribution from phase three. The first step was taken by Erik Lindahl when he planned the Swedish study refered to above. 2.2. "The national income in Sweden 1861-1930" At the end of the 1920’s there was very little guidance for Lindahl and his co-workers to be had from national income estimates elsewhere in the world17. He knew, of course, of the theoretical definitions of the concept of income in the standard economic liter­ ature, but when it came to translating these definitions into practical guidelines, he had to work out his own solutions. For this reason, and also because it gave direction to later work elsewhere in Scandinavia, a short summary of the theoretical discussion in this book is in order. Lindahl began his analysis by observing (chapter 1, volume 1), that the national income could be defined either as (1) total income produced within a country, or as (2) total income received by the inhabitants of the country, the difference consisting of net income received from abroad. Lindahl "for several reasons" preferred national income concept (1), amplified as ".... the total income from agents of production functioning within the country ..." (p. 2) In principle, the estimate should include ".... everything posessing an exchange value .... evaluated at current market prices ...." (p. 6) but, admittedly, the exact borderline would have to be drawn by convention. Thus, Lindahl’s concept of national income came close to what has become known later, in the SNA as "net domestic product at market prices", and in Scandinavia as "net national product". Having defined national income Lindahl proceeded with a theoretical discussion of how it could be measured. He described four different methods summarized as follows (p. 33):18 1) by adding together the net incomes of the different industries, 2) by adding together the net results of national production, i.e. the total consumption and total net investments, plus or minus, respectively, any export or import surplus, 3) by adding together the incomes accruing to the owners of agents of production in return for their contributions19; or

28

O. Aukrust/The Scandinavian Contribution

4) by adding together the incomes accruing to the inhabitants of the country, less their receipts from investments abroad, and plus the incomes of foreigners from investments within the country" ("national income received")." Lindahl’s detailed discussion of the four methods led him to consider a remarkably large number of problems which, later, have become standard in national income literature. Thus, in connection with method 1) he discusses i.a. problems of valuation of goods not traded on a market or traded at prices distorted through price control, monopolies and indirect taxation; the many-faceted problems connected with the treatment of the public sector20; the problem of drawing the borderline between current input on the one hand and consumption or investment on the other; and of the allowance to be made for the depreciation of durable goods (loc.cit. pp. 6-15). More or less identical problems are met with, according to Lindahl, when national income is estimated as a sum of consumptiom, investment and net exports by method 2). Method 3), estimating national income as an aggregate of functional income shares, meets with the problem of estimating the contributions of labour, entrepreneurship and (real) capital from observed payments of wages, interest on borrowed (financial) capital, dividends, rents on leased capital, undistributed profits of corporations etc. Again there are problems related to the public sector, for instance the treatment of interest on the public debt and measuring the contribution to the national income of publicly owned durable goods (pp. 18-22). Finally, estimating national income as an aggregate of incomes received (method 4), meets with well known data problems resulting from tax evasion and the concept of income used in tax assesment statistics (pp. 22-24). Lindahl’s discussion of the above issues is valuable even today. Leaving aside conceptual issues for the time being (some of them are returned to in section 6) we turn to a description of how the estimates were carried out in practice. The statistical material did not allow a free choice among the methods described above. Methods 1) and 2) based on production statistics were preferred, but method 3) had to be resorted to for industries where production statistics were missing. The calculations were carried out industry-wise. Ten activities were distinguished21. For each activity, first net output(or value added), as a step towards estimating national income by method 1), and then the amount of gross output flowing to consumption and investment, as a step towards estimating national income by method 2) were calculated. Different procedures were chosen for calculating net and gross output depending on the statistics at hand. For activities where production statistics were available (commodity producing activities (1-4), except handicrafts) the calculations could be carried out "from the top and down", i.e. starting by estimating the value of gross output and ending by net output after deducting the value of inputs and depreciation. For the remaining activities mostly producing services (5-10), the calculations had to be carried out "from the bottom and up", i.e. starting by calculating incomes paid out as a measure of net output (method 3), afterwards adding debit items to arrive at gross output. The most noteworthy feature of the Swedish inquiry, in some way heralding commodity flow accounting, was the method invented for computing the costs of raw

O. Aukrust/The Scandinavian Contribution

29

materials etc. used as inputs when passing from "gross output" to "net output" in the case of the commodity producing activities: These costs were not computed from the financial statistics of the activity in question (such statistics were generally not available) but, instead, as the sum of inputs received by that activity from other activities as computed from the production statistics of those other activities (adjusted for imports and exports). In order for this to be possible some form of commodity flow analysis had to be undertaken which would show, for each activity, the destination of raw materials etc. delivered from it. In most cases it was possible to judge from the nature of a com­ modity or a service how it was used. However, in two cases (bank services and public services to business) this was not found possible; instead, when estimating national income, the value of these services was deducted in total, as unallocated input, from the sum of net product of activities. In retrospect the Swedish study must be deemed an admirable piece of work for its time. It was well thought out theoretically, and resulted in coherent and detailed data on the national income by industrial origin, on consumption and investment by commodity groups, and on the industrial structure. It included an original methodological idea which a few years later in Denmark came to be developed into a fullfledged commodity flow approach to national accouting22. The study established Sweden as the first country in the world to have national income statistics on an annual basis spanning back to the middle of the nineteenth century23. The careful documentation of sources and methods, described in great detail in over 600 pages in a separate volume, set a high standard to strive for by others in similar studies in the future.

2.3. Denmark: Viggo Kampmann The Swedish study inspired a similar study by the Danish statistical office (Det statistiske Departement) around 1935, aiming at estimates for 1930-1939. Responsibility for the project during its critical stages rested in the hands of Viggo Kampmann, a competent economist who later (1960 -1962) became Danish Prime Minister. The plans for the study, together with some tentative results, were described in a paper read by Kampmann to the Economic Association in Copenhagen in 1942 [34], The discussion which follows is based on this paper which provides livelier reading than the official reports written by others later24. Kampmann’s methodological approach was clearly patterned on the Swedish experience. The object of the study, according to Kampmann, was to measure "the income of factors of production working within the country" including Danish ships in foreign waters. These incomes were to be estimated from production statistics and similar sources, industry by industry, in a way which would at the same time provide information on the national income by consumption and investment. The first step in­ volved, for each industry, estimating the value of gross product (gross output). Then:

30

O. Aukrust/The Scandinavian Contribution

"In order to move from the value of gross product to net product (or income) one has to deduct the value of goods and services used in production; this value may be found either directly through estimates based on costs statistics ["driftsstatistiske beregninger"] for each single industry, or indirectly by classifying the gross product of industries by uses. Furthermore, the last mentioned method allows consumption and investment to be found as a residual, namely as that part of gross output which is not used in production, either immediately or delayed, since it goes directly to consumption, or which is used in production only gradually in the form of buildings and machinery ..." ([34], p. 370, my translation). The "indirect method" which Kampmann describes here is what later has become known as "the commodity flow method". The method was applied by him with commen­ dable consistency. Industry by industry (ten industry groups)25 and commodity by commodity the value of gross output was estimated at market prices. Then, additions were made for imports, and exports were subtracted. The result was "total supply". Finally it was decided, for each commodity, how total supply was distributed by uses (consumption, gross investment, intermediate input by industries). Throughout a distinction was made between "durable" and "non- durable" goods and services. The final result was, for each year, an input-output table showing the supply of commodities by origin, and the absorbtion of these commodities by categories of uses. Through suitable aggregations, the table would provide estimates of gross domestic product, total final consumption, and total gross investment. As a by-product, the method automatically resulted in much useful information on the commodity composition of the output of industries, and on the composition of consumption and investment by commodity and by industrial origin. The Danish input-output tables for 1930-1939 may have been the first to have been produced routinely in connection with national accounting work. For this reason they deserve being reproduced, in English translation, for one year in Appendix B. As regards methodological details Kampmann deviated from solutions chosen by Lindahl et.al. on a number of points, for instance in the treatment of durable consumers’ goods and of domestic and public services. (These deviations are discussed in a wider context in sections 5 and 6.) In many respects, in my opinion, Kampmann’s study was a significant advance on its Swedish model. Thus, an attempt was made for one year to estimate the distribution of the national income by functional shares26, and there were tables giving the net product of industries both at market prices and at factor cost. The greatest advance was, however, that the Danish study included estimates of the national aggregates at constant prices, something which had not been attempted in the Swedish study nor, to my knowledge, anywhere else. Clearly, Kampmann was one of those who earliest and most consistently exploited the opportunities offered by the commodity flow approach to national accounting, and his work deserves unreserved admiration.

O. Aukrust/The Scandinavian Contribution

31

2.4. Norway: The Central Bureau o f Statistics While Lindahl’s work in Sweden was followed quickly by a study along similar lines in Denmark, empirical work in the Lindahl tradition was not initiated in Norway until 1943 when a project was started at the CBS as a result of rather remarkable events27. The project was assigned to a team of young economists including former assistants of Frisch, among them Petter Jakob Bjerve. The aim was to provide data needed for the planning of post-war reconstruction work, not primarily to copy what had been done earlier in Sweden and Denmark. Yet the estimation methods used to a large extent copied those used in Sweden. As Bjerve has recounted: "The Swedish study was the Bible". However, true to the Frisch tradition, Bjerve, responsible for the theoretical part of the final report, conducted his discussion of concepts and methods within an accounting framework, something which must have been rather unusual for its time28. The study covered the period 1935-1943. It aimed primarily at an estimate of net (domestic) product by industry. The estimates were carried out industry-wise, for 13 main industries and 23 subgroups. For each industry, and for the economy as a whole, the net product was split into "wages" and "other incomes", and according to its origin in private or public activity. (These were distinctions not attempted in the Swedish study, and they were clearly inspired by similar distinctions made in Frisch’s experimental national accounting system.) Market prices were applied throughout. Estimates at constant prices were found by deflation with suitable price indicies, industry by industry. Commodity flow accounting was not attempted; thus, no estimates of investment and consumption were forthcoming automatically. Instead, a crude direct estimate of gross investment, exports and imports was made for selected years, and con­ sumption for those years was found as a residual. Other points of interest are returned to in sections 5 and 6. The study described above was published in 1946. In the meantime national budgeting work (planning) had been started in the Ministry of Finance with Bjerve, who had left the CBS, in charge as head of the national budget unit. To cater for the immediate and long-term data needs of the national budget a division of work was arranged with the CBS: while the Ministry of Finance, with only limited assistance from the CBS, would update and continue on a year to year basis the national income series just published, the CBS would be given the time and resources needed for developing a revised and extended national accounting system suitable for the needs of the future29. This will be considered in the next section.

3. The beginning of modern national accounting (1945-1955) 3.1. Influences from outside Interest in national income work in the Scandinavian countries got a strong boost at the end of World War II when post-war economic problems greatly increased the demand

32

O. Aukrust/The Scandinavian Contribution

for national income statistics. It was natural, therefore, that work done during the war on national accounting in the UK and USA attracted attention. In particular, Sir Richard Stone’s work for the League of Nations [60], with its focus on the interdependence of sectors, soon became an important source of inspiration; this was an aspect of national accounting which had been completely neglected in the systems of Frisch and Lindahl. Leontief s work on input-output provided evidence on the usefulness of commodity flow accounting. Obviously the task now facing Scandinavian economists was to construct a national accounting system which would incorporate the best ideas from the West into established Scandinavian traditions. The new wave of interest in national income work started earliest in Sweden where Konjunkturinstitutet, at the initiative of its chief Erik Lundberg, started work on national income statistics in 1944 under the direction of Ingvar Ohlsson. The CBS in Norway followed with plans for a revised national accounting system in 1946 with Odd Aukrust in charge. In Denmark the push forward came in 1948, when better data were needed for national budgeting purposes in connection with the Marshall plan. Responsibility for national income work in Denmark had by then been taken over by Kjeld Bjerke, with Poul Milhoj a close co-worker. The differing starting dates should be noted. Ohlsson and Aukrust completed their national accounting systems by the late 1940s, before international recommendations were available; they therefore had to work out their own solutions, starting from the conceptual foundation inherited from Lindahl and Frisch. This necessitated some original thinking on their part, and resulted in two dissertations [50] [5]. Bjerke and Milhoj, on the other hand, continued their work well into the 1950s and had to pay attention to the international recommendations of OEEC [49] and United Nations [62] and to a considerable degree were influenced by the solutions indicated there30. Methodologically the developments in the three countries during the post-war period took different paths. Denmark continued to use as their main tool the commodity flow method which had already been succesfully used before the war. Norway followed Denmark’s example, but with considerably more commodity and industry detail. Sweden, on the other hand, discontinued completely her industry by industry estimates of the national product finding it more convenient to build her estimates from the expenditure side [50], p. 245. 3.2. Sweden: Ingvar Ohlsson Ohlsson’s activities at the Konjunkturinstitutet from 1944 to 1953 combined responsibility for the empirical national income work of Sweden with theoretical studies. The first couple of years were spent on preliminary studies. Figures based on a new approach, applying a national accounting framework for the first time in Sweden, was published in 1946 [44], The following years saw successive improvements of methods and data until, by the autumn of 1951, the empirical work was crowned with the publication of "Nationalbokf0ring 1946-1950" [45], Two years later followed Ohlsson’s main work, a dissertation under the title "On national accounting" [50]. This was a

O. Aukrust/The Scandinavian Contribution

33

major treatise in which Ohlsson, against the background of existing international literature, justified in great detail the principles which he had adopted for his empirical work. Ohlsson’s main thesis was that in the design of a national accounting system attention should be paid to the institutional conditions of the country in question (a market economy and a centrally planned economy needing different systems), the types of primary statistics available and, in particular, the purposes for which the resulting statistics were to be used. Four different purposes were singled out for discussion: Measurement of results, (production and/or welfare); analysis of income behaviour (including business cycle analysis); analysis of economic structure; and national budgeting. The bulk of Ohlsson’s book was a systematic discussion of how the choice of purpose ought to influence the solutions chosen to key problems in national accounting: boundary problems, sectoring and accounting design, principles for recording transactions, valuation, treatment of the government, banking and insurance sectors, definition of national aggregates. Ohlsson’s theoretical argument was reflected in his empirical work, most notably in the fact that the national accounts for Sweden in [45], were given in two alternatives, each constructed according to its own principles with a particular purpose in mind: alternative 1 was designed for measuring output and welfare and was drawn up so as to facilitate estimation of national aggregates; alternative 2 was intended primaryly for business cycle analysis. The two alternatives differed in important respects: while alternative 1, since it was intended for measuring results, had to contain a number of unrequited (imputed) real flows, alternative 2 recorded with few exceptions only realized monetary transactions. There were other differences also, relating for example to accounting structure and to the treatment of depreciation charges, of repairs and maintenance, of non-monetary benefits, etc. Nothwithstanding Ohlsson’s empirical work it is his theoretical studies, in particular his dissertation from 1953, which remains of interest to the profession today. This study continues to rank as a major contribution to the international national accounting literature. Indeed, most issues known in the literature were considered by Ohlsson in some context or other, and his profound analysis of them nearly always make rewarding reading even today. The book was widely read, and its message was quickly and generally accepted as valid. Those of us who participated at international meetings at the time will remember that, during the 1950s, hardly any discussion of national accounting concepts and practices could take place without somebody taking the floor to remind his fellow statisticians that the purpose would have to be agreed for a sensible conclusion to be possible. Yet, in statistical practice Ohlsson’s dictum - that alternative national accocunting systems should be published for different purposes was nowhere lived up to as far as I know. Rather, the trend has been everywhere to develop the national accounts into a general purpose economic information system by adding steadily new information and more detail31. With 350 large pages Ohlsson’s 1953 study is too rich for anything like a complete summary of it to be possible here. For our purpose we shall have to content ourselves

34

O. Aukrust/The Scandinavian Contribution

with a sample of ideas, selected with a view to their significanse for the development of national accounting theory and as expressions of Ohlsson’s originality. This, however, can be most conveniently done in connection with our discussion of particular issues in sections 5 and 6 below. 3.3. Norway: Odd Aukrust When Odd Aukrust took on responsibility for the national accounting project of the Norwegian CBS in 1946, it was fairly clear that the new Norwegian national accounting system would have to be based on concepts and definitions taken over from Frisch’s Eco-circ System, combined with an accounting structure along the lines proposed in Stone’s League of Nations paper [60], Empirically the data situation invited, equally obviously, an application of the Danish commodity flow approach, coupled with an analysis of public sector accounts according to principles already established by Bjerve at the Ministry of Finance. This approach promised to result in a national accounting system which would provide detailed mapping of flows of goods and services, including annual input-output tables, which were precisely the type of data most needed for policy-making purposes during the post-war reconstruction period32. By the autumn of 1948 Aukrust felt that he had arrived at a satisfactory synthesis of ideas adapted (mostly) from Frisch and Stone. The outcome was a paper contributed to the 1949 IARIW conference [3], where it shared the normal fate of contributed papers of not being widely read33. In the meantime empirical work speeded up under the supervision of Aukrust’s associate Otto Chr. Hiorth and resulted in "Nasjonalregnskap 1930-1939 og 1946-1952" [57J34. A theoretical discussion of the principles underlying the estimates was ready in manuscript form early in 1954 and was published as Aukrust’s dissertation one year later [5]. This publication, which was written in Norwegian, contained as an appendix an attempt to establish an axiomatic foundation for the theory of national accounting. A summary of the axiomatic system was made available in English at an IARIW meeting 10 years later [7], Again it is convenient to postpone a detailed discussion of Aukrust’s theoretical ideas to sections 5 and 6, where they will be considered in a wider context. 3.4. Denmark: Kjeld Bjerke and Poul Milh0j In Denmark work on national income statistics continued for some time along the lines laid down by Kampmann before World War II with only minor extensions and improvements. The break came in 1948 when a national accounting framework was applied for the first time and new concepts were introduced. Early results were published in 1951 with data back to 1947 [19]. A full description of the new system, written by Milh0j, appeared in 1955 [20]. The new system was strongly influenced by the international "standardized systems" which were under development. Thus, the accounting framework adopted for Denmark by Bjerke and Milh0j was nearly identical with the accounting framework recommended

O. Aukrust/The Scandinavian Contribution

35

by the United Nations in [62], The routing of transactions through the accounts was also identical, but with one important deviation: whereas the UN system treated payments of interest and dividends (except interest on public and consumers’ debt) as payments to factors of production, Bjerke and Milhoj remained lojal to the Scandinavian view of considering interest and dividends a special category of transfer payments, and recording them as such. Under foreign influence the Danish accounts strived to find a place for each of the bewilderingly many national product and income concepts so characteristic of the UN system: Using SNA terminology there were "domestic" and "national" totals (differing by the treatment of (net) interest and dividend payments from abroad), each of which capable of being measured "gross" or "net", "at factor cost" or "at market pri­ ces". For domestic use, however, Denmark kept to Scandinavian terminology: The term "product" was reserved exclusively for use with reference to real flows, i.e. as a measure of value added, while the term "income" was applied whenever financial income was also included. In their empirical work Bjerke and Milhoj continued Kampmann’s practice of applying the commodity flow method and input-output accounting, supplemented with an analysis of general government fiscal accounts. However, they improved the results by increasing the number of industries and commodities. Relatively detailed input-output tables (aprox. 30 industries) were estimated carefully for two selected years (1947 and 1949), and by summary methods for other years (1948, 1950-1952). The number of different categories of goods and services distinguished was not reported; from information given in the text, however, it appears that in the "make" table somewhere between five hundred and one thousand groups of goods and services were distinguished. The resul­ ting statistics were published partly within the framework of the accounting system mentioned above, partly as input-output tables, but mostly in the form of a large number of analytical tables.

4. Contributions to national accounting theoiy and method 4.1. "Real" versus "financial" phenomena. Definition of transactions and flows The work by Scandinavian economists on national accounting in the early post-war years was strongly influenced by the pionering work during the 1930s by Frisch and Lindahl. Two ideas, in particular, gave a direction to their work which distinguishes it from Anglo-American thinking on fundamental points. One was the insistence, by Frisch and Lindahl alike, that "real phenomena" were what mattered, and that real phenomena be kept clearly distinct from phenomena in the financial sphere. Real phenomena, Frisch used to say, would exist even in a world without property rights. Real phenomena, therefore, are capable of being described in purely real terms and, in the interest of clarity, they should be thus described. This did not mean, however, that the financial superstructure was not also of interest. On the contrary, important problems in economics concern precisely the interplay between the

36

O. Aukrust/The Scandinavian Contribution

two spheres. A good national accounting system should cater for both. The other idea was the observation that economic phenomena always are related to economic objects (real or financial), which suggests that the entries in the national accounts should be thought of as representing flows and stocks of real and/or financial objects. These ideas of Frisch/Lindahl were worked out in considerable detail by Aukrust in 1949 [3]. Aukrust took as his startingpoint the concepts of economic objects and sectors. Economic objects are of two kinds: Real objets, a category which includes not only tangible goods, but also services, and financial objects by which is meant money and claims and property rights of any kind, including the ownership of a real asset as distinct from the asset itself. The sector concept is used in a rather abstract sense, simply indicating some part (a factory, a company, an industry, a country etc.) of the economic totality. The concepts of economic objects and sectors are interrelated: "The sector will ... command certain real assets, and it will hold a certain amount of financial assets (positive or negative) against other sectors, and it will have a certain production and consumption capacity (positive or zero)" (p. 171). Knowing what is meant by objects and sectors, a transaction may be defined as the passing over from one sector to another of an economic object, real and financial transactions corresponding to real and financial objects respectively. Transactions in the same type of objects may be grouped together, thus constituting what we shall call flows (real or financial). Flows passing from one sector to another are intersectoral flows. We may be interested also in what is going on within a particular sector. This may be described by a system of intrasectoral flows, one example would be a flow of goods which is produced within a sector and consumed within that same sector. Very often a transaction, and consequently the flow which is made up of transactions of that kind, is intimately connected with another transaction moving in the opposite direction, that is, each of the two transactions presupposes the other. We shall talk in this case of (two) requited transactions and, consequently, (two) requited flows (real against real, or real against financial, or financial against financial). In contrast to this, we shall talk of transfer transactions and transfer flows when some objects (real or financial) are passing from one sector to another without any objects moving in the opposite direction. The idea that an economic transaction is something which happens to an economic object was adopted by Ohlsson in his work for Sweden: "Economic transactions occur when goods and/or financial assets and liabilities in one way or another change hands, change place, change functional type, and/or when a service is performed" [50] p. 11. Danish economists were less inclined towards theorizing than their Norwegian and Swedish collegues. When they started work on national accounting in the late 1940s, they seemed to be quite happy to accept the idea, taken over from Anglo- American literature, that national accounting had to do with measuring "transactions" - without caring much about the precise content of that concept. What little was said about the issue, however, was clearly of Scandinavian origin; 'Transactions may be classified as real (goods and services) and monetary (payments). Transactions are mostly two-sided in that goods are exchanged against goods, claims

O. Aukrust/The Scandinavian Contribution

37

against money, or goods against money. In addition there are one-sided transactions in the form of gifts, social grants, taxes etc." ([20], p. 28. My translation).'1 Thus, all Scandinavian countries define transactions in terms of "something happening to economic objects". With this intellectual legacy they found it difficult to accept the Anglo-American view that national accounting had to do with the recording of flows of "payables" and "receivables" (actual and imputed). "Payables” and "receivables" arose in the Scandinavian way of thinking as the ex post consequences of transactions and could not be clearly identified conceptually without a prior understanding of the more basic notions of a transaction and the objects involved in a transaction. 4.2. Towards an axiomatic foundation Aukrust returned to the ideas set out in 1949 a few years later. Following a suggestion by Frisch, he tried to deal in an axiomatic manner with the problems of definition, classification, and measurement in the national accounts in an appendix to his dissertation from 1955 [5], The main ideas may be summarized in the form of a lengthy quotation from a more recent work [7]: "We shall consider the elementary units to be classified in national accounting work to be economic objects (real and financial) rather than economic transactions ... We want to demonstrate that ... some of the most important aggregates in the national accounts can be defined as classes of such elements ..." "For a classification to be possible, we have to postulate that certain distinguishing characteristics (properties) attaching to the individual objects are given, which can form the basis for their classification. In selecting these characteristics we note that, typically, the aggregates to be defined have reference to particular transactors (sectors), and that they have a time dimension. This suggests that the categories of "sector" and "time" will have to be introduced into the system ... Finally, we shall have to assume that certain types of events (transactions or transformations) are given, to which objects may be subject and which are of interest to us in the national accounts, e.g. sales..." "We shall postulate, therefore, that for each individual object the following characteristics are given: (i) Information as to whether the object is a real or financial object, (ii) Information as to the points of time at which the object is in existence, (iii) Information as to which transactors the objects are related to at any particular point of time during their existence, (iv) Information as to which transactions (events of the given types) the object is subject to during its existence; and, for every such transaction, further information as to the time interval when it takes place. These types of information constitute the distinguishing characteristics which will make a classification possible ...". With this information given, a number of entries in the national accounts may be conceived of as classes ("baskets") of real and/or financial objects, either as sub-classes of objects existing at a certain point of time (stocks), or as sub- classes of objects subject to certain types of events during a certain period of time (flows). There are, however, important entries in the national accounts (value added, net worth and other "balancing

38

O. Aukrust/The Scandinavian Contribution

items") which cannot be conceived of as classes35, but which exist only in value terms: "Next a set of evaluation coefficients (prices) is postulated for all objects ... With such a set of evaluation coefficients given, the way is open for establishing, in the form of a scalar number, what we may call the value of a class; furthermore, it can be shown that those national accounting entries that can not be defined as classes ... such as value added or saving, can instead be defined as value concepts ... Finally, if we further postulate that "exchange of objects always occurs according to the prices that are postulated", it can be shown that simple relationships (...eco-circ relationships ...) will exist between the value concepts established...". Aukrust started work in the belief that realizing these ideas would be a simple job. The task turned out to be not all that simple. By the time he was through he found that it had taken no less than twenty postulates to construct what was no more than the rudimentary framework of a viable national accounting system. Three postulates served to introduce the basic categories of sectors, economic objects, and time (points of time, and intervals). A subset of six postulates described the real circulation by introducing the idea of "ownership" of real objects, and by defining, and dating, the types of events ("transactions") which real objects could be subject to (production, consumption, change of ownership). A further sub-set of six postulates described the financial circulation by introducing the ideas of "creditors" and "debtors" of financial objects, and by defining, and dating, the types of events ("transactions") to which financial objects could be subject (creation, cancellation, change of creditor or debtor). Three postulates described the interplay between the real and the financial circulation and introduced the distinction between requited and unrequited transactions. Two postulates related to evaluation and measurement. They stated, in essence, (i) for every object a number ("price") is given, which expresses the value of that object, (ii) two requited transactions (flows) have always the same value (the postulate of "the preservation of values in exchange")36. The most important new insight gained from the exercise may have been the realization that most entries in the national accounts can be conceived of as representing stocks and flows ("baskets") of real and/or financial objects, but also that there are others (the balancing items) which cannot be thus defined and which exist only in value terms. This distinction is relevant i.a. for the problem of deflation in the national accounts. Another recognition brought out clearly is that numbers representing entries in the national accounts are functions equally of quantities and prices, and the same goes for relationships derived froms such numbers. Once this is realized it is no longer surprising that the growth rate of GDP computed at constant prices tend to change whenever the base year is changed, or that an export surplus at current prices may turn into an import surplus when measured at the prices of some other year. Finally, certain terminological requirements are suggested: In formulating verbal definitions the terminology chosen should clearly reflect the quantity/price dimensions of the aggregates to be defined37.

O. Aukrust/The Scandinavian Contribution

39

4.3. The purposes o f national accounting Ingvar Ohlsson’s most noteworthy contribution to the theory of national accounting, no doubt, was his analysis of the extent to which, and how, national accounting practice should be influenced by the purposes which the resulting statistics are intended to serve. Admittedly, Ohlsson was not the first to see that different purposes might call for different statistics. Only a few years earlier the famous market price-factor cost discussion between J.R. Hicks, Simon Kuznets and I.M.D. Little, in Economica, had taken place. But Ohlsson was the first one who studied systematically how different purposes might call for different answers to problems selected from the entire field of national acounting theory. In general, Ohlsson found that different purposes called for different solutions to a number of problems encountered in national accounting work. The most obvious conflicts were between accounting principles suitable for the measurement of results on the one hand, and those suitable for analysis of economic behaviour on the other. When measurement of results was the purpose, there was a further conflict between the production/productivity and welfare aspects. Analysis of economic structure called for much the same type of data as does analysis of production. National budget work used elements of all the other purposes. Here is a selection of Ohlssons conclusions: i. Goods and services should be valued at factor cost for the purpose of analyzing output and productivity, but at market prices when the purpose is measurement of welfare. If national accounts data are to be used for analyzing economic behaviour, the valuation should reflect the valuations set by the economic units themselves. ii. The sector division should be strictly institutional when the national accounts are to be used for analyzing economic behaviour; for measurement of results a functional sector division is preferable. iii. For analysis of results the boundary of production should be drawn so as to include non-market production whenever feasible; for analysis of economic behaviour it is better, as a rule, to select only monetary transactions. iv. Expenditures on maintenance and repairs, since they add to capital values, should be included in gross capital formation for the purpose of measuring results, but should be treated as current input, in accordance with business practice, for the analysis of behaviour. v. While depreciation charges in the enterprise sector should be calculated at current market prices for measurement of results, they should be measured in accordance with actual business practice (which may be different) for the purpose of economic behaviour analysis. Faced with such findings Ohlsson concluded, in a summary chapter ([50], pp. 320-321), that the national accountant has to choose between three alternatives for the form of presentation of official national accounting statistics: i. the construction of a general purpose NA-system from which to extract the special purpose systems ii. the construction of different NA-system for different purposes, and

40

O. Aukrust/The Scandinavian Contribution

iii. the construction of one special purpose NA-system with a list of corrections for the main items for which different treatment for different purposes is required It was "at least conceivable", Ohlsson admitted, that a general purpose system (alternative 1) could be constructed that might be used "for all reasonable purposes". However, this would be "an extremely complex matter". For his empirical work Ohlsson prefered different systems for different purposes (alternative 2), as has been described earlier. However, Ohlsson’s experiment in that direction was nowhere followed up and remained an episode even in Sweden. 4.4 The nature o f government and financial institutions Another original contribution by Ohlsson, also arising from his discussion of purposes, was his analysis of the nature of general government and financial institutions and their role in economic life. According to Ohlsson, general government engages in four different functions: production, consumption, income redistribution, and investment. The services produced are of three types depending on whether they "promote" (Ohlsson’s term) (i) production directly, such as maintaining ports and roads for the benefit of identifiable producers; or (ii) consumption directly, such as supplying medical attention and educational facilities for the benefit of identifiable consumers; or (iii) consumption collectively, such as maintaining a national defence or a foreign service ([50], pp. 21-23). The treatment of these services in the national accounts, Ohlsson argues, ought to depend upon the purposes for which the accounts are to be used. His discussion of how they should be valued is fairly standard and offers few new insights. His discussion of how they should be allocated to users, however, is original. Two alternative treatments are suggested, each suited to a particular purpose: In a national accounting system intended for business cycle analysis and national budget work it is convenient to look upon government simply as itself being the consumer of the services produced ("govern­ ment consumption"). In a system intended for analysis of production and welfare a more complicated treatment is preferable: Services benefiting producers (type i) should be considered "intermediate products" and charged as (non-marketed) input to the enterprise sector and thus excluded from the national product. Services benefiting identifiable consumers (type ii), though paid for by government, should be imputed to households and considered part of private consumption. Finally, services of type (iii) might be treated as consumed collectively and considered a special category of final expenditure ("collective consumption"). Part of this discussion sounds very modern! However, the ideas were only partly followed through in empirical work. Ohlsson observes that banks and insurance companies are like general government in that they produce services the costs of which are recovered only in part through sales on a market. This similarity makes him think that the services rendered by banks and insurance should be treated in the national accounts in exactly the same way as services rendered by general government: "In my opinion banking, insurance, and pure government are on an equal footing as regards the assignment of their services. It seems then rather peculiar to me that

O. Aukrust/The Scandinavian Contribution

41

nowadays in official estimates government services are assigned to the government sector, but banking and insurance services are assigned to the benefiting sectors. They should all be treated in a similar manner" (loc. cit. p. 148). Thus, in a system intended for business cycle analysis, banking and insurance services should be considered as being consumed by the bank and insurance companies them­ selves, as "enterprise consumption" (a kind of collective consumption in analogy with government consumption)38. For the purpose of analyzing production and welfare bank and insurance services should be imputed in part to enterprises and in part to households as recommended by Stone in [60]. (As in the case of government services the recommendations were only partly followed through in Ohlsson’s empirical work). 4.5. Accounting design. Quadruple accounting In the years immediately following World War II, national accounting pioneers invested considerable effort in searching for the design of an "ideal" national accounting system. The design should be analytically useful, and it was sometimes believed that the logic of the design might help in deciding how national accounting aggregates should be defined. In Sweden, Ohlsson [45] designed his two alternative accounting systems with an explicit view to the two different purposes which they were to serve. The first set of accounts was intended for measurement of results; consequently, the design was chosen so as to allow various national aggregates to be easily calculated. The second set of accounts was intended for business cycle analysis, and the design in this case aimed at giving a picture of purely monetary transactions taking place between economic subjects. The same set of four institutionally defined sectors was used in both alternatives, but the accounting structure within sectors was different. In Denmark, the first estimates by Kampmann for the 1930’s were presented within an input-output like framework. Later, Bjerke and Milh0j [20] adopted for their work the accounting structure recommended by the international organizations, with some minor changes. In Norway, sets of accounts were used by Bjerve partly as a pedagogical device, partly as a means for controlling the internal consistency of historical data [56] or of forward-looking data in the national budget [12]. Aukrust [3] experimented with accounting designs based on a three-fold classification of flows into real-financial, requited - unrequited (transfers), and current-capital. The logic of the system was formal: Real flows were shown on real accounts (current or capital), requited financial flows on current or capital financial accounts, while transfers were recorded on income accounts. Aggregates not representing flows (value added, saving etc.) showed up as ba­ lancing items. This meant "double book-keeping" in respect of each sector, e.g. a financial transfer received would be recorded as an inflow to a financial account and as an income element on income account. In a system with two or more sectors "quadruple accounting" would result since inter-sectoral flows would give rise to two entries in the accounts of each of the pair of sectors affected. This principle o f quadruple

42

O. Aukrust/The Scandinavian Contribution

accounting has been adopted and made explicit in the SNA Mark III (ch. 2, para 60). Whereas Ohlsson in Sweden used accounts as a convenient framework for presenting statistics this was not the practice followed by his collegues in Denmark and Norway. In these countries the accounts were considered merely a convenient means of making the system of definitional equations known to the readers. For this purpose a highly aggregated set of account for the national economy as a whole was deemed sufficient, with some details added in order to depict the interrelationship between general government and the rest of the economy. Detailed systems of "accounts" or "work-sheets" were used exclusively as office tools during the estimation process. For the purpose of presenting the result to the public the format of traditional tables was prefered. 4.6. Valuation: The idea o f factor cost In the standard literature the two consepts of "national product at market prices” and "national product at factor cost" used to be thought of as two measures of the same total, only valued differently. Ohlsson in his discussion of valuation went along in the same direction, as did Bjerke and Milhoj in Denmark. Lindahl in Sweden and Frisch and his followers in Norway took a different position. They had, of course, no problem with the idea that goods and services, in principle, may be valued at alternative sets of prices. They argued, however, that valuation "at factor cost" was neither practical nor very useful in an economy where taxes and subsidies are shifted back and forth within a complicated network of sales and purchases. (The compromise idea of valuation at "basic prices" might have been more acceptable.) They observed, further, that in actual practice the alternative evaluation was never undertaken: "National product at factor cost" was generally nothing more than total factor income which was found by deducting (net) indirect taxes from the value of product measured at market prices. The same applies at the level of individual industries. Lindahl [42] expressed the thought very clearly. "Factor income is that part of value added which is left for distribution to factors of production when government has taken its share in the form of indirect taxes" (my translation from the Swedish original, p. 103). Thus, Lindahl considered "product at factor cost" not the same total as "product at market prices", valued differently, but as a part of the latter and as such, in a sense, itself a market price concept. The same standpoint had earlier been taken by Frisch and others in Norway. In the national accounts of Norway, Aukrust made "factor income" a key concept both at the industry level and for the economy as a whole, while aggregates "at factor cost" were never refered to. 4.7. Definition o f product and income aggregates. The concept o f residence The Scandinavian countries have agreed to choose the concept of (gross or net) "nationalprodukt" as their main national accounting aggregate. It is defined as the value

O. Aukrust/The Scandinavian Contribution

43

added by resident factors of production. It is nearly, (but not quite) identical with what is called "domestic product at market prices" in the SNA. The difference has to do with the different content given to the idea of "residence" in the two systems. In the SNA residence relates to economic units, and "domestic product" is a measure of values created by resident producers. In Scandinavian thinking, residence is a characteristic of factors of production, and "nationalprodukt" is a measure of values created by resident factors39. The difference may be quantitatively unimportant but it becomes conceptually significant when a factor of production (e.g. a border worker) resident in country A works in a producing unit resident in country B. The values created (added) by this fac­ tor would be counted as part of the "nationalprodukt" of A according to Scandinavian practice, while according to SNA rules it would be part of the "domestic product" of B. As a corollary, a payment to A from B in respect of the services rendered by that factor would be treated as a payment for export of services in Scandinavia but as factor payment in SNA. Whereas the Scandinavian countries agreed at an early stage to adopt (gross or net) "nationalprodukt" as their main measure of product (and of "income generated" or "originating") they went different ways in their choice of additional national accounting totals. In Denmark Bjerke and Mi 1h 0j (1955) chose to include the large number of measures of product and income, at market price and at factor cost, so characteristic of the UN system (section 4.4). In Norway Aukrust (1952), believeing that confusion was likely to result from presenting the reader with too many nearly identical concepts, went to the opposite extreme. In addition to national product, gross or net, which could be interpreted not only as a measure of output but also as a measure of "income origi­ nating", he restricted his choice of income totals to two: (1) "Factor income" (wages + operating surplus), a measure of "primary income" left for distribution to factors of pro­ duction after government had taken its share in the form of indirect taxes. At the national level "total factor income" could be shown by industrial origin or, alternatively, by receiving institutional sector. (2) Disposable income, a measure of "redistributed income" available for consumption and/or saving, defined at the national level as the sum of "nationalprodukt" (income originating) and net income received from abroad in the form of interest, dividends and other transfers. "Disposable national income" could be broken down by institutional sectors. The concept of "national income" as defined in SNA was rejected as being ananalytically not very useful: "A concept which includes net interest and dividends from abroad, but excludes net mutual aid, etc., received, is too narrow as a measure of welfare (where Disposable National Income is the better concept), and at the same time too broad as an indicator of productivity (where Geographical Product "nationalprodukt" is more illuminating" ([3], p. 181). The main concepts introduced by Ohlsson [50] for Sweden were in many respects similar to those chosen by Aukrust for Norway. In particular, the concept of "disposable national income" is common to both systems except that the terminology is different. To summarize: In comparison to the SNA it appears that the Scandinavian countries tended to replace the concept of "national income" with the concept of "total factor

44

O. Aukrust/The Scandinavian Contribution

income". Two of them (Norway and Sweden) also made use of the concept "disposable national income", a concept which later came to be accepted for the SNA with the 1968 revision of the system. 4.8. Commodity flows, input-output, and double deflation From a methodological point of view the most noteworthy contribution by Scandinavian national accountants has probably been their development of the commodity flow method. It started when Lindahl and his collegues applied a rudimentary form of commodity flow analysis in their national income estimates for Sweden in the early 1930s. The credit for having demonstrated the full power of the commodity flow app­ roach to national accounting must, however, go to Kampmann in Denmark who in a project initiated around 1935 applied the method with great consistency i.a. publishing annual input-output tables (section 3.3). Kampmann’s work in Denmark became known to Norwegian statisticians in 1946 when Norway embarked on a project aiming at a revision of the existing national accounts. It seemed natural for many reasons to make the commodity flow method the cornerstone of Norwegian work also. The estimates published in [57] were based on an analysis of probably near 1000 different groups of commodities, including 350 categories of consumers’ goods. The commodity flows were shown in condensed form as a sector-sector input-output table distinguishing some 35 industries. The commodity flow method has continued to serve as the cornerstone of national accounting work in all later years in both Denmark and Norway. More details have of course been added as improved statistics have become available. A major overhaul of the Danish system was initiated by Bent Thage in 1968 patterned on the make and use matrix framework of the SNA Mark II. According to Thage, the Danish system at present (1990) contains about 2 500 groups of commodities, 130 industries, and about 75 categories of final demand. In Norway the same numbers are 1 750, and 180 respectively. With input-output tables available both at constant and current prices, "double­ deflation" of the estimates of product (value added) at the industry and national level followed more or less automatically. Norway has used double deflation since 1952. In Denmark the method seems to have been used for the first time in [20] for the purpose of revaluing the 1947 input-output table at 1949 prices. Given the availability of input-output tables on a regular basis Scandinavian economists, quite naturally, were among the first to explore their potential for economic analysis. Studies based on Norwegian and Danish data were presented at international conferences on many occations in the early 1950s (see e.g. [4], [10]). In Norway, chief responsibility for input-output work was in the hands of Per Sevaldson who, with annual data at his disposal, did pioneering research on the stability of input-output coefficients ([53], [54]), and on the incorporation of input-output structures in macro-economic models. Aukrust as a consultant to the UN Economic Commission for Europe assisted in exploring the use of input-output methodology in economic analysis in general [21],

O. Aukrust/The Scandinavian Contribution

45

4.9. Flows o f financial objects While input-output accounting very early became a regular part of national accounting in Scandinavia, accounting for flows of financial objects took longer to develop. The need for such data was felt first in Norway in connection with national budget work. Early attempts made it clear that financial statistics would have to be improved for financial accounting to be possible. Consequently, during the 1950s work proceeded in parallell on theoretical and practical issues in Norway and to some extent in Sweden and Denmark as well. A paper by Bjerve and Selsjord [13], describing the principles guiding work in Norway, was presented to the 1959 IARIW conference, in many ways anticipating ideas which some years later became part of SNA Mark II. Numerical work was a different matter, however, and an integration of accounts for financial objects (stocks and flows) with the already existing national accounts could only be achieved gradually over a number of years. 4.10. Reliability indicators The national accountant faces the problem of conveying to his readers a picture of the reliability of his estimates. The best way of doing this is through a detailed description of estimation methods. In addition some authors (e.g. Geary in Ireland) experimented with attaching to individual series a "reliability figure" expressing the subjective evaluation of the quality of that particular serie by the person responsible for the estimate40. Such reliability figures, on a scale from 1 to 5, were used for the first time in Scandinavia in Sweden in 1949. Noting that some series may be reliable with respect to level (because they are based on reliable bench-mark data) but doubtful with respect to year-to year movements, while the opposite may hold true for others, Ohlsson in Sweden and Aukrust in Norway tried to improve on the standard practice by using a combination of two reliability figures for each statistical serie: One in respect of level and one in respect of trend. Ohlsson in addition warned his readers by adding a "± x" to particular weak figures such as estimates of depreciation and saving. The use of reliability figures was never attempted in Denmark41. Since the use of reliability figures, perhaps regrettably, did not catch on internationally, Sweden and Norway also soon decided against continuing their practice.

5. Definition and classification by convention There are many controversial issues in the national accounting literature which has little to do with principles and which, therefore, require conventional solutions. It may be of interest for the record to note how some of the more well-known issues were treated in Scandinavia in the early days when international recommendations were yet to be agreed.

46

O. Aukrust/The Scandinavian Contribution

5.1. The boundary o f production Early Scandinavian studies adopted a broad concept of production. Lindahl et al [39] set the standard by including "... everything possessing an exchange value...'1. Denmark and Norway followed, Aukrust [57], for instance, even going so far as to include in national product the value of wild berries picked for own consumption. The increase in the value of standing forests was included in some studies [45], [57] but excluded in others (Denmark, earlier studies for Sweden and Norway). The exception to defining production broadly was Ohlsson [45] who in one of his two alternative systems (the one intended primarily for business cycle analysis) included only "realized monetary transac­ tions". Considerable attention was paid to the treatment of consumer durables. All Scandinavian studies included the services of dwellings in the national product. Lindahl et. al [39] included also the services rendered by all kinds of other consumer goods, ranging all the way from household china and glass to automobiles; the value of the services was calculated as the annual amortization of the goods plus a yield of 5 per cent on their remaining value. Ohlsson [50] followed Lindahl part of the way by including, in principle, the services of automobiles (in practice putting the capital costs of the services equal to the value of acquisitions). Kampmann [34] agreed with Lindahl in theory, but decided in practice to take the output of consumer durables in a given year as a measure of these services. CBS [56] and Aukrust [57] in Norway and Bjerke and Milh0j [20] in Denmark considered consumer durables, apart from dwellings, as consumed when bought. Practice with respect to the treatment of unpaid domestic work varied amongst countries and over time. In Sweden, Lindahl et.al. [39] published two alternative measures of the national product, one excluding and one including the value of non-paid domestic work. Preference was given to the former. In Ohlsson’s system for business cycle analysis from 1951 [45] only monetary transactions were considered, thus the issue did not arise. In his alternative system intended for analysis of results, the value of the services of housewives etc was "unfortunately" also excluded due to "problems of valuation". In Denmark both Kampmann in 1942 and later Bjerke and Milh0j in 1955 followed Lindahl in excluding unpaid domestic work in their principal system yet offering, as an option, an estimate of the values of these services42. In Norway, CBS in 1946 chose to include the value of unpaid domestic work in the national product. Aukrust in 1952 excluded it. 5.2. General government All Scandinavian studies made a distinction between government enterprises, which were included in the enterprise sector, and general government which was treated always as a producer, and sometimes also as a consumer, of government services. The value of the output of government services was estimated at costs in all studies. Costs were always taken to include maintenance and amortization costs of government

O. Aukrust/The Scandinavian Contribution

47

buildings and plants. In Denmark, but not in Norway and Sweden, an imputed interest on government buildings was also included, as required by OEEC [49]. The different practice in the three countries presumably reflected the fact that the fiscal accounts of Denmark included such an entry while the fiscal accounts of Norway and Sweden did not. When it came to allocating government services to users, different solutions were chosen. In the earlier studies, before the idea of considering general government a (collective) consumer had been generally accepted, the issue was the distribution of these services between private consumption and intermediate business input and, if the latter, between individual industries. Lindahl et. al. [39], by convention, routed one half to consumers and the other half as (unallocated) current input to business, to be deducted as a correction item from the sum of value added by individual industries when estimating national product. CBS in Norway in 1946 [56] tried to judge the allo­ cation of government services between consumption and current input to business on the basis of the types of services rendered, following Lindahl in treating the business part as a correction item rather then charging it to individual industries. Kampmann [34] in Denmark followed an argument by Kuznets of treating 15 per cent of government services as current input to business and the rest as components of consumption; the part treated as current input was allotted to individual industries in proportion to the value of their output. With the arrival of Stone’s study from 1947 for the League of Nations it became usual to consider the output of government services to be consumed by government itself as "government consumption". This treatment was adopted by Aukrust [57] for Norway and by Bjerke and Milh 0j [20] in Denmark. In Sweden Ohlsson argued, as explained in 5.4, that, in theory, a more differentiated treatment was called for [50]. In practical work, however, he agreed to treat the total of govenment services as consumed by government itself, adding a sub-classification of government consumption into "intermediate" and "final" consumption in the variant of his system intended for the analysis of results. Data for the general government accounts could only be derived from a detailed analysis of the fiscal accounts of central and local government, aiming at regrouping the various entries of receipts and outlays into categories as defined in the national accounts. This was achieved in different ways, and typically required the co-operation of the competent authorities. In Norway, for instance, a major budgetary reform was carried through in the middle of the 1950s, one purpose of which was to make the format and specifications of the fiscal budget conform to the needs of the national accountant; similarly, directives laid down by the central authorities for the budgeting practice of municipalities were revised with the same purpose in mind. 5.3. Banks and insurance companies The output of banks and insurance companies was valued at costs in all Scandinavian studies, the exact method by which this was done varying somewhat in practice.

48

O. Aukrust/The Scandinavian Contribution

Different methods were used for allotting the services of banks and insurance companies to users. As for government services, the issue was the distribution of these services between consumption and intermediate business input and, if the latter, between individual industries. Lindahl et al [39] charged, conventionally, all banking services, unallocated, to business, and all insurance services, unallocated, to consumption, well aware that a finer division might have been preferable. In Norway, CBS in 1946 followed Lindahl in charging all banking services to business, but differed by charging insurance services partly to business and partly to consumption. Kampmann [34] made an attempt to distribute the services both of banks and of insurance companies between business and consumption. As with government services the part charged to business was allotted to individual industries in proportion to the value of their output. Later authors were faced with the recommendations in Stone’s League of Nation paper [60], but the recommendations given there had little impact. Ohlsson [45], true to his view that banking and insurance services were similar to government services and should be treated analogously, treated all the services rendered by banks and insurance companies as consumed by these institutions themselves as "enterprise consumption", i.e. as a separate category of "final output". Aukrust [57] and Bjerke and Mi 1h 0j [20] took a pragmatic view: They disliked the idea of having to show "imputed bank charges" as part of consumers’ expenditure and as a corresponding imputed element of household income, and they objected even more to having to chase imputed bank charges through their input-output tables, thus distorting the estimates of value added of all industries. Rather than introducing fixtitious flows, difficult to explain, at numerous points in the system, they prefered to concentrate all difficulties in one single entry by treating the value of the services of banks and insurance as a correction item to be deducted from the sum of value added by individual industries when estimating national product. (This was the solution adopted later in SNA Mark II.) This correction item equalled Ohlsson’s "enterprise consumption" in magnitude and, like the latter, it served to balance the accounts. 5.4. The concept o f "gross-gross" capital formation A standard problem in national accounting relates to the borderline to be drawn between current input and capital expenditure. The issue primarily concerns the treatment of expenditures on daily upkeep, maintenance, and minor and major repairs. The Scandinavian countries agreed shortly after World War II to adopt what became known internationally as the "gross-gross" concept of investment. (If I recall correctly the term was suggested, humourosly, by Geary). According to Scandinavian practice, gross investment included, in addition to new equipment and new construction, all kinds of repair and maintenance to equipment and structures. Only "daily upkeep" was treated as current input. The arguments leading to the choise of the wide "gross-gross" concept had to do partly with statistical convenience and partly with analytical usefulness. The following quotation from Bjerke/Milh 0j [20] may be taken as representative of the

O. Aukrust/The Scandinavian Contribution

49

Scandinavian view (my translation): "The reason why it has been decided not to follow international practice and the proposals of OEEC and UN, according to which only new investments, major alterations etc. are considered as investment, has to do with the difficulties involved in drawing a borderline between those (major) repairs which should be included and those (minor) which one would like to exclude. While any borderline will have to be drawn arbitrarily it is considered easier to distinguish between repairs on the one hand and ordinary daily upkeep etc. on the other. To this may be added that investment in the narrow sense and repair works overwhelmingly call for the same sort of materials etc. which suggests that they should both be included as investment. Finally, scarce supply of goods, e.g. in time of war, may lead to much increase in repairs activity, which in such a situation in a sense replaces new investments" (p. 15-16). The argument was not accepted when the issue was up for discussion at OEEC in 1951 in connection with the proposed "OEEC Simplified System": The gross-gross concept lost narrowly when a vote was taken amongst the experts present. 5.5. The classification o f transfers Scandinavian authors have tended to use the word "transfer" as a synonym for "unrequited", i.e. "without a quid pro quo". The problem of classifying transfers has been approached by asking: Why are they paid? Ohlsson and Aukrust, in particular, have argued that a classification based on the motives of the payer and/or the reactions of the receiver (e.g. "current" versus "capital") is unlikely to result in well-defined categories; therefore more formal criteria are needed. Ohlsson [45] proposed a four-fold classification of transfers based on their economic and legal background (essentially on extension of a classification scheme borrowed from Ralph [52]): (i) voluntary transfers, e.g. gifts, grants, non-obligatory fees, dividends etc., entrepreneurial withdrawals, (ii) contractual transfers, e.g. interest, wages, rents, insurance claims, pensions, (iii) compulsory transfers, e.g. taxes, fines, war damage payments between nations, (iv) "conditional" (or "indirect") transfers e.g. indirect taxes and subsidies ".... closely connected or conditioned by market transactions ..." Aukrust [5] reasoned along similar lines: "In traditional literature one criterion used for grouping transactions has been the supposed motives or repercussions that gave rise to, or resulted from such transactions. Pivotal national accounting items such as national income, private saving, exports, etc, are so defined that their exact content is dependent upon the distinction drawn between "direct" and "indirect" taxes, between "current transfers" and "capital transfers", etc. In other words, the definitions presuppose that distinctions are drawn on points for which the possibility of drawing precise distinctions may not exist.... In order to achieve the sharpest possible definition of the national aggregates, we have decided to use a system of classification which allows us to define the national aggregates solely by formal criteria .... Classifications based upon the "repercussions" of or "motives" behind the transactions appear only at a later stage in the classification system" (p. 109).

50

O. Aukrust/The Scandinavian Contribution

It is noteworthy that in Ohlsson’s classification scheme "factor payments" do not appear as a separate category. It is also noteworthy that wages are considered transfers: "If the production result is the center of interest, wages (and salaries) ought to be treated as transfers on a contractual basis on the income redistribution accounts.... The treatment of wages involve a question of assigning production. The value added in a sector generally means the value added by all factors of production which are organized in the sector. ... In that case the wages ought to be treated as transfers, in order to obtain certain additive characteristics of the accounting design ..." ([50], p. 209, text and footnote). This becomes especially desirable when, as in Norway, the system has production broken down by industries. Aukrust thought of employees, in one capacity, as factors of production to be classified by industry, transferring to themselves in their capacity as consumers, as wages, their share in the value added of that industry, in much the same way as entrepreneurs, as consumers, make entrepreneurial withdrawals. 5.6. Capital transfers The principal view of Ohlsson and Aukrust on the classification of transfers was reflected in their attitude to capital transfers. Ohlsson admitted that a distinction between current and capital transfers had a natural place in a national accounting system intended for business cycle analysis; in such a system it might even be correct "that a transaction should be considered as a current transfer in one sector and a capital transfer in the other" [50], p. 152). In national accounting systems intended for analysis of results ("R- statements") however, the distinction between current and capital transfers could not be maintained: ".... All transfers must be treated as current, i.e. must become items in the calculation of final income. The transfer .... has altered the economic position of the recipient sector as much as savings out of other income have. The transfer in question should consequently be taken account of in the income calculation for R-statements" ([50], pp. 152-153). In the system of Aukrust, which was primarily intended for analysis of results, the category of capital transfers did not exist. He was aware, of course, that transfers of a capital nature would have to be treated differently from other transfers in the analysis of many problems. He prefered, however, to provide the necessary data in the shape of suitable sub-classifications of transfers in general. 5.7. The classification o f taxes With respect to taxation Aukrust [57] decided to discard the traditional two-fold classification of taxes into "indirect" (assumed to be passed on) and "direct" (assumed to be carried as levied). Instead, he looked for a classification based on objective criteria which in this case could only be a classification by the "object" of taxation. The result was as follows (p. 192):

O. Aukrust/The Scandinavian Contribution

51

1. Taxes on certain plant and equipment, production and trade ("indirect taxes"). a. Customs duties, b. General sales tax. c. Taxes on certain plants and equipment (mainly automobile duties paid by business) d. Special excise taxes, e. Conscession duties and taxes on the establishment of new businesses, f. Export duties, g. Stamp duties (paid by business). 2. Taxes on income and property, social security premiums etc. a. Taxes on income and property, b. Taxes on consumer goods, c. Income from lotteries etc. d. Inheritance tax. e. Fines and confiscations, f. Special tax on property increase during the war. g. Extraordinary taxes levied in the early post-war years, h. Social security premiums. A classification remarkably similar to the above from 1952 has now been proposed for use in SNA mark III. (See "Accounts and Classifications"). 5.8. The treatment o f fees and charges A controversial issue in the early days of national accounting, inherited from the theory of public finance, was the treatment of fees and similar payments from the private sector to general government. When paid by business such fees and charges were treated by OEEC [49] and UN [62] always as indirect taxes. When paid by persons and organizations, they were treated in one of three different ways (presumably following Anglo-American practice): either as payments for goods and services (catalogues and postcards from museums), or as indirect taxation and considered part of consumers’ expenditure (real estate taxes, motor vehicle duties, dog licences), or as "transfers" to be treated like direct taxes (passport fees, school and hospital fees). The Scandinavian countries followed the international recommendations in treating fees and charges as indirect taxes when paid by business. Practice varied with respect to fees paid by consumers. Sweden (Ohlsson), as far as can be seen, followed the international recommendations on all points, including treating real estate taxes, motor vehicle duties and dog licencies as indirect taxes and part of consumers’ expenditure. Denmark (Bjerke and Milh0j) deviated by treating real estate taxes not as an indirect tax, but as a direct tax to be paid out of personal income. In other respects the international recommendations were followed. Norway (Aukrust) took the position that, logically, consumers never pay indirect taxes. Fees and charges paid by consumers, therefore, had to be considered either as payments for goods and services, or treated as direct taxes paid out of income.(This position, incidentally, has been adopted for the SNA, Mark III). The former category was taken to include passport fees etc. and also, again anticipating SNA mark III, school and hospital fees. The latter category included motor vehicle duties, real estate taxes, fines etc.

52

O. Aukrust/The Scandinavian Contribution

6. Epilogue: towards SNA, and beyond The "simplified" and "standardized" systems of national accounts of the early 1950’s [48] [49] [62] were largely the work of English and American economists. Understandably, they reflected mainly Anglo-American thinking at the time; there are no traces, for instance, of Scandinavian ideas43. On the other hand the international standards came too late to influence to any significant degree the developments in Scandinavia. Anticipating that the international recommendations some day would have to be revised, Sweden, Norway and Denmark decided to retain their recently adopted national accounting systems conceptually unchanged for the time being. However, a number of extensions were introduced successively during the 1950s and 1960s by various authors in answer to specific analytical needs. For example: Unbroken historical series covering 100 years or more were established in all three countries in order to allow studies of economic growth44. International comparisons where facilitated when Denmark and Norway were found to be among eight countries with sufficient data allowing inclusion in the Gilbert-Kravis international comparison of national product and price levels. [30], Quarterly national accounts data were estimated in Norway for the first time in 1953 and in Sweden on an experimental basis from around 1960. In Norway estimates of the stock of real capital, allowing studies of economic growth, became available in 1958 [6]; regional national accounts were added in 1965, and estimates of holdings of financial assets and liabilities were developed gradually over a long span of years. In Sweden, where the national accounts originally designed by Ohlsson did not contain any industry detail, major efforts were made from 1960 onwards to fill the lacuna. A university initiative [32] resulted in an input output table for one year with 127 commodity groups/industries. A major extension of the official Swedish national accounts aiming at providing data needed for analysing industrial structure was initiated in 1963/64 when responsibility was taken over by the CBS from Konjunkturinstitutet: estimates of GNP by 31 industry groups in current and constant prices, and with wages and gross operating surplus shown separately, were published in 1966; employment data followd in terms of hours worked (1968) and persons employed (1970); data on real capital stock were added in 1974. (Employment data had been available in Norway since 1952.) To accomodate the data needs of "the Scandinavian theory of inflation" (known in Sweden as "the EFO report") a classification of industries as "exposed" or "sheltered" (relative to competion from abroad) was introduced as an optional classification in the national accounts of Norway in 1966 and in those of Sweden shortly after. Meanwhile, at the international level, the feeling was growing that the United Nations standardized system from 1953, with amendments in 1954, would have to be revised. Scandinavian economists were amongst those who most eagerly requested revisions; they hoped, naturally, that an overhaul of the system might bring it more in line with Scandinavian thinking. The final result (SNA Mark II) came after year-long discussions in which Scandinavian economists took active part (Aukrust, for example, was a member of the UN committee of experts supervising the work), and not without a certain amount of success: while the revised system obviously combined ideas drawn

O. Aukrust/The Scandinavian Contribution

53

from a number of origins, making it hard to decide who contributed what, it is a fact that the outstanding features of SNA Mark II - input-output tables, dual sectoring, production accounts for industries, transactions in financial assets and liabilities - had existed for many years as characteristics of the national accounting systems in use in Scandinavia. Influences from Scandinavia might be traced also on other points, for example, in the treatment of banks, in increased consideration being given to realsphere phenomena implying some shift of interest away from "national" to "domestic" totals, in the introduction of the concept "operating surplus" (a concept known in Norway since 1952 under the name of "owners income") allowing interests, dividends etc. to be treated as special categories of transfers, and in the introduction of the concept "disposable national income". As a whole, the new system [63] was greeted in Scandinavia as a blend of ideas to which the Scandinavian countries had contributed significantly and as a compromise solution with which they could easily live. A switch to the new system started immediately. It was completed in Sweden in 1970, in Norway in 1973, and in Denmark in 1978, with complete numerical revisions carried out simultanously as needed. Although these revisions, and the developments inside and outside Scandinavia since 1970 are not part of the present story, it is gratifying to note that the next version of the SNA now about to be published will include additional features consistent with Scandinavian thinking, e.g. quadruple accounting (chapter II), a revised classification of taxes ("Accounts and Classifications"), and an improved treatment of factor incomes (to be renamed "primary income") and property incomes (chapter 7) - even coming close (chapter I) to endorsing the old Scandinavian view that transactions should be defined in terms of events happening to economic objects.

Endnotes 1. The first work of this kind in a Scandinavian country was A.N. Kiser’s estimate for Norway for the year 1891 [35]. It is possible that Kiasr’s interest in the field dates further back; Frisch [24] has a reference to 1874: "...the year when Kiser wrote his seminal treatise on the national income”. I have found no traces of such a study by Kiaer. According to information given by W ederwang [64], the earliest studies for Denm ark and Sweden are from 1917 and 1924 respectively. According to W ederwang both studies were rather rough and not to be taken very seriously. In Scandinavia’s neighbouring country to the East (Finland) an estimate of national income seems to have been carried out as early as in 1869 by K.E.F. Ignatius [47]. 2. Sir John Hicks reports that he m et Lindahl during his stay in London in 1934 and 1935. Hicks must have been impressed; in a retrospective article in Economica, 1973 (p. 8) Hicks in a passing sentence refers to Lindahl as “the father of social accounting theory". Had Hicks known Scandinavian economics better, he might well have included Frisch in his appraisal. 3. It has been speculated (Andvig, [1] p. 203) that at some stage Frisch’s ambition may have been to construct a system for economics comparable to Linne’s system for botanies. Bjerve [14] denies this, but thinks that Frisch may well have hoped that his concepts, terminology, and symbols could be accepted as an international standard. A pointer in this direction, according to Bjerve, is Frisch’s choice of mnemonic notation for his variables (Appendix A) and also the fact that Frisch in 1940 prepared, in English, a draft of 93 typewritten pages entitled "Suggestions for a coordinated system of notations in monetary theory”. (As so often with Frisch this manuscript was never sent off to the printers.)

54

O. Aukrust/The Scandinavian Contribution

4. In writing the above I have relied heavily on a review article by Petter Jakob Bjerve tracing the history of Frisch’s work in this field from 1928 to 1948. [14] During a great part of this period, from 1939 onwards, Bjerve was actively involved in the work, as Frisch’s trusted assistant, and later, as his close associate. Bjerve’s impressions of his co-operation with Frisch, then and later, are recalled vividly in [15]. According to Bjerve the first sketch of what was to become the Eco-circ System was given in a series of lectures on economic theory which Frisch gave in 1928-1929. Frisch returned to the them e in lectures on "macrodynamics" in 1933-1934 and also in the introduction to his famous article in the Cassel "Festschrift" [22] where the term s micro- and macro-economics may have been used for the first time in print. An important milestone was A general system o f concepts and symbols (mimeographed, 1935) prepared as part of lectures on "modern monetary theories" with the idea of facilitating the comparison of theories of different authors. Frisch’s work on conceptual issues took on an additional aspect when his institute, some years later, started experimental work on national accounting. During this period the conceptual system from 1935 was revised and presented not only algebraically, but also graphically and in accounting terms. The graphic presentation caused problems. Bjerve recollects [14]: "I took part in num erous discussions lasting for hours of what in the end became the Eco-circ graph. Unbelieveably many drafts ended in the waste-basket before Frisch at long last felt happy with the result". The result was the Eco-circ system published in a mimeographed text in October 1942 [25]; a printed, but abbreviated version of the text appeared in a Swedish journal the next year [26]. After a break Frisch, assisted by Bjerve and Aukrust, returned to the problem four years later. The result was [2], which may be considered a modified version of the original system from 1942. 5. The 1948 version of the Eco-circ system clearly reflected the data requirem ents of the recently initiated national budgeting work. (Bjerve, [15], p. 42). The system no longer related to sectors in general. Instead, it referred directly to the national economy as a whole, focusing, in particular, on the interactions between the public and the private sector. The influence from contem porary British White Papers is noticeable. 6. Bjerve [14] recalls the origin of the first financial graph (probably in 1942): "When the graphic representation of concepts and relationships relating to the real circulation was finished, Frisch suddenly got the idea that there ought to be a corresponding graph describing the financial circulation and he hurriedly drew such a graph. The idea was certainly not bad. Surely, the analytical usefulness of concepts such as e.g. "financial value added" (E) and "financial depreciation" (D) may be doubted ... but ... it helps avoiding the confusion of real and financial concepts ..." (p. 27). 7. A t the national level, where real income (in Frisch’s terminology) is the national product, this implies defining national income as "the national product plus net financial income from abroad". 8. Thus Stone is clearly wrong when he suggests [47] that the term "Social accounting" was used for the first time in 1942 by Professor Hicks. 9. For instance, in the experimental accounts for the sector "public and semipublic banks” (for which Bjerve was responsible) there were 127 individual accounts which could be aggregated hierarchically into 14 m ajor accounts. The two-fold classification of sectors into "organizational" and "industrial" corresponds to what is nowadays term ed "institutional" and "functional" sectors. 10. It may perhaps be claimed that Frisch’s creativity was paired, on occasion, with a certain lack of realism. Yet, Andvig [1] in his study of Frisch clearly exaggerates when he writes as follows: "Frisch’s major work on theoretical national accounting from the first half of the 1930’s ... was a curious mixture of useful abstractions such as the sector-concept, overambitious attem pts to classify everything in economics, and a surprising num ber of analysis of various types of non-operational concepts" (p. 237). In another connection, Andvig, with loving criticism, characterizes Frisch’s research style: ”... he was carried away by his usual enthusiasm for whatever he was thinking about and jum ped to strong conclusions about the real world significance of his thinking as a kind of afterthought..." (p. 102). And again: "His research style, here as elsewhere, was erratic - more like a firework spreading in all directions than a carefully managed, singlepurpose rocket" (p. 107).

O. Aukrust/The Scandinavian Contribution

55

11. In distinguishing between these two aspects of Lindahl’s work I follow Ingvar Ohlsson [50]. According to Ohlsson, Lindahl’s second line of work (his theoretical studies of concepts) ”... is of interest for the methodology of national accounting primarily because it contains the germ of many lines of developments in later works" (p. 266), obviously including the works of Ohlsson himself. 12. It is fairly clear that Lindahl [41] to some extent was inspired by Frisch [23]. (Andvig [1], p. 203). However, it is equally obvious that Lindahl’s ideas on conceptual issues evolved gradually over a long period of time dating back to the late 1920s. W hen finally written out and included as a separate section in [41] this was done with the following introduction: T h e "Algebraic Discussion” which concludes this part is an endeavour to give a general formulation of some basic economic concepts, in such a way as to make them directly applicable to real situations; and to give a systematic account of the relations between these concepts. Although the task may seem elementary, it is of fundamental importance, and my exposition shows that it presents certain difficulties ...” (Preface, p. 11-12). 13. It has not yet becom e usual in national accounting to distinguish between a) "real capital”, which is a factor of production consisting of concrete production facilities, together with certain rights attaching to their exploitation, such as patent rights, the name of the firm, goodwill etc., and b) "financial capital", which is a distribution factor and comprises claims (and debts) usually referring to monetary transactions. "The traditional view is that what is here called financial capital can be traced back to some underlying real capital and that financial income can be looked on as income derived from this real capital. Thus when the shareholder collects his dividends this is considered as income from the company’s real capital in which he participates through his shares. In this case it is possible to make the imputation, but in many other cases it is difficult to find the real capital that is supposed to correspond to the financial capital. For instance, it is not very acceptable to consider interest on the national debt as income from the real capital at the disposal of the government..." (p. 83). Lindahl continues by saying that he agrees "whole-heartedly" with the following statem ent by Ohlsson: "The insistence on treating realized interest as a type of factor remuneration is the reason why national income is often presented with interest as an income share. This procedure seems rather meaningless. Interest payments are contained in an income redistribution process based on the ownership of financial capital. It may be of interest to record the distribution of income before or after this redistribution and to register net interest payments for each sector" (p. 84, note 25). It seems that what Ohlsson and Lindahl are asking for is what the 1993 SNA intends to do, 40 years later, in the primary and secondary distributions of income accounts. 14. "This definition of the national income concept corresponds most closely to what is usually (internationally, e.g. in [49]) called "national income at market prices". However, it is more common to speak of "national income at factor cost", and m ore common still to understand the concept of national income implicitly in this sense. I have long been critical of this latter concept ... convinced ... that ... we ought to use some m ore succinct terms, such as "total factor incomes" or perhaps "total factor incomes plus total net financial income". This entirely avoids the troublesom e distinction between estim ates "at market prices" and "at factor cost” ... "I am also of the opinion that we ought to try if possible to avoid the distinction between "gross national income" and "net national income" and content ourselves with the latter. A gross calculation seems more natural for national product than for national income ...” (loc. cit. pp. 91-92). 15. The study was part of a m ajor project "Wages, cost of living and national income in Sweden 1860-1930" by the Institute for Social Sciences at the University of Stockholm. The project was initiated by Professor Gosta Bagge and financed by a long-term grant from the Rockefeller Foundation. Over the years the project came to engage a num ber of the best, young economists of Sweden. In addition to the study of national income, the project resulted in "The cost of living in Sweden 1830-1930" by Gunnar Myrdal, and in "Wages in Sweden 1860-1930", in two volumes, by G osta Bagge, Erik Lundberg, and Ingvar Svennilson (1933-1935). 16. In Sweden the historical series of Lindahl et al. were updated in several steps by Einar Dahlgren [16] [17], then working within one of the ministeries. In Denmark (around 1935) and in Norway (1943)

56

O. Aukrust/The Scandinavian Contribution

studies similar to the Swedish one, but covering much shorter spans of time, were initiated by the respective national statistical offices. 17. It is symptomatic that the bibliography included in [39] lists only one empirical study from a foreign country dating further back than 1926; this honour goes to "Income in the U nited States, its amount and distribution 1909-19" by the staff of the National Bureau of Economic R esearch, (Vol. 1-2,1921-22, New York). 18. It is of some historical interest to compare Lindahls list with a similar listing in Studenski [61] describing the situation in national income literature in the early 1930s. According to Studenski national income might be m easured (1) as an aggregate of net output of various branches ("national income produced", "Wertschopfung”), (2) as an aggregate of incomes earned by individual producers (“national income paid out", "Verteilung”), (3) as a sum of final goods and services ("national income consumed", "Verwendung”). (The English and Germ an terms given in brackets seem to have been used in an American study by Mitchell and Kuznets [46] and a G erm an study by the Statistisches Reichsamt [55] respectively.) - The three m ethods described here by Studenski are the same as m ethods l)-3) of Lindahl, only listed in a different order. In addition Lindahl adds one method (no. 4) not mentioned by Studenski, observing that incomes can be measured not only when paid out (Studenski no. 2, Lindahl no. 3) but also when received. 19. Note Lindahl’s choice of words in his description of method (3): W hat he means by "agents of production” is labour (including work by self-employed), land, and real (physical) capital. M ethod (3), therefore, in Lindahl’s interpretation am ounts to measuring income as a sum of "functional" income shares, in the tradition of classical economic theory. 20. General government is treated by Lindahl as producing services for the benefit partly of producers, partly of consumers; thus government is not itself a consumer of the services. "The point is here to determine how much of the current public expenditure should be included in the total production costs of agriculture, manufacturing and other industries” (p. 11). On the valuation of government services he reasons: If a public service is supplied against payment that payment m easures its value. In other cases the value of the services is measured by the costs of providing them, including imputed ("real”) interest on the (real) capital used while producing them. - Lindahl mentions the possibility of measuring the value of services rendered to private producers by the rates and taxes that the producers pay. This procedure (which he possibly may have seen discussed in the literature) is dismissed by Lindahl as unrealistic "at least in Sweden". 21. 1. Agriculture, 2. Forestry, 3. Manufacturing, mining and handicrafts, 4. Building and construction, 5. Transport and communications, 6. Commerce, hotels and restaurants, banking and insurance, 7. Professional services, 8. Domestic work, 9. Services rendered by durable consumers goods, 10. Public administration. - Note activity 9: Not only dwellings, but all kinds of consumers durables were treated as capital, i.e. treated as invested when bought and producing services when used. 22. Outside Scandinavia the commodity flow approach was used probably for the first time by Kuznets [37]. There were, of course, also the related early works by Wassily Leontief. 23. Admittedly of a somewhat shaky quality by standards of m ore recent research. See, for instance Krantz [36]. 24. Official results from the project were released for the first time by Statistisk D epartm ent in its news bulletin in the autum n 1945. (Statistiske efterretninger, no 52,1945). The final report of the project was "Nationalproduktet og Nationalindkomsten 1930-1946” [18] which also included some preliminary estimates for the war years. 25. 1) Farming, 2) horticulture, forestry, fishery, 3) manufacturing, 4) handicrafts, 5) building and construction, 6) commerce and banking, 7) transport, 8) liberal professions, 9) public services, 10) services of dwellings. 26. "By source ... incomes ... may be split into interest, wages and operating surplus. If interest is charged on all capital used in production, irrespectively of whether it is owned by the entrepeneur or not, the surplus becomes payment for the entrepeneur’s labour plus what is called profit. An estim ate of

O. Aukrust/The Scandinavian Contribution

57

imputed interest for 1935 gives as a result slightly m ore than one billion kroner or one fifth of total national income; and a very rough estim ate of wages amounts to a total wage bill of about 2.5 billion kroner. T here rem ains as income for the entrepeneurs 1.7 billion kroner, a very considerable part of which must be seen as rem uneration of their labour ..." (loc.cit. p. 378). 27. The project was triggered when the Germ an controlled Quisling government asked Statistisk SentralbyrS to provide an estimate of the costs of the war to the Norwegian economy, the estimates to be used after the war sis a basis for claiming war damages - from England! The request provided a pretext for Statistisk Sentralbyri to start an inquiry for a much different purpose, namely (1) to provide annual estim ates of the national income from 1935 to 1943, (2) to estim ate the value of the stock of Norway’s real capital in 1939, the capital losses during the war, and the costs of the Germ an occupation, which would provide a basis for claiming damages from Germany. The project team included, in addition to Bjerve, E. Erichsen, O.D. Koth Norby, S. Rossen and F. Wedervang. The resulting publication from 1946 [56] was the outcome of the joint efforts of the team-members. 28. The accounting structure consisted of a set of four accounts which could be used for any sector (a person, an industry, the total economy). It showed how the (real) income of the sector could be estimated, by three different methods, as balancing items on three independent accounts, namely (1) as the sector’s net product (value added), (2) as the sum of the services provided by the sector’s factors of production (labour, organization, real capital), (3) as the sum of the sector’s consumption, real investment, and export surplus (corresponding to Lindahl’s m ethods 1, 3 and 2 respectively). Intersectoral flows were not shown. 29. With the needs of the national budget in mind Bjerve in 1946 designed a revised national accounting structure highlighting the interrelationship between the government and the private sectors, adding new variables and new accounts and at the same introducing certain terminological changes. Responsibility for current national income estimates for Norway remained with the Ministry of Finance through 1950. 30. Their decision may have been influenced by the fact that Bjerke and Milhoj probably did not take the same strong interest in theoretical and conceptual issues as Ohlsson and Aukrust. They were pragmatists, more attracted by empirical work than by theoretical refinements. Bjerke, furthermore, had to divide his interest between national accounting work and the collection of primary statistics within fields assigned to the statistical division of which he was chief. The greater acceptance by Denm ark of international recommendations did, however, give rise to some differences in the developments of national accounting in the three countries. 31. For instance, Aukrust in Norway argues in his dissertation: "In our work the guiding principle has been to aim for a national accounting system which will be suited "best possible" to the many different uses for which the data are to be used and - equally important - to the primary statistics available. There has been no explicit discussion of which purposes should be given priority in the event of conflict between them (to construct two or more alternative systems is hardly practical), since it has been a basic premise of our work that most reasonable demands may be met by making the system sufficiently detailed" [5], 32. While having some knowledge of Frisch’s Eco-circ System Aukrust was in 1946 a total new-comer to the field of national accounting. At an early date, he joined on a three weeks study-trip to Copenhagen and London which provided him with an opportunity to learn about the commodity flow approach used in Denm ark by Viggo Kampmann, and about the work in England by Jam es M eade and Richard Stone. The visit to England came about at the initiative of Bjerve with the consent of the Minister of Finance. The group visiting London included i.a. Ohlsson from Sweden. This is a welcome opportunity to recall, with gratitude, the many sessions when Stone gave freely of his time, guiding his Scandinavian visitors through the intricacies of his League of Nations paper [60] which was available in 1946 in mimeographed form. The m eeting with Stone is described with great humour, unfortunately in Swedish, by Ohlsson [51], and also by Bjerve [15]. - Aukrust, on his way home stopped briefly in the Hague to have a look at what was being done in the Netherlands. The visit provided additional evidence in favour of commodity flow accounting and was the start of a lasting contact between the Netherlands and Norway.

58

O. Aukrust/The Scandinavian Contribution

33. An early version of the paper presented at a Seminar at London School of Economics during a stay there in the early months of 1949. Advice and criticism received from professor M eade on that occasion is recalled with gratitude. I rem em ber, for instance, that the term s "requited" and "unrequited" were suggested by him. 34. This publication, edited by Aukrust, contained some 250 pages of tables together with a non-technical description of concepts and methods. Input-output tables (35 industries) were included for the postwar years. On the whole the amount of detail was rather large for its time and comparable to that of the corresponding US publication which in some ways served as a pattern. In fact, one of the tables in the Norwegian publication, with data on "Full-time equivalent employment, by industry”, m easured in manyears, was a direct copy of a similar table for the United States. (The popularity fo this table makes one wonder why tables on labour input have not been included among the standard tables of the new SNA). O ther tables analyzed receipts and expenditures of general government in detail. Tables rather than accounts were preferred as a format for publishing the resulting statistics. A small set of accounts in the text served mainly to explain the interrelationships between im portant flows and aggregates, and as a framework for describing their operational definitions. 35. The reason why balancing items cannot be defined as classes is that while, in logic, classes may always be added, subtraction is only allowed when the subtrahend is a sub-class of the minuend: W e may conceive of a basket containing three apples and two oranges, but not one containing three apples minus two oranges. 36. Alternative systems of postulates are of course possible. Some might lead to national accounting systems different from the one described here, in much the same sense as non-Euclidian geometries are different from Euclidian geometries. An example is given in [8], 37. The definition of gross value added suggested in an early draft of "SNA mark III", for some reason not included in the final text, reflected this fact. It read: 'T he concept of gross value added is defined as the increase in value between a set of inputs of goods and services and the set of outputs of goods and services which results from transforming, or consuming, those inputs in the process of producing those outputs." 38. In this context Ohlsson observes, quite logically, that once the concept of "enterprise consumption" is allowed into the system, this category ought to be broadened so as to contain ” ... not only the value of banking and insurance services, but also the cost of collective arrangem ents for the employees in the enterprises, such as holiday places, medical attendance, sports fields, lunch rooms. Consumption of this kind is similar to government consumption and could also be assigned to the producing sector. Many imputations would be avoided by such a compromise" (loc. cit. p. 149). This interesting suggestion was not followed up in Ohlsson’s empirical work, however, presumably because of lack of data. 39. The Scandinavian view has its roots in the works of Frisch and Lindahl. Frisch, noting that the factors of production of a sector may be active in production outside the sector as illustrated in the Eco-circ graph (Appendix A), argued that the concepts of "residence" and "sector" are logically interrelated: "The logical startingpoint ... [for defining a sector] ... is a criterium which decides whether a person or a piece of capital (any element which can generate ... value added, in short any factor) is resident "by us", i.e. in the sector which we are studying, E, R, J etc. relates to these resident factors. Thus, the residence criterium gives content to the concept of the sector”. ([26], p. 113) Frisch’s reasoning, at the national level, logically led to the product (value added) generated by resident factors being considered "the product of the nation". Frisch did not discuss how the borderline between "resident" and "foreign" factors of production should actually be drawn. In the empirical work of the Scandinavian countries the issue is resolved by treating land and stationary real capital always as resident of the country in which it is located, while special conventions have been laid down in respect of labour and movable equipment (ships, airplanes etc.) (Com pare the theoretical discussion by Lindahl ([42], p. 88). 40. The use of reliability indicators was recommended by Frisch as early as in 1939: "As far as possible ± limits should be indicated for the individual entries in the accounts, e.g. as done systematically by Sir Josiah Stamp in his estimates from 1919 of the national income of different countries" ([24], p. 143, my

O. Aukrust/The Scandinavian Contribution

59

translation). 41. The decision against adopting this practice was defended by Bjerke and Milhoj with the following words (my translation): "Such a sytem, in which data are characterized according to a m ore or less detailed scale, ranging from highly reliable to rather doubtful, no doubt gives more information than indications of the reliability of estim ates given at scattered places in the text. We have, nevertheless, hesitated in adopting such a system, mainly for fear that it might add further arbitrariness. In reality, it would be no m ore than an application of personal judgement, admittedly, the persons responsible for the estimates are best placed to exercise, but which can nevertheless never be exercised according to some objective norm. To this may be added the fact that the majority of aggregates will be composed of reliable and less reliable components to such an extent that the reliability of the total will be even more difficult to judge. Finally, there is the difficulty that the reliability of the estimates may vary over time ... new or revised primary statistics may change the reliability of estimates fundamentally from one year to the next" ([20] p. 68*-69*). 42. The arbitrariness necessarily involved in this kind of exercise is illustrated by a quotation (my translation):".... it is assumed, first, that the value of the work of a wife may be set equal to the average income of domestic servants increased by 10 per cent. Thereafter, this income has been imputed with its full amount to the following categories of persons from the population census: M arried women assisting in the activities of their husbands; widows below the age of 65; divorced women, and daughters active in domestic work at home. M arried women with an outside activity have been imputed one half of this income ...” ([20] p. 44-45*). 43. Together with, among others, Jean Marczewski from France, I worked as an assistant to Stone at the "O EEC national accounts unit" in Cambridge, England, during the autumn months of 1949 when the "OEEC simplified system" [48] was drafted. None of us had the slightest influence on the outcome. A ttem pts by Ingvar Ohlsson, Kjeld Bjerke and myself to influence the system in 1951 when a later version of it came up for discussion by O E E C m em ber countries were equally unsuccesful. - The standardized system of the United Nations from 1953 was agreed without contribution by Scandinavian experts. 44. In Norway this was done by extending existing series backwards in two steps first to 1900 and then to 1865 [58] [59], and in Denm ark also in two steps to 1870 [11] and 1720 [31]. Sweden obtained series back to 1860 as a result of work by O. Lindahl [43] and Johansson [33] bringing up to date the series given in Lindahl et al. [39]. Some of these studies were financed as parts of Simon Kuznets’ "international growth project".

References [1] Andvig, Jens Chr. (1986): Ragnar Frisch and the G reat Depression. A study in the interwar history of macroeconomic theory and policy. Mimeographed. Norsk utenrikspolitisk institutt. 467 p. Oslo. 1986. [2] Aukrust, Odd, Petter Jakob Bjerve, Ragnar Frisch (1948): A system of concepts describing the economic circulation and production process. Stencil-memo from University Institute of Economics, 8 March 1948. Oslo. [3] Aukrust, Odd (1949): On the theory of national accounting. In: Review of Economic Studies. Vol 16, no. 41 (1949/1950), pp. 170-188. [4] Aukrust, Odd (1950): Input-output studies in Norway. In: Input-Output relations. Proceedings of a conference on inter-industrial relations held at Driebergen, Holland. Ed. by The Netherlands Economic Institute, Leiden 1953, pp. 111-121. [5] Aukrust, Odd (1955): Nasjonalregnskap. Teoretiske prinsipper. (National accounts. Theoretical principles). Samfunns 0 konomiske studier no. 4. Statistisk Sentralbyri, Oslo. [6] Aukrust, Odd and Juul Bjerke (1958): Realkapital og 0 konomisk vekst 1900-1956, (Real capital and

60

O. Aukrust/The Scandinavian Contribution

economic growth 1900-1956). Oslo, Statistisk SentralbyrS, Artikler no. 4. English edition in: Income and W ealth series VIII, ed. by Raymond Goldsmith and Christopher Saunders, London 1959. [7] Aukrust, Odd (1966): An axiomatic approach to national accounting: An outline. In: Review of Income and W ealth, ser. 12, no. 3, pp. 179-190. [8] B6nard, Jean (1972): C om ptabilitl nationale et moddles de politique ficonomique. Paris. [9] Beretning (1940): Beretning om D et 3. nordiske statistikerm0te i Oslo den 28. og 29. juni 1939. (Proceedings of the 3 Nordic M eeting of Statisticians, Oslo, 28-29 June, 1939), Oslo 1940. [10] Bjerke, Kjeld, Poul Milhoj and P. N arregird Rasmussen (1954): National Experiences: Denm ark. In: The Structural interdependence of the economy. Proceedings of an international conference on input-output analysis, Varenna, 27 June-10 July 1954, pp. 233-257. [11] Bjerke, Kjeld and Niels Ussing (1958): Studier over Danm arks nationalprodukt, 1870-1950. (Studies of D enm ark’s national product 1870-1950). [12] Bjerve, Petter Jakob (1947): Nasjonalbudsjettering. (National budgeting) In: Ekonomisk tidskrift, 1947, Srg. 49, pp. 99-112. [13] Bjerve, Petter Jakob and Mikael Selsjord (1959): Financial accounting within a system of national accounts. In: Income and W ealth Series VIII pp. 60-79. [14] Bjerve, Petter Jakob (1986): Ragnar Frisch og 0kosirksystem et. (Ragnar Frisch and the Eco-Circ System). In: Sosial0 konomen 1986, Srg. 40,no. 9, pp. 17-23 and no. 10, pp. 26-30. Also as Statistisk sentralbyri’s reprint series no. 19, Oslo. [15] Bjerve, Petter Jakob (1989): 0konom isk planlegging og politikk. (Economic planning and policy) Oslo. 228 p. [16] Dahlgren, Einar (1936): Produktionsstatistisk uppskattning avSveriges nationalinkomst Sren 1930-1934 ("Production statistical estimates of the national income of Sweden 1930-1934”). Statens offentliga utredningar, no. 18, 1936. Appendix 1. [17] Dahlgren, Einar (1941): PM med vissa berakningar ang&ende nationalinkomstens storlek ("PM with some estimates relating to the size of the national income"), Statssverkspropositionen, 1941, Bilaga B. [18] D et Statistiske D epartem ent (1948): Nationalproduktet og nationalindkomsten 1930-46. (National product and national income 1930-1946). Statistiske M eddelelser 4. rackke, 129. bind, 5. haefte. Kebenhavn. [19] Det Statistiske D epartem ent (1951): Nationalproduktet og nationalindkomsten 1946-49. (National product and national income 1946-49). Statistiske M eddelelser 4. raskke, 140. bind, 2. haefte. Kebenhavn. [20] Det Statistiske Departem ent (1955): Nationalindkomsten 1938 og 1947-54 (National accounting 1938 and 1947-54). Statistiske Meddelelser 4. raskke, 160. bind, 2. hxfte. Kobcnhavn. [21] Economic Commission for E urope (1956): Input-output tables: Recent experience in W estern Europe. In: Economic Bulletin for Europe, Vol. 8, no 1, May 1956, pp. 36-60. [22] Frisch, Ragnar (1933): Propagation problems and impulse problems in dynamic economics (In Economic Essays in H onour of Gustav Cassel). London. [23] Frisch, Ragnar (1935): E t generelt monetasrt begreps- og symbolsystem (A general system of concepts and symbols) Mimeographed. Oslo. [24] Frisch, Ragnar (1939): (Com ments on a paper by Erik Lindahl) I: Beretning om D et 3. nordiske statistikerm 0te i Oslo den 28. og 29. juni 1939. Oslo 1940, pp. 141-149. [25] Frisch,

Ragnar

(1942):

Noen

innforingsmerknader

om

0kosirksystem et.

D et

0konomiske

sirkulasjonssystem. (Introductory rem arks to the Eco-circ System - the economic circulation system). M imeographed. Oslo. [26] Frisch, Ragnar (1943): 0kosirk-system et (det 0konomiske sirkulasjonssystem) (The Eco-sirc System - the system of economic circulation). In: Ekonomisk tidskrift, irg . 45,1943, pp. 106-121. Stockholm. [27] Frisch, Ragnar (1948): W ork on national accounting done at the University Institute of Economics. Stencil-memo, D ecem ber 1948. Oslo. [28] Frisch, Ragnar (1949): Axiomatic remarks on some national income concepts. Copy of a letter to

O. Aukrust/The Scandinavian Contribution

61

[29] Frisch, Ragnar (1955): M arket price versus factor cost in national income statistics. In: Sankhya, vol. 15, M arch 1955, pp 1-8. Calcutta. [30] Gilbert, Milton and associates (1958): Comparative national products and price levels. OEEC, Paris [31] Hansen, Svend Aage (1972-1974): 0konom isk vaekst i Danm ark 1720-1970. (Economic growth in Denm ark 1720-1970). Vol. 1 and 2. Copenhagen. [32] Hoglund, Bengt and Lars W erin (1964): The production system of the Swedish economy. An input-output study. Stockholm economic studies, new series IV, Stockholm. [33] Johansson, 0 ste n (1967): The gross domestic product of Sweden and its composition 1861-1955. Stockholm economic studies, new series VIII, Uppsala. [34] Kampmann, Viggo (1942): Danm arks nationalindkomst i trediverne. (D enm ark’s national income in the 1930s). Lecture to National0konomisk Forening 19. October 1942. In: National 0 konomisk Tidsskrift, vol. 80, 1942 pp. 367-384. [35] Kiacr, A.N., (1893): Indtacgts og formuesforhold i Norge. (Income and W ealth in Norway.) In: Tillaeg til Statsokonomisk Tidsskrift 1892 ogl893. [36] Krantz, Olle (1988): New estim ates of Swedish historical GD P since the beginning of the nineteenth century. In: Review of income and wealth, ser 34, 1988 no.2, pp. 165-181. [37] Kuznets, Simon (1938): Commodity flow and capital formation. Nat. Bureau of Ec. Research, New York. [38] Kuznets, Simon (1941): National income and its composition, 1919-1938, Volume I and II, New York. [39] Lindahl et al (1937), Erik Lindahl, Einar Dahlgren and Karin Kock: The National income in Sweden. 1861-1930. Parts one and two.University of Stockholm, Institute for Social Sciences. London 1937. [40] Lindahl, Erik, (1939): Om de problem er som knytter seg til utarbeidelsen av en statistikk over nasjonalinntekten og nasjonalformuen. In: Beretning om Det 3. nordiske statistikermote i Oslo den 28. og 29. juni 1939. Oslo 1940, pp. 104-125. [41] Lindahl, Erik (1939): Algebraic discussion of the relations between some fundamental concepts. In: Studies in the theory of money and capital, pp. 74-138. London. [42] Lindahl, Erik (1954): Nationalbokforingens grundbegrepp. (The basic concepts of national accounting.) In: Ekonomisk tidskrift, Srg. 56, 1954, pp.87-138. Reprinted in English translation in International Economic Papers no. 7, 1957. London. [43] Lindahl, Olof (1956): Sveriges nationalprodukt 1861-1951. (The gross domestic product of Sweden 1861-1951). M eddelanden fr&n Konjunkturinstitutet. Serie B:20 Stockholm. [44] Meddelanden Serie A:14 (1946): Konjunkturlaget hasten 1946. Meddelanden frSn Konjunkturinstitutet. Serie A:14. Stockholm. [45] Meddelanden Serie B:13 (1951): Nationalbokforing 1946-1950. Berakningsmetoder och anvandningen for nationalbudgetering (National accounting 1946-1950. Estimation m ethods and uses in national budgeting work). M eddelanden frSn Konjunkturinstitutet. Serie B:13. Stockholm. [46] Mitchell, Wesley C. and Simon Kuznets (1934): Current Problems in M easurem ent of National Income, Bulletin de l’lnstitut International de Statistique, vol. 28, No. 2, pp. 280-298. [47] Niitamo, O.E. (1982): A survey of the recent and current state of Finnish national accounts. In: Tutkimuksia N:o 83, Central Statistical Office of Finland, pp. 13-24. [48] O E E C (1950): A simplified system of national accounts. Organisation for European economic co-operation, Paris, April 1950. [49] O E E C (1952): A standardised system of national accounts. Organisation for European economic co-operation, Paris, 1952. [50] Ohlsson, Ingvar (1953): On national accounting. Konjunkturinstitutet. Stockholm. [51] Ohlsson, Ingvar (1987): Nationalrakenskaperna och den ekonomiska statistiken. (National accounts and economic statistics). In "Konjunkturinstitutet under Erik Lundbergs tid”, pp. 35-68, Stockholm. [52] Ralph, E.R. (1948): The concept of transfers in national income estimates. In: Quarterly Journal of Economics, vol. 62, 1948, pp. 327-361. [53] Sevaldson, Per (1954): National Experiences: Norway. In: The Structural interdependence of the

62

O. Aukrust/The Scandinavian Contribution

economy. Proceedings of an international conference on input-outpur analysis. Varenna 28 June-10 July 1954, pp. 295-311. [54] Sevaldson, Per (1961): Changes in input-output coefficients. In: Structural interdependence and economic development. Proceedings of an international conference on input-output techniques. Geneva, Septem ber 1961. Ed by Tibor Barna, London 1963, pp. 303-328. [55] Statistisches Reichsamt (1932): Das deutsche Volkseinkommen vor und nach dem Kriege, Berlin. [56] Statistisk SentralbyrS (1946): Nasjonalinntekten i Norge

1935-1943. Realkapitalen 1939 og

kapitalreduksjonen under krigen. Okkupasjonskostnadene. (National income of Norway. Real capital 1939 and the reduction of capital during the war. Costs of the occupation). NOS X 102. Oslo. [57] Statistisk SentralbyrS (1952): Nasjonalregnskap 1930-1939 og 1946-1951. (National accounts 1930-1939 and 1946-1948). NOS XI, 109. Oslo. [58] Statistisk SentralbyrS (1953): Nasjonalregnskap 1900-1929 (National accounts 1900-1929) (NOS X I 143) [59] Statistisk Sentralbyra (1965): Nasjonalregnskap 1865-1960. (National accounts 1865-1960). (NOS XII 163). [60] Stone,Richard, (1947): Definition and m easurement of the national income and related totals. Appendix to M easurem ent of national income and the construction of social accounts. Studies and reports on statistical methods No. 7. United Nations, Geneva. 1947. (Stone’s mem orandum was avai­ lable in mimeographed form in 1946). [61] Studenski, Paul (1958): The Income of Nations, New York. [62] United Nations (1953): A system of national accounts and supporting tables (Studies in methods, series F, no. 2) New York. (Refered to as "SNA mark I" [63] United Nations (1968): A system of national accounts. (Studies in methods, series F no. 2, rev. 3). New York. (Refered to at "SNA mark II".) [64] Wedervang, Ingvar (1926): Nasjonalinntekten i Norge. (The national income of Norway.) In: Statsokonomisk Tidsskrift 1926, pp. 149-201, 237-298.

O. Aukrust/The Scandinavian Contribution

63

Appendix A The Eco-circ graph The Eco-circ graph is constructed according to the following principles: Every line represents an economic flow. All lines curve the same way, positive flows mooving clock-wise. At each junction point the sum of the flows moving into the point equals the sum of the flows moving out of it. The graph as a whole may be interpreted as a closed system of definitional equations, or as a system of accounts. The graph may be made to represent a closed sector (as is shown for the real circulation on this page), or an open sector (as is shown on the page opposite) by splitting all relevant flows into two, those internal to the sector and those external to it, e.g. A = Afa + A", where A“ would represent (net) exports out of the country and A " = £ “ + H".

Ordinary letters A, C, T etc. denotes real magnitudes, heavy letters A, C, T etc. denotes financial magnitudes + and - used as superscripts indicate the positive and the negative component given magnitude, e.g. J = J* • T

respectively of a

N, K, O used as superscripts indicate functional income shares RN, RK, R° in = intern and ex = extern are added to distinguish internal flows from flows out of/into the sector, e.g. A = A " + A“

The real circulation in a closed sector A

=

"All" the product, gross output

C

=

"Consum ption", "C onsom ption",

T

=

"Com eam o" "Tax", "Taxe"

I

=

"Investm ent”, "Investissem ent", "Investition”, "Investim ento"

J

=

G ross investm ent (the neighbouring

D

=

letter to I) "D epreciation", "D epreciation", "depreziazone" H

=

Interm ediate input ("H ardw are")

E

=

V alue added ("E arnings", "Ertrag",

R

=

N et incom e ("R evenue", "Revenu",

S

=

"Savings", "Spargeld"

N

=

L abour ("N um ber", "N um erus")

K

=

C apital ("K apital")

O

=

"O rganization", "O rganisation"

F

=

"Excedent", "Entrata") "Reinertrag", "Reddito")

"Free", "Frei". W indfalls, exogenous changes in asset holdings

G

=

"Gain", "G ain", "G ewinn", "Guadogne" (revaluation, holding

M =

gains and losses) M oney

O. Aukrust/The Scandinavian Contribution

The real and financial circulation of an open sector

Real circulation

Total domestic consumption of goods and services 1935. Million kroner at market prices. Agriculture

Gardening forestry, fishing

1 604,8

172,1

Plus: Imports non-durables.................. d u ra b le s ......................... Total ................................

277,9

18,6

277,9

18,6

Less: Exports non-durables................. d u ra b le s ......................... Total ................................

866,4

39,3

866,4

Domestic u s e s ..................

1 016,3

b. Gross investments in: A griculture.................. Gardening etc.............. Manufacturing .......... Handicrafts ............... Build. & Constr........... Commerce, banks etc. T ra n sp o rt.................... Professions.................. Public s e rv ic e s .......... Dwellings .................. Total gross investm. .

Building and constructions

Commerce Banks Insurance

Transport

Professions

Public goods and services

Rent of devellings

All trades

I R*

-

172,1 -

-

-

266,4 -

316,7 3,7 -

-

-

-

-

203,7 104,0 307,7

-

39,3 151,4

3 219,7

894,7

688,3

1 643,3

486,3

178,3 8,3 764,3 266,9 229,6 196,3 67,6

27,0 1,3 114,5 25,3 69,3 68,4 38,6

35,9 1,5 20,5 5,7

74,9 5,7 112,9 71,3 69,2 80,5 38,9

33,9 3,6 42,7 20,5 32,8 92,9 49,1

-

-

21,4 3,3 3,9

28,6

1 222,2 87,7 1 309,9

209,0 -

-

-

1 318,4 104,0 1 422,4

.

209,0 277,9

.

523,1

14,4 1,5 22,5 8,1 6,3 14,7 6,2 2,5

572,2

.

9 473,2

-

-

-

-

-

18,7 6,1

4,1

-

79,9 74,1

-

-

5,1

-

279,6

-

81,3

-

3 824,0

-

. . -

_ . -

. . -

-

-

-



69,8 9,7 148,3 33,0 7,8 37,2 145,8

_

.

_

_

-

25,5 35,0

.

.

. . _ -

1 708,9

357,8

278,1

478,2

11,9 1,5 29,9 2,0 1.5 3,0 15,3

31,7 2,0 44,1 6,6

2,1 0,2 6,1 10,4 0,4 2,3 2,6

_

0,8

-

62,6 202,8

-

-

-

-



67,1 202,8

221,3

65,1

410,2

24,9

-

-

-

-

721,5

1 124,6 164,9

414,1 57,7

1 053,2 87,0

206,7

-

1 289,5

471,8

-

1 140,2

206,7

-

.

.

3, 7.

-

-

-

-

122,8

272.3

13,4

-

404,8

523.1

-

24,1 6,0 68,2 14,0 5,9 14,9 84,5

Total consumption . .

277.9

8 538,3 1 047,4 9 585,7

37,6

-

-

572.2

. -

-

122,8

523.1

630,8 21,9 1 415,5 364,8 411,1 478,5 237,8 2,5 175,8 85,3

_

-

277.9

-

695,3

-

_

404,8

695,3

-

-

c. Personal consumption: non-durables ............ d urables......................

1 531,4 111,9 1 643,3

-

-

611,5

278,1 410,2 688,3

-

22,1

-

771,9 122,8 894,7

925,7 87,7 1 013,4

-

-

0,2 2,4

2 111,5 402,5 2 514,0

-

-

17,0 43,4

. -

-

277,9 -

277,9

441,8 -

441,8

572,2

4 045,9 881,8

572,2

4 927,7

O. Aukrust/The Scandinavian Contribution

1 604,8

Total m a te ria ls ..........

Handicrafts

cc

Production value non-durables................. d u ra b le s ......................... Total ................................

of which: a. Materials etc. for: A griculture.................. Gardening etc.............. Manufacturing .......... Handicrafts ............... Build. & Constr........... Commerce, banks etc. T ra n sp o rt.................... Professions.................. Public s e rv ic e s .......... Dwellings ..................

Manufacturing

>

1

66

The Accounts o f Nations Z. Kenessey (Ed.) IO S Press, 1994

DUTCH NATIONAL ACCOUNTS: A HISTORY Gert P. den Bakker* National Accounts Research Division, Netherlands Central Bureau of Statistics

Abstract. This paper describes the history of Dutch national accounting. After two early estimates in the beginning of the nineteenth century, m odern national accounting started in the 1930s in the context of the Tinbergen model for the Dutch economy. Work surged after W orld W ar II to provide data to the government for economic planning purposes. In the 1980s, the development was towards a flexible and institutional approach.

Introduction This paper deals with the origin and development of the Dutch national accounts. In the past, part of this history was written in Nooteboom [1] and Oomens [2]. Both articles have been written in Dutch. For the non-Dutch readers, there is of course Studenski [3]. In Studenski’s study the developments in the Netherlands were not fully represented (perhaps as a result of language difficulties). For instance, Studenski only deals with the official estimates prepared during the 1930s. The description of developments between 1940 and 1955 is cursory. This is not to say that the present paper gives an exhaustive description of the history of national accounting in the Netherlands. Its purpose is to provide more details about the early developments in the Netherlands and to extend this description to the present. In the Netherlands, the first known national income estimates go back to 1803 and 1804. The purpose was to compare the economic power of the Netherlands with other countries. Strangely enough, these estimators were not followed by any others in the nineteenth century, whereas in nineteenth century England and France the development of both the theory of national income and the estimation methods continued. In the first decade of the twentieth century, the thread was resumed with national income estimates on the basis of tax statistics. National income estimates on a regular basis started in the second half of the 1930s. A project was set up to improve the estimates which played an important role in the Tinbergen econometric model of the Dutch economy. During and after World War II, the Dutch Central Bureau of Statistics (CBS) undertook the compilation of a * The author would like to thank Frits Bos for his stimulating comments. The views expressed in this paper are those of the author and do not necessarily reflect the views of the Netherlands Central Bureau of Statistics.

G.P. den Bakker/Dutch National Accounts

67

system of national accounts for the purpose of planning economic recovery. This development accelerated in the early 1950s. Countries which wished to qualify for the Marshall aid were obliged to submit requests and to provide macro-economic data cast in a national accounts framework that followed the OEEC guidelines [4,5], In the 1970s, the CBS improved and extended the basic statistics for the national accounts and this led to a substantial qualitative and quantitative expansion of the Dutch national accounts. The paper is organized as follows: sections 2-4 contain a chronology of Dutch national accounting work. Section 2 presents two early national income estimates. Section 3 describes the developments that started with isolated national income estimates in the beginning of the twentieth century and led to a true system of national accounts in 1950. Section 4 deals with the period from 1950 up to the present. In section 5 a summary and conclusions are presented.

1. Early Estimates 1.1. Introduction A description of the history of national accounting in the Netherlands should start in the early 1800s with studies by Keuchenius [6] and Metelerkamp [7], Up to now, these studies are virtually unknown in the international literature on the history of national accounting1. For instance, Studenski’s dates the first national income estimate in the Netherlands back to 1910. The ’discovery’ of these two Dutch studies seems to imply that the Netherlands were the fourth country in which national income estimates were made2. As is well known, the very first estimates were made for England by Petty in 1665 and King in 1696. They made their estimates to compare the wealth and power of England with other countries. Therefore King [8] estimated national income for France and Holland as well (for the reference year 1688). Like Petty and King in England more than a century earlier, Keuchenius and especially Metelerkamp were concerned with determining the resources of their country in relation to other countries and its capability to survive from an economic and military viewpoint. 1.2. The First Dutch Estimator: Keuchenius Dr. W.M. Keuchenius, member of the city council of Schiedam, was the first Dutchman who estimated national income for the Netherlands. He developed a hypothetical national income estimate, based on a situation that might become reality if war-stricken Europe were to conclude peace. Very little is known of Keuchenius except that he wrote his book just before his departure to the Dutch East Indies. He made a ’national balance’ for the Batavian Republic (1795-1806)

G.P. den Bakker/Dutch National Accounts

68

which in modern terms amounts to a review of the supply and disposition of goods and services. In estimating national income, Keuchenius distinguished four economic activities which yielded income: agriculture, fishing, international trade and industrial activities. Besides, he estimated income from foreign properties and interest on national debt. In Keuchenius’ view domestic trade was not a productive activity. In addition, he did not distinguish construction, government and other services. However, interest on government debt was part of national income. Elements of Keuchenius view can be found in the Material Product System which has been adopted by the (former) centrally planned countries as the basis for their national accounting figures. Keuchenius estimated national income at 221 million guilders, see table 1. His population estimate was 1,882,891 persons, so that he arrived at a per capita income of 117 guilders. The contribution of agriculture and fishery to the national income was 45% while 27% of national income was property income from abroad (including colonies). Keuchenius was rather vague about the expenditure side of his balance. Final consumption expenditure and government expenditure were 166 (that is per capita 88 guilders) and 55 million guilders respectively. Keuchenius did not mention fixed capital formation and did not present explicit estimates for exports. He only stated that imports (71 million guilders) should not exceed exports. Table 1. National income according to Keuchenius, circ a 1800 (mln gId) A griculture Fi shery In te rn a tio n al trade, cargo trade, insurances Income from properties in the Dutch East Indies Income from properties in the Dutch West Indies Factories and processing in d u strie s Income from foreign property Inte rest on government debt National income

84 15 30 11 9 3 40 29 221

Keuchenius also presented a multi-industry analysis, which resembled an input-output analysis in some respects. He estimated the contribution of all industries to national income and employment and calculated the effects on the national economy if the relative significance of a certain industry was to change (as a result of government policy). In addition, he discussed the advantages and disadvantages of the presence of foreign workers in the Netherlands and he presented regional figures and estimates of the government debt and tax revenues.

G.P. den Bakker/Dutch National Accounts

69

1.3. The Estimates of Metelerkamp Shortly after Keuchenius study, a second one on national income and wealth was published by R. Metelerkamp [7]. Metelerkamp, a jurist, was a member of a society for art and science. He was a man of importance who, because of his political views, was not allowed to hold a position in the government of the Batavian Republic. Instead, he set to study and travel, collecting information for his book. Later on, after the French period, he became chairman of the Lower House of the States General (the Dutch parliament of the time) and member of the Council of State. Metelerkamp, as his predecessors, aimed to study the relative power of his country. His estimation method resembled that of W. Pitt and H. Beeke for England. He mentioned the work of Keuchenius and he corrected some of his figures. In addition to national income estimates for the Netherlands, Metelerkamp presented figures for France, England and Saxony. Besides, he gave figures for national wealth, imports and exports, income and outlay of the government and size and composition of the military force for eight European countries. For the Netherlands, Metelerkamp arrived at a national income estimate of 250 million guilders in 1792. He estimated the population at two million people, which implies a per capita income of 125 guilders. This is only slightly higher than Keuchenius estimate. Metelerkamp’s estimate of national wealth was 3,500 million guilders3. On the basis of his estimates, Metelerkamp concluded that the economic position of the Netherlands was quite good, relatively speaking. However, he judged the militairy situation as very bad.

2. Towards a System of National Accounts 2.1. The start: Bonger and Bakker Systematic national income estimates in the Netherlands were introduced by Prof. W.A. Bonger. Bonger, by training a jurist, based his estimates on tax statistics. In his first publication [9] he presented figures for 1908. Afterwards, he extended these estimates up to 1920. The first official national income estimate (for 1929) was published by the CBS in 1933 [10]. This estimate was also based on the income method and was prepared under the direction of the CBS’s accountant, Dr. O. Bakker [11]). The estimate was part of a study on the influence of wages on the cost price of products, carried out by the CBS on request of the Supreme Labour Council (dd. 4 July 1931). The 1929 figures are given in table 2. Afterwards, Bakker made estimates for 1914, 1919 and 1929-36.

70

G.P. den Bakker/Dutch National Accounts

Table 2. National income according to Bakker, 1929 (mln gld) Income (income tax fig u re ) Undeclared income (10X) Income not taxed Professionals Small re n tie rs Persons abroad lia b le to Dutch income tax Net-income to abroad Retained p r o fits Social c o n trib u tio n s

4367 437

National income of which Uages and s a la rie s P r o fits

5467

686 40 37 -400 200 100

4000 1450

2.2. The First System o f National Bookkeeping: Van Cleeffs 1938 Estimates The very first ’system of national accounts’ for the Netherlands was compiled by Van Cleeff [12,13]. He published his system in the monthly ’De Economist’ and called it a ’national book-keeping system’4. Van Cleeffs intention was to demonstrate that, in principle, a national bookkeeping could be compiled in analogy with the methods used in ordinary bookkeeping. According to Van Cleeff a national bookkeeping was important for a centrally planned economy. In 1945, he published an article [14] on this subject (written in 1942) and joined the newly established Central Planning Bureau (as many other Dutch national income pioneers). Van Cleeff distinguished four sectors (trade, enterprises, government and consumers) and five main groups of transactions (accounts) (commodity flow, financing etc.). He assumed that total income equals total expenditure in every sector and that all transactions were paid in cash. The results of Van Cleeffs study are presented in table 3. Although Van Cleeff used crude figures, his national income estimate (5100 million guilders) was virtually the same as Derksen’s later estimate: 5153 million guilders [15]. Van Cleeff made various recommendations which were carried out later, such as an industry break-down and volume figures. He also emphasized the need of a central register which should link the existing (non-CBS) registers. Besides, he wrote: ’... of course, a national balance sheet has to be added, or better, two of them, one for the beginning and one for the end of the year.’ The Dutch CBS has indeed compiled balance sheets (for instance for 1938), but in the course of time the balance sheets disappeared. Recently, the CBS started again with the compilation of balance sheets.

G.P. den Bakker/Dutch National Accounts

71

Table 3. A summary annual survey for the Netherlands, 1938 (mln gld) Trade

1.

II.

PURCHASING POWER FLOW Cash Receipts and payments COMMODITY FLOW Goods Purchase and sale Manufacture Profit Capital goods Purchase and use Capital stock Investment

III. FINANCING Share holding, etc. Issued Share capital, etc. Subscribed Undistributed profits IV.

V.

INCOMES AND OUTLAYS Paid incomes Various income components Surplus to profit account Surplus to commodity account Received incomes Various income components Outlays Purchase of goods Taxes CALCULATION OF PROFIT Profit account Surplus paid incomes Surplus commodity account Distributed profits Retained earnings Total

Production

D

C

D

C

6550

6550

4950

4950

5750 800

6550 -

Government

Consumers

D

C

700

700

-

-

-

-

-

-

-

-

-

-

-

-

D

c

5100

5100

4350 2850 1500

_

_

1150

1150

-

-

400

-

-

-

_

_

.

.

.

.

-

-

200 200

.

-

-

-

-

700 -

-

-

-

-

_

700

_

.

200

450 -

450 -

2100 -

_

_

_

_

_

_

_

_

-

-

-

-

-

-

4200 700

450 350 -

800 -

1300 200

1500

2100 _

5100 _

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

14350 14350 14450 14450

1400

1400 10200 10200

2.3. Derksen and Tinbergen in the 1930s and 1940s In the second half of the 1920s, the CBS started with business cycle research. After Prof. dr. J. Tinbergen had joined the CBS in 1927, econometric research accelerated and in 1936 Tinbergen presented his famous macro-econometric model for the Dutch economy [16]. The Tinbergen model gave the construction of models and the development of national accounting in the Netherlands a head start. In 1937, a project was set up to improve the national income estimates to serve as a better statistical basis for econometric studies. Prof. dr. J.B.D. Derksen, later chief of the National Income Unit of the United Nations, was responsible for these estimates. The first results refer to 1921-1936 and are the first official national income estimates for the Netherlands [17]. This publication also contained a detailed description of definitions and methods.

72

G.P. den Bakker/Dutch National Accounts

The figures were estimated on the basis of both the income and the production method. However, the former method was preferred: ’... for the time being, a better method to arrive at a fairly reliable national income estimate is not available [17, p.8]. The (improved) production method was used for plausibility checks on the results of the income method. In the latter method a fraud of 10% was assumed. In the production method thirty-one industries were distinguished. The Dutch approach to national income measurement was quite advanced. For example, services of owner- occupied dwellings were included in national income5. The results of both methods for the years 1921-36 show about the same annual growth rates of national income. The differences between both methods, expressed as a percentage of their average results, ranged from 1.4% in 1931 to 11.3% in 1923. The estimates according to the production method were always lower than those according to the income method. In the 1930s, the gap between the outcomes of both methods narrowed substantially. For these years more and better information (e.g. production statistics) became available. In the following years, estimates for the late 1930s were added, the estimates according to the production method were improved [19,20] and research on definitions and methods continued [21], In 1941 [22], national income estimates for 1900-20 were published which were based on the income method and consistent with the one from 1921 onwards. The fraud for these years was assumed to be 20% (copied from Bonger). This is twice as high as for 1921-39. In the latter years, the verification of the tax declarations was much improved. In 1948 [23], the first revision of national accounting data (referring to the years 1921-39) was published. This publication replaced the first official national income estimates [17]. Both the estimates by the income and the production method were corrected. It goes without saying that the outbreak of World War II and the successive occupation of the Netherlands greatly influenced work on national income. During the war, there was hardly any contact with statisticians in other countries. At CBS, research went on and in September 1945, very shortly after the end of the war, Derksen visited the National Institute of Economic and Social Research in London to exchange ideas with leading foreign national accountants. It turned out that during the war research in the Netherlands had been focused on the statistical aspects, whereas in the English-speaking countries research was more theoretically oriented. Derksen had various discussions with Stone who in 1946 prepared his famous monograph on social accounting for the United Nations [24], In order to make available to English-speaking readers the results of the investigations that had been carried out in the Netherlands during the war, Derksen prepared a paper ’A System of National Book-keeping - Illustrated by the experience of the Netherlands Economy’ [15], In the paper he presented tables for 1938 and national income estimates for 1900-1938. In addition, Derksen used the national bookkeeping system to show the impact of the war and the German occupation on the Dutch economy. The figures roughly

G.P. den Bakker/Dutch National Accounts

73

show which part of the national output of goods and services had to be put at the disposal of the Germans and which part was left to the Dutch people and how production and government expenditure were financed. Derksen wrote in his concluding remarks: ’The method of the national bookkeeping system also has the advantage that it offers in a concise form a programme of research. By studying the tables and figures it is easy to see where statistical information is still unsatisfactory and further research is needed. It is possible to organise a detailed research programme, using the national book-keeping system as a general guide’ [15]. At the CBS, this train of thought has been further worked out. The development of the system of national accounts has had a great influence on both the statistical programme and the organization of the CBS [25]. Derksen stated as well: ’Much statistical work remains to be done before it can be said that the national book-keeping system constitutes a valuable tool in economic analysis’ [15]. Indeed, this work has been done and the first results were published in 1950 [26], This publication is briefly discussed below, in section 3.5. The next section elaborates the developments at the CBS during the war (on which Derksen reported in London) and immediately after the war. 2.4. Progress at the CBS: The Commission on National Bookkeeping. On 19 January 1943, a ’Commission for the National Bookkeeping’ was established. This Commission was in charge of the research which should result in a ’National bookkeeping’6. The establisment of this Commission should be considered as the starting point for the development of a regular system of national accounts at the CBS. For the first time, the construction of a national bookkeeping was explicitly formulated as a purpose on its own. The background of the members of the Commission, their job description and the sequence of work done by the Commission all indicated that the concepts co-ordination1 and integration should play a key role in the compilation of national accounts. The Commission was made up of 6 members, the chairman was Idenburg, Director-General of the CBS, and amongst the members were Tinbergen and Derksen (secretary). The Commission met irregularly, and this is probably owing to the state of occupation. In her first meeting, the Commission formulated the importance of studies on national bookkeeping. Especially after the war, there would be an urgent need of such general statistical surveys. In the first place, on behalf of the necessary reconstruction of the country. In the second place, for economic policy purposes whereby an indicative economic planning system would be necessary for some years to come. In addition to an extension and deepening of detailed statistical information, an overall view of the economy would be required. Such overviews would also serve to inform the business world and the educated people in general, as their support for the many radical changes required would be indispensable. The Commission’s starting point was the working paper ’Schema’s voor een nationale boekhouding’ (Outlines for a national bookkeeping). This paper [28],

74

G.P. den Bakker/Dutch National Accounts

which was never published, has been written in the Division for Business Cycle Research and Mathematical Statistics. The paper discussed the purpose and usefulness of a national bookkeeping and as well the theoretical and statistical problems in compiling these figures presented examples of tables and a research programme that encompassed the whole statistical office. Some passages in the minutes of the first meeting of the Commission are worth mentioning. The Commission underlined the importance of the work on national accounting. However, one of the members noticed that the national accounts should be considered as a ’luxury statistic’ and that the ordinary statistics should not be discriminated against. The Commission concluded that those statistics which the CBS was obliged to compile (e.g. statistics on food supply) or which would be necessary for the national bookkeeping should have priority. But with regard to less important statistics (also in view of the paper shortage at the time!), persons could be transferred to work on national accounts. In April 1949, the Commission presented an outline to estimate national wealth. At the beginning, national balance sheets were an integral part of the Dutch system on national accounts. In the 1950s, the importance of national wealth estimates for plausibility checks of national accounting data was mentioned in a memorandum of the Division of National Accounts and Statistical Analysis. Experiences had shown that separate estimates of changes in assets and liabilities, adjusted for the effects of price changes, provide useful internal checks of the national accounting estimates. However, over time, national wealth disappeared from the statistical programme. The lack of separate data on the stock of capital goods by type and by major branches of economic activity was an important reason for this. Until now, no overall estimates of national wealth are available8. Capital stock figures are used in the perpetual inventory method to estimate fixed capital consumption. The Commission mentioned two methods to estimate national wealth. The first one estimated national wealth as the total value of all physical capital (e.g. buildings and equipment). This value should be equal to the replacement value less fixed capital consumption. The second method used the total value of claims on national wealth (e.g. money, shares and bonds). The first estimates for 1938 arrived at 28,500 million guilders. The Commission organized the research as follows. Working groups were set up to study different aspects. The groups were supposed to survey available data sources, such as production statistics, foreign trade statistics and accident statistics. A very useful data source were the figures gathered by ’rijksbureaus’ (government offices). These data became available as a consequence of the war as they were necessary for rationing purposes. They concerned stocks, production, intermediate consumption, primary costs and imports and exports. The lack of statistical information in those days forced the statisticians to develop methods to make up for these gaps when compiling national accounts. Gradually the idea of an input-output approach arose. This means that for every industry information was needed about: - The value of gross production, intermediate consumption and value added

G.P. den Bakker/Dutch National Accounts

75

components; - Which goods were produced and which goods were used up (commodity flow!). In addition, specific problems were studied: e.g. imputations for owner-occupied dwellings, government production, sector classification, military expenditures, the treatment of banking services, gains and losses on stocks, valuation of gross production (inclusive or exclusive of transport margins). All estimates related to the year 1938. In 1944, CBS was informed that in England important work had also been done in the field of national accounting, and that this was published in a Budget White Paper ’An Analysis of the Sources of War Finance and estimates of the National Income and Expenditure in the years 1938 to 1944’. A Commission of the Dutch government in exile in London recommended similar studies for the Netherlands after the liberation. The 1944 Annual Report of the CBS proudly mentiones that much work had in fact already been done. This was evidenced by the paper ’Nationale boekhouding, doeleinden, problemen en resultaten’ (National Bookkeeping, objectives, problems and results) [30]9. After the liberation, the work on national accounting was swiftly resumed, but the effects of the final war years on the statistical organisation had been so devastating that it was not until the end of 1948 before a reliable set of national accounts for a post-war year (1946) could be completed. A brief survey of these results was published in April 1949 [32], Detailed figures, together with revised accounts for 1938 were published in 1950. The latter publication is the subject of the next section. 2.5. A True System o f National Accounts (1938, 1946 and 1947) The first phase in the development of the Dutch national accounts ended in 1950 when accounts for 1938, 1946 and 1947 were published. This publication [26] was the crown of a development of about fifteen years. In addition to detailed figures, a description was presented of data sources, the theoretical framework and the estimation methods. For the first time, the estimates were fully based on the method of net production. A complete system of national accounts was presented. In this system six sectors were distinguished: enterprises, credit institutions, insurance companies and pension funds, government, households and the rest of the world. Notably, in the present national accounts the same sectors are distinguished; only now the sector social security funds is also distinguished separately. The accounts of the sectors were presented as T-accounts. The enterprise sector was the heart of the system. For twenty-five industries value added had been calculated by subtracting intermediate consumption from gross production. This was carried out within an input-output tables framework. The input-output table for 1938 was the first one ever compiled for the Netherlands. With this, the foundation was laid for the Dutch ’input-output tradition’: complete input-output tables are compiled on an annual basis and even the quarterly accounts

76

G.P. den Bakker/Dutch National Accounts

and the quarterly flash are based on the input-output method. The 1938 input-output table was not published until 1984 [33]10. Remarkably, the publication presents clear indications of the reliability of the estimates. This was one of the rare occasions that such indication have been presented. For every item of the sector accounts the uncertainty margin was given. This margin ranged from 2-5% (reliable estimate) to > 20% (crude estimate). The CBS reporting policy concerning the reliability of national accounting figures changed substantially in the course of time. The start was very promising: in the first national accounts publication [26], for every item of the sector accounts an uncertainty margin was indicated. Later on, the information became vaguer and qualitative and at present hardly any information in this respect is given. It is self-evident, that this does not mean that the reliabily as such has been worsened. At the end of the 1940s, the statistics underlying the national accounting figures were a rather chaotic collection of uncoordinated individual statistics [25]. There was no system of economic statistics. On the other hand, there existed no national accounting system to provide the basic conceptual framework. The emergence of the latter provided, for the first time, a systematic framework that could be used to judge the consistency, comprehensiveness and co-ordination of the economic statistics. In the Netherlands, the national accounts have had a strong impact on economic statistics because almost all statistical work was and still is organized in a single, centralized office whereby the national accounts are compiled in that same office.

3. Elaborations and Improvement of the System 3.1. A Review o f the Developments after 1950 After the breath-taking development of national accounts in the 1930s and 1940s, an era of elaboration and quality improvement followed11. Besides, Dutch national accountants actively participated in the development of the international guidelines on national accounting (see section 5.2). This era bears the stamp of Dr. C.A. Oomens, who was head of the National accounts division, director of Co-ordination and director of Economic Statistics successively. More than anyone else he has long been the driving force behind the Dutch national accounts and their influence on the development of the system of economic statistics. Oomens joined the CBS in 1936 where Tinbergen and Derksen were his masters. Under the guidance of Oomens, progress was made in the field of: input-output systems, financial accounts, regionalization of national accounts, calculation of real national income and product, estimation of national wealth, integration and co-ordination, realisation of the central business register and modernization of the European System of Integrated Economic Accounts. As often, history repeats itself: as in the 1930s, in the 1950s model-builders asked

G.P. den Bakker/Dutch National Accounts

77

again for adequate data which were not available. In 1953, Tinbergen (at that time director of the Central Planning Bureau) and Idenburg (Director-General of the CBS) took the initiative for a large-scale project: An Econometric Analysis of the Netherlands Economy. The Netherlands Economic Institute in Rotterdam and the Mathematical Center in Amsterdam participated. The project was financed by ’ZW O’, the Dutch acronym for ’Zuiver Wetenschappelijk Onderzoek’ (Pure Scientific Research). A first stage of the project was the construction of an adequate data base for the interwar period. A considerable number of series were compiled which have been used in the construction of the econometric models of the Central Planning Bureau until the mid 1970s. Informally, the Central Planning Bureau also provided a part of the series to academic researchers. But the data set was not published as such. Only recently, in the revision of the interwar period, national accounting data for the interwar period have been published. At the Dutch CBS, statistical co-ordination was considerd to be a sine qua non for integration of basic statistics into a national accounts framework. This is illustrated in a publication about the organization and activities of the CBS [35]. The part about the work of the statistical departments starts with the subject statistical co-ordination. The following citation tells this story: ’The need for basic statistics which - even if collected from separate enquiries and with unequal frequencies - are comparable so that they can be integrated in a system and used together in various forms of analysis, is steadily growing. This requires more attention to this subject from the statisticians. Co-ordination in this sense concerns the choice of definitions, of statistical units, of statistical classifications and the contents of questionnaires. As for definitions, international recommendations are to be followed as far as possible, in particular those of the United Nations System of National Accounts and its Common-Market version ’Systeme Europeen de Comptes Economiques Integres’. The fact that international recommendations are made by more than one international agency and that they are not always compatible, poses some problems. Another difficulty is that the broad definitions of the international agencies must be converted into more specific texts which, if applied in many departments of the Bureau, result in full comparability of the data obtained on that basis. As to the choice of statistical units it must be considered essential to limit the number of units to be used in statistical enquiries. In certain cases this may mean a less ideal adaptation to specific purposes, but in general this loss can be accepted in exchange for greater comparability of the statistics. Here the main problem is the choice of units in the business sector of the economy. The Bureau has chosen two basic units: a) the kind-of-activity-unit, to be used is statistics relating to the production process and to investment; b) the institutional unit, viz. the unit which receives and disposes of income, owns and accumulates property, lends or borrows, for financial statistics. For the first of these units a regional classification is added. Also for classifications

78

G.P. den Bakker/Dutch National Accounts

(of activities, of functions, of goods and services) international recommendations exist and sometimes differ. In setting up its classifications the Bureau tries to construct them in such a way that they fit into international classifications, serve national needs and are still feasible. Co-ordination as described here requires three phases that can be described as: definition, execution and maintenance. In the first phase the ultimate goal must be described. In the second phase the existing statistics are adapted or amplified in order to become part of the co-ordinated system. The third phase starts immediately after that: the organization of the work must be such, that the system, with all the changes that characterize the economic process, remains a co-ordinated one. Mainly in view of the third phase it is considered essential that the way in which decisions about the delineation of the statistical units and about the classification of these units are made, is recorded in registers for each of the two units.’ After the first modern national accounts publication in April 1950, national accounting figures for 1948 (including an input-output table) and provisional ones for 1949 and 1950 were published in the quarterly journal ’Statistische en Econometrische Onderzoekingen’ (Statistical and Econometric Studies) [36]. Owing to a lack of space only the most important tables were presented! Partly as a followup of discussions between the CBS and the Central Planning Bureau aimed at harmonizing the definitions, there were some changes compared with the preceding publication, such as: - The state and subsidized education was transferred from the enterprises to the sector government; - Expenditure on maintenance (to maintain the assets in their original condition) by enterprises was booked as fixed capital formation. National accounting figures for subsequent years were published in various volumes of ’Statistical and Econometric Studies’. A regular, yearly publication of national accounts started in September 1953 [37] with figures for 1938 and 1946-50 and provisional ones for 1951 and 195212. In the next publication [38], provisional 1953 figures were published, together with revised ones for 1938 and 1946-52. This revision encompassed both new estimates (for instance, for the sector insurance companies better basic data were available) and changes in definitions. For example, taxes were no longer measured on the basis of tax assessments but reflecting governmentally recovered amounts. In the introduction to the publication it was mentioned that the development of the system of national accounts had reached a phase, both nationally and internationally, in which no great changes were to be expected in the coming years. Explicitly, the CBS revision policy was mentioned: comparability over time of the figures has a higher priority than the accuracy of the level estimates. The system was not fully in accordance with the 1953 SNA but the intention to adapt to the SNA was stated. In the National Accounts 1954 [39], a revision of the figures from 1948 onwards was announced. The purpose was to improve both the comparability over time and the usefulness of the data. In this publication, for the first time data at constant

G.P. den Bakker/Dutch National Accounts

79

prices (of 1949) for the macro-economic aggregates were given13. Also for the first time, long time series of national accounting figures, from 1900 onwards, were published. The number of variables was limited to volume index numbers of national income and per capita national income. In CBS [39], corrections of earlier published figures were carried out. These corrections concerned, amongst others, recalculations of fixed capital formation and the change in stocks. In this publication, the royalties etc. received from the rest of the world were booked as exports of services (so far, they had been booked as income from abroad). In 1958 [43] a complete revision of the national accounts for the years 1948-57 was published. The years 1946 and 1947 were left aside because they were too much influenced by the war. Therefore it was considered that these years should be deleted from a time-series for use in economic model-building. The explicit purpose of the revisison was to obtain time-series which were completely comparable as regards accuracy, definitions and classifications. New basic statistics made it possible to improve upon the estimates. The international definitions as given in the publications of the United Nations [44] and the OEEC [5] were now completely implemented. Finally, new and more detailed classifications were worked out. These were needed for purposes of domestic economic policy and for submission to international agencies. The most important changes in definitions were: - general government expenditure was broken down into consumption and fixed capital formation for civilian purposes. Government expenditures for military purposes were considered as consumption expenditure. Corresponding estimates were made for fused capital consumption and imputed rent on civil government buildings; - the increase of work in progress in construction was regarded as fixed capital formation instead of as an increase in stocks. On the other hand changes in the value of livestock were included in the increase in stocks; - before the revision all social security funds, pension funds and life-insurance companies were treated as one sector. In the revision, a separate sector ’social insurance’ was introduced as part of the general government. A new sector was created for life insurance and pension funds; - the interest on public debt was not added anymore to national income; - a seperate sector for credit institutions was distinguished. The 1948-57 revision resulted in higher growth rates of national income and final private consumption expenditure. In the National Accounts 1959 more figures on deflators and volume indices were presented and employment figures, were given too. Unrequited transfers to and from abroad were moved from the primary to the secondary income distribution account. In the National Accounts 1961, unemployment figures were presented for the first time. They were, together with employment figures, estimated within a national accounts framework and were fully consistent with the national accounting data. The

80

G.P. den Bakker/Dutch National Accounts

publication of these data stopped after the National Accounts 1979; from that year on, only aggregate employment figures by industry were estimated within a national accounts framework. The reason for this is that in 1980 the adequacy of the data sources was doubted and moreover the assumption of a simple dichotomy of the labour force, into working and non-working participants, was no longer considered sufficiently realistic, due to emergence of many part-time jobs. In the National Accounts 1971 the Dutch Standard Classification of all Industrial Activities was introduced. In the National Accounts 1972, the 1968 SNA [45] definitions were adopted. In the National Accounts 1974 the input-output table for 1972 was published as an annex to the publication14. From the National Accounts 1977 on, regional figures were published. Earlier, regional accounts for the provinces and a few other regions were prepared for the year 1960 by splitting up the totals of the national input-output table into regional input-output tables. For this purpose an intensive use was made of the available basic statistics, supplemented by additional information and special estimates. Five-yearly regional input-output tables were published between 1960 and 1975. The compilation stopped because it absorbed too many resources. The National Accounts 1980 presented the 1977 revision. This revision was needed because the years prior to 1977 were ones of major extensions and improvements to the statistics on which the national accounts are based. The extensions primarily involved a number of new statistics in services (e.g. medical services, business services, hotels, cafs and restaurants), while improvements particularly concerned a far better co-ordination of existing statistics (for instance, the general implementation of the Dutch Standard Classification of all Industrial Activities). The 1977 base year revision involved: - a revision of the data in the light of the availability of many new sources; - improvement of a number of methods, including a more consistent treatment of a number of goods and services transactions; - a revision of data in the light of changes in the classification of statistical units by economic class. The 1977 revision resulted in an increase of gross domestic product at market prices by 13,500 million guilders, or 3.5%. The most extensive adjustments were for services. In 1985 [46], revised time series for 1969-76 were published, fully comparable with the revised estimates for 1977. In the National Accounts 1981 the number of published industries expanded from 35 to 57. In the National Accounts 1983 the publication of an experimental inputoutput table for 1981 in prices of 1980 was mentioned. In the National Accounts 1984 consistent figures were given from 1970 onwards. The revised input-output tables for these years then also became available on machine-readible tape. From this publication on, price and volume mutations are published in the forms of chain indices. In 1990, an extensive description of methods in English language became available [47], From the very beginning, input-output tables played an important role in the

G.P. den Bakker/Dutch National Accounts

81

Dutch national income estimates. The Netherlands were one of the first countries in the world in which routinely, yearly input-output tables were compiled as an integral part of the national accounts. In the early 1980s, the method of compiling inputoutput tables was changed so that during the entire statistical process, from the processing and analysis of the basic data up to and including the phase of balancing the tables, data in current prices and deflated data are gathered and analyzed simultaneously. This leads to a substantial improvement of both the figures in current prices and regarding the volume and price developments. Several years ago, the CBS embarked on a project dealing with the compilation of long, consistent time series of national accounting data. In this project, each period is tackled separately, the results being published whenever they become available. The first period to be revised was the interwar period. One of the reasons to choose this period was that in the second half of the 1980s the CBS still employed people who, many years ago, had compiled the original figures for that period. In the revision of the 1921-39 data, the concepts and methods that are in current use have been used as far as possible. Moreover, additional series have been compiled in order to be able to draw-up the SNA ’Consolidated accounts for the nation’. One example of a change in method is that the national income is fully compiled by means of the net production method. Two examples of conceptual changes are: a recalculation of consumption of fixed capital by means of the perpetual inventory method and a recalculation of yearly price and volume indices by means of a Paasche-Laspeyres pair. For 1938, relatively much statistical information is available, which made it possible to compile a Social Accounting Matrix [48]. The first results of this revision have already been published [49]. In 1992, the results of the most recent revision have been published. This revision with 1987 as benchmark year is characterized by a greater emphasis on the institutional aspects of the system, particularly in the recording of goods and service transactions and in the generation of income. Starting with this revision, aggregation and integration of the basic data will be based on the supply and use matrices, estimated both in constant prices and in current prices. For example, this allows for a better link to basic statistics and for a more sophisticated comparison and mutual balancing of data from different sources. Over the period 1978-1987 a number of new sources have become available and some existing sources have become more complete. For instance, in this revision for the first time direct use is made of data from CBS surveys on household consumption expenditure and gross fixed capital formation. It is also the first occasion in which the production statistics for trade have been utilized. With respect to income transactions, considerably more information has become available since the previous revision. This particularly concerns the Business Finance Statistics compiled over the years 1977-1985. This statistic provides detailed data on the profit and loss accounts and the balance sheets of non-financial enterprises. Independent income estimates were arrived at by combining these data with information on sole proprietorships and partnerships obtained from production

82

G.P. den Bakker/Dutch National Accounts

statistics. In fact, this meant a partial rehabilitation of the income method in compiling the Dutch national accounts. The 1987 revision has been the occasion for a critical examination of all kinds of existing and recording methods which led to a range of adjustments. In the course of time, several major revisions (see table 4 for some quantitative results) have been carried out in the Netherlands. These revisions are, in chronological order of publication: - Revision 1948-57, comparable data for 1948-57, published in 1958; - Revision 1968 SNA, comparable data for 1960-72, published in 1973; - Revision 1977, comparable data for 1977-80, published in 1981; - Revision 1921-39, comparable data for 1921-39, published in 198715; - Revision 1987, comparable data for 1987-91, published in 1992. In the course of time, several studies have been carried out to the magnitude of the hidden economy. For instance, Begeer and Van Tuinen [50] reported that the black economy is 10 to 15% of the national product. However, more than half of it is covered by the official national accounting figures.

Table 4. The e ffe c t of revisions of n atio nal accounting fig ure s Gross Domestic Product at market prices

Net National Income at market prices

Before revision

Before rev ision

A fter revison

mln gld Revision 1948-1954 1948 1954

%

After revison

mln gld

%

15440 26030

15013 26738

-2.8 +2.7

14240 24310

13535 24657

-5.0 +1.4

1968 SNA revision 1969

103359

101715

-1.6

95289

93797

-1.6

Revision 1977 1977

261410

274930

+5.2

237760

251100

+5.6

6236 5624

5679 5446

-9.0 -3.2

5780 5395

5777 5399

-0.1 +0.1

430170

440580

+2.4

383960

390890

+1.8

Revision interwar period 1921 1938 Revision 1987 1987

G.P. den Bakker/Dutch National Accounts

83

3.2. The 1980s: Towards a Flexible and Institutional Approach In the Netherlands, the idea that a system of national accounts serves various purposes has a long tradition. Already in the working paper in which the framework for the Dutch national accounts was formulated (written in 1942, see section 3.4), the need for flexibiltiy was emphasized (p. 7): the structure of the national bookkeeping will entirely be determined by the needs. It is to be expected that in such a system changes have to be made continually. In the Netherlands, much attention has been paid to the structure of the next, 1993 SNA. Since the first proposal in this field [51] many studies16 have been published, which led gradually into a line of thought, sometimes referred to as the Dutch school [64], The theoretical foundation of the Dutch view is presented in [65]. They stated that the many purposes of the SNA frequently conflict with one another and concluded that an alternative structure of the system is necessary in order to achieve greater flexibility. They recommend a SNA with a general purpose core supplemented by special purpose modules. The core is to be a complete, consistent, integrated and detailed system of national accounts with a greater institutional content than the 1968 SNA and a more elaborate description of the economy at the meso-level. The autors mentioned three principles the core has to fulfill (Van Bochove and Van Tuinen, p. 140-143): - The intersection principle. This principle consists of two parts. In the first place the core should contain as few special purpose elements as possible. Secondly, the core should contain the essential structural components that are the basis for the construction of major alternative descriptions of the economy as a whole; - The parsimony principle: The core should, as far as possible, reflect the economic agents’ perceptions of themselves and of their transactions. Its concepts should be free of the influence of hypotheses that are based on theoretical analysis instead of subjective experience; - The consistency principle: The core has to provide a consistent, self-contained system of national accounts, just as the 1968 SNA is. The modules are more analytic and reflect special purposes (e.g. imputations, attributions and re-classifications) and specific theoretical views, while maintaining an explicit link with the core. Unlike the 1968 SNA, the core is not an attempt to meet all demands on national accounting, but the core opts for a number of clear-cut choices, e.g. with respect to the production boundary, consumption by the payee and so on. This means that the core as such is insufficient to satisfy all major purposes of national accounts. In turn, this implies that some of the modules are not optional: a number of modules should be a standard part of the system, in addition to the core. In Gorter [57] a systematic review of imputations to be removed from the core system was presented. The idea of a core-module system is further given shape in Gorter and Van der Laan [61]. In addition to the core they distinguished five types of modules. Some of the tables mentioned under the supporting modules are

84

G.P. den Bakker/Dutch National Accounts

mandatory (e.g. data at constant prices). Their system is shown in figure 1. Nowadays, environmental issues are in the limelight17. In order to analyse the environmental problems and to formulate policies there is an urgent need on statistical information in this field, especially concerning the impact of environmental phenomena on the economic system and vice versa. At a very early stage, Dr. R. Hueting, working on environmental statistics in the CBS, stressed the importance of connecting these statistics with the national accounts [67,68]. An alternative approach consists of an environmental module in physical terms which is linked to the core national accounts, so that a complete and systematic account of all changes caused by production and consumption processes can be given. This module is therefore not the outcome of a model, but a statistic [60,69]. Recently, the CBS has started a pilot project to fill in such an environmmental module with actual data.

Fig. 1. A System of Economy-related Statistics. Source: G orter and Van der Laan ([61], table 4).

Kazemier and Exel [63] present a supplementary module on non-market production, thus incorporating non-market production within the concepts of the SNA. The relation between market and non-market production is shown in the framework of a Social Accounting Matrix. They add the informal sector to a SAM of the formal economy. This extended SAM can be considered as a prototype of a SNA-module on total production. The module on total production can serve many purposes. For example, it can be used to study the effects of increased female labour force participation on consumption and national income. A study on theoretical and statistical problems with respect to a research and development module [62] and a human capital module is carried out too. Anticipating on the realization of a full-fledged core and module system, the 1987 revision stressed the institutional aspects of the system. Before the revision, trade was described from a functional viewpoint. This means that the data related to all

G.P. den Bakker/Dutch National Accounts

85

trade activities (main activities and side activities), no matter in which (major)group the units were classified. After the revision, the trade estimates relate to units which are classified according to their main activity. In the publication ’National accounts 1991’ [70], a detailed National Accounting Matrix for 1989 and aggregated ones for 1987-89 are presented. This can be seen as the first step towards the publication of a complete Social Accounting Matrix (SAM) and to the use of a SAM as a framework in which finally all three methods to compile national accounts, the income method, the production method and the expenditure method, are integrated at a meso-level. 4. Summary and Conclusions 4.1. Development o f Dutch National Accounts closely related to Economic Policy and Model-building In the course of time, the development of national accounts has been stimulated enormously by model-building. The Dutch national accounts originated from the need to provide data for the Tinbergen model. In the 1950s, the model-building efforts of the Central Planning Bureau again required new national accounting data. National accounts are now widely used in economic analysis and for the formulation of economic policy. The Central Planning Bureau uses national accounts as a basis for its work and the preparation of the annual ’Central Economic Plan’(the official economic forecasts for the Netherlands). Therefore, a close co-operation is maintained between the Central Planning Bureau and the CBS, e.g. to ensure consistency in the definitions used by the two Bureaus. The national accounts are also used by the Ministry of Finance. For instance, in the government budget national accounts classifications applied. In the Netherlands, national accounting figures play an important role to set norms for and to evaluate the outcomes of economic policy. For instance, the magnitude of Dutch foreign aid is tied to net national income. The target government deficit is expressed as a percentage of national income. As mentioned before, the national accounts are now being developed into an institutional system. In this way, the national accounts serve as an organizing framework for a system of basic statistics. This system of economic statistics provides data on production, income distribution, financing and expenditure and their international and intertemporal relations. In the course of time, the national accounts have also served as a framework to structure the division of work among basic statistics. The organization of the work at the CBS and the division of the Directorate of Economic Statistics in departments is modelled to the types of basic statistics used in the National Accounts. In the Netherlands, the national accounts have had a strong impact on economic statistics because almost all statistical work was and is still organized in a single, centralized office whereby the national accounts are compiled in that same office.

86

G.P. den Bakker/Dutch National Accounts

In compiling the national accounts figures, detailed cross-checking of the data are carried out in an input-output framework. This means that the final national income estimate is not just an average of two or three methods. At a low level of aggregation data are compared and plausibility checks are carried out, which lead to corrections. The national accounts and its figures, compiled by the Dutch CBS are very authoritative. It happens very rarely that competitive estimates are made [71].

4.2. The Dutch Contribution to National Accounting The development of the Dutch national accounting system started in the 1930s and 1940s. From that time on, Dutch statisticians have contributed to the international development of national accounting systems. Rostow [72] mentions the importance of the Dutch contribution to the development of national accounting. In the chapter on the emergence of national income accounting he states: ’In the course of the 1920s the statisticians, conscious of their growing skills and the proliferation of government statistical series, conscious also that new and puzzling problems were shaking the complacency of the mainstream economists, unfurled their flags and asserted with confidence a major role in the further development of the social sciences. One could quote various Americans and Britons to this effect (as well as Norwegian, Dutch, and other econometricians)...’ This might be seen as an underestimation of the contribution of Tinbergen and Derksen in those years. For example, Arvay [73] notices: ’... the economic theory of Keynes and Tinbergen developed in the 1930s which served as a fundamental basis for the SNA-type national accounting system.’ In December 1945, Derksen represented the Netherlands at a meeting (Princeton, NJ.) of the subcommission on national income of the League of Nations Commission of statistical experts. There was much discussion on international guidelines. The national bookkeeping system, as had been developed in the Netherlands, basically complied with these guidelines. In 1952, the OEEC ’Standardised System of National Accounts’ [5] was published. Dr. G. Stuvel was closely involved in drafting this report. Stuvel was one of Tinbergen’s assistants and was later in charge of the national accounts work at the OEEC in Paris. The 1953 SNA reflects the development of the national income accounting framework in a number of countries during and after World War II. Ruggles and Ruggles [74] mention four countries: the United Kingdom, the United States, Norway and the Netherlands. Prof. dr. D.B.J. Derksen and Dr. C.A. Oomens played an important role in the realization of international standard conventions. The former became head of the national income unit of the United Nations. In the mid 1960s, Dr. C.A. Oomens was member of the Expert Group, convened by the United Nations Secretary-General to assist and advise in drafting the 1968 SNA. Derksen and Oomens also advised other

G.P. den Bakker/Dutch National Accounts

87

countries on the development of national accounts, the former, for instance, advised together with Kuznets and Stone the Indian National Income Committee [75] and the latter contributed to the development of the Mexican national acounts. In the forthcoming revised SNA, several ideas which were promoted by Dutch national accountants, will be incorporated (setting aside imputations, three-dimensional value-added table, applying chain indices, matrix presentation, institutional sector classification, meso approach, etc.). In addition, the chapter on Social Accounting Matrices [76] has been written by a Dutch national accountant. These matrices fit well in the flexible approach to national accounting as advocated by the Dutch school. The CBS is also closely involved in the realization of the ongoing revision of the European System of Integrated Economic Accounts. For this purpose, one of its staff members is seconded to Eurostat.

Endnotes 1. These studies have been m entioned before in articles in the Dutch language [2], 2. This implies that Studenski’s tables 9-2 and 10-5, on p. 141 and 156 respectively, have to be modified. Studenski already m entioned that it was likely that Dutch national income estimates were made before 1910 (p. 145, 537). 3. Remarkably, M etelerkam p estim ated national wealth fourteen times as high as national income. This is much higher than, for instance, in the 1940s (the form er being five times as high as the latter). Probably, different definitions and estimation methods play a role here. 4. Van Cleeff (not employed at the CBS) deserves the credit of having introduced this term. 5. In, for instance, the U nited States these services were excluded [17,18]. 6. The term ’National accounts’ was only used at a later stage. 7. A deep discussion ’avant la lettre’ on the theory and need of co-ordination on behalf of a statistical description of the ’economic blood circulation’ was already given in Roet [27]. 8. However, in this field work is in progress at CBS [29]. 9. This paper was not published until 1978 [31]. 10. In the recent revision of the national accounts for the interwar period this table has been revised (see section 4.1). 11. This paper presents the development of the definitive national accounts. Alongside with this development short term statistics have played an important role in the Dutch national accounts [34]. In the national accounting framework, short term statistics cover annual (provisional and revised provisional), quarterly and monthly figures. Already in the interwar period, CBS presented short term statistics (an economic barom eter). Starting from the reference year 1977 quarterly accounts were published, based on complete quarterly input-output tables are compiled. Since 1985, a monthly business cycle report is published. 12. This pattern continues up to the present times: final estimates are published two and a half year after the end of the reference year. 13. Dutch national accountants have always been interested in the deflation of national accounting figures [40]. In the 1980s, an in-depth study has been carried out related to the objective to compile a complete input-output table in prices of the previous year, see Al, Balk, D e Boer and Den Bakker [41], The authors are in favour of using chain indices with the Laspeyres index for volume developments and the Paasche one for price movements. In a study by Den Bakker [42] the sensitivity of long term growth rates to variations in index number formulae and weights was analysed.

88

G.P. den Bakker/Dutch National Accounts

14. In addition, input-ouput tables and input-output coefficients are published in a separate publication ’The production structure of the Netherlands’ economy’. 15. This concerns macro-economic data. Detailed figures may be published in 1994. 16. See e.g.: A1 [52], Van Bochove and Bloem [53], A1 and Van Bochove [54], Bloem [55], Van Bochove [25], Van den Bos [56], G orter [57], Bloem [58], Keuning [59], D e Boo, Bosch, G orter and Keuning [60], G orter and Van der Laan [61], Bos, Hollanders and Keuning [62], Kazemier and Exel [63]. 17. However, years ago, national accountants were already aware of these issues. For instance, Derksen [66] wrote: ’As in other countries, a discussion has developed on the need for changing the concept of national income by making allowances for the decline in ’social welfare’ caused by air pollution, noise caused by m otor vehicles and airplanes, etc. The Bureau shares the views of those who claim that national income is not and never was intended to be a comprehensive measure of social wellbeing and social welfare, though it is closely related to these concepts.’

References [1]

L. Nooteboom (1978) Veertig jaren Nationale rekeningen (Forty years of National Accounting).

[2]

C A . Oom ens (1985) De ontwikkeling van de Nationale Rekeningen in Nederland (The

[3]

P. Studenski (1958) The Income of Nations. New York University Press, New York.

[4]

O E E C (1951) A simplified system of national accounts. OEEC, Paris.

In: Denken en m eten - statistische opstellen. Staatsuitgeverij, ’s-Gravenhage, pp. 1-17. development of the National Accounts in the Netherlands). Statistiscli Magazine 2 5-15.

[5]

O E E C (1952) A standardized system of national accounts. OEEC, Paris.

[6]

W.M. Keuchenius (1803) D e Inkomsten en Uitgaven der Bataafsche Republiek, voorgesteld in

[7]

R. M etelerkam p (1804) De toestand van Nederland in vergelijking gebragt m et die van eenige

eene Nationaale Balans (Income and Outlay of the Batavian Republic). W. Holtrop, Amsterdam. andere landen van E uropa (The state of the Netherlands in comparison with other European countries). Cornel en Van Baalen, Rotterdam. [8]

G. King (1936) Two tracts. The John Hopkins Press, Baltimore.

[9]

W A . Bonger (1910) Vermogen en inkomen in Nederland (W ealth and income in the Netherlands). De Nieuwe Tijd, March 1910.

[10] CBS (1933) Onderzoek naar den invloed van het arbeidsloon op den kostprijs der producten (An inquiry to the influence of wages on the cost price of products). Algemeene Landsdrukkerij, ’s-Gravenhage, part II. [11] O. Bakker (1939) H et nationale inkomen en de berekening ervan (National Income and its Estimation). De Economist 3 (1939) 220-234. [12] Ed. van Cleeff (1941) Nationale boekhouding: Proeve van een jaaroverzicht Nederland 1938 (National bookkeeping: an A ttem pt at an Annual Survey Economist 7 /8 (1941) 415-424.

for the Netherlands, 1938). De

[13] Ed. van Cleeff (1941) Beteekenis en inrichting eener nationale boekhouding. (On the Significance and the Organization of a National Book-keeping System). De Economist (1941) 608-623. [14] Ed. van Cleeff (1945) Theoretische planhuishoudkunde - Hoofdlijnen en problem en van een leer der planmatig geleide economische orde (Theoretical Planeconomy). De Economist, N ovem ber/ Decem ber (1945). [15] J.B.D. Derksen (1946) A System of National Book-keeping, illustrated by the experience of the Netherlands economy, Occasional Paper X, National Institute of Economic and Social Rescearch. Cambridge University Press, London. [16] J. Tinbergen (1936) Kan hier te lande, al dan niet na Overheidsingrijpen, een verbetering van de binnenlandse conjunctuur intreden, ook zonder verbetering van onze exportpositie? Welke lering

G.P. den Bakker/Dutch National Accounts

89

kan ten aanzien van dit vraagstuk worden getrokken uit de ervaringen van andere landen? (Is a Recovery in the Domestic Economic Situation of this Country Possible, with or without Action on the Part of the Government, even without an Improvement in our Export Position? W hat Can Be Learned about this Problem from the Experience of other Countries?). Pre-adviezen voor de Vereeniging voor de Staathuishoudkunde en de Statistiek. Martinus Nijhoff, ’s-Gravenhage, pp. 62-108. [17J CBS (1939) Enkele berekeningen over het nationale inkomen van Nederland (Some Computations of the National Income of the Netherlands). Speciale Onderzoekingen van de Nederlandsche Conjunctuur, no. 2. Centraal Bureau voor de Statistiek, ’s-Gravenhage. [18] C.S. Carson (1975) The History of the United States National Income and Product Accounts: the Development of an Analytical Tool. Review of Income and Wealth, June 1975 pp. 153-181. [19] J.B.D. Derksen (1940) Eenige recente gegevens over het nationale inkomen van Nederland (Some recent estimates of Dutch national income). Maandschrift Centraal Bureau voor de Statistiek (1940) 701-707. [20] J.B.D. Derksen (1940) Berekening van het nationaal inkomen volgens de objectieve methode (Estim ation of national income according to the production method). Maandschrift Centraal Bureau voor de Statistiek (1940) 837-838. [21] J.B.D. Derksen (1940) H et onderzoek van het Nationale inkomen (An Inquiry to National Income). De Economist (1940) 571-594. [22] J.B.D. Derksen (1941) Berekeningen over het nationale inkomen van Nederland voor de periode 1900-1920 (Com putations of the National Income of the Netherlands for the period 1900-1920). Speciale Onderzoekingen van de Nederlandse Conjunctuur, 1941, no. 4. [23] CBS (1948) H et nationale inkomen van Nederland 1921-1939 (National income for the Netherlands 1921-1939). No. 7 der Monografien van de Nederlandse Conjunctuur. Uitgeversmaatschappij W. de H aan N.V., Utrecht. [24] R. Stone (1947) Definition and M easurem ent of the National Income and Related Totals. M em orandum to the Sub-committee on National Income Statistics of the League of Nations Com mittee of Statistical Experts. In: M easurem ent of National Income and the Construction of Social Accounts. Studies and Reports on Statistical Methods, No. 7. United Nations, Geneva. [25] C A . van Bochove (1987) The Micro-Meso-Macro Linkage for Business in an SNA-Compatible System of Economic Statistics. National Accounts occasional Paper Nr. NA-020. Central Bureau of Statistics, Voorburg. [26] CBS (1950) D e Nationale Jaarrekeningen: doeleinden, problemen, resultaten (The National accounts: objectives, problems, results). No. 8 der Monografien van de Nederlandse Conjunctuur. Uitgeversmaatschappij W. de H aan N.V., Utrecht. [27] I. Roet (1936) Coordinatie van statistieken (Co-ordination of statistics). De Economist (1936) 377-399. [28] CBS (1942) Schema’s voor een nationale boekhouding (Outlines for a National Bookkeeping). Centraal Bureau voor de Statistiek, ’s-Gravenhage. Unpublished. [29] J.

Frenken (1992) How to

M easure Tangible Capital Stock?. Paper

presented at the

Twenty-Second G eneral Conference, August 30 - September 5. Flims, Switzerland. [30] J.B.D. Derksen (1978) Nationale boekhouding, doeleinden, problemen en resultaten (National bookkeeping, purposes, problems and results). In: CBS, 1978, Denken en m eten - statistische opstellen. Staatsuitgeverij, ’s-Gravenhage, 1949, pp.XIX-XXVI. [31] CBS (1978) Denken en m eten - statistische opstellen (Statistical Essays). Staatsuitgeverij, ’s-Gravenhage. [32] CBS

(1949)

National Accounts of the

Netherlands,

1946 and

1947. Special Statistical

Communications, No. 4001. Netherlands central Bureau of Statistics, ’s-Gravenhage. [33] C A . van Bochove and G.M. Zijlmans (1984) D e Nederlandse produktiestructuur in 1938 en 1981 (The production structure of the Netherlands’ economy in 1938 and 1981). De produktie-structuur van de Nederlandse volkshuishouding. Staatsuitgeverij/CBS-publikaties, ’s-Gravenhage, 1984, deel XII, pp.39-63.

90

G.P. den Bakker/Dutch National Accounts

[34] S.B. Algera and R .JA . Janssen (1991) The development of short-term macro-economic statistics. In: CBS Select 7 - Statistical Essays. SDU/Publishers/CBS-publications, The Hague, pp. 99-118. [35] CBS (1969) The Netherlands Central Bureau of Statistics - Organization, function and activities. Netherlands Central Bureau of Statistics, The Hague. [36] CBS (1951) De Nationale Jaarrekeningen 1948 (National Accounts 1948). Statistische en econometrische onderzoekingen (1951) 38-43. [37] CBS

(1953)

N ationale

Jaarrekeningen

1948-1950

(N ational

A ccounts

1948-1950).

Uitgeversmaatschappij W. de H aan N.V., Utrecht. [38] CBS (1954) Nationale rekeningen voor Nederland 1938 en 1946-1953 (National Accounts for the Netherlands 1938 and 1946-1953). Statistische en econometrische onderzoekingen, 2e kwartaal 1954. [39] CBS (1955) Nationale rekeningen voor Nederland 1938 en 1947-1954 (National Accounts for the Netherlands 1938 and 1947-1954). Statistische en econometrische onderzoekingen, 2e kwartaal 1955. [40] C A . Oomens (1949) H et reele nationale inkomen van Nederland (The evaluation of real national income in the Netherlands. Statistische en econometrische onderzoekingen 1 (1949) 14-25. [41] P.G. Al, B.M. Balk, S. de Boer and G.P. den Bakker (1986) The use of chain indices for deflating the national accounts. Statistical Journal o f the United Nations ECE 4 (1986) 347-368. [42] G ert P. den Bakker (1991) The Choice of Index Number Form ulae and Weights in the National Accounts - A sensitivity analysis based on macro-economic data for the interwar period. National Accounts Occasional Papers Nr. NA-044. Netherlands Central Bureau of Statistics, Voorburg. [43] CBS (1958) Nationale rekeningen 1948-1957 (National accounts of the Netherlands 1948-1957). Statistische en Econometrische Onderzoekingen, 2e en 3e kwartaal (1958) 101-118. [44] United Nations (1953) A Sytem of National Accounts and Supporting Tables. Studies in Methods, Series F, No. 2, Rev. 1. U nited Nations Publications, New York. [45] United Nations (1968) A Sytem of National Accounts. Studies in Methods, Series F, No. 2, Rev. 3. United Nations Publications, New York. [46] CBS (1985) Nationale rekeningen 1969-1981 met herziene reeksen voor de jaren 1969-1976 (National accounts 1969-1981 with revised series for the years 1969-1976. Staatsuitgeverij / CBS-publikaties, ’s-Gravenhage. [47] C.N. G orter et al. (1990) Methodological Description of the calculation of the National Income (gross, at m arket prices) of the Netherlands for 1986. BPA-no.: 9904-90-STD.E8/Intern. Central Bureau of Statistics, Voorburg. [48] G ert P. den Bakker, Jan de Gijt and Steven J. Keuning (1992) A Historical Social Accounting Matrix for the Netherlands (1938) - including a comparison with the 1987 distributions of income and outlay. National Accounts Occasional Paper, Nr. NA-55. Netherlands Central Bureau of Statistics, Voorburg. [49] G ert P. den Bakker, Theo A. Huitker and Cornelis A. van Bochove (1990) The Dutch Economy 1921-39: Revised Macroeconomic D ata for the Interwar Period. Review o f Income and Wealth 2 (1990) 187-206. [50] W. Begeer and H.K. van Tuinen (1985) Statistische operationalisering van het begrip informele economie (The concept unofficial economy put into statistical operation). Statistisch Magazine 1 (1985) 5-16. [51] R. van Eck, C.N. G orter and H.K. van Tuinen (1983) Flexibility in the system of National Accounts. National Accounts Occasional Paper, Nr.NA-001. Netherlands Central Bureau of Statistics, Voorburg. [52] P.G. Al (1985) Dual sectoring in National Accounts. National Accounts Occasional Paper Nr.NA-10. Netherlands Central Bureau of Statistics, Voorburg. [53] C A . Bochove and A.M. Bloem (1985) The structure of the next SNA: review of the basic options. National Accounts Occasional Paper Nr.NA-09. Netherlands Central Bureau of Statistics, Voorburg.

G.P. den Bakker/Dutch National Accounts

91

[54] P.G. A1 and C A . van Bochove (1986) A proposal for the synoptic structure of the next SNA. National Accounts Occasional Paper, Nr. NA-14. Netherlands Central Bureau of Statistics, Voorburg. [55] A.M. Bloem (1988) M icro-macro link for Government. Review o f Income and Wealth, (1988) 289-311. [56] C. van den Bos (1988) The institutional sector classification. National Accounts Occasional Papers, Nr. NA-28. Netherlands Central Bureau of Statistics, Voorburg. [57] C.N. G orter (1988) Im putations and re-routeings in the National Accounts. National Accounts Occasional Paper, Nr. NA-001. Netherlands Central Bureau of Statistics, Voorburg. [58] A.M. Bloem (1990) Units in National Accounts and the Basic System of Economic Statistics. Review o f Income and Wealth 3 (1990) 275-288. [59] S J . Keuning (1991) Proposal for a Social Accounting Matrix which Fits into the Next System of National Accounts. Economic Systems Research 3 (1991) 233-248. [60] A J . de Boo, P.R. Bosch, C.N. G orter and S J . Keuning (1991) An Environmental Module and the Com plete System of National Accounts. National Accounts Occasional Papers, Nr. NA-046. Netherlands Central Bureau of Statistics, Voorburg. [61] C. G orter and P. van der Laan (1992) An Economic Core System and the Socio-economic Accounts M odule for the Netherlands. Review of Income and Wealth 2 (1992) 199-223. [62] F. Bos, H.J.G.M. Hollanders and S.J. Keuning (1992) A Research and Development Module Supplementing the

National Accounts. National Accounts Occasional Paper, Nr.

NA-51.

Netherlands Central Bureau of Statistics, Voorburg. [63] B. Kazemier and J. Exel (1992) The Allocation of Time in the Netherlands in the Context of the SNA; a Module. National Accounts Occasional Paper, Nr. NA-52. Netherlands Central Bureau of Statistics, Voorburg. [64] U-P. Reich (1989) Essence and Appearance: Reflections on Input-Output Methodology in Terms of a Classical Paradigm. Economic Systems Research 4 (1989) 419. [65] C A . Bochove and H.K. van Tuinen (1986) Flexibility in the next SNA: the case for an institutional core. The Review o f Income and Wealth 2 (1986) 127-54. [66] J.B.D. Derksen (1973) Problems of determining and measuring the reliability of the national accounts: Netherlands experiences. Paper presented at the Thirteenth Conference of the International Association for Research in Income and Wealth, 31 August-5 September 1973, Balatonfured, Hungary. [67] R. Hueting (1974) Nieuwe schaarste en economische groei - M eer welvaart door minder produktie? (New Scarcity and Economic Growth - more W elfare through less Production?). Agon Elsevier, Am sterdam /Brussel. [68] R. Hueting (1980) New Scarcity and Economic Growth: more W elfare through less Production? North-Holland Publishing Company, Amsterdam. [69] S.J. Keuning (1992) National Accounts and the Environment: the case for a System’s Approach. National Accounts Occasional Paper, Nr. NA-53. Netherlands Central Bureau of Statistics, Voorburg. [70] CBS (1992) Nationale rekeningen

1991 (National Accounts 1991). SDU

/

Uitgeverij /

CBS-publikaties, ’s-Gravenhage. [71] H.K. van Tuinen (1985) Heden en toekomst van de Nationale Rekeningen (Present and Future of National accounts). Statistisch Magazine 2 (1985) 17-22. [72] W.W. Rostow (1990) Theorists of Economic Growth from David Hum e to the Present - With a Perspective on the Next Century. Oxford University Press, New York/Oxford, 1990, pp. 209. [73] J. Arvay (1992) The M aterial Product System (MPS) - A Retrospection. Paper presented at the Twenty-Second G eneral Conference, August 30 - September 5, 1992, Flims, Switzerland. [74] N. Ruggles and R. Ruggles (1970) The Design of Economic Accounts. National

Bureau of

Economic Research, Num ber 89, General Series. Columbia University Press, New York/London.

92

G.P. den Bakker/Dutch National Accounts

[75] S.G. Tiwari (1992) Development of National Accounts in India. Paper presented at the Twenty-Second G eneral Conference, August 30 - Septem ber 5, 1992, Flims, Switzerland, 1992, pp. 8. [76] UN (1992) D raft Revised SNA, No. ST /E S A /ST A T /S E R .F/2/R ev.4, Chapter XX. United Nations, New York.

The Accounts o f Nations Z. Kenessey (Ed.) IOS Press, 1994

93

KEYNES’ HOWTO PAY FOR THE WAR AND ITS INFLUENCE ON POSTWAR NATIONAL ACCOUNTING Yoshimasa Kurabayashi Toyo Eiwa Women’s University, Yokohama, Japan Abstract. In this article the author attem pts to inquire how the idea incorporated in Keynes’s How to Pay for the War founded the M eade and Stone system of national accounts, which, in turn, was eventually developed into the U nited Nations System of National Accounts. Moreover, the author indicates an im portant contribution to this area by a Japanese scholar.

If we define m odern macro-economics as beginning with Keynes, it is more than his General Theory of 1936 that we should be meaning ... This is his booklet How to Pay for the War (1940) [John Hicks (1990), p. 528]

Introduction So wrote Hicks in his final paper that was published posthumously in June 1990 issue of the Economic Journal. In the paper he intended to show that the Keynesian method is founded on the tables of what Hicks called Social Accounts and he set that method in broad historical perspective. Greatly inspired with his perspective, in what follows I attempt to follow up the idea incorporated in Keynes’s How to Pay for the War from the viewpoint of constructing a system of national accounts and inquire how Keynes’s idea was succeeded and developed by his associates, in particular by Meade and Stone, under his strong influence and patronage. In section 2 my attention is drawn to the intellectual environment of the late thirties and I ask how the concept of national income and expenditure emerged from that environment. In section 3 I shall make an issue of the accounting structure of How to Pay for the War and consider characteristics of that accounting structure. Having shown in the section that the accounting structure of How to Pay for the War is incomplete to the effect that the external transactions are not explicitly presented, my inquiry in section 4 goes into the scope and methodology of the 1941 White Paper on National Income and Expenditure on whose publication Keynes exerted substantial influence. Among others, I shall show that the Meade and Stone system of national accounts on which the 1941 White Paper is founded paved a novel way to the formulation of national accounts which finally culminated in the 1952 Edition of the United Nations System of National Accounts after the end of the Second World War. My paper concludes with some observations about the connections of the Meade and Stone system with the Keynesian method of How to Pay for the War and the 1952 edition of SNA. The Keynesian method of How to Pay for the War, which formed the cornerstone of future developments of a system of national accounts, deserves to be very carefully

94

Y. Kurabayashi/Keynes’s How to Pay fo r the War

scrutinized, as I intend to show in this paper. But this paper is further motivated by my inquiries into the progress of SNA including both 1952 and 1968 Editions and the revision of the latter Edition. I have been concerned with this matter in my studies published a few years ago [Kurabayashi (1989)]. During my reserearch a Japanese scholar, who virtually pioneered the national accounts studies in Japan in the early forties, have caught my attention. The name of this early author is Yuzo Morita, who held the chair of statistics at Hitotsubashi University and also served almost ten years immediately after the end of the second World War as Director of the Statistics Bureau of the Japanese Government. In the middle of the War he published three papers on national income studies [Morita, 1942A), (1942B), (1943)]. Among these, the last work is his magnum opus and is definitely relevant to my current article, because the work was an extensive and careful study not only of Keynes’s How to Pay for the War but also of the 1941 White Paper on National Income and Expenditure. Morita’s paper was reaching almost 120 pages. I have been greatly impressed by this masterly work and regret that his great contribution has been neglected by Japanese specialists after the end of the Second World War. Now it’s time to restore his duly deserved honour. Thus my paper is, in part, a sincere homage to this magnum opus of a great Japanese prede­ cessor. For this reason, a brief introduction to Morita’s contribution to the progress of national accounts studies in Japan is appended to my article for the information of those readers who are unfamiliar with his works. 1. The Late Nineteen Thirties: The Advent of National Income and Expenditure Four factors, Paul Studenski once made a point, were responsible for the rising interest in measurements of national income and expenditure in the inter-War period. They are: a) the colossal problems of economic reconstruction facing the nations vitally affected by the war, b) the equally major shifts that had taken place in the relative economic power of the different nations as a result of the war, c) the staggering problems created by the world-wide economic depression of the nineteen thirties, and d) the need for mobilization of economic resources to meet the threats of another world war. [P. Studenski (1958), p. 149] Factor c) essentially accounts for Herculean efforts at measuring national income for 1929-32 of the United States by Simon Kuznets. As Studenski also stated, we have to note that eminent Scandinavian economists such as Erik Lindahl and Ragnar Frisch also substantially contributed to the measurements of economic aggregates for their countries. It is interesting to see a fundamental difference in the approach to compile and represent these economic aggregates, which foreshadowed future differences in the methodology of measuring and representing economic aggregates in a system, that lies between Kuznets and Scandinavian economists. While Kuznet’s efforts were directed to the measurements of national income and its components, the attention of Scandinavian economists were commonly focused on the presentation of economic aggregates in a coherent accounting system. Following the terminology invented by Richard Stone, this development of methodology may be characterized as a progression from national income to national accounts. But

Y. Kurabayashi/Keynes’s How to Pay fo r the War

95

I do not intend to go further into the national income measurements which took place either in the United States or in Scandinavian countries in the late nineteen thirties, as some authors had already given succinct accounts of them. [Ingvar Ohlsson (1953), Odd Aukrust (1955), P.J. Bjerve (1959)]. It is factor d) that Lord Keynes was faced with when he wrote How to Pay for the War. Though Studenski touched upon a few points, we have to recall the witness borne by Roy Harrod and D.E. Moggridge so that we may understand why the economic aggregates got caught the attention of Lord Keynes in his booklet. Roy Harrod explained the situation in his biography of Lord Keynes: "Gradually, however, as autumn [of 1939] wore on, his mind became focused upon the central problems of war finance ... Meanwhile he was at work on a lengthier exposition of his ideas in the form of a small book, which appeared under the title of How to Pay for the War, in February 1940.” [Harrod (1951), p. 490] Harrod set forth the significance of this publication, stating: "This little book is of great theoretical as well as practical interest. Complaints have been made that, in the elaboration of his General Theory, Keynes dwelt to a preponderating extent on the problems of unemployment. Deficient demand was the prevailing malady of the inter-war period, and Keynes kept his eye on the practical situation. In How to Pay for the War Keynes applied his technique of analysis to the economics of excessive demand. The book thus fills a gap." [Harrod (1951), p. 490] It goes without saying that the central focus of How to Pay for the War was the question of inflation. In this work Keynes revolutionalized the ways of thinking of economists on inflation. The older school view was obscured by the idea that there would be no inflation if the budget deficit was financed by "voluntary lending". The central point Keynes made was that borrowing was essentially inflationary whenever it exceeded the amount that people would have been willing to lend without an emerging rise of prices. It was necessary for Keynes to estimate the inflationary pressure on a certain scale. The concept of national income and expenditure could play a pivotal role for the estimation, where Keynes faced a problem. Moggridge, in his recent biographical work on Lord Keynes, accounted for the way how Keynes dealt with the problem: "In 1939 the British Government did not produce national income statistics. In collaboration with Erwin Rothbarth, an extremely able statistical assistant in the Cambridge Faculty of Economics whose pre-war research project had ended when David Champernowne entered government services, Keynes set about providing the necessary numbers, working from estimates previously published by Colin Clark. He published his estimates in the Economic Journal for December 1939, with supplementary notes in March and June 1940, as well as a privately printed Budget o f National Resources in March 1940." [D.E. Moggridge (1992), p. 631] Incidentally, we have to note that the articles to which Moggridge refers in the aforementioned quotation are included in Vol. XXII of The Collected Writings of John Maynard Keynes. The revolutionary idea by Keynes of analysing the war finance and inflation was not confined to the use of national income and expenditure; he went further on to the notion of national accounts.

96

Y. Kurabayashi/Keynes’s How to Pay fo r the War

It is now appropriate to consider the accounting structure that was established by Keynes in the How to Pay for the War. But, before dealing with that task, it is important to stress that Keynes predicted that voluntary savings would be insufficient for the government to borrow. As a result, he estimated that there would be an inflationary gap which could only be filled with increased taxation or with a rise in prices, the latter being related to increasing money incomes. Keynes’s fiscal measures were designed to reduce money incomes during the war, with the government promising to make partial repayments after the war. The advantages of the fiscal measures that Keynes envisaged are double. Firstly, he thought that, if ’deferred pay’ were released at the beginning a post-war slump, it would be possible to maintain aggregate demand without recourse to loan-financed public works. Secondly, on the basis of the reduced disposable incomes, the public would be able to underwrite a greatly increased National Debt. It would be the wage earners, not the capitalists who would emerge from the war as the main holders of the increased National Debt. Thus, by a process that could lead to the smooth working of the war economy, a wider and more equal distribution of property, which had been the dream of social reformers, could be attained as well. Opinions whether the main idea in How to Pay for the War was put into effect in reality differ according to authors. Harrod was less convinced of the effect by saying that "in effect the post-war credit scheme was an interesting experiment, but only played a minor part in the whole situation." [Harrod (1951), p. 494] But Moggridge highly regarded Keynes’s idea stressing, that the policy, "the importance of Keynes’s role in devising it which Harrod underrated, also added coherence to the Government’s anti-inflationary policy." [Moggridge (1992), p. 647]

2. A National Accounting Structure of How to Pay for the War In the Appendices of How to Pay for the War Keynes presented actual figures concerning what he termed the national output and taxable income for the financial year 1 April 1938 - 31 March 1939 in terms of the prices of that year. As I referred to the point in the preceding section, he compiled these figures on the basis of statistical estimates made by Collin Clark. Keynes claimed that these two concepts are fundamental and serviceable for general use. [JMK, IX (1971), p. 429] Without being entangled with further details of conceptual arguments for and against the use of Keynesian concepts of macroeconomic aggregates, those figures which were compiled by Keynes are easily translated into macroeconomic aggregates that are in current use of a system of national accounts and they can be immediately placed into an accounting structure. Five accounts are designed to form the accounting structure. They are: 1) the aggregate production amount, 2) the formation of income account, 3) the income and outlay account for the private sector, 4) the income and outlay account for government, 5) the aggregated capital formation account.

Y. Kurabayashi/Keynes’s How to Pay fo r the War

97

It is of extreme interest to see that the breakdown of sectors for the income and outlay account is essential for the accounting structure of How to Pay for the War so that Keynes could delve into the distribution and redistribution of incomes between the sectors for the management of war economy. The following notations are introduced so as to simplify the presentation of the accounting structure of How to Pay for the War. - Y, net national product valued at factor prices which was termed national output by Keynes; - IDT, (net) indirect taxes; - PC, private consumption expenditure which was termed (private) consumption at market prices by Keynes; - I, net capital formation which was termed net new investment by Keynes; - DT, direct taxes; - PS, private saving which might be implicitly termed national saving by Keynes inasmuch it is derived from the figures referred to saving as the sum of ‘net new investment’ and ‘government deficit’; - PY, private product valued at factor prices which was termed (private) wages and salaries derived from current output; - GY, government product valued at factor prices which was assimilated with government trading profits by Keynes; - TRY, (net) transfer incomes; - GS, government current surplus which was negatively represented as government deficit by Keynes. For recognizing a salient feature of the accounting framework of How to Pay for the War I wish to draw attention to the following remark by Keynes: "Undistributed incomes of companies, etc., are included as part of the incomes of individuals owning them." [JMK, IX (1972), p. 431] Now we can present an accounting framework of How to Pay for the War in a system of equations indicated below. Figures (in terms of million pounds), as compiled by Keynes, are indicated within parenthesis of relevant entries for information and as a concise reference. 1) the aggregate production account Y (4,850) + IDT (670) = PC (4,380) + GC (850) + I (290) 2) the formation of income account PY (4800) + GY (50) = Y (4850) 3) the income and outlay account for the private sector PC (4,380) + DT (550) + PS (370) = PY (4,800) + TRY (500) 4) the income and outlay account for government account TRY (500) + GC (850) = DT (550) + GY (50) + IDT (670) + (-GS) (80) 5) the aggregate capital formation account I (290) + (-GS) (80) = PS (370) The accounting framework of How to Pay for the War as presented here needs to featured in respect of three points. In the first place, the accounting framework constitutes a fully articulated system to the effect that each entry in the system appears

98

Y. Kurabayashi/Keynes’s How to Pay fo r the War

on both sides of the accounts. A fully articulated system ensures the logical consistency of an accounting framework in such a way that all entries of the framework are consistently recorded on both the debtor and the creditor sides. So a set of accounts could show the balancing relationship between the debtor and the creditor sides. Secondly, it should be noted that the accounting framework is based on net concepts. Keynes was specific in explaining the reason why he adopted a net system in a short note which appeared in The Economic Journal, March, 1940. In the note he argued for the concept of national output (product) which should be definitely distinguished from gross national output (product) proposed by Colin Clark. Keynes’s argument followed: 11 It [gross national output as by Colin Clark] differs from national output as defined above in two respects. In the first place, my national output is measured in terms of current cost of output. This is open to the objection that prices are ordinarily understood - the prices, that is to say, which enter into the usual index numbers - are not equal to cost in this sense, but are market prices ... Mr. Clark’s procedure is open to the objection that his gross national output can be changed merely as a result of a change in the character of taxation. But there is a second respect in which Mr. Clark’s gross national income [gross national output and gross national income are alternately used in Keynes’s note] is misleading, if it is meant to indicate (as some current writers seem to be supposing) the potential amount of current consumption......It follows that his gross national income equals the current factor-cost consumption and net new investment (exclusive of the current cost of making good wastage) plus indirect taxes and rates plus twice the current cost of making good wastage (once because the last item is included in the market price of consumption goods and once because it can be diverted to some other purpose). Thus gross national income might be, and often is, taken to give us the potential rate of current consumption and investment measured at market prices. ... If Mr. Clark’s gross national output pretends to provide a serviceable clue to this, it is impostor." [JFK, XXII, "The Concept of National Income: A Supplementary Note", pp. 69-72] It seems to me that Keynes was concerned with recommending (net) national output for gauging not an actual (ex-post) but a potential (ex-ante) level of uses of resources. It is one thing that the valuation of national output by market prices could produce double counting and thus effect the gauging of potential level of uses of resources as Keynes claimed. But, it is quite another thing, to my mind, to question whether the valuation of national output might actually generate double-counting in ex-post figures. It appears the Keynes’s argument is liable to be out of place regarding the distinguishing of the measurement of ex-ante and ex-post figures. Thirdly, we have to stress that external transactions are not explicitly presented in the accounting framework. In reality, they are reflected in relevant entries. This can be seen as Keynes presented another formulation of the aggregate capital formation account in How to Pay for the War as indicated below. 5)’ the aggregate capital formation account by Keynes I (290) + DEP (420) = I (290) + GS (80) + (NB (340)) Note that the last entry in the right-hand side of 5)’ did not originally appear in this

Y. Kurabayashi/Keynes’s How to Pay fo r the War

99

formulation by Keynes. But it is necessary for this account that a balancing entry, which might be termed net borrowing from abroad (NB), should be introduced so that this account could remain in a balanced relationship. DEP in this account stands for good wastage and depreciation according to Keynes’s terminology. The fact suggests that the accounting framework of How to Pay for the War is incomplete to the effect that external transactions are not explicitly presented in the accounting framework. The task for incorporating external transactions in the scope of a system of national accounts was finally settled by the joint efforts by J.M. Meade and J.R. Stone on which Keynes exerted strong influence.

3. The White Paper on National Income and Expenditure The publication of a White Paper which beared the title An Analysis o f the Sources of War Finance and an Estimate o f the National Income and Expenditure in 1938 and 1940 on 7 April, 1941 after the Budget speech marked an epoch in the future studies in national accounts. Keynes greatly contributed to the prepartion for the publication of the White Paper. Moggridge gave a concise account of the progress in the preparation in this biography of Keynes. "How to Pay for the War and Keynes’s private attempts to estimate the national income led the successor to the Stamp Survey, the Central Economic Information Service of the Offices of the War Cabinet, to attempt an estimate and through Austin Robinson to persuade the authorities that a more systematic exercise was essential. James Meade, then with the League of Nations in Geneva, was asked to take charge of the project. After returning across France during the German invasion, Meade began work in June 1940. By the end of July he had evolved the underlying accounting framework. He was then joined by Richard Stone, a young Cambridge economics graduate then working in the Ministry of Economic Warfare, who became responsible for organising and perfecting the statistics. As Keynes settled into the Treasury, he took this project under his wing, easing it through obstacles and departmental jealousies, and persuading Sir Richard Hopkins to be a friendly supporter of the exercise. By early December 1940 the first results of Meade’s and Stone’s work were available. ... The Meade-Stone white paper went through a series of proof after 1 March and met the cautions, but constructive criticism of Dennis Robertson and the much less helpful attitude of Hubert Henderson. Final decisions had to be taken on the exact details of the taxation proposals and other aspects of Budget, and, of course, the Chancellor needed a Budget Speech. Keynes was involved in all these: he even tried his hand at a draft of the Budget Speech which included an excerption of the relevant White Paper figures and quotations from the Madge surveys." [Moggridge (1992), pp. 645-646] Parallel with the preparatory works of the White Paper by Meade and Stone Keynes suggested to Francis Hemming that the results of the work appear as a technical article in The Economic Journal. The celebrated article by Meade and Stone entitled "The

100

Y. Kurabayashi/Keynes’s How to Pay fo r the War

Construction of Table of National Income, Expenditure, Saving and Investment" thus appeared in the June-September, 1941 issue of the Journal. The Meade-Stone paper presents the basic accounting stucture on which the White Paper is founded. Now it is appropriate to give a brief account of the accounting structure that Meade and Stone proposed in their paper. Briefly speaking, the accounting structure is tabulated in a set of tables, which consists of Tables A, B, C, D, and E respectively. Table A represents three aspects of the production, distribution and expenditure of net national product at factor cost (NNP at FC for simplification). The distribution of NNP at FC are itemized in Column AI as (A l) rent, (A2) profit and interest, (A3) salaries, and (A4) wages, the sum of these items in the column AI being net national income at FC. The production of NNP at FC is broken down according to different categories of industry such as (A6) agriculture, (A7) mining, (A8) manufacturing, (A9) transportation, (A10) distribution, (A ll) personal, (A12) government services, and (A13) net incomes from abroad. This is shown in Column AIL The expenditure of NNP at FC is displayed in Column AIII and divided into (A15) personal consumption, (A16) government expenditure on goods and services, (A17) subsidies, (A18) (minus) indirect taxes, (A19) domestic investment, and (A20) foreign investment, the sum of these expenditure components being termed net national expenditure at FC. It is axiomatic to see that net national income at FC (A5) is equal to NNP at FC (A14), which is again equal to net national expenditure at FC (A21). Table B indicates the formation, distribution and expenditure of personal income. While in Column BI the personal income is separated from net national income (B1 and A5) by adding transfer income from government (B2) and subtracting such items as direct taxes (A3), government income (A4), and undistributed profits (B5), the size distribution of personal income according to different income brackets is exhibited in Column BII. Then the expenditure of personal income is broken down into personal consumption at MP (B9) and personal saving (BIO), both of which are shown in Column B ill respectively. Table C represents sources and uses of the total saving. While personal saving (Cl and BIO), undistributed profits (C2 and B5), and government surplus (C5 is in negative sign and it is represented as government deficits in the Table) form the sources of total saving, its uses are allocated to either domestic capital formation or net lending to the rest of the world, both of which are lumped together and are shown as domestic and foreign investment (C4). Thus a balancing relationship is maintained between sources and uses of the total saving as indicated below: C l + C2 = C4 + C5 Table D indicates international transactions between the national economy and the rest of the world. In brief, the items appeared in the left-hand side of the Table stand for national exports, which is defined as the sum of (i) domestic exports of good and services, (ii) consumption of foreign residents, and (iii) the receipts of factor income from abroad, the items in the right-hand side of the Table consisting of national imports and net lending to the rest of the world. National imports can be defined analogous to national exports. Thus we have:

Y. Kurabayashi/Keynes’s How to Pay fo r the War

101

D1 = D3 + D4 D l, D3 and D4 being national exports, national imports and net lending to the rest of the world respectively. Table E is virtually reduced to the aggregate production account in terms of national concepts expressed at market prices (MP). Indeed, the left-hand side of the Table stands for NNP at MP minus national exports as a single item (A1 at MP - D l), whereas the right-hand of the Table is indicated by such entries as national consumption, net capital formation and negative imports. We can immediately transform the entries of both sides of the Table into the aggregate production account in terms of national concepts as indicated below: NNP at MP + D3 = (A15 + A16) + NDCF (A19 + A20 - D4) + D l Therefore, this aggregate production account differs from the balancing relationship derived from Column AIII in that the former is expressed in MP but the latter is valued at FC. As Meade and Stone did not explicitly provide for it, the aggregate production account in terms of a domestic concept should not be derived from these entries which appeared in Table A through Table E. In fact, as I noted before, national exports could be transformed into domestic imports if the former were separated from consumption of foreign residents and the receipt of factor incomes from abroad. It would become a similar exercise to transform national imports into domestic imports. But we cannot put the transformation into practice for the Meade and Stone system because the receipt and payment of factor incomes between the national economy and the rest of the world is netted out and is recorded as a single entry (Did) in Talbe D. This demonstrates that the Meade and Stone system was, in essence, a system of entries consisting of national concepts. Moreover, it should be noted that the explicit formulation of sectoral accounts within the Meade and Stone system was not brought to is completion, although the formation and the appropriation of income accounts for the private sector were built into the system. Lack of the entries which express financial flows of lending and borrowing of financial assets and liabilities and of the interaction between real and financial flows within the scope of the Meade and Stone system could be responsible for this incompleteness. Hence, it naturally follows that the Meade and Stone system did not form an entire or fully-articulated system. In order to carry out the task that the sectoral accounts should be explicitly fitted into a fully articulated system of national accounts, we have to anticipate the United Nations system of national accounts of both 1952 version and, as the culmination of the system, the 1968 SNA standard. Returning to the publication of the White Paper, it is interesting to see that Tables A and B of the Meade and Stone system were used for the presentation of available data. Harrod commented on the methodolgy and data availability for national income estimates at the time and went on to the significance of the White Paper. "This was indeed a great revolution. The Treasury had hitherto confined itself to figures for actual, known transactions. This account included estimates, and certain figures had to be obtained by the method of difference from other estimates - all of which was very dangerous. Yet this kind of national accounting has come to be regarded

102

Y. Kurabayashi/Keynes’s How to Pay fo r the War

as the essential tool of any economic planning, whether of an Individualist or Socialist variety, and other nations have followed Britain in presenting such accounts. The initiative must be attributed to Keynes, but for whose active interest the compilation would not have been made and published at that time." [Harrod (1951), P. 502] The publication of the White Paper was the culmination of the work. Stone recalled the moment stating: "It was a great day. We drank champagne that night and felt we accomplished something. [J.R.N. Stone (1951), P. 85]

4. Concluding Remarks: Towards the 1952 SNA and the Aftermath The advent of the White Paper on National Income and Expenditure and the invention of the Meade and Stone system of national accounts upon which the White Paper founded thus formed the cornerstone of the future development of national accounts studies. It was Richard Stone who stood in the centre of cultivating international guidelines of national accounts after the end of the Second World War. In a sequel to an earlier report for the League of Nations entitled Measurement o f National Income and the Construction o f Social Accounts which was prepared by Stone and published in 1947, an international group of experts was formed on the basis of recommendation provided by the United Nations Statistical Commission, at its 4th and 5th sessions, to prepare "a standard national accounting system in order to provide a framework for reporting national income product statistics which is of general applicability". [Preface to First (1952) Edition of SNA] At the invitation of the group Stone acted as Chairman and the 1952 Edition of SNA resulted. It is relatively easy to find out points of influence from the national accounts studies that was seed by Keynes in How to Pay for the War and developed further by Meade and Stone in the earlier version of SNA. Among others, I can suggest the following three points. In the first place, major economic aggregates on which Keynes and his associates focused were net national income and net national product expressed at factor cost. This sharply contrasted with the conceptual framework adopted by the United States National Income and Product System whose aggregates were expressed in gross terms. The 1952 edition of SNA referred to this point saying: "Since this [consumption of fixed capital] is similar to the consumption of raw materials and the like, it is desirable to have a concept of net product in which a provision for the consumption of fixed capital is deducted from the gross total." [1952 Edition of SNA para. 42] But 1952 SNA envisaged the net concepts as complicating the matter: "In a stationary economy, in which the quantity of fixed assets in use remains the same and does not change in quality, these provisions can be defined so that in each year they are equal to the replacement needed. In an economy, however, in which the stock of assets is changing in quantity, it is not equally simple to set up corresponding rules and if the rules appropriate to a stationary economy are applied it may be found that the

Y. Kurabayashi/Keynes’s How to Pay fo r the War

103

provision will no longer be equal to the required replacement. The problem of defining provisions for the consumption of fixed capital are further complicated in an economy which is characterized by changes in demand or by technical change which lead to obsolescence and the replacement of assets by different assets." [1952 Edition of SNA, para. 43] It seems to me that this point made by 1952 SNA still hits the crux of the logical foundation of any system of national accounts. I shall again come back to this point at the end of the section. Secondly, Keynes was well aware that changes in prices greatly affect the estimates of output capacity and the national income. I quote the How to Pay for the War, where he argued: "The money measure of our output capacity will, of course, vary according to the levels which are reached from time to time by wages and prices. To avoid this complication the following figures are all given in terms of pre-war prices." [JMK, IX, p. 381] But the figures indicated in the White Paper are values not in constant prices but in current prices. The 1952 Edition of SNA also inherited the principle that the figures in a system of national accounts will be expressed in current prices. The principle was clearly witnessed by the Preface to 1952 Edition of SNA: "This report is confined to national accounts expressed in current money terms. For many purposes it is necessary to make comparisons over time in terms of constant prices. The many conceptual and statistical problems involved in obtaining the information for such comparisons are not examined in this report." [1952 Edition of SNA P- ix] One had to wait for another contribution by Stone who raised the question whether a system of national accounts in constant prices can be constructed without ambiguity and inconsistency by using available price and quantity data, Stone proposed a solution to this question in his Quantity and Price Indexes in National Accounts. It should be noted that the issue about the construction of a system of national accounts in constant prices were often made an issue in the agenda of the United Nations Statistical Commission at the sessions of the late fifties and early sixties. Thirdly, it was in the 1952 Edition of SNA that sectors in the economy were well defined so that sector accounts might be geared into a fully articulated system of national accounts. It is interesting to see that the sectoring in 1952 Edition of SNA was based on the institutional basis. In the 1952 Edition of SNA three basic sectors naturally followed: i) enterprises, ii) households and non-profit institutions, and iii) general government. The sectoring of the 1952 Edition of SNA advanced a set of large steps in comparison with the Meade and Stone system of national accounts on which the White Paper on National Income and Expenditure founded. In conclusion we have to come back to Sir John Hicks’s paper a passage of which was quoted at the beginning of this article. At the end of the same paper Hicks wrote: "It will make no difference to the accounts of the current period, what is the form in which it leaves its terminal stock. But in practice it may make a considerable difference.

104

Y. Kurabayashi/Keynes’s How to Pay fo r the War

Keynesians have not (to my knowledge) done much on this Continuation theory; perhaps that is why their doctrine has got the reputation of being short-sighted. Ricardo however did attend to it, in the difficult chapter on machinery which he added to the last edition of his book which he himself prepared. Though some of his immediate followers did consider it, in the days of the neo-classics (and Keynesians) it got lost." [Hicks (1990), p. 537] The dynamic interdependence between the closing and opening balance sheets still remains as an open question at the centre of a system of national accounts, even in the revised system of national accounts recently adopted by the United Nations Statistical Commission.

Appendix:

Professor Yuzo Morita’s Contribution to Studies in National Accounts in Japan of the Early Nineteen Forties

As mentioned earlier, Professor Yuzo Morita played an essential role in introducing the idea of national accounts in Japan still in their stage of infancy. This article is an appropriate place to give a brief account of his contribution to the progress of national accounts in Japan particularly in the nineteen forties. But, first let me put in a few words about Morita’s academic career. He studied at Tokyo University of Commerce (now Hitotsubashi University) specialising in economic statistics and graduated in 1925. After having taught at Yokohama Commercial College (now Yokohama National University) for several years, he studied abroad in Berlin and Vienna, where he witnessed the Anschluss by Hitler (March, 1938) in Vienna. He obtained the degree of Doctor of Economics from Tokyo University of Commerce on the basis of a master­ piece entitled The Analysis o f Demographic Changes in Japan (Jinko Zoka no Bunseki), in Japanese, 1944. After the end of the Second World War, he was among the central figures for restructuring the statistical system in Japan. He assumed the Directorship of the Bureau of Statistics of Japanese Government and served in that post for 10 years (1947-57). During his Directorship Morita took the main initiative in carrying out the 1950 Population Census which became the first full-fledged Population Census after the end of the Second World War. He developed a wide variety of economic and social statistics as well. He held the chair of statistics at Hitotsubashi University from 1949 until his retirement in 1965. Morita published numerous books and articles concerning the theory of statistics, as well as economic and social statistics. It should be noted that he was one of the founding fathers of the Japan Association of Statistics, which was inaugurated in April 1931, of which he later became its President. He also served as the Chairman of the Statistical Council of the Japanse Government. So much for the academic and governmental career of Yuzo Morita. Now let us turn our attention to his contribution to the studies in national accounts in Japan during the early nineteen forties. As I noted in the earlier section, during the war years he published three national income studies. Among them, [10] is the most substantial work in regard contents and length. In what follows I shall try to summarize the content of

Y. Kurabayashi/Keynes’s How to Pay fo r the War

105

this study [10]. This work, which is quite exhaustive and lengthy, consists of four parts and two appendices, running 121 pages. It is convenient to show the sections of this study: I. The Concepts and Definitions of National Income and their Variety II. The Scope of National Income and its Statistical Measurement III. The Circular Flow of National Income as a Tool for its Statistical Measurement. IV. The Statistical Framework of the Circular Flow of National Income - the Network System of National Income Statistics. Appendix I. An Estimate of National Income and Expenditure of the United Kingdom in 1938 and 1940. Appendix II. An Estimate of National Income and Expenditure of the United Kingdom in 1938, 1940 and 1941. In the first part of this study, Morita starts with examining a variety of concepts of national income and related economic aggregates such as social product (Sozialprodukt according to his usage), national output and national dividend. Some German economists at the time (for example, J. Schumpeter, A. Ammon and K. Diehl are quoted by him) had some general reservations about the validity and usefulness of national income as a theoretical concept. But he makes a point that there are multiple concepts of national income from the standpoint of statistical measurement. Hence, they are, he argues, worthwhile to study more closely. In this context, three major aspects of national income are explored by him. They are: (i) gross versus net national income, (ii) private (individual according to his usage) versus national income, and (iii) realized versus potential (or capacity) national income. Having recognized the existence of a variety of national income concepts, he argues in the second part of his study that they can be sorted out according to three basic activities: (i) production, (ii) distribution, and (iii) expenditure. While he notes that the production approach to national income is more appropriate to measure aggregate production, the income approach is good for the measurement of the functional distribution of incomes. In this connection three major points for the measurement of national income are raised by him: (a) the treatment of services, (b) he boundary between intermediate and final products, and (c) the problem of double counting. It is interesting to see from his discussions that they faithfully reflect those issues which concurrently with this article were discussed at such academic forum as the Conference on Research in Income and Wealth organized by the National Bureau of Economic Research. However, in this part of his study, reference to this Conference is limited to the work by Simon Kuznets on National Income which appeared in Encyclopedia o f the Social Science, vol. 12. The limitation is quite understandable, because earlier series of Studies in Income and Wealth could scarcely be available for Morita when he wrote this piece. The major portions of this study are in subsequent two parts. In the third part, the author considers a diagrammatic presentation of the circular flow of national income. It is not only necessary but also useful for the statistical measurement, he argues, that the circular flow of national income is mapped out in a consistent way according to the

106

Y. Kurabayashi/Keynes’s How to Pay fo r the War

different facets of the basic activities, e.g. production, distribution, and expenditure of national income. In carrying out this exercise, he claims, not only the flow of goods but also the flow of services should have a diagrammatic presentation. Secondly, it is stressed that the redistribution of incomes through the transfer of incomes in the diagrammatic presentation should play an essential part. Thus the structure of the diagram for representing the circular flow of national income proposed by him cannot be reproduced here because of its complexity for reproduction. However, in essence, it is very close to the Norwegian Ecocirc-System developed by R. Frisch, P.J. Bjerve and O. Aukrust. It is a great pity that the diagrammatic presentation of the circular flow of national income could not be completed and filled in with statistical figures for the Japanese economy by Morita. Having borne in mind the diagrammatic presentation of the circular flow of national income, Morita attempted to scrutinize exhaustively Keynes’s estimates of U.K.’s national income and expenditure for the 1938-39 fiscal year in Part IV of his study. By the time Morita wrote his work, Keynes had established a great name in Japan, particularly after the publication of his General Theory. Some active Japanese economists pored over the Keynes’s revolutionary works and those of his followers. For example, such names ought to be mentioned as Yasuma Takat (Kyoto University), Ichiro Nakayama (Tokyo University of Commerce), Nisaburo Kito (Tokyo University of Commerce), Takuma Yasui (Tokyo University) and Hideo Aoyama (Kyoto University). In this light, it was quite natural that Keynes’s work on national income caught Morita’s attention. In the last part of his study, he compared Keynes’s estimates not only with corresponding figures subsequently published in two White Papers on the War Finance (Cmd. 6261 and Cmd. 6347) but also with those figures which appeared in the Meade-Stone system of national accounts which I have reffered in the text. It is interesting to see from Morita’s article the meticulous attention he paid to the internal consistency between the entries of national income estimates or those of national accounts. In addition, he provided critical comments on the derivation of figures. It should also be noted that in his work he attempted to transcribe U.K.’s estimates of national income and expenditure for 1938 and 1940 into his diagrammatic presentation of the flow of national income. After performing this task, he concluded his master­ piece by considering the essential aim for national income studies in War time. He defined it as delving into the interacting network of the flow of incomes between the government which is necessarily swelling up by excessive war demands on the one hand, and the rest of the economy on the other, so that an appropriate balance between them may be restored. Two appendices are attached to his study: (i) a Japanese translation of a part of the government paper (Cmd. 6261) which is related to the estimates of national income and expenditure in 1940 and 1938 and (ii) another Japanese translation of the corresponding part of the other government paper (Cmd. 6347). Both appendices provide useful background information concerning the structure of U.K.’s national income and expenditure. As I have referred to in the text, Morita’s study has sunk into long oblivion by Japanese economists as well as the national accounts specialists in Japan. But, since it

Y. Kurabayashi/Keynes’s How to Pay fo r the War

107

is acknowledged that the ideas developed by Keynes in his How to Pay for the War had paved the way to the 1952 version of SNA, it seems the time has come to evaluate Morita’s contribution to the development of national accounts in Japan. This way his contributions could be looked at in its wider and international perspective. As I have said earlier, this appendix intends to pay homage to a master-piece by a great Japanese scholar.

Acknowledgements The author wishes to express great indebtedness to Sir Jack Hibbert, Professors Thomas K. Rymes and Christopher T. Saunders, and the editor of this volume for their comments on the earlier version of this article.

References [1]

Aukrust, Odd (1955) Nasjonalregnskap, Teoreliske prinsiper.

[2]

Bjerve, Peter J. (1959) Planning in Norway.

[3]

Cuyvers, Ludo (1983) "Keynes’ Collaboration with Erwin Rothbarth", Economic Journal, Vol. XCIII.

[4]

H arrod, R.F. (1951) The Life o f John Maynard Keynes.

[5]

Hicks, John (1990) "The Unification of Macro-economics”, Economic Journal, Vol. C.

[6]

Keynes, J.M., Collected Writings; Vol. 9, Essays in Persuation, 1972, Vol. 22, Activities 1939-45, 1978,

[7]

Kurabayashi, Yoshimasa (1989) Progress o f the United Nations System o f National Accounts, (SNA no Seiritsu to H atten), in Japanese.

[8] [9]

Moggridge, D.E. (1992) Maynard Keynes: An economist’s biography. M orita, Yuzo (1942) "National Financial Resources and National Income" (Kokka Shiryoku to Kokumin Shotoku), Business Administration, (Shogaku), No. 34, in Japanese, 1942A.

[10]

M orita, Yuzo (1943) "The Circulation of National Income" (Kokumin Shotoku no Junkan), National Income and its Distribution (Kokumin Shotoku to sono Bunpu), ed. by Japan Association of Statistics, in Japanese.

M orita, Yuzo (1942) "Measurement of National Income" (Kokumin Shotoku no Suikei), Annals o f Japan Association o f Statistics (Nihon Toukei Gakkai Nenpo), N o .ll, in Japanese, 1942B. [12] Ohlsson, Ingvar (1953) On National Accounting. [11]

[13]

Patinkin, Don (1976) "Keynes and Econometrics: On the Interaction between the Macroeconomic Revolution of the Interwar Period", Econometrica, Vol. XLIV.

[14]

Peden, G.C. (1988) Keynes, the Treasury and British Economic Policy.

[15]

Skidelsky, R obert (1992) John Maynard Keynes: The Economist as Saviour 1920-1937.

[16]

Stone, Richard (1951) The Role o f Measurement in Economics.

[17]

Stone, Richard (1951) "The Use and Development of National Income and Expenditure Estimates”,

[18]

Lessons of the British War Economy, ed. by D.N. Chester. Studenski, Paul (1958) The Income o f Nations. U nited Nations (1952) A System o f National Accounts and Supporting Tables, Studies in Methods,

[19]

Series F No. 2 1952.

108

Y. Kurabayashi/Keynes's How to Pay fo r the War

[20]

United Nations (1968) A System o f National Accounts, Studies in Methods, Series F No. 2 Rev. 3,

[21]

An Analysis of the Sources of W ar Finance and an Estim ate of the National Income and

[22]

An Analysis of the Sources of W ar Finance and an Estimate of the National Income and Expenditure

1968. Expenditure, Cmd. 6261, 1941. in 1938, 1940 and 1941, Cmd. 6347, 1942.

109

The Accounts o f Nations Z. Kenessey (Ed.) IOS Press, 1994

AMERICAN CONTRIBUTIONS TO THE DEVELOPMENT OF NATIONAL ACCOUNTS Zoltan Kenessey* Voorburg, The Netherlands Abstract. This article follows up earlier historical studies about U.S. national accounting by Carson and by Kendrick. It extends these studies on some points, such as the importance of Irving Fisher, Morris Copeland, R obert Martin, Milton Gilbert, Edward Denison, and George Jaszi. In particular, the early Am erican emphasis on the potentials of double-entry bookkeeping for national accounting is explored. Yet, the actual introduction of this technique in U.S. statistical practice came about in the context of contacts with the national accounts work in the U.K.

Introduction The evolution of national accounting in the U.S. was perhaps a less dramatic development than in the other three principal countries about the same time: the U.K., Norway, and The Netherlands. While the establishment of the U.S. national accounts was also related to the great war efforts, still circumstances in America were considerably less trying during those years than in England, Norway, and Holland (the latter two even being occupied). Yet the importance of the national accounts undertaking were no less in the U.S. than in the other countries; in their detail, scope, and policy use the work was at least as important, and arguably the utilization of national accounts for the large U.S. economy (a main engine of the war) played an unequalled role in the statistical underpinnings of the entire international war effort. It also seems that the U.S. national income work before the establishment of the national accounts was exceptionally extensive for the 1930’s. In general, statistical work on the economy was extended very much in the U.S. already in the 1920’s. The Great Depression accelerated this process, not in the last regarding aggregate statistical measures such as national income. Both matters of statistical methodology and the exploration of data sources received much attention in the interwar era. The work of Simon Kuznets is the best known, but certainly not the only example of the progress made. While the focus of our review is on the 1930’s and the 1940’s, it is useful to reflect a bit on the U.S. statistical developments in the 1920’s. This is done not only for historical reasons. It seems also to be relevant to certain statistical concerns of our days.

* The author is Director o f the International Statistical Institute in Voorburg, The Netherlands. The views expressed are personal.

110

Z. Kenessey/American Contributions

Interestingly, in the U.S. the large spurt in economic statistics in the 1920’s occurred under fairly conservative administrations. Indeed, this very productive era in statistics was closely connected with the name of Herbert Hoover, who served as Secretary of the U.S. Department of Commerce, where an important part of the progress in economic statistics occurred. Hoover’s failings and his Presidency during the fateful 1929-1932 Depression years are too well known here to be recounted. Yet Hoover deserves credits not only for matters of statistics. To mention one, his role at the peace talks after the First World War was much appreciated by Keynes, who was the famous critic of the disastrous treaty forged at Versailles regarding Germany (an exercise that was to be repeated in an even worse way at Trianon regarding Hungary). In Keynes’ judgement "Mr. Hoover was the only man who emerged from the ordeal of Paris with an enhanced reputation...[his] atmosphere of reality, knowledge, magnanimity and disinterest...if they had been found in other quarters, also would have given us the Good Peace." Keynes (1920) p.274,fn. Interestingly, according to the recent penetrating analysis of Professor William Barber, subtitled ’Herbert Hoover, the Economist, and American Economic Policy, 1921-1933’ "As President, Herbert Hoover understood one of the basic insights of a theoretical system later to be identified as Keynesianism: that the behavior of the macroeconomic system is determined by the aggregate volume of spending. Indeed, he grasped this point as well as did anyone of his time - and better than did most of the economists in that period. This was the basis for his initial effort in depression fighting...". Barber (1988) p.189. It would go beyond our survey to refer to the political and cultural constraints (real and self-imposed) that prevented Hoover from being effective in that fight. I submit, that Herbert Hoover - conservative as he was - would be amazed seeing today’s arguments against money to be spent on economic statistics, which he understood to be very important for business purposes. The shortsighted view of not wanting statistics because they may lead to regulation or greater governmental role most likely would have stroked his agile mind as a bad case of non sequitur indeed.

1. References to the 1930’s and the beginnings of the 1940’s As intimated in the abstract above, U.S. national income and product accounts developments between 1932 and 1947 were analyzed in considerable detail by Carol Carson, who prepared her dissertation about the subject under the direction of Professor John W. Kendrick at George Washington University in Washington D.C.. Kendrick himself also contributed to the historical evaluation of U.S. developments (see Kendrick, 1970). Their work, but particularly Carson’s, facilitated very much my effort to review the major American contributions to the development of national accounting in the broader international context. Indeed, my first brief comments are based on Carson’s account of U.S. developments in the 1930’s and 1940’s, without repeating the many threads treated so ably in her work. Carson (1975). As already suggested it appears that in terms of data work the American efforts

Z. Kenessey/American Contributions

111

regarding national income estimating (as well as in exploring certain methodological matters) were more diverse and probably also more comprehensive than elsewhere. Also, the circumstance, that official governmental work on the subject, carried out in cooperation with the National Bureau of Economic Research, resulted in a report entitled ’National Income, 1929-32’ that was submitted to the U.S. Senate on January 4, 1934 signalled a major achievement. Simon Kuznets, who in 1929 joined the National Bureau where he started to work on expanding national income work carried out there earlier, in 1933 went to the U.S. Department of Commerce and took charge of the project that resulted in the Senate submission mentioned. I already mentioned that similarly to Europe, the 1930’s witnessed extensive work and debate in the U.S. regarding the clarification of concepts and methods relevant to the estimates. Also, just as in Europe, the first main impetus to the work came from the need for information in the midst of the Great Depression. The second, of course was the Second World War itself. The above-mentioned report which was submitted to the U.S. Senate was a great success, almost 4,500 copies were sold, about "800 more than sales of the ’Statistical Abstract’ [the U.S. Statistical Yearbook] of that year." Carson (1975) p.159. While Simon Kuznets returned to continue his research at the National Bureau of Economic Research, the work at the U.S. Department of Commerce was continued first under Robert F. Martin, and thereafter under Robert N. Nathan, who was a student of Simon Kuznets. Among the other efforts in the 1930’s Carson relates first Clark Warburton’s estimates started in 1932, developed in connection with the Brookings Institutions studies on ’America’s Capacity to Consume’. She mentions that a December 1934 article by Warburton in the Journal of the American Statistical Association already used the term ’gross national product’ in print. Second, Simon Kuznets also continued his research and prepared further estimates, in particular regarding durable capital and capital formation. In his report about his results in 1937 Kuznets already used the term ’gross national product’ as well. Thirdly, Laughlin Curry - who was the Assistant Director of Research and Statistics at the Federal Reserve Board - also initiated work relevant to the topic, by focusing on the ’pump priming deficit’ which was a tool of considerable interest, as governmental stimulus became important for fighting the Depression. Kuznets’ wide-ranging contributions to U.S. national accounting cannot be reviewed here. He tirelessly worked on many aspects of income and product estimating and was willing to challenge prevailing views. For example he considered whether government services should be accepted as final products entering GNP or rather as intermediate services to production of other elements of the total. He also discussed with Hicks the true connections between national income and societal welfare. He was also critical, or less than enthusiastic, about certain results of the 1947 revision of the U.S. official national accounts. While his views did not necessarily prevail on all points, the high standards he set himself and others regarding methodological clarity and exploration of data sources were of deep influence and lead to many fruitful results in the U.S. national accounting work.

112

Z. Kenessey/American Contributions

It is also important to underline Kuznets’ attitude to theory. It was a thoroughly positive stance, but also critical of certain attitudes - unfortunately still widely spread among theoretically oriented researchers. Fogel summarized Kuznets’ viewpoint on this matter as follows: "Kuznets was, however, impatient with theorists who knew so little about the institutions or processes about which they theorized that they could not distinguish between metaphors and reality and so failed to consider the logical implications of assumptions that violently distorted the real world. One of Kuznets’ repeated contributions was the demonstration that certain so-called pure theories embodied false assumptions about empirical matters, assumptions that critically affected the conclusions derived from the theories. In so doing he helped to counter the view that in theoretical work, cleverness and elegance were all that mattered." Fogel (1987) p.10. Speaking of the wider impact that economic theories may have, presentations by Alvin H. Hansen and Curry at hearings in the Congress in May 1939 were "instrumental in introducing Keynesian thinking to popular Washington economics." Carson (1975) p.166. And later in 1939 Curry went to the White House to become one of the administrative assistants of Roosevelt. Other developments included the first ’product side’ estimates of the gross national product by V Lewis Bassie, who earlier worked with Curry at the Federal Reserve and joined the Department of Commerce later. Bassie also prepared an important memorandum on the "Effects of the Defense Program on the Economy" in 1940. By this time the attention turned to the expected needs of the likely involvement of the U.S. in the War. In mid-1940 Robert N. Nathan, who was in charge of its work over the preceding four years, left the National Income Division of the Department of Commerce, which carried forward the work that first resulted in the publication of ’National Income, 1929-32’ in 1934. Nancy Ruggles’s account summarized the macroeconomic data needs of the 1930’s well for many countries, including for the United States: "The great depression of the 1930’s led to the next major step. In response to pressing policy needs and taking advantage of concurring developments in economic theory, statisticians began to convert the aggregate national income time-series into a system of national income statistics, where the components - consumption, investment and saving, on the one hand; wages and profits, on the other - were looked upon as necessary ingredients in explaining the whole, and of as much interest as the aggregate figures. The statistical developments in some cases preceded the theoretical - Kuznets’ first data came well before Keynes’ theory - but both reflected the same need and they reinforced and stimulated each other." Ruggles (1987) p.377. It should be noted, that earlier national income work focused largely on establishing the total income (and product) of nations. Naturally, this required attention at least to the major components of income and product. However, with the emergence of Keynesian thinking - as well as Frisch’s system - the detailed statistical portrayal of the whole economic reproduction process was put on the agenda. And this task was to be accomplished in an integral manner, in a systematic and organic connection of rather detailed accounts to be set up for this purpose.

Z. Kenessey/American Contributions

113

2. The beginnings of official national accounts in the 1940’s In the history of the establishment of the official national accounts it was an important step when the then editor of the ’Survey of Current Business’, the today still existent publication of the U.S. Department of Commerce, Milton Gilbert became the new chief of the Division charged with national income work. Milton Gilbert stressed Keynesian orientation, and strengthened the Division by additional professionals (Edward F. Denison and George Jaszi were among his new recruits). The work on national income, product and expenditure was significantly expanded and speeded up. "The GNP series was effectively launched in May 1942: estimates extending from 1929, tables showing the interrelationships among segments, and a few notes on the sources and methods of estimates were provided...In August 1942, gross national product and national income were first presented on a quarterly basis...Also ...[there] was a table showing GNP in constant prices." Carson (1975) pp.172,173. Special significance may be accorded to the cooperation with other statisticians, especially from the U.K.. During the War, at the 1944 consultations of national income experts from the United States, the United Kingdom, and Canada, the conceptual and statistical work on national income in the three countries was compared. Indeed, Edward Denison felt in a 1992 communication to the author that on that occasion "Agreements were reached on all major issues, as summarized in my report in Vol. 10 of the Income and Wealth Series." However, the postwar history of national accounting at the broader international level started in 1945 on a meeting held in Princeton, U.S.A. The prelude to this meeting occurred in April 1939, when the League of Nations Committee of Experts decided to include the statistical measurement of national income into its program. Following the interruption caused by the war the Sub-Committee on National Income Statistics was set up to study the matter under the chairmanship of Richard Stone. George Jaszi represented the U.S. at the League of Nations expert group meeting where Stone’s seminal report was discussed. (The final report, in 1947, was already issued under the imprint of the United Nations.) Jaszi’s own summarizing reference to the early developments, offered at the time of his retirement in his famous "An Economic Accountant’s Audit" was the following: "The very idea of income and product accounting, which emerged in the early 1940’s, grew out of the practical needs of economic mobilization for World War II and made its debut almost simultaneously in the statistical offices here and in Canada and England. A great deal of responsive listening was taking place, and it continued." Jaszi (1986) p.412. Finally, at the time of the 1947 basic revision of the U.S. estimates (which was supervised by Milton Gilbert, assisted by Denison, Jaszi and Charles F. Schwartz) several improvements were introduced, as well as the detailed accounting framework was accepted for the presentation of the estimates.

114

Z. Kenessey/American Contributions

3. American proponents of double-entry techniques in the 1930’s Carson’s account appears to imply that the American acceptance of the double-entry accounting format was under the influence of the tripartite consultations and the work of Stone, in general. While I have no reason to question this implication, I wish to touch upon the influence of Irving Fisher on the matter, which I also mentioned in the introduction to this volume. I accept that the direct influence very well may have come from Stone. However, the role of Fisher and Copeland may be expanded in order to round out the story. First it is useful to recall, that Stone’s famous League of Nations study includes the following important reference: "...This report approaches the study of national income from the point of view of social accounting." The footnote attached by Stone to the term ’social accounting’ reads as follows: "It is believed that this term was first used by J.R. Hicks in his book "The Social Framework: An Introduction to Economics" (1942). For an explicit recognition of the accounting aspect of national income work see the following articles by M.A. Copeland: "National Wealth and Income - an Interpretation" in Journal of the American Statistical Association, Vol. XXX, No 190,1935, pp. 377-386 and esp. 379 and 386, and "Concepts of National Income" in Studies in Income and Wealth, Vol I (1937), pp 3-63 and esp. p. 63." Stone (1947) p.23. Copeland’s (1935) article, which was referred to by Stone, includes extremely clear suggestions on the matter such as these: "Although the process of measurement of these basic concepts is partly a statistical one, it is also in an important sense an accounting process, an attempt to portray the economic condition and operation of a society in terms of double-entry bookkeeping, in terms of a set of controlling accounts...(p. 378)...It seems probable that we are on the threshold of a much fuller use of national balance sheets and income statements and their detailed breakdowns, a use which may take a direction for the accounts of our national economic system, analogous to the progress from general accounting to cost accounting for the individual business." (p. 386) Copeland’s (1937) article, also referred to by Stone, includes a quotation from a letter he wrote in January 1936: "May I offer some suggestions regarding possible lines of inquiry which I believe would be profitable? Several of these emphasize the need for studying wealth and income together, setting up what amounts to a consistent scheme of social capital and income accounts for each major industrial grouping in our economic system....This should be an experimental study for sample years, which would attempt to work over available data into the form of a double entry system of accounts on a rough accrual basis appropriate for use in national income and wealth measurements. Such a study should throw light on a number of problems - the handling of government interest, relief payments, government budget deficits, etc. in national income estimates; ..." (p.63) Undoubtedly Copeland knew what he was talking about and for what purposes he wanted such accounts to be utilized. Indeed, his 1937 article concludes - after the quotation from his 1936 letter - with the following statement: "I now wish to urge this proposal again." (p. 63)

Z. Kenessey/American Contributions

115

Robert F. Martin, who was already mentioned in connection with his work on national income at the U.S. Department of Commerce after Kuznets returned to the National Bureau, later joined the research staff of the National Industrial Conference Board in New York, and in 1936 published a study there with the title ’National Income and its Elements’. Martin (1936). This study, in its first chapter, contains a first section entitled ’An Accounting System for the National Economy’. Among others, this section contains statements such as the following: "In an accounting system for the national economy, the balance sheets would list the national resources as assets and the claims of owners as liabilities. Its income account would show on the income side receipts for goods and services, and on the outgo side payments to contributors to production." Further Martin talks about "...a consolidation of the accounts of economic enterprises to obtain a master statement for the national economy...". Also, he states that "In a consolidated income account of the economic system...receipts would correspond to the expenditures of individuals buying the goods and services, while the disbursements for capital, labor, and management services would represent the income of all individuals, or national income." (These quotes are from p.l in Martin, 1936). On the second page of the study Martin expresses the view that "The complete setting up of such a comprehensive bookkeeping system is impossible at present because of the limited state of available information....The best that can be done is to establish a simple framework for the system and to expand it and fill in requisite data as they become available." Martin (1936) p.20. In the footnote attached to the word ’system’ in the above sentence, Martin mentions that "For a preliminary suggestion of this nature see Morris A. Copeland, "Some Problems in the Theory of National Income", Journal of Political Economy, Vol.XL, February, 1932." Martin (1936) p.2,fn. Martin’s own suggestion is the following: "The basic framework of the aggregate income accounts of the individuals in its simplest form might be summarized as follows: Receipts Expenditures Salaries and wages Shelter Entrepreneurial income Food Dividends, interest, net rent Clothing Other items Other items. ...It is obvious that the national income statements simply represent a tentative and only partially complete accounting of our economic operations." Martin (1936) p.2. Since my interest here lies in the proposed use of double-entry accounting, I also quote the footnote added by Martin to the above shown presentation of ’receipts’ and ’expenditures’: "If the account were set up for the national economy as a business enterprise, the receipts and expenditures would be reversed. For example, salaries and wages would be listed as an expense of the business enterprise and the expenditures of individuals for shelter as a receipt." Martin (1936) p.2,fn. Martin was clearly aware of the work of Irving Fisher, and of course of Simon Kuznets. Regarding Fisher he refers to his 1906 book "The Nature of Capital and Income", while for Kuznets he cites his "National Income" article in the Encyclopedia of the Social sciences in 1933.

116

Z. Kenessey/American Contributions

In conclusion I quote Martin’s statement, that "No one point of view is adequate to cover all the different phases of income in the national economy. The fact that there are different totals, which are useful for different purposes, has led to some confusion concerning the meaning and relative merits of the various estimates now available. Much of this confusion disappears when these estimates are viewed in the light of a complete accounting system...'1. Martin (1936) p.3. Thus Martin’s advocacy of double­ entry accounting (and apparently he well understood what technicality he was talking about) went well beyond suggesting a new form for presenting aggregate economic data. Clearly, he favored a systematic way of looking at such measures, one that could serve many purposes yet one that was providing an integrated system. In the light of the foregoing (and running ahead of some additional evidence cited below) it seems justified to place the ’intellectual fatherhood’ of the idea of double entry accounting for purposes of national income, product, and expenditure into America, where this idea apparently grew out of Irving Fisher’s work at the beginning of the century, and was expounded by Morris Copeland and Robert Martin in the 1930’s. Naturally, such a suggestion does not reduce the importance of Meade and Stone in our story. Indeed, in the United States - as discussed below -the introduction of double entry national accounting in 1947 apparently had little to do directly with Fisher, Copeland or Martin but a great deal with the UK experience regarding this the matter. Also, we should again recall, that during the 1930’s the first input-output studies of Leontief matured in the United States. They exercised much influence both at home and abroad after the Second World War, but only later got directly related to official national accounts work. Notwithstanding these references, it is fully understandable, that it was Sir Richard Stone, who "was awarded the 1984 Nobel Prize in Economic Sciences ’for having made fundamental contributions to the development of systems of national accounts’. His death in December 1991 came as the United Nations was revising its System of National Accounts (SNA), the development of which is largely attributable to him, and as the United States was embarking on the path of modernizing and extending its national accounts along the lines of the revised SNA." Statistical News (1992). As mentioned elsewhere, among the American contributors for work related to national accounting Simon Kuznets and Wassily Leontief were recipients of Nobel awards.

4. Irving Fisher’s influence Returning to the late 1920’s and the 1930’s, and to Copeland’s activities at the time, it should be mentioned that in 1927 Copeland joined the Division of Research and Statistics at the Federal Reserve Board. At the Federal Reserve Copeland’s first monetary analysis was a statistical study concerning Irving Fisher’s ’equation of exchange’. Copeland (1929). Later, in his 1935 article, he again returned to Fisher and stated: "Statistical determinations of national - or broadly of social - wealth and income offer economists the chance to make their basic concepts accord with scientific method.

Z. Kenessey/American Contributions

117

Irving Fisher perceived this clearly a quarter of century ago..." (p. 377). In his footnote to this statement Copeland cites Fisher’s famous work of 1906 about capital and income. We know of the attention of Erik Lindahl to Fisher’s accounting thoughts; and Aukrust related that Hicks "met Lindahl during his stay in London in 1934 and 1935. Hicks must have been impressed; in a retrospective article in Economica, 1973 (p.8) Hicks in a passing sentence refers to Lindahl as ’the father of social accounting theory’..." Aukrust (1992) p.4. Thus both via Copeland (as seen in Stone’s quote of 1947) and possibly via Lindahl and Hicks (the latter also figuring in the same footnote by Stone) we can surmise the ’seeping in’ of the Fisherian thoughts into the conceptual arsenal of Stone. It would carry us too far if we entered into an analysis of the complexities of Irving Fisher’s thought about the use of accounting in treating capital and income. In order to show, that he meant to extend the accounting treatment from individuals and single businesses to society at large, we restrict ourselves to a single quote from his book ’The Nature of Capital and Income’: "If, then, we suppose balance sheets so constructed as to include the whole world of real and fictitious persons, with entries in them for every asset and liability, even public parks and streets, household furniture, persons themselves, and other possessions not ordinarily accounted for in practice, it is evident that we shall obtain ...a complete account of the distribution of capital value among real persons; and a complete list of the articles of actual wealth thus owned." Fisher (1919) p.96. And he continues in the chapter entitled ’Income summation’: "By combining the net incomes of all persons, the net income of society may be obtained." Fisher (1919) p. 141. In his time Fisher was very widely read, not only in English but in German as well. Therefore his admonishment in his article entitled "The Income Concept in the Light of Experience" (first published in German) in the section about ’Income accounting’ must have reached a wide audience: "The student of income can scarcely go astray if he will take the trouble to learn the ordinary bookkeeper’s art of crediting and debiting...This is simple ’double entry bookkeeping’." Fisher (1928) pp. 4-5. The overall importance of bookkeeping for economic development, of course, was not a secret to other economists and economic historians. For example Fisher’s contemporary, Werner Sombart - who was at the time very widely read in Germany and elsewhere - accorded an important role for the evolution of capitalism to the role of measurement, commercial arithmetics, the ’Arabic notation’ (numerals), the ’Art of Calculation’, and last but not least to bookkeeping. (Kenessey, forthcoming). Before Sombart, as Koshimura pointed out, Marx "...in his letter of June 18, 1862, to Friedrich Engels he wrote of his intention of studying the Italian method of bookkeeping. Marx, however, did not adopt this double entry system of bookkeeping, either in the analytical study of Tableau Economique or in the development of his own schema of capitalist reproduction." Koshimura (1984) p.l. Perhaps as a reaction to the relative oblivion Irving Fisher has fallen among most contemporary economists, sometimes exaggerated claims are made about his contribution to national accounting. In Fisher’s new biography Robert C. Allen (1993) flatly suggests that "He created the national accounting systems of the United States and

118

Z. Kenessey/American Contributions

other countries" (p.13). In respect of Fisher’s ’Nature of Capital and Income’, published in 1906, however Allen’s formulation is more careful and balanced, yet still somewhat hyperbolic: "Economic statistics and accounts, including the national accounting arrangements used by every country in the world...rely on it...Fisher was the originator and founding father of all national accounting systems." (p.95). I believe that Fisher’s originality and impact on other thinkers (including Copeland in the U.S. and Lindahl in Sweden, who in turn influenced still others in the U.S., U.K., and Norway) is beyond doubt. But the creation of the national accounts themselves ought not to be attributed to him. Fisher was quite explicit in 1927 advising those interested in income research: "...the student of income can scarcely go astray if he will take the trouble to learn the ordinary bookkeeper’s art of crediting and debiting. Unfortunately there has been too much haste to get at the total income received by an individual from all sources to permit a patient study of each item of income flowing from each item of source." Fisher (1927) p.4. There can be no doubt, that Fisher was concerned not only with income of individuals (or single business entreprises). While most of his examples and explanations are on the micro-level, his macro-references are also there. A particularly clear one is the following: "...the entire income of the community. This is social income." Fisher (1919) p.U6.

5. Morris Copeland’s moneyflows Finally, it seems important to say a few words about Copeland’s money-flow or flow of funds system as well. This system is usually not included in the survey of historical developments in national accounting (for example in the rich survey of Carson). Yet, and not only because of Copelands just mentioned championing of accounting frameworks for national income estimating, this work seems to belong to the broad subject under review. Certainly, Jaszi (1986) thinks so: "By the mid-1950’s, it was widely recognized that alternative systems, such as flow of funds and financial balance sheets as well as input-output, might be useful for policymaking, and there was great concern that these systems should be integrated with the NIPAs [National Income and Product Accounts]." Jaszi (1986) p.412. It would go beyond the framework of this article to chronicle the establishment of flow of fund accounts by Copeland in the U.S.A. Yet it has to be mentioned that the National Bureau of Economic Research in 1944, at Wesley C. Mitchell’s suggestion (Millar, 1991), asked Copeland to study moneyflows of the American economy. As reported by Stephen Taylor "Copeland had done the actual statistical work for his book, in the later 1940’s, in office space of the Federal Reserve Board that had been donated to the project by the Board....Copeland’s work had strong Fed support from Winfield Riefler, who wrote an introduction to Copeland’s Moneyflows, and from Ralph Young, then Director of the Fed’s Division of Research and Statistics." Taylor (1991) p.l.

Z. Kenessey/ American Contributions

119

Copeland, in his famous study published in 1952 indicated that he became intensely interested in accounting in graduate school at the influence of Walter W. Stewart, who was his undergraduate teacher at Amherst. John C. Dawson reports about this matter the following: "Years ago when I was Copeland’s graduate student at Cornell, he told me he had asked his undergraduate teacher at Amherst, Walter W. Stewart, what he should learn as a graduate student. Stewart had replied, ’See if you can learn the difference between a debit and a credit.’ I believe this story to be significant...". Dawson (1991) p .l)1 Jacob Cohen provided in 1972 a retrospect on Copeland’s moneyflows on the basis of a quarter-of-a-century experience with them. Cohen summarized his evaluation of the expectations and the results related to Copeland’s innovations as follows: "The new accounting system was hailed as expanding the potential frontier of model-building.... Moneyflows accounting was expected to be as popular with the next generation of economists as the GNP accounts were with the current one."..."From the vantage point of a quarter-of-a-century, the fruits of Copeland’s work in moneyflows are impressive if not sensational. His interpretation of the moneyflows accounts exerted a strong influence on monetary theory. The later Federal Reserve flow-of-funds accounts are now a statistical staple. Nonetheless, it appears that Riefler and Mitchell were too optimistic in their forecast. What has happened is that the flow-of-funds accounts have attained a popularity comparable to the Commerce accounts but only as sources of financial data and as a basis for financial model-building." Cohen (1972) p.l and p.17.

6. Growth accounting by Denison As already mentioned earlier, Denison played an important role in setting up the modern national accounts in the U.S. In addition, he also established a related discipline, usually called ’growth accounting’. Indeed, his scholarly fame today is mainly related to the work undertaken by him in this direction. Our review cannot go into the development of this branch of research, which in the recent decade focused especially on the possible factors explaining the slowdown of productivity growth in some major industrial economies. Denison, who was a meticulous researcher, and who of course knew the data sources, methodology and scope of national accounts information, made the most out of these materials for establishing the weight of the main sources of economic growth. Similarly to Denison’s work, here we cannot go into the various proposals regarding the desirable improvements of the U.S. accounts. However, one of them, the book of Nancy Ruggles and Richard Ruggles (1970), at least should be mentioned in our overview.

120

Z. Kenessey/American Contributions

7. George Jaszi’s contributions

On his retirement from the U.S. Department of Commerce as Director of the Bureau of Economic Analysis, which has been responsible for the work on the U.S. national accounts, the following listing of the enhancements were highlighted in connection with his quarter-of-century directorship of BEA: "* The national income and product account estimates were provided on a more timely schedule, in more detail, and with more attention to separating changes in value into changes in prices and changes in constant-price measures. The amount of quarterly presentation in the Survey now consists of about 50 table. The 130 tables usually presented in the July survey provide even more detail. * Wealth accounts were developed as an extension of the income and product accounts. * The concept of the Federal budget in the framework of the economic accounts was forged. * The input-output work at the Department of Commerce was started in 1962. Its hallmark is the conceptual and statistical consistency of input-output tables with the other branches of BEA’s economic accounts. * Regional accounts were expanded to provide, in addition to the annual income measures already available for States, quarterly measures for States and annual measures for local areas. * The U.S. balance of payments accounts were enhanced by conceptual improvements, methodological changes to keep pace with rapid changes in the international financial system, and presentation of more detail. * Pioneering estimates of pollution abatement and control expenditures were prepared. * Forward-looking analysis that supplements BEA’s work on the economic accounts was strengthened. Within the Federal Government, BEA took a leading role in developing macroeconometric models of the U.S. economy." Survey of Current Business (1985) pp.1-2. My review of key American developments were terribly incomplete without a reference to the vast amount of practical experience, methodological acumen, and indeed wisdom that was accumulated during the many years of work for the establishment of national accounts in the U.S. The best reference in this regard, I believe, is to George Jaszi’s lecture in the mid-1950’s at the time of his retirement as Director of the Bureau of Economic Analysis, significantly entitled: ’An Economic Accountant’s Audit’. Jaszi (1986). I wish to refer only a few of his key admonishments, such as the one that aggregate data of business units "can be defined only in terms based on business accounting concepts", (p.412) My second selection is also reflective of long experience: "Reliance on the by-products of administrative operations secures inexpensively a large quantity of high-quality data.” (p. 413) My favorite selection is the following: "...A sustained effort is needed to teach users that the NIP As are eminently useful in macroeconomic analysis if they are not regarded as a precision instrument and that they may be lethal if they are." (p. 415) Jaszi’s 1986 view on the growingly difficult tasks of forecasting also deserves attention: "...I believe that forecasting error is likely

Z. Kenessey/American Contributions

121

to increase...It is likely to become more and more difficult to foresee exogenous factors such as international economic developments, wars, and the risks of wars." (p. 416) The sue years that elapsed since this particular forecast seem to justify Jaszi’s power of foresight in this matter. Jaszi also suggested, "...that forecasting may not be necessary to make economic accounting useful.” (p. 416). He pointed out that "In the natural sciences, no one questions the usefulness of the theory of evolution just because it does not tell us about future evolution. For economic history, a similar statement can be made." (p.416) Finally, I consider it the expression of both great wisdom and of profound aesthetic sensibility when Jaszi declares: "...If we regard economic accounting as an art, nothing is wrong with taking pleasure in having painted a good picture of the economy. This is a highly subjective affirmation to which I am deeply attached." (p.416) Indeed, I think we can view the complex developments of national accounting in the United States, in which George Jaszi played such a prominent role, as a great surge in creativity that produced outstanding results by him and others for the benefit of the U.S. and of the world community in this field. Such creativity was certainly not only characteristic for the American personalities. We have ample evidence regarding the outstanding European developments which prove this point for the other side of the Atlantic as well. After his death, his successor as Director of BEA summed up Jaszi’s contributions to U.S. national accounting during the first years of the accounts, that is in the period which is the subject of my review in the following way: "He was one of a team of four the others were the late Edward F. Denison, the late Milton Gilbert, and Charles F. Schwartz - who roughed out a sketch of the two-sided economic accounts that were prepared during World War II to provide information needed for economic mobilization. The same team prepared the first precise formulation of the accounting system in 1947 and wrote the first detailed explanation of its conceptual framework in 1951." Carson (1993) p.225. As mentioned earlier, many related developments could not be explored in our review here. To mention one, the American contributions to international comparisons of national income, product and purchasing power were quite substantial, indeed of leading importance at the international scene. First the work undertaken by Milton Gilbert and Irving Kravis in the OEEC and OECD was of great import. Later, the wider UN International Comparison Program also heavily involved Americans, such as Irving Kravis, Robert Summers, Alan Heston, and the author of this review. An overview of this work can be found in Kravis (1984), and the first UN report in Kravis etal. (1975). Similarly, I did not relate to our story the extensive scholarly work undertaken in the U.S. regarding national wealth (except in the reference to BEA’s including such estimates in their work). In this work the name of Raymond Goldsmith shines with particular illumination. I trust, that a more comprehensive work will be able to cover his work, and many other important contributions, which could not be part of this short overview.

Z. Kenessey/American Contributions

122

Endnotes 1. The quotes from the papers of Dawson, Millar and Taylor are from respective unpublished reports prepared for the Session honoring Morris A. Copeland, at the 1991 August meetings of the American Statistical Association in Atlanta.

References [1] [2]

Allen, R obert C. (1993) Irving Fisher. A Biography. Blackwell. Barber, William J. (1988) From New E ra to New Deal. H erbert Hoover, the Economist, and American Economic Policy, 1921-1933. Columbia University Press, New York.

[3]

Carson, Carol S. (1975) The History of the United States National Income and Product Accounts: The Development of an Analytical Tool. The Review of Income and W ealth, pp. 153-181.

[4]

Carson, Carol S. (1993) In Memoriam: George Jaszi. Review of Income and Wealth, June issue, p. 225.

[5]

Clark, Colin (1940; second ed. 1951) The Conditions of Economic Progress. Macmillan. London.

[6]

Cohen, Jacob (1972) Copeland’s Moneyflows After Twenty-Five Years: A Survey. The Journal of

[7]

Copeland, M A . (1929) Special Purpose Indexes for the Equation of Exchange for the United States,

[8]

Copeland, M.A. (1932) Some Problems in the Theory of National Income. Journal of Political

Economic Literature, Volume X, Number 1. 1919-1927. Journal of the American Statistical Association, June 1929. Economy, Vol. XL, February 1932. [9]

Copeland, M.A. (1935) National W ealth and Income - an Interpretation. Journal of the ASA, Vol. XXX, No. 190.

[10]

Copeland, M.A. (1937) Concepts of National Income. Studies in Income and W ealth, Volume One, NBER, New York.

[11]

Dawson, John C. (1991) Copeland on Social Accounting. Grinnell College. (Prepared for the Atlanta ASA Copeland Session).

[12]

Fisher, Irving (1919; first ed. 1906) The Nature of Capital and Income. Macmillan. New YorkLondon.

[13]

Fisher, Irving (1928) The Income Concept in the Light of Experience. English reprint. The original publication in Germ an was in Vol. Ill of the W ieser Festschrift, Die W irtschaftstheorie der Gegenwart, 1927, Vienna.

[14]

Fogel, R obert W. (1987) Some Notes on the Scientific M ethods of Simon Kuznets. W orking Paper No. 2461. National Bureau of Economic Research, Cambridge, M.A.

[15]

Jaszi, George (1986) An Economic Accountant’s Audit. Lecture on Economics in Government. AEA Papers and Proceedings. American Economic Review, Vol. 76, No. 2.

[16]

Kendrick, John W. (1970) The Historical Development of National-Income Accounts. History of Political Economy, II. Fall 1970.

[17]

Kenessey, Zoltan (1993) Postwar Trends in National Accounts in the Perspective of Earlier Developments. In the Value Added of National Accounting. Netherlands Central Bureau of Statistics.

[18]

Kenessey, Zoltan (forthcoming) The Scope of Output and Productive Sectors: Germ an Contributions in the 19th Century. In the volume of the W erner Sombart Conference, July 1991, Heilbronn.

[19]

Keynes, J.M. (1920) The Economic Consequences of the Peace. H arcourt, Brace and Howe, New York.

[20]

Koshimura, S. (1984) Capital Reproduction and Economic Crisis in Matrix Form . The Wako University, Tokyo.

[21]

Kravis, I., Kenessey, Z., Heston, A. and Summers, R. (1975) A System of International Comparisons of Gross Product and Purchasing Power. John Hopkins University Press, Baltimore.

Z. Kenessey/American Contributions

123

[22]

Kravis, Irving B. (1984) Comparative Studies of National Incomes and Prices. Journal of Economic Literature, Vol. XXII, M arch 1984.

[23]

Martin, Robert F. (1936) National Income and its Elements. National Industrial Conference Board. New York.

[24]

Millar, Jam es R. (1991) Copeland on Economics: Institutionalism from a Natural Science Point of View. (Prepared for the Atlanta Session of the ASA in honor of M. Copeland).

[25]

Ruggles, Nancy (1987) Social Accounting. In the New Palgrave, Vol. 4. Macmillan.

[26]

Ruggles, Nancy and Ruggles, Richard (1970) The Design of Economic Accounts. National Bureau of Economic Research. Columbia University Press, New York.

[27] [28]

Statistical News (1992) Issue 97, p. 97.58 Obituary of Sir Richard Stone. Stone, Richard (1947) Definition and M easurem ent of the National Income and Related Totals. Appendix to the R eport of the Sub-Committee on National Income Statistics of the League of Nations Committee of Statistical Experts, entitled: M easurement of National Income and the Construction of Social Accounts. Studies and Reports on Statistical methods, No. 7. United Nations, Geneva, 1947. (Stone’s study - a revised version of his 1945 paper for the famous Princeton meeting is the largest part of the U N publication).

[29]

Survey of Current Business (1985), February Issue. (Untitled introductory statem ent on first two

[30]

Taylor, Stephen (1991) Moneyflows after Copeland: Forty Years of Flow-of-Funds Accounts.

pages). (Prepared for the 1991 Atlanta ASA Session in honor of M. Copeland).

The Accounts o f Nations Z. Kenessey (Ed.) IOS Press, 1994

124

DEVELOPMENT OF NATIONAL ACCOUNTS IN INDIA S.G. Tiwari New Delhi, India

Abstract. T he paper surveys the development of national income and regional accounts in India including statistics on income distribution and poverty. It presents the framework of regional accounts developed and the technique employed in obtaining income distribution statistics under conditions of non-availability of reliable data on personal income. Finally it presents techniques employed for assessing incidence of poverty over time and space.

Introduction In a developing country like India of half the size of Europe, a federal set-up, a large public sector, and an established planning process since 1950, the nature and type of national accounts have been conditioned to meet its concerns and growing needs. The objective of the present paper is to briefly indicate first the early efforts made before the establishment of an integrated system of national accounts, then the work on establishment of a framework of an integrated system of national accounts through the recommendations of the National Income Committee; and the permanent establishment of a National Income Unit (subsequently known as National Income Division and finally National Accounts Division) to compute regular series. Establishment of the Central Statistical Organization (C.S.O) and the National Sample Survey Organization, State Statistical Bureaus and a coordinating machinery between central and state statistical organizations immensely helped collection and improvement of data needed to meet the requirements for mational accounts as well as state income and other aggregates. The work done by individuals (including the present author), research institutions, government departments including C.S.O. over the years in the field of national and regional accounts is voluminous and cannot be described in a brief paper such as this. The paper, therefore, will refer only to select contributions in certain fields including areas of concern, such as income distribution and poverty.

1. Early attempts during nineteenth century Estimation of national income in India in the 1950’s was not a new development. As many as seven estimates of national income are available for the second half of the

S.G. Tiwari/National Accounts in India

125

Table 1: Estimates of National Income in India, 1857-1900*

Author

( i)

Title of source

(2)

Dadabhai Naorogi

Poverty and UnBritish rule in India, 1871

Atkinson, F J.

A Statistical Review of the Income & W ealth of British India, JRSS, June, 1902

M ajor Baring

Budget Speech, 1882

Richard, T.

A paper read before the Institute of Bankers, London, July, 1881

Horne, E.A.

Atkinson, FJ.

Digby, W.

An estim ate of India’s National Income, Bengal Economic Journal, January, 1918

A rea covered

Year of Ref.

(3)

(4)

(5)

186768

340

British India

-

National Income (Rs. crores)b

Per capita Income Rs. (6)

1875 375

20

1881

525

30.5

1881

223

27

1891

A Statistical Review of the Income and W ealth of British India, JRSS, June 1902

1895

877

"Prosperous" British India, 1901

189899

428

10 (not veri­ fied)

39.5

18.0

a Quoted from M. Mukherjee, A Preliminary Study of the Growth of National Income in India, 1857-1957, In Asian Studies in Income and Wealth, 1965. b Crore is equivalent to 10 million.

126

S.G. Tiwari/National Accounts in India

nineteenth century. These are given in Table 1. Although they were not comparable as they were on different source material based, some on detailed data, and some others on limited information. While some are over some others are under estimates. Efforts of the Indian thinkers was to use national income data to prove that during this period of British rule, economy was allowed to remain agrarian and plantation and no industrialization took place with the result that there was heavy drain on the economy resulting in the poverty of the people. The rulers on the other hand tried to prove through national income estimates (which were later found to be over estimates) that the country had progressed. Thus these estimates were used by Indians to discredit the government, while with additional data government tried to claim credit.

2. Estimation attempts up to the Second World War Estimation of national income work gathered momentum with the beginning of the twentieth century. As many as twenty seven estimates are available either for India as a whole or for British India (see Table 2). Of these, two relate to a province by the present author for two different years. These estimates were used for varying purposes, viz., assessment of economic conditions, and the performance of the government, assessment of taxable capacity, progress or decline of the country and to judge the standard of living of the people. It was during this period that V.K.R.V. Rao made the first scientific estimate of national income based on published material as well as ad hoc collection of extensive data for 1931-32 for British India. This pioneering estimate was based on the income method. The estimate for a province by the present author made use of both income and product approach to cross-check one estimate against the other.

3. Work up to the establishment of National Income Committee, 1949 This was the period when the office of the Economic Advisor, Ministry of Commerce commenced compiling annual estimates of national income from 1945-46 although a number of research workers continued to bring out their own estimates which are as given in Table 3.

4. Establishment of National Accounts Groundwork for authoritative series estimates of national income and its various components was laid with the appointment of National Income Committee by the Government of India in August 1949 with Prof. P.C. Mahalanobis as Chairman and Prof. D.R. Gadgil and Dr. V.K.R.V. Rao as members. In its work the Committee was

S.G. Tiwari/National Accounts in India

127

assisted by the advisory services of Prof. Simon Kuznets, Prof. J.R.N. Stone and Dr. J.B.D. Derksen. The Committee was required to prepare a report on national income and related estimates, to suggest measures for improving quality of available data and for the collection of further essential statistics, and to recommend ways and means of promoting research in the field of national income. For the secretarial work of the Committee, the National Income Unit was created in the Ministry of Finance. The Committee submitted its First Report in April 1951 and the Final Report in February 1954. In its First report the Committee presented estimates of net output of the Indian Union as well as the net output per engaged person for the year 1948-49. In this report the Committee also presented a framework of social accounts which was drawn up by Prof. Richard Stone. The Committee also presented accounts for the year 1948-49 in which all the entries were filled other than private consumption expenditure, private saving and depreciation. In its Final Report the Committee gave in detail the methods for compilation of sectoral estimates and presented national income estimates for 1948-49 (improved), 1949-50 and 1950-51. From this time onwards commenced the regular compilation of official estimates of national income in order to finally construct national accounts. For work on national accounts the Committee recommended the setting up of an Advisory Committee on National Income. In January 1957, a Preliminary Conference on National Income was convened in which research workers from various Government Departments, Central Bank, research institutions and universities participated and decided to form the Indian Conference on Research in National Income1, which after its third Conference in 1962 was converted into the Indian Association for Research in National Income and Wealth (IARNIW) which meets every two years and holds a seminar between the years of the Conference. Senior research workers as well as heads of certain research institutions such as the Indian Statistical Institute, Gokhale Institute of Politics and Economics, Delhi School of Economics, Reserve Bank of India, Central Statistical Organization, Ministry of Finance, some State Statistical Bureaus, etc. took it upon themselves of contributing papers from their staff to the Conference. In addition papers have been invited from known researchers in selected fields. Selected papers from among those read at the preliminary, First, Second and Third Conference were published in National Income and Allied Topics Vol. I, II and III brought out by Asia Publishing House. The Association has been holding conferences and seminars regularly and publishes selected papers in its half yearly Journal of Income and Wealth. For the development of work in national and regional accounts the Government of India appoints from time to time standing working groups, expert groups, etc., which have as their members very eminent persons. There is a standing Advisory Committee on National Accounts which had been advising on all aspects of national and regional accounts and measures for improvements of the quality of the data and for the collection of further essential statistics.

128

S.G. Tiwari/National Accounts in India

Table 2: Point Estimates of National Income in India, 1901-39®

Author

(1)

Title of Source

(2)

Area Covered

Year of Ref.

National Income (Rs. Crores)

Per capita income (Rs.) Current Prices

(3)

(4 )

(5)

(6) 30

l.Curzon

Budget speech 1901

British India

1901

675

2 .Giffen R.

Economic Enquiries and Studies

India

1903

900

30

3.Horne E.A.

An Estimate of India's National Income. Bengal Economic Journal 1918

British India

1911

989

42

4.Shirras G.F.

Report on an Enquiry into Working Class Budgets in Bombay

British India

1911

1920

80

5-Balakrishna

Industrial Decline of India, 1917

India

1911 -12

539

21

6.Vakil and Muranjan

Currency and Finance in India, 1926

India

1911 -14

1774

58.5

7.Wadia & Joshi

The Wealth of India, 1925

British India

1913 -14

1087

44.5

8.Lupton A.

Happy India, 1922

British India

1919 -20

2854.5

9.Slater G.

The Madras Year Book, 1923

Madras

1919 -20

••

lO.Sarkar B.N.

Proceedings of the Council of State, 1921

British India

1921

••

11.Shah K.T.

Trade, Tariff and Transport in India, 1924

British India

1921 -22

1470

46

12.Shah & Khambata

Wealth and Taxable Capacity in India, 1924

India

1921 -22

2364

74

13.Tiwari S.G.

Economic Prospe­ rity of U.P.1951

U.P.

1921 -22

••

107

114 112 Over 100

Continued.

S.G. Tiwari/National Accounts in India

129

Table 2: Point Estimates of National Income in India, 1901-30® (continued)

Author

Title of Source

Area Covered

Year of Ref.

14.Shirras G.F.

Poverty & Kindred Economic Problems in India, 1935

Brit. India

1921 -22

2866

116

15.Rao V.K.R.V.

An Essay on India's National Income, 1940

Brit. India

1925 -29

2068

80

16.Shirras G.F.

Poverty & Kindred Economic Problems in India, 1935

Brit. India

1926 -27

2824

115

17.

Central Banking Enquiry Committee, 1931

Agr. Popu­ lation

1928

••

42

18.

Simon Commission Report, 1929

India

1929

116

19.

Quoted by Davis: Population of India and Pakistan

Brit. India

1931

63

20.Rao V.K.R.V.

The National Income of British India

India

1931 -32

1689

62

21.Desai R.C.

The Standard of Living in India and Pakistan

Brit. India

1931 -32

2809

72

22.Tiwari S.G.

Economic Prospe­ rity of U.P.1951

U.P.

1931 -32

50.5

23.Grigg J.

Quoted by Davis: Population of India and Pakistan

Brit. India

1937 -38

56

24.Tiwari S.G.

Economic Prosperity of U.P. 1951

U.P.

1938 -39

••

56.1

25.Student

Commerce, Dec. 1943

Brit. India

1938 -39

1865

65

26.Natarajan B.

An Essay on National Income & Expenditure in India, 1949

Union Pro­ vinces

1938 -39

1482

68.5

27.

The Eastern Economist

••

1939 -40

1924

67

Natio­ nal Income (Rs. Crores)

Per capita income Current Prices

National Income in India, 1857-1957, in Asian Studies in Income and Wealth, 1965. b Crore is equivalent to 10 million.

S.G. Tiwari/National Accounts in India

130

Table 3: Point Estimates of National Income for 1940-41 to 1949-50a

Author

Title or Source

Area Covered

Year of Ref.

Government of India

British India

194243

3433

114

2.

Commerce Dec. 1943

British India

194243

4265

142

3.Adarkar and Tandon

Mimeographed Ministry of France

India

194445

5060

171

4.Saxena D.N.

Ph.D. Thesis London University

British India

194546

5430

224

5.Natrajan, B.

An Essay on National Income & Expenditure in India, 1949

Union Provin­ ces

194950

5858

229

1

.

National Income (Rs. crores)

Per Capita Income (Rs.)

a Quoted from M. Mukherjee, A Preliminary Study of the Growth of National Income in India, 1857-1957, in Asian Studies in Income and Wealth, 1965. b Crore is equivalent to 10 million.

S.G. Tiv/ari/National Accounts in India

131

5. Structure of National Accounts

As stated earlier the initial structure of national accounts was contained in the first report of the National Income Committee, April 1951. These are as below: (a) Domestic Product Account (b) Private Appropriation Account (c) Government Appropriation Account (d) Consolidated Resting Account, and (e) Rest of the World Account The structure of National Accounts in India has been based on the conditions and circumstances of a developing sub-continent with federal structure and varying resource endowment, climate, economic, social, environmental, political and cultural settings. Over the years this structure of national accounts got deepened and widened as a result of further thinking, research work done in the country, and international standards (SNA). In this process Indian national accounts got influenced by the United Nations recommendations and the latter in turn got shaped in several fields by Indian experience. An expert from India always worked as a member of expert group for 1953 SNA, 1968 SNA, and 1993 SNA. It is, therefore, natural that both the UN and Indian SNA should be governed by a mutuality of interest. After the adoption of the socialistic pattern of society, interest in India was also directed in working out some of the components of Material Product System in the context of planning process of the country. In fact it was also considered that material and non-material components be shown in Indian SNA. Thinking on the structure of national accounts continued after 1953 SNA through 1968 SNA and it finally got concretised in a Seminar organized in Mahabaleshwar in January 1973 on whose name the accounts are named, that is, Mahabaleshwar Accounts. Special mention may be made of a paper by M. Mukherjee and Prasada Rao[2], and these accounts were meant to meet immediate needs of planners and policy makers. Main features of these accounts were: (a) The accounts of the nation were divided into three sections viz. (i) Consolidated Accounts of the nation (ii) Accounts of the Public Sector and (iii) Supply and disposition of goods and services. The Consolidated Accounts of the Nation were adopted from 1968 SNA with some additional breakdowns. In order to demarcate non-market transactions separately, compensation of employees was shown both in cash and kind. As a new item income of the self employed was introduced under factor income which included primary incomes of own account workers as in their case labour income cannot be separated from operating surplus. In the subsequent work on the subject these authors in collaboration with Uma Dutta Roy Choudhury termed the income of the self-employed as mixed income in order to indicate the character of income[3]. Further private final consumption expenditure and gross fixed capital formation out of own production have been distinguished from that originating out of purchases from the market. (b) In respect of national disposable income and its appropriation that part of saving

132

S.G. Tiwari/National Accounts in India

which primarily takes place in the rural areas in the form of physical assets with the use of own labour and material and with no corresponding financial transaction is shown separately from saving in financial assets. (c) In the list of supporting tables more important additions were (i) break-downs of gross domestic product by rural and urban areas, (ii) details of labour force and employment, (iii) size distribution of household consumption expenditure, and personal income, and (iv) gross domestic product by character of enterprise. An important characteristic of Indian National Accounts is that it has given precedence to national income/net national product as conceived in 1953 SNA over Gross Domestic Product as propounded in 1968 SNA. In the national accounts in India, households were always considered not only consumers but also producers. In fact a very large volume of production activities are carried out by households in the field of agriculture, household and cottage industries, construction, road transport, restaurants and storage, retail trade, finance, community and personal services. The same is true in the case of most developing countries. However, neither 1953 nor 1968 SNA provided for it. In 1953 SNA bank service charge was treated as input in the user industries. 1968 SNA deviated from it and to simplify matters provided for imputation of bank service charge. This change has not been accepted in India where the 1953 procedures continued to be followed. The subject of reproducible tangible wealth and flow of funds accounts received the attention of research workers in India from late fifties for purposes of estimation of depreciation, capital output ratio, and savings. It finally culminated in the establishment of official estimates of capital stock in India[4], and their use in estimation of consumption of fixed capital in the official national accounts statistics. However, it appears that capital losses and destructions still remain to be allowed. So is also the case with destruction through natural calamities (affecting particularly agriculture and rural housing) as well as accidents (as in railways). A framework of flow of funds was established by a Working Group on Flow of Funds in early 1960’s which had before it a background paper from Prof. H.W. Arndt of Australian National University and the work was continued at periodic intervals on the financial flows of Indian Economy. Estimates for real flows were made only intermittently. Early compilation of an input-output table was undertaken in 1950’s in the Indian Statistical Institute and subsequently by Planning Commission with the help of individual research workers from the Ford Foundation, CSO, other official agencies and research institutions for use in determining Five Year Plan targets. The first inputoutput transactions table consistent with national accounts for 1968-69 was jointly prepared by the CSO and Planning Commission and published in 1978. Subsequently the CSO on its own prepared the 1973-74 table and decided to prepare such tables every five years. Thereafter 1978-79 and 1983-84 tables were published and work is in progress for the 1989-90 table.

S.G. Tiwari/National Accounts in India

133

6. Development o f National Accounts

Development of national accounts in India has been jointly undertaken by individual research workers, academic institutions and government - Center and States. This commenced with the convening of Preliminary Conference on National Income in 1957 on one hand, and the constitution of the Standing Working Group on State Income and the appointment of the Advisory Committee on National Income in 1958 on the other. It was recognized that for the development of national accounts it was necessary to expedite compilation of State income estimates. In fact national income estimates for goods producing sectors are first built at State level and then aggregated to obtain national totals. Pioneering efforts were made by research workers in the preparation of papers for the Preliminary2 and First3 National Income Conference, inspite of a weak data base. Over the years the data base improved permitting preparation of more reliable estimates. Some of the papers (with their authors) for Preliminary and First Conference are being mentioned here to give an idea of the fields covered. These are as below: 1. B.K. Barpujari: The National Accounts of India 1951-52 to 1955-56 2. Baldev Kumar: Estimates of Fixed Capital Formation in India, 1948-49 to 1954-55 3. K.V.R. Avadhani, A.K. Ghosh, R.M. Honavar and M.L. Trikha: Savings in the Indian Union 1949-50 to 1954-84 4. Uma Dutta: Increase in Inventories in relation to Fixed Capital Formation and National Income 5. Uma Dutta: An Estimate of Reproducible Tangible Wealth in India; 1949-50 6. S.P. Dhar, M.G. Kanbur, K.N.C. Pillai and V.S. Puri: Private Consumption Expenditure in India, 1949-50 7. M.J.K. Thavaraj: Capital Formation in the Public Sector in India, a Historical Study 1898-1938 8. S.M. Kansal: On the Deflation of National Income in India by the Final Expenditure Approach 9. S.K. Chakravarti, Uma Dutta and V. Srinivasan: Share of Urban and Rural Shares in the Domestic Product in India 1952-53 10. N.C. Arora and K.R.R. Ayengar: Long Term Growth of National Income in India; 1901-1956 11. K. Mukherjee: A Note on Long Term Growth of National Income in India, 1900-01 to 1952-53 12. M. Mukherjee: On the Appropriateness of the Price Average to be used in obtaining the value of output in Agriculture 13. M. Mukherjee: On the Available Estimates of the Breakdown of National Income by Distributive Shares in India 14. V.V. Divatia: On Computation of State Income 15. Bhanu Koti: Inter State and Inter Regional Balance of Trade: Methodologicaland Practical Problems with special reference to India 16. Indian Statistical Institute: Inter-Industrial Relations in the Indian Union, 1951

134

S.C. Tiwari/National Accounts in India

On conversion of the Conference into Association, contributions for the development of national and regional accounts continued unabated in each Conference and Seminar. The subjects covered are given in the Appendix. It is not possible to describe here the large volume of contributions by individual authors spanning over three decades. But all these materials were found extremely useful by the National Accounts Division, CSO, to improve the methodology and data base for the compilation of various national and regional aggregates and their breakdowns. A couple of contributions made through other organs at the time were: (a) M. Mukherjee and N.S.R. Sastry: An estimate of Tangible Wealth of India (Income and Wealth Series VIII, 1959). (b) V.V. Bhatt: Savings and Capital Formation (Economic Development and Cultural Change, Vol III Part 1 No.3, April 1959). (c) M. Mukherjee: A Preliminary Study of the Growth of National Income in India, 1857-1957, Asian Studies in Income and Wealth, 1965. In respect of official activities, the National Income (Accounts) Division has been very active in deepening and widening its area of activities, that is from national accounts to regional accounts, input-output table, capital stock, depreciation etc. In this process the data base is being expanded by the National Sample Survey Organisation, State Statistical Bureaus, a number of Government Departments and the Reserve Bank of India. An Advisory Committee on National Income was constituted in 1958 with Prof. P.C. Mahalanobis as Chairman, and five members, viz. Prof. VKRV Rao, Prof. D.R. Gadgil, Dr. I.G. Patel, Prof. M. Mukherjee and the author (as in-charge of National Income Unit) as Member Secretary. The terms of reference of the Advisory Committee were: (a) to advise on the preparation of national income and related aggregates, (b) to suggest measures for improvement of the quality of data and for the collection of further essential statistics, and (c) to recommend ways and means for the promotion and co-ordination of research in national accounts. The Committee has been functioning since then and has advised on the methodology of estimation of sectoral net output, revision of base, savings, capital formation/capital stock, non-profit institutions, system of national accounts, regional accounts, international/inter-state comparison of prices, net material product, research/ improvements, etc. An expert group on Savings under the chairmanship of the eminent economist Prof. K.N. Raj was appointed in May 1981 in order to (i) critically review the available estimates of investment and savings in the economy, both of the aggregates and their components (ii) to recommend improvements in the methods and procedures of estimation including suggestions for building up regional estimates. A Committee on Distribution of Income and Levels of Living under the Chairmanship of Prof. P.C. Mahalanobis worked during 1960 to 1969 in order to: (a) review the changes in levels of living during the First and Second Five Year Plans (1951-52 to 1959-60),

S.G. Tiwari/National Accounts in India

135

(b) study recent trends in the distribution of income and wealth and in particular, (c) to ascertain the extent to which the operation of the economic system has resulted in the concentration of wealth and means of production. With the advice available from various Committees, Expert Groups, the National Accounts Division undertook a detailed exercise to review the data base and methodology and discussed its proposals for a Revised Series of National Product for 1960-61 to 1964-65 in a Seminar, and after incorporating its suggestions came out with a revised series wit 1960-61 as the base[5]. Work on estimation of capital formation and saving continued. Among others references may be made to a paper by S.G. Tiwari, B. Kumar and J. Kumar on "Estimates of Capital Formation in India, 1950-51 to 1961-62", ICRNIW, 1963. On completion of this exercise and its discussions specially with the Reserve Bank of India which had been compiling estimates of saving over the years, the official saving estimates were finalised and incorporated into CSO publication on national accounts. Subsequently the series on the National Accounts was revised with 1970-71[6] as the base and lately with 1980-81 [7] as the base after reviewing the methodology and databases as well as undertaking several developmental studies. Each time a publication on Sources and Methods too was brought out.

7. Regional Accounts Estimates of regional income were initially developed by individual research workers for different states namely B. Natarajan for British Provinces in undivided India, through allocation method for 1938-39 and 1949-50; by the present author for United Provinces by rural and urban areas at current and constant prices for 1921-22, 1931-32 and 1938-39 based on extensive collection of data; by Mr. V.V. Divatia for Bombay State for 1938-39 to 1948-49 using extensive statistical material. A number of other authors also compiled estimates either based on limited data or obtained through allocation method. A number of research institutions such as National Council of Applied Economic Research (NCAER) etc. followed suit. Some of the state Govern­ ments e.g., Uttar Pradesh (U.P.), commenced compiling their own estimates. They were, subsequently taken up by various State Statistical Organizations with a view to produce and publish regional/state income estimates. As a result of recommendations of a Preliminary Conference in 1957, an active working group on State Income was formed whose function was to examine methodology and, source material of State income estimates and suggest ways for further improvements. The NCAER undertook comprehensive study of state income at state level and even computed income by districts to determine techno-economic feasibility and thus contributed greatly to development of methodology. The Indian Association for Research in National Income and Wealth also organized a special seminar in Poona on state income in 1962 where conceptual and data problems of state income estimation

136

S.G. Tiwari/National Accounts in India

were discussed in detail. In another Seminar of the Associations in Mahabaleshwar the subject was again discussed and a format of Regional Income and Accounts was drawn which included tables relating to (a) State Domestic Product and its component series (b) State Domestic Expenditure and its component series and (c) Important topics of price deflators, employment pattern and development indicators. The main gaps at the state level related to tables on (i) saving (ii) change in stock and (iii) external trade and finance. A Regional Accounts Committee was appointed by Government of India. It had worked during 1972-76 and drew heavily on the Mahabaleshwar Accounts of the State. It recommended the following accounts:(i) Consolidated accounts of the region (ii) Income and outlay account (iii) Capital finance account (iv) Household Accounts: (a) Income and outlay account (b) Total consumption and income of the population (v) Accounts of State and Local Government: (a) Production account of State Government Departmental Enterprises. (b) Income and outlay of State Government Administrative Departments and Departmental Enterprises. (c) Capital finance account of State Government Administrative Departments and Departmental Enterprises. (d) Capital finance accounts of State Government and non-Departmental Enterprises. Net State Domestic Product has been arrived at by use of depreciation provided for income tax returns, or percentages of output, etc. Recently with the estimation of depreciation for national income estimation with the help of capital stock estimates derived through use of perpetual inventory method, the State Statistical Bureaus too have been advised by the CSO to estimate depreciation with the help of perpetual inventory method by building up capital formation estimates over a long period of time. However, this has not been practicable and hence it is being examined if a national wealth survey could be undertaken for this purpose. States have been compiling estimates of State Domestic Product at current and constant prices on the basis of standard methodology laid down by Standing Working Group on State Income for over twenty years. Most of them have also compiled estimates of fixed capital formation by industry of use[8].

S.G. Tiwari/National Accounts in India

137

8. Sub-regional Accounts and District Income A number of States are very large, and have varying agro-climactic conditions and resource endowments. It has, therefore, been considered necessary to compile sub­ regional estimates of output and capital formation. Data on household consumption expenditure has already been collected annually through sample surveys. For balanced regional development data are being collected on various social indicators at district level. Of these indicators output/income is considered by districts for use in planning for reduction of regional and sub-regional imbalances.

9. Regional Disparities Regional disparities cannot always be assessed with the help of GDP per capita at current and constant prices. In a country of the size of India with varying agro-climactic zones, mountainous, plain, desert and coastal regions, there are differences in the purchasing power of Indian currency in different states. Hence much before the commencement of UN Comparison Project for estimation of purchasing power parity (PPP), methodological work was initiated by the author of this paper following the Gilbert Kravis approach. Considering that such PPP has relevance in areas with similar pattern of production/consumption, it was observed that valid comparison should be separately done for rural and urban areas. An examination of data revealed that rural Maharashtra was similar to rural Madras (now called Tamil Nadu) and urban Maharashtra to urban Bengal. The study was undertaken for two adjoining states, viz. Maharashtra and Madhya Pradesh for which data were available. The study revealed that the value of output of the two states at each others prices differed between 19 and 21 percent. In continuation with the work of UN International Comparison Project (ICP) in which India is not only a participating but a core country for regional comparison, it was decided to determine the inter-state variations in the purchasing power of rupee and experimental work was done for a number of years.

10. Income Distribution and Level of Living Extensive work has been done in India on developing statistics of income distribution and level of living. Data on income for income distribution studies has been found to be highly under-estimated. This happens not only in India but also in most developing countries where income tax is levied. M. Mukherjee and G.S. Chatterjee[9] in the study of trends in distribution of national income 1950-51 to 1965-66 first worked out productivity per worker by economic sectors. Other distributions related to distribution by industrial origin, distribution by size, distribution by final use, and regional distribution.

138

S.G. Tiwari/National Accounts in India

The measure of inter-sectoral disparity adopted was half the sum of the absolute difference between sectoral percentage composition of the national product and the labour force. In the study on the distributive shares on national income were taken as (a) income from assets (b) participation income (c) wages and salaries and (d) income of the selfemployed. Here income of the self employed constitutes about half of national income. Size distribution of consumption expenditure was based on National Sample Survey data and was also taken as distribution of income in view of non-availability of data on size distribution of saving. However, for determining indicators of real shifts, per capita expenditure in various expenditure classes was divided by price index numbers of cereals consumed by these classes and then calculated distributions afresh and obtained the measure of disparity. Regional disparities were also viewed with the help of rate of growth of nominal and real per capita income of the states. H.F. Lyall[10] in studying inequality of Indian incomes, 1961, obtained the distribution of personal income by integrating the size distribution of consumer expenditure obtained from NSS with size distribution of income before tax, obtained from the income tax data through graphical method. Mahfooz A hm ed[ll] further developed it by using a Pareto function. Subsequently Mahfooz Ahmed and Nikhilesh Bhattacharya did a vigorous and improved exercise using assumptions made by Lydall and Ahmed and brought all the data on income to correspond with the period of consumer expenditure. The Committee on Distribution of income and level of living put forward a new technique in assessing changes in the level of living with respect to consumption over two plan periods of ten years. It first estimated the demand for consumption of the present and increased population and the balance to be available for increased consumption and saving. The Committee made extensive use of the fractile graphical analysis developed by Prof. P.C. Mahalanobis[12] in assessing the level of consumption of persons in various size groups of expenditure. As regards level of living the Committee opined that there are many aspects of the level of living such as availability of food, clothing, housing, employment, education, health care, social and cultural amenities and viewed the subject with the help of all these indicators.

11. Poverty The subject of poverty has been extensively studied in India from the time of Dada Bhai Nauroji in the 19th century to the present day. Several economists and statisticians have studied the subject, determined the poverty line and estimated the number of poor for a year. The names of the authors and the poverty norm used by them is given in Table 4[13].

S.G. Tiwari/National Accounts in India

139

Poverty has been estimated by adopting a norm which either relates to per capita consumption of a given quantity of foodgrains and income needed for it (for absolute poverty) or income/expenditure needed to provide a given intake of calories per day. Some authors included other essential needs as well. In this exercise some authors used a general norm per person while some others allowed for lower needs for women and children. Some authors concentrated on rural areas alone where poverty is known to be rampant while some others on the country as a whole as well as for rural and urban areas. B.S. Minhas, S.M. Kansal and L.R. Jain studied incidence of urban poverty in different states (1970-71 to 1983), whereas B.S. Minhas and L.R. Jain studied incidence of rural poverty for the same period based on the use of NSS data for (NSS) years 1972-73, 1973-74, 1977-78 and 1983 with 1970-71 as the base. B.S. Minhas, L.R. Jain and S.D. Tendulkar studied the incidence of poverty in the 1980’s by constructing appropriate cost of living indices which allowed for the observed variations in consumption patterns and prices across states, the state specific rural and urban poverty norms as well as the corresponding estimates of the incidence of poverty in 20 states and all India in 1987-88 were presented[14]. Appropriately computed incidence of poverty in 1987-88 (affecting appropriate price adjustment at the state level for the relevant poor population) came to about 48.7 and 37.8 per cent respectively in rural and urban India. In work on this subject some persons used norms of calories from foodgrains as they alone are mostly available to the rural poor, while others have considered all items of food consumption. Over the years the standard norm recommended by the Indian Council of Medical Research (ICMR) for average calorie intake has changed from 2500 calories per person per day in 1945 to 2200 calorie intake for 1988 in view of change in the food habits, general improvement in health and the average age of the people. Work by different authors on poverty is related to one single year or to NSS computing survey years. In view of differences in methodology and sources data on income and expenditure which themselves may not be comparable because of various limitations, data on extent of poverty could not become available on time series basis. In view of widespread poverty in the country, the Government of India in the Planning Commission has been undertaking exercises to assess the effect of policy measures in providing employment opportunity to the poor through various programmes to lift them above the poverty line, and has tried to put up annual data on the percentage of people below the poverty line. This implies that in place of a single year estimate, the Government has been aiming at time series data. Eminent persons in 1962 had taken Rs. 20/- per capita expenditure at 1960-61 prices as the base presuming that free education and health facilities as well as an element of subsidy on urban housing for the poor would be provided by the state. This norm was used in preparing India’s perspective plan for 15 years (that is, 1961-62 to 1975-76). Subsequently too this norm has been carried forward to other years using the cost of living index.

140

S.G. Tiwari/National Accounts in India

Table 4:

Year

Estimates of Poverty - Percentage of Population below the Poverty Line

Poverty Norm

Population below poverty 1ine (%)

Source of Data

Author

NA

40 (R)

Village Surveys

H.H. Mann

1948

2. 1955 -56

Monthly Income less than Rs.20 per capita

44(1)

H.F. Lydall

1960

3a

19601961

Foodgrains con­ sumption of 518 grammes (1,800 calories) per day per person rural & 432 grammes (1,500 calories) urban

52 (R) 8(U) 44(1)

NSS

P.D. Ojha

1970

3b

19601961

Per capita con­ sumer expenditure less than Rs.15 per month rural, Rs.22.5 per month urban

38-40 (R) 51 (U) 41(1)

NSS

V.M. Dandekar & N.Rath

1971

Per capita con­ sumer expenditure less than Rs.15 per month rural, at 196061 prices

45 (R)

NSS

P.K. Bardhan

1973

1973

1 . 1918 &

1948

4. 196465

••

5a

19681969

-do-

54 (R)

NSS

P.K. Bardhan

5b

19681969

-do-

46

NA

Padam Singh

Year of estima­ tion or publi­ cation

1981 (unpub­ lished paper)

6. 197172

Energy intake V.M. 1981 less than 2,750 Dandekar calories per 46(R) NSS day per person NOTE: R = Rural; U - Urban; I = All India; NSS = National Sample Survey; CSO = Central Statistical Organization; NA = Not Available Continued...

S.G. Tiwari/National Accounts in India

141

Table 4: Estimates of Poverty - Percentage of Population below the (continued) Poverty Line

Year

7. 197778

Poverty Norm

Per capita consumption expenditure of Rs.49 per month rural (2400 calories per day) & Rs.75 per month urban (2,100 calories per day) at 1973-74 prices

Author

Year of esti­ mation or publi­ cation

Population below poverty line (%)

Source of Data

51 (R) 38 (U) 48(1)

CSO

Ministry of Program Implemen­ tation Govern­ ment of India

1988

8. 198384 (Provi­ sional )

-do-

40(R) 28(U) 37(1)

NSS

Planning Commis­ sion (7th Five Year Plan)

1985

9. 198788

-do-

44.88(R) 36.52(U) 42.70(1)

NSS

B.S. Minhas L.R. Jain S.D. Tendulkar

19871988

NOTE: R= Rural; 0 = Urban, I = All India; NSS = National Sample Survey; CSO = Central Statistical Organization; NA = Not Available

142

S.G. Tiwari/National Accounts in India

Some years back the Government of India constituted a high power expert Group to advise on a more recent base instead of 1960-61 for estimation of poverty norm. The Expert Group recommended 1973-74 to be the new base and suggested that in working out inter-state poverty, the price differences within states should be taken into account as was done by Minhas and Jain.

Appendix Subjects covered in Conferences and Seminars of the Indian Association for Research in National Income and Wealth during 1962-1991 - Regional accounts, regional disparities, measures of development, welfare and economic growth - Short-term forecasting - Public Sector accounts - Monitoring of Indian Economy - Data-base of national and regional accounts - National and regional accounts by institution - Review of national and regional sectors accounts by institutional sectors - ICP - Reliability of national and regional accounts - Non-monetised transaction - Service - public and/or private - Underground economy - Capital stock and capital consumption - Technology - Poverty and income distribution - Revaluation - Working force - District income estimates - U.N. SNA - Flow of funds accounts - Environmental accounting - Regional input-output table.

Endnotes 1. The constitution of the Conference was drawn up by Prof. Simon Kuznets along the lines of those of the International Association for Research in Income and Wealth. 2. Held on 28-30 Jan. 1957. 3. Held on 2-5 July 1958.

S.G. Tiwari/National Accounts in India

143

References [1] [2]

National Income Com mittee (1951) First Report, April 1951, Ministry of Finance. M. Mukherjee and D.S. Prasada Rao (1973) On an Economic Accounting framework for India and other developing countries, Indian Association for Research in National Income and W ealth Seminar, January 1973.

[3]

For a fuller account see M. M ukherjee (1975) Um a Dutta Roy Choudhury and D.S. Prasada Rao, Economic Accounting for Developing Countries, Published in the Review of Income and Wealth, Dec. 1975.

[4]

Central Statistical Organization (1988) Estim ates of Capital Stock of Indian Economy as on 31 M arch, 1981.

[5] Central Statistical Organization (1967) Brochure on Revised series of National Product for 196061 to 1964-65, Aug. 1967. [6]

Central Statistical Organization (1989) National Accounts Statistics - Sources and Methods.

[7]

Central Statistical Organization (1989) National Accounts Statistics - Sources and Methods.

[8]

Central Statistical Organization (1990) Estim ates of State Domestic Product and Capital Formation.

[9]

M. Mukherjee and G.S. Chatterjee, Trends in Distribution of National Income, 1950-51 to 196566, Economic and Political Weekly.

[10] H.F. Lyall (1960) The Inequalities of Indian Incomes, Economic Weekly, June 1960. [11] Mahfooz Ahmed (1965) Size Distribution of Personal Income and Saving in India, Papers on National Income and Allied Topics Vol. III. [12] P.C. M ahalanobis (1960) A method of Fractile Graphical Analysis, Econometrica, Vol. XXVIII, No.2. [13] Quoted from K.S. Krishnaswamy, Editor (1990) Poverty and Income Distribution, Oxford University Press. [14] B.S. Minhas, L.R. Jain and S.D. Tendulkar (1991) Declining Incidence of Poverty in 1980’s Evidence versus Artefacts, Economic and Political Weekly, July 6-13.

144

The Accounts o f Nations Z. Kenessey (Ed.) IOS Press, 1994

GERMAN NATIONAL ACCOUNTS BETWEEN POLITICS AND ACADEMICS Utz-Peter Reich Fachhochschule Rheinland-Pfalz, An der Bruchspitze 50, 55122 Mainz, Germany Abstract. This paper follows the history of Germ an thought about national accounts from its first beginnings with Leopold Krug in 1805 until the practical and official establishment of the national accounts in the Federal Republic after the Second W orld W ar. It is argued that national accounting has always been situated between the scientific community on the one hand and the political environment on the other, and it is shown how the Germ an national accountants found a successful path in the middle. One result is the contention that the often quoted resistance against national accounting by G erm an economic theorists in the nineteenth century was perhaps not true, but that it was strong in the 1920s.

1. The Purpose of Remembering The history of the German national accounts is not a straight line development. There are curves and breaks, inadvertent impediments and sudden slopes which make it difficult to compose them in one peaceful story. The history of German national accounts reflects the history of Germany itself, and just recently, that history has provided another surprise. At the time when the idea of this study was conceived, it would have been improper to speak of "the" German national accounts because there were two of them, and not just two sets of accounts but two opposing systems. By now the fighting twins have disappeared under a surface which looks as innocent as if there has never been anything but one unique set of accounts. The situation is strange, and while the author admits a certain satisfaction about the simplification of his task, he also feels a certain discomfort because of his failure to understand the history. What happened? And what happened before that happened? And then before that, and so on. The history of the German national accounts has still to be written. There have been occasional investigations, of course, the most comprehensive of which still seems to be the one done by Studenski early in 1958 [48], from which the others have drawn (e.g. Stobbe in 1980 [47]). A thorough study of the events, ideas and controversies behind different concepts and approaches is still missing. Probably we cannot expect a full account today. Such an account would have to be situated within a general history of thought in Germany, not just in the academic sense, the history of economic doctrines, but in the general sense of ideas which prevailed in public, and thus supported and structured socio-economic thinking

U.-P. Reich/German National Accounts

145

long as material history moves in leaps and bounds. Therefore in the following pages I will not attempt to explain what happened in the history of the German national accounts at large. I put before the reader only one small question: Why is it that the word "theory" is, half-consciously, avoided in the halls of the German national accounts? Not that the word is never used. But if it is, it is always in connection with some hidden reserve, with some expression that this is something a genuine national accountant does not side with, emotionally. I will try to explain the observation, because if it is true, it is not natural. My answer will be subjective, colored by my own personal experience, and others will probably have other experiences. Still, it has some general application, perhaps. Consequently, I do not attempt to give anything like an objective historical record of the development of national accounts in Germany. The space accorded in this volume does not permit an elaborate exposition of facts and people which such a record must necessarily provide. Although I do hold the proposition that the history of national accounts in Germany still needs to be written there is a lot of material already available, in any case much more than I can present here [see 8, 18, 20, 29, 42, 47], Instead I want to point out a certain long term development trying to relate the thinking of the post-war generations back to that of their parents and grand­ parents. I regret that this leaves little room for a proper recognition of the post-war achievements which we actually profit from today, but I hope that the idea of spinning a longer historical thread is worth this price.

2. Act One 2.1 Early Exposition The first one generally reported as having compiled an estimate of national income in Germany is Leopold Krug in the year 1805. Typically his was not an estimate for Germany, but for one of its states, Prussia [24]. Krug was followed by Dieterici in 1846, again with an estimate on Prussia [13]. Then such efforts became more frequent and culminated in several studies covering the whole of the German empire for the time before the First World War (see [47] p. 174 for names). This paper being of a conjectural type it will propose some unproven hypotheses. Let this be the first: If indications are correct, the project of national accounting by Leopold Krug in 1805 was a stroke of a genius. A "Royal Prussian Privy Registrar" (koniglich preuBischer geheimer Registrator) by title and profession, he spent his "hours of recreation", as he confesses, compiling lots of data from thirty provinces of the state of Prussia, arranging them in a systematic way to which the concept of national income provided guidance, and discussing principles of government budget policy and taxation, in his two volumes of "Betrachtungen iiber den NationalReichtum des preuBischen Staates und iiber den Wohlstand seiner Bewohner" (Treatise on the national wealth of the Prussian state and the welfare of its

146

U.-P. Reich/German National Accounts

inhabitants). He does not make reference to other writers. Perhaps his is an autochtonuous achievement grown out of the culture of diligence established within the Prussian civil service. It is known that thirty years earlier Frederic II, the king of Prussia, took great and personal interest in the statistics of his state [49]. Unless research proves otherwise the depth of Krug’s arguments, the refinement of his theoretical abstractions, and the caring scrutiny with which he collected his data allow us to call him the father of German empirical economic research. Krug is not a pure statistician. He has economic principles in mind and a driving ideology, which he pursues with intellectual sophistication. Not to further happiness ("Gliickseligkeit") of the people, he writes, is the task of government as proclaimed by theory, but "to remove all obstacles which are in the way of cultivating the physical and economic, the moral and intellectual wealth of a nation." [24, p. 2] Krug chooses the negative circumscription because he finds the positive formula unoperational. In similar dialectics he answers the question of the content of national income. Is it to measure wealth? No, says Krug, it measures poverty because "the degree of poverty can be determined more accurately than the degree of riches of a people." After all the debate we had on the meaning of national income a truly modern point of view. We have no room to do justice to Krug’s economic investigations, which seem not to have been recognized by his readers. Let it just be said that Krug employs the concept of national income in two ways. He compiles the figures in order to prove facts and relationships, and he uses the system of his figures in order to argue economics. Here we are interested in the compilation part. The figure is 261,000,000 Reichsthaler. The Reichsthaler is a monetary unit which from the examples Krug gives in his writing may be estimated to have a purchasing power of roughly 100 $ in order of magnitude. The break-down of the figure is as follows (table 1):

Table 1 Yearly national income of the Prussian state 1. 2. 3. 4. 5. 6 7. 8. 9.

Cultivated fields Meadows, pastures, and commons Woods Horticulture, viticulture etc. Mines Fishery Hunting Factories, crafts and arts Commerce

Total Source: [24] p. 224.

126,643,000 84.340.000 13,000,000 16.463.000 3,000,000 2.996.000 1.997.000 6.333.000 6.228.000

Rthlr.

261.000,000

Rthlr.

U.-P. Reich/German National Accounts

147

Although the table is called national income the figures refer to gross output of the chosen sectors. They are based on Krug’s definition: "The yearly national income of a state consists of the sum of the yearly produced consumer goods to which is added the part which the inhabitants of a state gain by means of their industry from the income of other nations." (p. 26) In modern terms Krug’s national income is basically private consumption plus the trade balance. For the world as a whole, so he observes, the latter would be nought, and there would be just consumption. The trade balance requires a compilation of exports and imports, but as the latter are unwanted by government they are hardly reported, and Krug is satisfied by incorporating exports alone. Actually it is not the export value which figures in his table but the value added in producing these goods. Intermediate consumption has been subtracted on the basis that it is contained in consumption (e.g. wool contained in exported cloth). If we consider that investment was still low at the time we have a concept of income similar to modern domestic product taken from the expenditure side, with the difference that government consumption is not included. It is the domestic product of the market, not a bad achievement for a start. In his methods Krug is no less modern. Considering the output of agricultural production (lines 1 - 7) as destined for consumption he derives the figures from production statistics, using quantities and mutiplying by prices. For cultivated fields, for example, he names 14 products, the smallest of which is caraway; tobacco is not missing, and the difference from today’s statistics is only that the variety of products has increased: In fact determining private consumption by supplying sectors’ sales is the basic method by which the Federal Statistical Office of Germany arrives at that figure today. A fundamental distinction in Krug’s theory is made between real and unreal income. National income is real income because it stems from nature. Unreal income is due to circulation where "not nature pays the wage as in agriculture and partially in mines, but where national income must pay the wage." (p. 28) This "circulation income" is the sum of all payments made in compensation of labor and services to each other. We ignore the idea of drawing income from nature in our national accounts today, but environmentalists might find a partner in Krug. We are sorry to leave Krug here. A proper presentation of his thought to the modern reader deserves an article of its own. Just one more statement may be quoted. Having said that the degree of poverty of a people can be determined with more precision than the degree of their wealth, he continues that therefore it is necessary in treating the wealth and the welfare of a nation to consider each class and trade separately, "in order not to fall into the often committed error of taking a people or a province for rich, because due to a disproportionate distribution of national income and lucrative trades some classes own great properties while others live in neediness."(p. 3) If there is a German tradition in national accounting, and if it has been founded by Krug, it is the tradition of national accounting as a distributive concern [26]. Krug, were he to judge the world today, would perhaps classify it as poor.

148

U.-P. Reich/German National Accounts

Krug’s work lasted 40 years. In 1846, ten years later than Krug had anticipated himself (p. viii), C.W.F. Dieterici felt that in his capacities of Royal Prussian Privy Senior Government Counselor ("Koniglich PreuBischer Geheimer OberRegierungsrath"), Director of the Statistical Bureau in Berlin, Ordinary Professor of Political Science at the University of Berlin, Member of the National Economic Associatian etc., - all these distinctions being neatly enumerated behind the author’s name in his book - it was his task to investigate whether wealth in Prussia had increased since the beginning of the century [13, p.xi]. Perhaps knowing some of Krug’s personal circumstances he seems to have judged him inferior in rank, and therefore, not trustworthy in his theoretical ideas. Dieterici is a learned man. He surveys the international literature, above all Adam Smith, but his writing is less attractive than Krug’s. Typically he does not criticize the Prussian government as Krug’s partial intention has been but wants to prove "the great sensitivity, mature consideration and true benevolence" of the Prussian government resulting in increasing growth of national welfare (p. xvii). His assessment of Krug’s studies might have determined the assessment of all researchers thereafter. To make long computations short, Dieterici arrives at a figure of 147,250,000 Reichthalers for 1805 (p. 41), which is little more than half of Krug’s figure. We must leave it to others to judge whose estimate was closer to reality. In his dictionary article on national accounts Stobbe [47] notes a general sceptical attitude of German economists regarding the possibility of actually measuring national income. F. Kleinwachter, he adds, even declared it impossible. The view was, in fact, first uttered by Studenski [48, p. 143], who wanted to explain the fact that Germany was a latecomer in national income estimating compared to other countries. But the matter needs investigation. It is true that at the beginning of the twentieth century such an attitude prevailed, but for most of the 19th century this is not certain. Some research has to be done here. Just to give some counterexamples: "One of the most important, but also most difficult goals of statistics in the sense of accounting for nations is national income". Roscher writes in 1854 [37, p. 263], explaining that such figures were needed not only for the comparison of the welfare and power of different peoples, but also for just taxation. Mangoldt in 1868 does not see an urgent practical need for compiling national income, but recognises its necessity for comparing the economic state between nations or in time "with some definitive results" [25, p. 305], He also notes the "inclination" of the theory at that time to deal with the investigation of income and the most appropriate method of measuring it. Finally, the famous dictum by Kleinwachter has been quoted out of context. Kleinwachter observes that income stands in the center of economics as a science, and it is in opposition to the main stream of literature that he writes his book on income, presenting a personal view: "If one looks sharply and tries to penetrate just a little beneath the surface, it becomes clear that it is absolutely impossible to circumscribe the concept of income so precisely that it may even be used in daily life"[22], Kleinwachter exaggerates, claiming a special and unique scientific position. What he is concerned with is not so much a practical but a

U.-P. Reich/German National Accounts

149

theoretical impossibility. He argues against an income concept too extensive to be operationalized. After all, Kleinwachter’s whole book is on income and he must have had some concept in mind in order to write it. To conclude, that period needs to be worked through more thoroughly. Empiricial work continued after Krug and Dieterici, first in different states of Germany, then for the empire as a whole [27, 45]. It culminated in Helfferich’s study on German national wealth 1888-1913 published one year before the war. The booklet found a large audience, its 7th edition appearing in 1917 [16]. There is no room here for even a short presentation of these studies. They testify to growing public interest and statistical capabilities. Although the first world war interrupted smooth development, the problems with which Germany emerged from the war intensified the demand for a national income figure. But now there was a formidable obstacle, the chairs of economics. 2.2 Clash with Theory In the year the 1926 rank and file of German speaking economics departments assembled at Vienna for a conference of their professional organisation, the "Verein fur Sozialpolitik" in order to discuss and pass judgement on national income and national wealth. Karl Diehl headed the meeting and edited the proceedings [12], We have only the reports and a summary of the dicussion as records, but they depict a lively meeting. In fact, national income was the first issue presented before a newly constituted sub-committee of the association, the subcommittee for theoretical problems [12, p. 4], Karl Diehl set the stage. In his introduction he expresses a certain ambivalence with respect to the subject. He criticises the use of "such numbers" in difficult questions concerning the fate of the German people, such as the issuing of war loans during the war, and later the problems caused by the treaty of Versailles, or the Dawes plan of reparations. He states as a "not very pleasant" result of his investigation that the possibilities by the statisticians allow only a concept of national wealth and national income in relatively unimportant and irrelevant matters, and that the questions most interesting and most relevant for the economist, namely the problems of productive power and economic capacity, cannot be treated by means of such a "summations concept" (p. 12). On the other hand he pronounces his opposition to Gerhard Colm in stating the he, Diehl, considers the compilation of national wealth for certain purposes of government finance and taxing as "indispensable despite all difficulties and sources of error" (p. 16). The roles of the play fitted the stage. There were five papers invited from Alfred Amonn, Prague; Gerhard Colm, Berlin; Ernst Schuster, Tubingen; M.R. Weyermann, Bern; and F. Zizek, Frankfurt am Main. All were university professors, but of the five, one considered himself a statistician; the other four were economists. And, probably not by chance, the statistician Zizek was the last one to speak. Conflict was programmed.

150

U.-P. Reich/German National Accounts

Amonn opens up a purist position. Noting that income in terms of use value, or its subjective significance, is not interindividually comparable, he retreats to measurement in terms of exchange value. Here again the mechanism of exchange means that income can be meaningful only as a relative not as an absolute measure similar to prices. Above all, income is an inherently individual category. To Amonn the idea that national income is distributed among its members is an illustrative conception rather than a reality. "Actually there is neither a distribution of some unified whole nor a summation of parts to a whole. In the individualistic economy there are only individual incomes of the members of the economy. The economy itself is only a demonstrative notion, not a reality, and has as such no income" (p.23). Gerhard Colm assumes a more differentiated stand. While he distrusts all attempts to measure national wealth, the measurement of national income to him is indispensable, in spite of several conditions and restrictions to which the use of the term is tied. But "no other concept is equally qualified to point out the significance of individual economic facts in the framework of the overall economy." (p. 44) This corresponds to today’s view, but at the conference it was the minority. The third author, Schuster, steers the vessel back into the main stream. He begins by separating two concepts of income which are both included in the word, a quantitative and a qualitative concept. The first is measurable, the latter not, but the latter is the relevant one, because it contains an idea, a norm of behavior. As to the first, he focusses on the distributional aspect: "Can we reason that by means of national income we grasp the potential for the living of a nation? No, because for national income there is no comparison. Income is the attribution of a share, and the size of the share signifies its meaning; it can be compared with other shares. National income is the attribution of all shares, attribution of the total, and the absolute figure says nothing about the meaning of this total sum,"(p. 86) The customary comparison of income at the international level "I consider the most dangerous und unpermissable. If the concept of income is indeterminate for one country, if the individuals in each country differ, as well as the price system, and the income system, there is only one thing identical between the national income of two nations, the name", (p. 88) The fourth paper, presented by M.R.Weyermann, is more modest in its zeal while more acute in method. It takes up the distinction between two concepts of wealth, one measurable, one not, and proposes to label by the term wealth only the first because "wealth is undoubtedly a sum of economic goods." The second might be called economic potential or perhaps national welfare, which necessarily must include free goods, non-disposable goods and other non-economic goods (p. 146). Instead of joining in such useful clarification the president of the session replies by referring to the chaos which bolshevism brought to Russia (p. 147). Called upon as the statistician, F. Zizek is on the defensive from the beginning. He compensates by launching a moderate offensive. Noting that decisive guide-lines from the part of the economic theory are missing he claims that, the other way around, statistical controversies might be used as critical and constructive material

U.-P. Reich/German National Accounts

151

for developing concepts of economics (p. 131). "In formulating problems and forming concepts the leading role belongs to economics, but this must take into account the actual possibilities of statistical surveying; these again imply a test of the concepts of economics from the point of view of real life and thus may contribute to corrections and more precise definitions" (p. 141). The mingling of questions of science with questions of status was one of the reasons why the debate failed to produce a consensus. The economists had to defend their stand: "We are about to terminate our discourse, which has not been useless. After our critique, the last bit of respect may have vanished for proposals of naming a simple sum for national income and national wealth. Such proposals are of value only for political agitation and so on." That was the summary by which the president Diehl concluded the conference. No doubt, the statisticians had fought bravely: "If it appears from the discussion that what we statisticians up to now have understood and compiled as national income and national wealth has no value for the economist, we are in no position to abandon the process of apprehending, and compiling it further in the same way for the simple reason that wide circles demand these figures, and our intention must be directed towards unification of concepts in order to communicate among ourselves in an unambiguous way," remarked Dr. Winkler from the Austrian Statistical Office. The effort failed. The Vienna verdict was felt as a slap in the face of the national accountants, and influenced the minds of the German public until the second world war producing a certain helplessness with respect to the topic of national income [19]. The negative, if not hostile, assessment of the young art of national accounting on the part of professional economists cast a spell on the national accounts profession in Germany which it could not break for a long time. 26 years after the Vienna dictum, and after the Second World War, one of the pioneers of national accounts still felt the blow and passed it back. As late as 1952 Paul Jostok, the then director of the Statistical Office of Baden-Wiirttemberg could not refrain from noting triumphantly that: Today we need not be surprised that the "utterly critical position which the Verein fur Sozialpolitik adopted at its Vienna meeting in 1926 with respect to the problem of national income has remained without influence on the construction of these statistics, even if it has not been widely refuted." [20, p. 136]. Had it been up to the economics profession in those years, the German national accounts would not have been born at that time. The divergence between theorists and practitioners of national income in the Twenties was deplorable because it impeded acceptance of the new economic tool. But this is only one side, and has been overcome. Worse although less noted is the fact that the theorists who fought the national acccounts had a lot to tell them in terms of substance. The subtitle of the proceedings of the Vienna meeting is Critical Conceptual Investigations ("Begriffskritische Untersuchungen"). The Germans of the time, the majority of them belonging to the so called historical school, were trained in conceptual reasoning, in defining qualitative distinctions, in trying to grasp reality by means of proper concepts. True, this sometimes led them into speculation and

152

U.-P. Reich/German National Accounts

individual partisanship, as long as empirical operationalisation was missing. But the national accounts were about to supply just this empirical substance, and the theorists did not see that it was their train that passed by in history. Had they boarded it, and entered into constructive discussion, some bad terminology and improper reasoning would probably have been avoided in the accounts of today. For example, Schuster begins by referring to individual income as the basis for national income because "the quantitative expression of the economy is that of a mechanism, i.e. income occurs only in the form of individual income.” (p. 55) He then distinguishes between income and return because an economy must solve two principal problems. It must "individualise a portion of economic matter, and it must attribute a reward to performance". The first has to do with the control of production, and is carried out within each economic enterprise; the other is control of distribution and concerns the living subjects of the economy. Return belongs to juridical; income, to natural subjects. The distinction is necessary as long as there is a distinction between business and household. In modern accounts the distinction between income and return is clouded. If income is based on transactions, Schuster’s definition as "the realisable purchasing power over which a person gains control as a consequence of his participation at the economic process" [12, p. 57] is valid. From this Schuster easily infers that government employees earn income in the strict sense of the definition. This is a much more straight-forward argument for the inclusion of government activity than the one ususally found in textbooks of today that the government sector is included in spite of its being of a non-market nature, i.e. as an exception to the rule of market output. In Schuster’s argument the inclusion is not to be excused, but naturally justified as part of the initial definition of income. In this way today’s national accountants may learn a lot from their old enemies. The failure to make use of the historical chance of cooperation is due to several mishaps. As said before, status problems were involved between economists and statisticians. This again may have been prompted by the fact that quantitative economics was advancing in general, trying to catch up with the idol of physics, so that certain conceptual methods were on the defensive and needed status support. Further it must be remembered that the measurement of stock variables like wealth or capital was of more concern than the measurement of income in the theoretical tradition of the nineteenth centuiy, and most people still thought so at the conference. This reflected perhaps the bourgeois view of economics that the capital of a private person determines economic success, "while today people think more in terms of social economics, of labor and ist dynamic employment" [21, p. 8]. As history has shown wealth is more difficult a variable to measure at the national level and is less important for running an economy, and in this sense the scepticism of the theoreticians proved justified. A fourth factor may have been political. Germany had produced a great theoretician at the macro level whose writings had decisive influence on the distribution struggle between the capitalist and the working classes. Exploitation is a concept of the macro-level, easily quantified by means of national

U.-P. Reich/German National Accounts

153

accounts although in fact it has never been introduced there. But the smell of politics, a smell of redistribution might have led the sober university employees and civil servants to follow the development with a certain reluctance. As Schuster notes: One cannot argue about the productive character of government employees in terms of quality but of quantity. If the costs of goverment are out of proportion in relation to its performance then "a part of the income of the civil servants is derived income."[12, p. 62f] Was it in the interest of civil servants to have their share in national income be read in the newspaper? All these are hypotheses, of course, to be corroborated or refuted by proper in-depth studies. 2.3 First Official Figures When at the Vienna meeting Mr. Winkler stood up and pointed out that the public demanded "these figures" he could refer to quite a few recent studies devoted to the national income of Germany. The ardent problem of reparations for the war called for an objective measure. What amount of reparations could duly be expected to be paid by the German economy? And more difficult, how was the load to be distributed within the economy? Industry argued that taxes were too high for the economy to recover from war and from inflation. In a memorandum on German economic and budgetary policy they called for restraint of government expenditure and lowering of taxes [34], In a brief paragraph they stated that "it is not possible to determine the size of the national income precisely, and one must rely on estimates" (p. 11). From tax statistics they estimated for the year 1925: 1. Income of employees 2. Income of self-employed 3. Capital income 4. Other income

33 - 36 bill. Mark 8 -1 0 " " 3/4 - 1 " " 1 " "

National income

43 - 48 bill. Mark

The tax load paid from this income was assessed at 25 - 30% compared to 14,2% on an income of 42 - 43 billion Mark before the war. A similar back-of-an-envelope estimate was presented by Elsas on two pages of an economics magazine [14]. By comparing the weekly wage between 1913 and 1925 and multiplying by an adjusted number of employees he arrived at an estimate of 1. Mass income 2. Capitalist income 3. Public income

31.9 4.5 3.0

bill. Mark " " " "

National income

39.7

bill. Mark

154

U.-P. Reich/German National Accounts

The purchasing power of the Mark of 1925 standing at 3/4 of the Mark of 1913, real income of the year 1925 was estimated at 30 billion Mark. In contrast to these quick estimates a more detailed investigation was carried out in a dissertation by Rogowski [36]. His interest was of the same nature as the others: "May the present study contribute to a deepening of insight into the impoverishment of our nation, and pave the way for a just assessment of its economic state" (p. III). For the year 1924 Rogowski estimated: 1. Income of employed 26.0 - 28.0 2. Employers'contribution to social security 1.0- 1.1 3. Net revenues of public corporations 1.0 - 4.5 4. Additional public income 3.5 - 4.5 5. Other private income, not assessable, hypothetical 11

bill. Mark "

National income

bill. Mark

42.5 - 45.9

Rogowski was detailed in his report, but did not advance to greater clarity, rather he seemed to be troubled by the uncertainties in concept and the lack of statistics. He made the point that the sources of error should be made explicit. His item 5 contained all income of the self-employed, income of the employed from non-wage sources, income of the non-working population from property, and the undistributed profit of companies. He found no statistics for these flows, and clearly showed this in his table. A theoretical result of his is worth quoting: "The expectation of gaining by presently available means a higher national income in the future than is possible at present must not itself be considered a present new income" (p. 38), a warning against naive present value computations. In comparing the results of this research it is frightening how divergent they are, and the reaction of the theorists may be partially understood. How useful could a figure be whose compilation proved so uncertain? Not only in quantitative terms, but even in structure and terminology of the accounts, no homogeneity seemed to rule. Each author employed a different terminology, and hence a different method of compilation although in principle all worked with the income approach. At any rate, it was time that some standardisation took place. The German "Statistisches Reichsamt" undertook the project of compiling official national income figures, and in 1932 published a comprehensive study of German national income, which it was to continue as a time series henceforth. Since this work is the culmination of decades, if not of a century of statistical and theoretical efforts, it is being presented in some detail, following Jostok [19].

U.-P. Reich /German National Accounts

155

Table 2 Structure and development of the German national income Billion Reichsmark Sources of income 1913 1928 1932 1938 Agriculture and forestry 5.7 5.8 6.0 5.8 Industry and trade 9.2 12.2 6.0 14.8 Wage and salary 20.7 42.6 25.7 42.7 Capital wealth 5.7 2.8 2.3 3.0 Renting and letting 0.9 0.8 0.8 1.1 Pensions and reliefs 1.4 8.5 9.3 7.6 Private income plus Undistributed company profits Public entrepreneurial income Employers' contrib. to soc. sec. Taxes missing in private income minus Double counting due to public income transfers

43.6

72.7

47.8

75.0

1.2 1.1 0.5 0.1

1.3 2.5 2.2 3.0

-0.4 1.0 1.7 2.6

3.4 1.5 2.9 2.3

0.8

6.3

7.5

5.4

National income

45.7

75.4

45.2

79.7

Source: [19, p. 53]

The table appears strange at first sight. In the first two lines we find the terminology of the product approach, in the third and fourth line there is the income approach; in the fifth line, the product approach; and finally there is something we would not count in national income, pensions and reliefs. While the resulting sum of private income is clear, the additions for arriving at national income are only partially so. The last two items, in particular, missing taxes, and double counting are incomprehensible as such. We take up each entry in turn. In the tradition followed by all German national accountants beginning with Krug, the Office relied on the income approach. Its basic source was income tax statistics. Consequently, the categories of private income reflected the classification of the income tax law. The figures in the first line refer to the income of the self-employed in agriculture and forestry. Since here tax statistics were poor, income compilation was based on the production approach beginning with gross output and deducting costs. Among these were interest and insurance premiums. Adjustments were made for owner occupied housing and own account production. Overall the difficulty of estimating this income led to the observation that "in this oldest branch of the economy which is tied so fully to organic life, commercial accounting is not really

156

U.-P. Reich /German National Accounts

applicable" [19, p. 58]. Entrepreneurial income in trade and industry was estimated in the second line from income statements. It is said that "the oscillations of property values are foreign to the notion of national income and must be eliminated" [19, p. 60] because they do not reflect the value of the goods produced in the year. Wages and salaries were again taken from income tax statistics in the same way as capital income. Inclusion of earnings abroad signifies that the national income concept was aimed at in modern terms, and not domestic production. The figures on capital income include only capital income rendered to natural persons, not to corporations. Income from renting is taken from income tax statistics, adjustments for owner occupied housing are left aside due to lack of data. Pensions and reliefs comprise all social transfer income in the modern sense. Including them here is contrary to the notion of factor income underlying modern national income. It seems that "private income" includes elements of what today we call disposable income. The last line in table 2, "double counting", clarifies this treatment. It comprises all social transfers except pensions for civil servants and farmers. Deducting them means that in national income social payments are excluded, only these pensions are included. Their inclusion is justified by saying that similar to social contributions by employers, civil servants and farmers earn their pension while working, although they do not see it. Instead of assigning a fictitious supplement to income for reflecting this earning, as we do today, the actual payments are taken as a measure. If all is taken together, the figure shown for national income is pretty much in line conceptually with what we would understand by it today, with one exception, the entry of so-called taxes missing in private income. This is an interesting point. The idea behind is what today we call individual consumption paid for by government. This consumption, it is argued, is partially financed by direct taxes, and thus included in the other figures, which represent income before tax. But direct taxes do not suffice to pay for such consumption. A part of other government finance goes here, too. The item of missing taxes is to correct for this undervaluation. It is equal to the balance between individual consumption paid for by government, and direct taxes received by government. A glimpse of what later became to be called the Economica debate between Hicks and Kuznets shines through here. Because of this imputation the concept of national income of the Reichsamt is situated between national income (at factor cost), and domestic product (at market prices) in today’s terminology. The figures were developed on the basis of an explicit and well defined theoretical concept: "Under national income or national product is understood the total of goods and services valued in money, which are at the disposition of a national economy after keeping intact the initial stock of wealth." [43, p. 11] This was later called the Hicksian definition. At the time and in its detail and structure the system represented an innovative achievement. Although it was not a system in the modern sense combining different approaches to one economic circuit, it was comprehensive in combining all available statistics (census data, government finance etc.) into one system which was a

U.-P. Reich/German National Accounts

157

realisation of the income approach. The public interest in this approach is demonstrated by an extensive analysis of income distribution carried out in the study. It is part of the irony of German national accounts history that later, after the war, the two other approaches, production and expenditure, were used, and the income approach neglected. While before the war German national accounting was based on distribution of income, not only as a matter of interest, but as the most reliable road towards compilation, the system developed after the war reconciled the production and the expenditure approach, and income disrtibution was neither quite integrated into the system nor coherently used in determining national income. It has never regained the prominence in determining national income it had in this period.

3. Catastrophe The year 1932, the year of the Statistisches Reichsamt’s first publication of national income of Germany was followed by the year 1933 in which the national socialists came to power. The event demarcated a radical change of politics and economics in Germany, so radical that people who survived could not, or would not, even remember. The Statistisches Reichsamt continued publishing its figures until 1938. Publication of the Statistical Yearbook in 1939 was interrupted by the war. There was a volume for the two years 1939/1940, but it was declared strictly confidential although nothing new was in the volume except a map on the peoples and territories seized in the war [44], For 1941/1942 another volume appeared, again confidential. What happened thereafter is not known. National income was compiled in the way it had been compiled before including analysis of income distribution, but were the figures ever used? We can pose some hypotheses. On the one hand, the new political system was highly planned. It was meant to be a war economy from the beginning [33, p. 3f]. Resources were mobilized on a macroeconomic scale, centralized decisionmaking on economic affairs was normal. Industry was organized in trusts, partly controlled by the government, labor was forced, production was directed. One would have thought that planning the war and fighting it would have encouraged development of the national accounts as it did in England. On the other hand, there are few published records of such activity. Later, H. Bartels notes that in contrast to other countries, German work in national acccounts was hindered by the war, and stopped by the capitulation [3, p. 141]. But the war was not the government. Who decided about the sort of the national accounts, and what? It seems that there was no development in terms of methodology, but compilations were carried on. Was this just pure bureaucratic routine? Or were national accounts ignored, their preparation impeded? A third hypothesis is that there were no economists among the national socialists, so that their economics was dominated by engineers who ran the big heavy industry trusts. They had no theory to apply to the national income figures. Tons of steel may have seemed a more

158

U.-P. Reich/German National Accounts

figurative indicator than millions of some monetary value. Two personalities stand out and can be recognized as having lived through these upheavals, Paul Jostok and Ferdinand Griinig. Jostok headed the national income section of the Statistisches Reichsamt in Berlin, apparently until the end of the war. He later was President of the Statistical Office of the state of Baden-Wiirttemberg, wrote about national income, but had no active part in the national accounts work after the war. His writing after the war was keen in national income theory and full of political zeal, but strangely void of notices about the time under Hitler. For example, in comparing the national product of the Federal Republic in the 1950s he turned to the corresponding figure in 1936 without explaining anything about the figures of 1937 - 45, which he must have produced himself [21, p. 22ff|. And in his account of the history of national accounts [20], Krug, and Dieterici are mentioned, a sign of his intimate, academic knowledge of the field, - so is the publication of the Statistisches Reichsamt in 1932, but thereafter only foreign studies are quoted. Ferdinand Griinig had an interesting career. A trained mechanical engineer, he was prompted to study macroeconomics by his observations of the great depression in 1931, and as a result published a pioneering volume on the macroeconomic circulation of value. In 1945 he was made one of the directors of the German Institute of Economic Research ("Deutsches Institut fur Wirtschaftsforschung") in Berlin, and immediately began working on national accounts. This is how his life reads in the obituary [9], Actually, Griinig had been active in the field before [18, p.88 ]. From 1936 to 1944 he was head of the division of central economic oberservation of the Imperial Economic Chamber ("Abteilung fur Zentrale Wirtschaftsbeobachtung der Reichswirtschaftskammer"). Grunig had no reservations constructing a time series of national income from 1891 through 1941, pointing out the questionable precision for the war time estimates. He also cited publications about macroeconomics based on national accounts during the war [18, p. 90], and refers to interesting conclusions regarding government debt and payment for the war, using them as witness for his thesis that "the great relevance of national accounts for responding to the problems which have been posed to the German economic researchers by economic policy, was fully recognised in Germany before the war and during the war." [18, p. 90]. Still the question remains in what sense were the accounts used by the national socialists. Bombach, in his obituary statement, regretfully comments that the "rich experience which Grunig was able to gather in his function of director of the Imperial Economic Chamber and which had to be kept strictly secret will be witheld from us forever." [9]. What was this experience? We do not know. The history of the national accounts turns here into a history of the people involved, and they seem to have disappeared without a trace. In the German writings after the Second World War the Nazi period is usually ignored, following the general custom of treating the period as abnormal, out of bounds of German history. Authors report about the publication of the Statistisches Reichsamt, and then jump to 1949 when the Federal Republic was created (for an exception to this rule see [42]). Moral questions apart, it needs digging in archives, following up

U.-P. Reich/German National Accounts

159

the lives of individuals in office in order to write the history of the German national accounts of these years. After the crush of and liberation from national socialism, Germany was divided. One part became the Federal Republic, and the history of its national accounts is fairly well known. The other part became the German Democratic Republic, which also started out on a system of national accounts, and about it very little is known. The national accounts became one of the planning devices of official economic policy in the GDR, but were soon made secret, and in terms of methodology as well as results, publication was restricted. It is impossible, at present, to draw a coherent picture of the national accounts of that part of Germany for that time. Where did the ideas and expertise come from? Was there a link between the national accounts produced in Berlin after, and those produced before, 1945, either in substance or in person? This is another area where research and documentation are urgently warranted.

4. Act Two 4.1 Support from Mathematicians In the year 1935 a new journal of economics, called "Archiv fiir mathematische Wirtschafts- und Sozialforschung" was founded in Germany. Apparently it represented the German equivalent to the English "Econometrica". The organisation of responsibilities is interesting. The journal was edited by two professors with the cooperation of another professor, and three young lecturers ("Privatdozenten") Hans Peter, Erich Schneider, and Heinrich von Stackelberg. Later all three became famous economists in the Federal Republic. The first two in particular were preoccupied with the concept of economic circulation ("Wirtschaftskreislauf), which they used for developing macroeconomic theory. These people were partly trained as engineers, so that a formal approach to economics to them was a natural way of thinking. They wrote the first books on economic circulation, Peter in 1937 [30], Schneider in 1947 [39]. Schneider’s book "fulfilled a mission: the students of the first post-war years, coming home from war and detention in thirst for knowledge owe it their penetration into the theory of economic circulation." [00, p.218]. The book became the model for writing on national accounts in the Federal Republic. Independently of them Carl Fohl, another engineer, worked on the same topic producing his book in 1937 [15]. The mathematical approach had an advantage at the time. Not only did it offer a way towards solving intricate theoretical problems, but it also was a natural means of taking economics and its theory out of politics. Here was pure theory, presented in equations and graphs, and there was applied politics which could use the mathematical results but at its own risk. A mathematical approach to economics was neutral to the political terms in which an economic circulation was placed. Income

160

U.-P. Reich /German National Accounts

was income, and consumption was consumption independent of whether it was produced under fascist or democratic rules. If only the economic circulation was complete, the rest of society did not matter. National accounts were a natural complement to the idea of mathematisation, producing tha data needed for the models. They were supported, therefore, by the new theorists in Germany after the war. In this spirit Wilhelm Krelle, a disciple of Schneider, produced a set of national accounts of Germany which included even an input output table [23]. However, he did so only once. More continuous and substantial work was undertaken in the economic research institutes which were partly founded then or revived from before the war. Griinig worked in the German Institute for Economic Research in Berlin, Ifo-Institut began in Munich, and Peter founded his Institute for Applied Economic Research ("Institut fur angewandte Wirtschaftsforschung") in Tubingen. To the mathematical economists national accounts were interesting as a problem of structure. The accounts contained what later became termed balance equations in economic models. National accounts were seen as a means of providing these equations, and the required data for the study of behavioral equations. It was not recognised that national accounts in their terminology, in their verbal representation and statistical operationalisation contained matters of economic substance. The title of Krelle’s study is indicative of this view: "National accounts including input output-analysis with figures for the Federal Republic of Germany". National accounts were conceptualised as a set of equations, and the figures were a means of showing a general, and comprehensive mathematical structure of the economy, but in themselves they were not a part of the problem. They were used for illustration. Consequently, national accounts were not considered an area of research for economics. In the "Zeitschrift fur Sozialwissenschaften", the journal of the economics profession, not a single article appeared on national accounts, nor was the topic ever treated in one of the conferences. The contrast to the times of the Vienna meeting could not be more pronounced. At that time, critical and verbal methods ruled in Germany, and the national accounts were rejected. Now that mathematics had taken over, national accounts were taken for granted, and ignored. In this way national accounts found support from the newly established academic culture because of their formal and quantitative appearance. At the same time, a division resulted between the theory and the practice of national accounts. The statisticians who began working out the future standard accounts did so with diligence but published little about methodology. The theoretical studies on national accounts were written by university staff, usually removed from practical concerns, controlling as means of production a blackboard and a piece of chalk. To them an approach to national accounts as a system of abstract variables in equations was handy. Schneider’s pioneering book clearly shows this bias, which determined the German academic tradition of teaching national accounts for quite some time. These academic activities seem to have been little related to the projects which began under offical auspices in the Federal Statistical Office. H ere its president

U.-P. Reich/German National Accounts

161

Gerhard Fiirst gathered a crew of young economists and statisticians at the city of Wiesbaden in order to build the national accounts for the new republic. The first signs of the activity of the office in the field of national accounts already appeared in 1949. One title is significant: "National income, social product and balance of payments of the Federal territory in the first year of the Marshall plan 1948/49" [41]. US secretary of state Robert Marshall was, so to speak, the political father of the official German national accounts. But Germany was not yet ready for the child. Another struggle still had to be fought. 4.2 Trouble with politics In spite of the courage and dynamics with which the building of national accounts was pursued in the Federal Statistical Office, public opinion did not watch these activities only with trust. Germany, having just lost a wrong ideology was badly in need of a new, and right one. In the West the market economy and its political sister liberalism took over while socialism which found its paradise in the East was soon discredited and became the ideological enemy. Wilhelm Ropke, one of the ardent ideologists of the new system, was outspoken: "If we deplore the centristic-mechanistic inclination of contemporary economic thinking ... we are reminded in the first place of a tendency which is inevitably associated with the name of Keynes, and roams about under the name of ’macroeconomics’. It treats the economic process as an objective-mechanical operation, moving in lump quanta, and believed to be determinable and even calculable in advance by means of corresponding mathematical- statistical methods. The economy then easily appears as a kind of monstrous pumping station, and, consequently, its science becomes more and more a kind of engineering art, in which equations sprout exuberantly, and the real harvest of one and a half century of intellectual labour, namely the teaching of price movements, almost falls into oblivion." [35, p. 333]. Calling upon the many errors and false forecasts experienced before, Ropke expresses distrust in so called "data". He is highly sceptical as regards national accounts, which he considers a main framework for expressing mechanical, centralistic thinking. Using it as a simple means of description of economic history, he does not object to, but the eagerness with which the advantage of national accounts is defended makes him suspect another goal. "It is the claim for power of the economocrat openly intending to employ national accounts as an instrument of ruling economic circulation." (p. 340) It brings with it "the constant temptation to force resistant reality into a plan and to adjust the estimates afterwards." (p. 340) In supporting his argument Ropke points to the fact that "the healthiest countries in terms of economic and monetary policy are without national budgets, while the countries which have specially fostered this method are countries of eternal sickness, and under unusually high pressure of inflation (Scandinavian countries, Netherlands, Great Britain, France)....Most likely the national budget as the essential instrument

162

U.-P. Reich/German National Accounts

of economocracy is intrinsicately related to the falsities which this philosophy usually commits." (p. 341). "Is it necessary to recall the sinister prophecies with which the opponents to the German market economy accompanied its rise, and which have been disproved again and again?" (p. 338) A footnote puts a final illumination on this point. Under the topic of dehumanisation of economics Ropke writes: "In the light of these remarks the statement may not appear risky that there is an inner relationship between Keynes and Picasso. Even if we did not know that they belong to the same epoque, the dehumanisation characteristic to both would prove it, and even in their alternating between classicism and avantgardism they are strangely similar. Thus we are not astonished about Keynes having an express preference for Picasso, while Picasso himself is known as a communist." (p. 356) This was not an intellectual climate in which a young discipline like the national accounts could grow freely. Ropke’s view was that of the ruling conservative and liberal political parties, and their economics minister Ludwig Erhard. In their fight against socialist opposition, national accounts became a highly political issue indeed, and they were pronounced by the socialists as an instrument of socialist economic policy [27], [11, p. 41], important enough to be explicitely included in the principal program adopted by the Social Democratic Party in Bad Godesberg 1958 [32, p. 191]. It was not until 1967, when the social democrats came to power, and the law on economic stabilisation was passed, committing the government to the achievement of full employment, external equilibrium, price stability and growth, that the national accounts were securely settled in the political environment of the Federal Republic of Germany. The political problems were sensed by the experts [31, p.189]. The danger was not that national accounts would be altogether abolished, but political sensitivity produced negative consequences for scientific exchange in the field. It seemed unwise to discuss in public all the problems and errors which inevitably accompanied the new enterprise. There was a tendency to publish only uncritically certain results. Rather than turning to the scientific community the experts of national accounts kept problems within themselves, and preferred to have the doors closed when errors and methods were discussed. This again raised distrust even on the part of well-meaning university collegues, and when their expectations to be involved in the methodological process turned out to be futile they lost interest. "In the future, co-operation with the universities is to be reinforced," Hildegard Bartels, the founder of the offical accounts, once said, indicating in the polite language of the government official that there might be a deficiency [3, p. 142]. The alienation of the economists working in and with the national accounts from those working with mathematics was a natural result of division of labor. But it was reinforced by the adverse political climate in which the new and fragile structure of the national accounts had to be erected.

U.-P. Reich/German National Accounts

163

4.3 Initiation by the International Community There is a place for wonder here, and possibly for future research. While the ruling political parties fought national accounts for reasons of ideology their government participated in the European Recovery Program, the so-called Marshall plan introduced by the United States in order to contain communism in Europe. The aid granted on the basis of the program was coupled with a demand for macroeconomic figures, and the newly founded OEEC promoted the setting up of national accounts for that purpose. We do not know how the two opposing stances were reconciled within the political establishment. This again is a question for research in the archives. To the statisticians reference to the expectations of the international community formed the main leverage for defending their goal. Mrs. Bartels, in an appearance before the statistical association, once launched an offensive. Her subject, "National accounts as an an instrument of economic policy", was one where statisticians do not venture easily. But from the beginning she called attention to national accounts as a field which was relatively little known in the Federal Republic and widely discussed internationally: "Unfortunately, it cannot be overlooked that many European and non-European countries have progressed much further in the practical compilation of national accounts as well as in their application and use for economic observation and policy than the Federal Republic. The question comes up repeatedly, of course, on what the advantage of these countries is based." [4, p. 325], The development of what today are the German national accounts can best be reconstructed by following the journal of the office "Wirtschaft und Statistik" where regular progress reports were published. Interestingly enough, the first article which appeared there on national product was born out of political inspiration. Monetary reform having just been passed, prices were rising, and unions demanded an increase in wages. In that situation Hildegard Bartels compared the development of wages to that of national product, and showed that industrial workers had improved their distributional position [1], The next article brought the first explanation of national product figures, taking the census year 1936 as base year [41]. Schorry explains the three possible methods of compiling national product, calling them the "local", the "personal", and the "indirect" method. While the first and the third method are appplied abroad and demanded from ERP countries Germany had up to now relied on the second method. In order to comply with the new demand the national product of 1936 was recompiled according to the new methods: it was projected on the new territory of the Federal Republic and then the value added was extrapolated by means of production indices. This method was continued for estimating national product over the next few years. The old German concept of dividing government expenditure into intermediate and final consumption, and including only the latter in national product was preferred by Schorry, but he yielded to the international standard. Two years later, the position was repeated: "The Federal Statistical Office intends to adhere to the approach, in

164

U.-P. Reich/German National Accounts

principle, because the method proposed by the OEEC results in an artifical blowing up of national product in all countries with extensive administrative activity which does not benefit private consumption really. In this way it diminishes international comparability..." [2, p. 223]. At the time Germany could not influence the standards of the OEEC, but some may remember the argument in the context of the concept of individual consumption acknowledged in the forthcoming SNA. In 1957, later than anticipated, the extrapolation method was abandoned and compilation of national product was based on comprehensive data of the Federal Republic, mainly statistics of turnover tax and statistics of cost structure [7, p. 124], The unanticipated lag of publication was explained by difficulty of reconciling the results of the different statistics, and it is here where the lack of public discussion and documentation is seriously felt. It is clear that the long awaited data were surprising to the national accountants in their incompatibility; it is also clear that under the circumstances such divergances could not be made public; it is deplorable that hence an explicit, systematic, and theoretically elaborated method of dealing with such phenomena was not developed. There must have been arguments and differences of opinion in the process which to today would be valuable to know. The recompilation showed that the early estimates were not bad, after all. Gross National Product had been overestimated by 8.3% for 1950, and by 5.8% for 1954. For net product the differences were 4.2% and 4.0%. Even lower were the differences in growth rates, 62.0 % was the growth of the German GNP in four years according to the old method as against 58.4 % acccording to the new one. [7, p. 13 If] Together with greater precision the new compilation brought with it more detail and led the way towards a full-fledged system of national accounts ("Volkswirtschaftliche Gesamtrechnungen") for which the OEEC had worked out the standard. The term "Rechnung" was chosen on purpose. It was to signify description of the past as opposed to planning for the future. This was the political compromise, and is the reason why in Germany no economic projections are made in the statistical office. Projections were considered political, and in Germany they are now carried out by the economic research institutes. "In order to eliminate even in the terminology any notion of instruments of total planning and obligatory execution, the expression "national budget" has been avoided in all official attempts of economic projection." [5, p. 321], With the foundation of the European Economic Community, and its statistical office EUROSTAT new demands for national accounts figures had to be met. This time even an iput-output table could not be avoided, which the office had experimented with before, but had to discard by administrative order. In 1965 the "Possibilities and limits of compilation of input output tables for the Federal Republic of Germany" were discussed in public and a new project undertaken with the help of EUROSTAT [6]. Here we leave our historical recital, because from now on it is clear that the national accounts were there in Germany, and the period runs into our times with the SNA of 1968, of which the criticisms, debate, and innovations concern us at present.

U.-P. Reich/German National Accounts

165

5. Are there Lessons?

The history of the national accounts in Germany has not yet been written. The facts we know raise more questions than they answer. From this paper some hypotheses emerge which would be interesting to pursue. It seems that in the dealing with the German national accounts right people did the wrong things, and vice versa. The verbal economists who by their training would have been well qualified to give guidance in elaborating the logical frame-work of the accounts opted for opposition, while in the second trial the mathematicians supported the accounts but their methodology did not give them any incentive for discussing how the accounts were constructed. Perhaps this is one reason why the word "theory" has a disturbing reminiscence in the statistical office. The division of practice and theory had its effect. Today the distribution list kept at the Federal Statistical Offfice of university economists interested in national accounts consists of fewer than 20 names. This must be compared with an environment in France, in the United States, and in other countries where academic interest in national accounts is strong enough to support a whole national accounts association of their own. In terms of organisation a similar restraint was followed. In contrast to France or the Netherlands, the German statistical office for a long time had no section for theory or concepts or similar work. This was considered inexpedient, and it is only now, in response to the European demand for an explanation of methods, that such a section was founded. Such management can be criticised, but it is understandable as a reflection of the historical experience of the national accounts being tied between the two poles, politics and academics. When the political side wanted the accounts the academics objected, and when academics were supportive, politicians smelled a rat. Even in 1977 the spirit of Vienna was still present. When some young researchers presented a paper on "A critique and extension of the national accounts" at a meeting of econometricians, Krelle grew angry, and called the paper unscientific because it had characterised government activity as an activity which assured a "specific" public order, and investigated the notion of reproduction of labour power.[10] Trained as a mathematician Krelle could not imagine that the content of a concept might be examined with other motives than purely ideological ones. History is to be remembered. It is most truly remembered if it is not any longer similar to the present. Fortunately, the present situation of German national accounts is not like it was in the past. The figures are well established and widely used. A great credit in this respect has to be given to Alfred Stobbe whose comprehensive textbook appeared in 1966 and, in revised editions, has trained generations of West German economists in the national accounts ever since [47]. The fact that in contrast to the unreliable estimates experienced by Ropke, German economists of today were united and correct in predicting economic disaster in case of an immediate exposure of the GDR to the world market, marks among others a success of the validity of the figures of national accounts.

166

U.-P. Reich/German National Accounts

Thus the problems recited in this paper are not those of today, anymore. Their shadow will vanish definitively when the university economists again recognise national accounts as a valid area of research, and the national accountants insist on discussing economic theory as part of their business. The two belong together.

Acknowledgements I am grateful to Hans Adler, Hildegard Bartels, Rainer Fremdling, Gunter Hamer, Heinrich Liitzel, Harry Schimmler and Rainer Staglin for helpful comments. Special thanks go to Bela Adler who skillfully weeded my English.

References [1] [2]

Bartels, Hildegard (1949) Reallohn und Sozialprodukt, in: W irtschaft und Statistik, p. 66-69. Bartels, Hildegard (1951) Probleme der Volkswirtschaftlichen Gesamtrechnungen, in: Wirtschaft und Statistik, p. 222-227.

[3]

Bartels,

Hildegard

(1952)

Volkswirtschaftliche

Gesamtrechnungen

und

Sozialprodukts-

berechnungen in der Bundesrepublik, in: Allgemeines Statistisches Archiv, Band 36, p. 141-152. [4]

Bartels,

Hildegard

(1954)

Volkswirtschaftliche

Gesamtrechnungen

als

Instrum ent

der

Wirtschafts-politik, in: Allgemeines Statistisches Archiv 38, p. 325-335. [5]

Bartels, Hildegard (1960) Das Kontensystem fur die Volkswirtschaftlichen Gesamtrechnungen der Bundesrepublik Deutschland, in: Wirtschaft und Statistik, p. 317-344,

[6]

Bartels, H., Hanisch, G., and Lauckner, W. (1965) Moglichkeiten und G renzen der Berechnung von Input-Output-Tabellen fur die Bundesrepublik Deutschland, in: W irtschaft und Statistik, p. 69-81.

[7]

Bartels, H., Raabe, K.-H. and Schorry, O. (1957) Die Neuberechnung des Sozialprodukts fiir die Bundes-republik Deutschland. Ergebnisse fur die Jahre 1950 bis 1954, in: Wirtschaft und Statistik, p. 123-149.

[8]

Bombach, Gottfried (1960) Kreislauftheorie und Volkswirtschaftliche Gesamtrechnung, in: Jahrbuch fur Sozialwissenschaft, B d .ll, Teil 1 H.2, p.219-242 und Teil 2 H.3, p. 331-350.

[9]

Bombach, Gottfried (1960) Ferdinand Griinig in memoriam, in: Jahrbuch fur Sozialwissenschaft, B d .ll H.2, p. 243.

[10]

Blum, Reinhard (1978) Bericht iiber die Diskussion zum Referat U.-P. R eich/ P. Sonntag/ H.-W. Holub, in: Bombach, G., Gahlen, B., and Ott, A., N euere Entwicklungen in der Theorie des Konsumenten-verhaltens, Tubingen, p. 321-325.

[11]

Deist, Heinrich (1959) Wirtschaft von morgen. Beitrage zur Wirtschafspolitik der SPD, Berlin und Hannover.

[12]

Diehl,

Karl,

(Hrsg.)

(1926)

Volkseinkommen

und

Volksvermogen.

Begriffskritische

Untersuchungen. Schriften des Vereins fur Sozialpolitik, Band 173, Miinchen und Leipzig. [13]

Dieterici, C.F.W. (1846) D er Volkswohlstand im preuBischen Staate in Vergleichungen zu den Jahren vor 1806 und von 1828 bis 1832 sowie aus der neuesten Zeit, Berlin.

[14]

Elsas, M oritz (1925) Das deutsche Volkseinkommen, in: Magazin der Wirtschaft, 1. Jg. 1. Halbjahr, H.21, p. 764-765.

[15]

Fohl, Carl (1937) Geldschopfung und Wirtschaftskreislauf, Miinchen und Berlin.

[16]

Helfferich, Karl (1917) Deutschlands Volkswohlstand 1898-1913, 7. Aufl., Berlin.

167

U.-P. Reich/German National Accounts

[17]

Griinig, Ferdinand (1933) D er Wirtschaftskreislauf, Miinchen.

[18]

Griinig,

Ferdinand

(1950)

Die

Anfange

der

Volkswirtschaftlichen

Gesam trechnung

in

Deutschland, in: Beitrage zur empirischen Konjunkturforschung. Festschrift zum 25jahrigen Bestehen des Deutschen Instituts fur Wirtschaftsforschung, Berlin, p. 71-103. [19]

Jostock, Paul (1941) Die Berechnung des Volkseinkommens und ihr Erkenntniswert, Stuttgart,

[20]

Jostock,

Berlin. Paul

(1952)

V on

den

ersten

V olkseinkom m ensberechnungen

bis

zur

Volkswirtschaftlichen Gesamtrechnung, in: Allgemeines Statistisches Archiv, Bd 36, p. 130-140. [21]

Jostock, Paul (1955) Das Sozialprodukt und seine Verteilung, Paderborn.

[22]

Kleinwachter, Friedrich (1896) Das Einkommen und seine Verteilung, Leipzig.

[23]

Krelle, Wilhelm (1959) Volkswirtschaftliche Gesamtrechnung einschlicBlich input-output-Analyse mit Zahlen fur die Bundesrepublik Deutschland, Berlin.

[24]

Krug, Leopold (1805) Betrachtungen iiber den Nationalreichtum des preuBischen Staats und uber den Wohlstand seiner Bewohner, Berlin.

[25]

Mangoldt, Hans von (1868) Volkswirtschaftslehre. Die Lehre von der Giitererzeugung, von der Vermogenserhaltung und der Verteilung der Giiter, Stuttgart.

[26]

May, R.E. (1899) Das Verhaltnis des Verbrauchs der Massen zu demjenigen der "kleinen Leute", der W ohlhabenden

und der Reichen

und die Marxistische Doktrin, in: Jahrbuch fur

Gesetzgebung, Verwaltung und Volkswirtschaft, Bd.23/1, p. 271-314. [27]

May,

R.E.

(1909)

Das

deutsche

Volkseinkommen

und

der

Zuwachs

des

deutschen

Volksvermogens im Jahr 1907, in: Jahrbuch fur Gesetzgebung, Verwaltung und Volkswirtschaft im Deutschen Reich, Band 33, p. .85ff. [28]

Nemitz, K. (1960) Sozialistische Marktwirtschaft. Die wirtschaftspolitische Konzeption der deutschen Sozialdemokratie, Frankfurt am Main.

[29]

Organisation for Economic Cooperation and Development, Structure and Sources of National Accounts in Germany, Paris, 1952.

[30]

Peter, Hans (1937) D er Wirtschaftskreislauf, Tubingen.

[31]

Peter, Hans (1958) Problem e der Volkswirtschaft, Tubingen.

[32]

Program me der deutschen Sozialdemokratie, Bonn, 1963.

[33]

Reichsamt fur wehrwirtschaftliche Planung, Die deutsche Industrie. Gesamtergebnisse der amtlichen Produktionsstatistik, Schriftenreihe des Reichsamts fur wehrwirtschaftliche Planung, H. 1, Berlin 1939.

[34]

Reichsverband der Deutschen Industrie, Deutsche Wirtschafts- und Finanzpolitik, H. 29, Berlin,

[35]

Ropke, Wilhelm (1958) Jenseits von Angebot und Nachfrage, Erlenbach-Ziirich und Stuttgart.

[36]

Rogowski, Erich (1926) Das Deutsche Volkseinkommen, Berlin.

1925.

[37]

Roscher,

Wilhelm

(1854)

Die

Grundlagen

der

Nationalokonomie.

Ein

Handbuch

fur

Geschaftsmanner und Studierende, Stuttgart, Tubingen, Bd.l. [38]

Schaffle, Albert E berhard Friedrich (1873) Das gesellschaftliche System der menschlichen Wirtschaft, 2 Bande, 3. Aufl. Tubingen.

[39]

Schneider, Erich (1947) Theorie des Wirtschaftskreislaufs, Tubingen.

[40]

Schorry, Otto (1949) Volkseinkommen und Sozialprodukt des Vereinigten W irtschaftsgebietes im

[41]

Schorry, O tto (1949) Volkseinkommen, Sozialprodukt und Zahlungsbilanz des Bundesgebiets im

[42]

Statistisches Bundesamt (1972) Bevolkerung und Wirtschaft 1872 bis 1972, Wiesbaden.

[43]

Statistisches Reichsamt (1932) Das deutsche Volkseinkommen vor und nach dem Kriege,

Jahre 1936 und im zweiten Halbjahr 1948, in: Wirtschaft und Statistik, p. 94-99. ersten Marshallplanjahr 1948/49, in: Wirtschaft und Statistik, p.256-261.

Einzelschriften zur Statistik des Deutschen Reichs Nr.24, Berlin. [44]

Statistisches Reichsamt, Statistisches Jahrbuch fur das Deutsche Reich 1939/40, Berlin, 1941.

U.-P. Reich/German National Accounts

Steinmann-Bucher, Arnold (1909) 350 Milliarden deutsches Volksvermogen. Das Volksvermogen Deutschlands, Frankreichs, GroBbritanniens und der Vereinigten Staaten von Amerika, Berlin. Stobbe, Alfred (1966) Volkswirtschaftliches Rechnungswesen, Berlin, Heidelberg, New York. Stobbe,

Alfred

(1980)

Volkswirtschaftliche

Gesamtrechnung,

Wirtschaftswissenschaft, Band 8, Gottingen usw., p.397ff. Studenski, Paul (1958) The Income of Nations, New York. W agemann, E. (1938) Die Zahl als Detektiv, Hamburg.

in:

Handworterbuch

der

169

The Accounts o f Nations Z. Kenessey (Ed.) IOS Press, 1994

THE SNA: 1968-1993 AND BEYOND Anne Harrison* O.E.C.D. Abstract. This paper describes the history of the System of National Accounts from the second to the third m ajor revision; the initiatives that led to suggestions for change and the forces resisting change. Recognition of these pressures is interesting not only because of the influence they had on the present revision process but because they are likely to continue into the future and reaction to them will determ ine the evolution of the SNA to the end of this century and beyond.

1. The first decade of the 1968 revision The second revision of the System of National Accounts, [1], was approved by the Statistical Commission at its Fifteenth Session. It was published in English in 1968, in French in 1970 and in the other official languages of the United Nations in the years that followed. In the late seventies a number of handbooks were published developing different aspects of the SNA. In particular, M60, [2], dealing with balance sheets and reconciliation accounts and M61, [3],dealing with income distribution were published in 1977; M64, [4], the manual on national accounts at constant prices, and M68, [5] giving guidelines on statistics of tangible assets, were published in 1979. This work developed against the background of radical changes in economic structures and analyses following the oil price increases of the early nineteen-seventies that brought levels of inflation to the developed market economies that had not been experienced since the war. These had immediate implications for the view of national accounting. Conventional constant price accounts where the prices of a single base year were applied to the whole of the existing time series were suddenly seen to be questionable, and the process of chain linking came to be widely practised. Like commercial accountants, national accountants were concerned with inflation accounting and the measurement of holding gains and losses in the economy. Analyses based on input output tables began to be questioned. Previously it was assumed that the co-efficients changed sufficiently slowly that a table one or two years old was adequate * The views in this paper are the views of the author and should not be attributed to the Secretariat or M em ber countries of the O ECD. Involvement in an event makes it easier to recount, but impossible to be entirely objective. As well as this formal disclaimer, therefore, I wish to make clear that the views expressed in this paper are purely personal and should not be attributed to other m em bers of the InterSecretariat W orking G roup or mem bers of the Expert Group in either personal or institutional capacities.

170

A. Harrison/The SNA: 1968-1993 and Beyond

for current usage, but a table dating from before the oil price increase in 1973 could clearly not be used for analysis in the immediate aftermath of the oil shock. With the growth of monetarism there was increased interest in the development of financial statistics, often not in statistical offices directly, but frequently in central banks. Just at the time when these changes might have been expected to lead statistical offices to be more open to new statistical development, including development of the SNA, another effect of the oil price increase became apparent. This was the retrenchment in government that took place in the downturn in the economies which occurred in the mid-seventies and which affected statistical offices as well as government generally, and in some cases disproportionately so.

2. Statistical Commission initiatives, 1979-1985 Publication in 1968 and after of the Blue Books led to implementation of the revised system in many countries, and in 1975 an inter-regional seminar was held in Caracas to discuss the problems of implementation. Regional meetings were subsequently held in Africa in 1975 and 1979, in western Asia in 1978, and in Europe in 1978 and 1980. A report on the findings of this review of the progress on implementation was submitted to the Statistical Commission at its Twentieth Session in 1979 and the Commission concluded that an Expert Group should be convened to consider the status of work on the SNA and to discuss the direction of future work. In making this proposal the Statistical Commission stated that they "did not consider the time had come for basic changes in the structure of the SNA and that the usefulness of the System for the purposes it now serves well should not be disturbed." At the same time, they considered that it should be reviewed from the point of view of needed clarification, updating and modification. An Expert Group consisting of ten country representatives and observers from the UN regional commissions and international organisations met from 28 April to 2 May, 1980 in New York. The meeting discussed a number of particular points of concern with the SNA. It began with a general consideration of the status of the SNA and concluded,[6], that ''..the meeting expressed general satisfaction with the basic structure of the revised SNA and stressed its importance as a point of reference for the integration of basic statistics in both developing and developed countries ... different views were expressed as to the appropriate level of complexity of the System, some participants stressing the need for a simple, easily understood presentation intermediate between the highly aggregated accounts for the nation as a whole and the very detailed and complex underlying accounts, while others thought that the complexity was essential if the System was to serve for basic data developm ent... Different views were also expressed on the limits of appropriate change in SNA at this time. The importance of continuity was stressed, both for compilers and for users of the statistics. It was therefore considered that the burden of proof was upon those proposing changes to demonstrate that they

A. Harrison/The SNA: 1968-1993 and Beyond

171

entailed substantial advantages". There was, however, unanimous agreement that some clarifications in SNA were required. The view was expressed that there was a prima facie case in favour of changes that would increase consistency among the international statistical systems. The Statistical Commission again considered the question of a revision to the SNA at its Twenty-first Session in 1981 and a further Expert Group was convened in New York in March, 1982. On this occasion it consisted of eleven country experts (some of them the same as the previous Expert Group meeting) and representatives of the International Organisations. The Expert Group was asked to consider "the adequacy of the present architecture of SNA how priorities should be established for future work and the timetable appropriate for their pursuit, and finally how results should be effected, whether interim changes should be contemplated and whether there should be a major revision at about the end of the decade." Again the meeting had an agenda with a number of specific points for consideration, but began by discussing the more general issues about how and when a revision should be undertaken. The meeting agreed, [7], "on the need for clarification and updating of SNA in order to remove ambiguities, ensure consistency and adapt new uses of data, and for harmonisation of SNA with other statistical standards used by the United Nations and other international organisations". However, it is interesting that they went on to say "The general view, however, was that a revision of the 1968 magnitude would not be needed in the coming decade". There was discussion about whether a revision could be implemented in stages and as a series of incremental improvements, but this was rejected in favour of a revision which might be accomplished by 1990. A significant development at this meeting was the confirmation by the Statistical Commission, of the initiative that work on the SNA should be shared among the international organisations. "It is apparent that those international organisations, regional commissions and specialised agencies that are directly involved in the development of systems and standards need to be brought together on a co-operative basis. The burden of work can then be shared and conflicting or duplicative efforts can be avoided. In terms of implementing a common set of standards, furthermore, representation of those directly involved is essential. Finally, an important requirement is continuity of participation. The Expert Group felt that future work could be accomplished more efficiently if a continuous working group were established by the secretariats of the international agencies generally concerned, and if this group met once a year. This group should be supported by appropriate experts and consultants and would have the task of directing the work on the major developments in SNA." The Inter-Secretariat Working Group on National Accounts set up in the light of this report in 1982, originally consisted of the United Nations Statistical Office (later the Statistics Division), OECD and Eurostat, but was later extended to include the IMF and

172

A. Harrison/The SNA: 1968-1993 and Beyond

the World Bank. The Statistical Commission meetings in 1983 and 1985 confirmed the need for a review of the SNA to be undertaken but stressed that this should concentrate on clarification and simplification and harmonisation among different international statistical systems.

3. Initial planning of the review, 1983-1985 The Inter-Secretariat Group decided that in order to cover all areas of the SNA a series of eight Expert Group meetings would be desirable. The first of these was to look at the structure of the System and this was to be followed by a series of six specialised meetings covering prices and volumes, the external sector, the household sector, the public sector, production accounts and input/output tables, and the financial accounts and balance sheets. The last Expert Group meeting would be a co-ordinating meeting to pull together all of the discussion of the previous meetings. The first meeting was planned for 1986, and at the rate of two per year, the last meeting was envisaged for 1989, leaving adequate time, it was thought, to prepare the revised draft of the Blue Book for the 1991 Statistical Commission. It was decided the Expert Group should consist of twelve people. Six of these would represent core experts and be present at all meetings; the other sue would be subject specialists drawn in for each meeting in turn. Of the core specialists three would represent developed countries, and three developing countries. For the subject specialists a global representation would also be sought. However, in view of the limited number of people who could participate in the Expert Group meetings, it was decided that the process should start by having an inter-regional meeting with representation exclusively from developing countries so that the core experts from these countries would be able to represent a wide range of views, it being made clear to all of the experts involved that they attended the meetings in a personal capacity rather than as representatives of their government or institution. It was also decided that it would be desirable to have the revision process fully documented but that the time each Expert Group spent reviewing an account of the meeting should be limited. Thus it was decided that the key decisions of the Expert Group should be written down in a terse and factual manner, and agreed by the experts at the meeting and that a full report describing the discussion that led up to these decisions should be prepared subsequently and circulated for comment.

4. The Expert Group meetings begin, 1986 Although a number of meetings of interested parties had already taken place at OECD, Eurostat and IARIW for instance, the first meetings formally planned as part of the revision process took place in June 1986. An inter-regional meeting took place, as foreseen, immediately followed by the first Expert Group meeting on the structure of

A. Harrison/The SNA: 1968-1993 and Beyond

173

the SNA. As mentioned above, the inter-regional meeting was confined to people from developing countries in part so that they would be neither overawed nor bulldozed by more outspoken statements from their colleagues from developed countries, more used to expressing their opinions in a forthright manner amongst groups of colleagues from other countries. It became clear at the inter-regional meeting that some of these fears were misplaced. There was unanimous agreement that the concerns of developing countries needed to be reflected in the SNA and there was a very clear message that all developing countries were not alike; there were greater differences between developing countries than between some developing countries and some developed countries. With this in mind, therefore, it was felt very strongly that there should not be two versions of the SNA, a "full" version for developed countries and a cut-down version for less developed countries, but that a single version should be expounded, capable of meeting the requirements of the conditions in countries at all stages of development. It would be for individual countries to decide how much of the System they could and would implement, given resource constraints and policy considerations. Another very clear message of the inter-regional meeting, and one shared by many others, was that one reason why the Blue Book needed to be revised was that the 1968 version is not easy to read and understand. Further, by concentrating on theoretical issues there is little practical guidance given to compilers. This deficiency is especially important in countries where there may be only one or two national accountants, and meetings with colleagues from other countries extremely few and far-between. This problem had long been recognised by the United Nations and other international organisations and by 1982 the idea of a series of handbooks to accompany the Blue Book had also been accepted and the first the these that subsequently emerged, on Production Accounts and the Public Sector, were already in draft at that time. These handbooks were to contain specific recommendations for practical implementation. The Inter-regional Group welcomed this initiative and suggested a list of handbooks, including an ordering of priority, which it was hoped would be completed and published at the same time as the revised Blue Book. The first Expert Group began on 23 June 1986 with a reminder from the Director of the United Nations Statistical Office of the mandate given by the Statistical Commission in respect of the review of the SNA; that is, that it was to concentrate on clarification and simplification of the System, and harmonisation with other international statistical systems. The Group willingly accepted this mandate. The very first conclusion written at one of the Expert Group meetings reads in full: "The revision will not involve major changes to the present SNA". If this was so clear, and agreed so easily by the experts at their first meeting, how is it, that over a period of time, a number of changes have indeed been suggested for the SNA to the extent that many of those less intimately connected with the revision feel that the process has gone far beyond the original mandate and has indeed suggested major changes not originally foreseen or agreed in principle by the Statistical Commission?

174

A. Harrison/The SNA: 1968-1993 and Beyond

Much turns on the two words "clarification" and "simplification". They have an obvious and immediate appeal, but even in that first Expert Group meeting a short discussion led to the conclusion that these two objectives were not always, or necessarily, complementary. One of the great strengths of the SNA is that it provides a co-ordinating framework for all economic statistics. Such comprehensiveness militates against simplicity. The System is by its nature complex, and must remain so if it is to fulfil its co-ordinating role. However, even a complex system can be explained simply; thus, at the outset the Expert Group interpreted the goal of clarification as a need for lucid and well-reasoned justification of the intricacies of the System rather than a removal of the intricacies. The need for simplification was seen primarily as a requirement of compilers who might have fewer human resources to compile the accounts, or fewer data sources than were ideal. At the beginning, therefore, clarification seemed to mean an expanded but easier to understand Blue Book, whereas simplification entailed the production of handbooks giving practical guidance on how corners might be cut if needed.

5. Clarification In many ways the discussions at the various Expert Group meetings could be seen as pursuing the goal of clarification. The experts aimed to unravel the System to make sure it was clear, consistent and appropriate in order that it may be explained without ambiguity. A working-through of this process in a number of instances is illustrative. In Paragraph 6.19 of the 1968 SNA there is a definition of the extent to which production of goods by households should be included within the production boundary. It reads: "It is also desirable to include in gross output (i) the output by producers of other commodities which are consumed in their households and which they also produce for the market, and (ii) the processing of primary commodities by the producers of these items in order to make such goods as butter, cheese, flour, wine, oil, cloth or furniture for their own use though they may not sell any of these manufactures." An interpretation of this instruction, with its emphasis on commodities being produced for the market, using primary commodities and moreover primary commodities of the producer, leads to the interpretation that if a tailor makes ten shirts, nine of which he sells, one of which is for his family, the value of ten would be included. Similarly, if he sells one and uses nine for his family, all ten would be included. But if he only makes nine for his family they would be excluded. Baskets woven from grass would be included; baskets woven with synthetic materials would be excluded. Cheese made from the milk of one’s own cow is included; cheese made from the milk of a neighbour’s cow is excluded. The experts concluded that such demarcations were inappropriate and confusing and that a clearer and simpler interpretation was that in general, goods manufactured by households should, in principle, be included within the production boundary, though in practice an attempt

A. Harrison/The SNA: 1968-1993 and Beyond

175

would only be made to quantify these if they were significantly large for the economy as a whole. This proposal has provoked no adverse comment. It is generally recognised as constituting clarification and simplification, but it does of course also constitute a degree of change in the 1968 System. A second instance concerns intangible assets. It became clear from discussions in the Expert Group and with experts outside this circle that the treatment of intangible assets in the 1968 Blue Book is not well-understood and the rationale for the treatment is far from clear. In the 1968 SNA, intangible assets are synonymous with non-produced assets and therefore payments arising in connection with them, such as patent and copyright income, is treated as property income. An interpretation of this means that, for example, the act of writing a book by an employee constitutes production within the System; the act of writing a book by someone who is self-employed is excluded from the production boundary. The experts felt that the production boundary should be defined in terms of the type of activities concerned and not in terms of the conditions of employment relating to those undertaking the activities. Thus, the production boundary was extended to include the activities of producing literary artistic works, whether on own account or not. Once such work on own account was admitted, the question arises of how to deal with major activities whose production may span several periods and whose income arises over a number of periods into the future, for example, the production of films. The conclusion seemed obvious. Such expenditures should continue to represent intangible assets but they would be treated in future as produced assets rather than non-produced assets, with the consequence that income arising from them was treated in the same way as income generated by any other productive activity and not as property income. This time the change suggested was rather more substantive, but arose again from seeking to clarify what it was the SNA recommended and to explain why this was so. This emphasis on rationalisation arose not just from a pedantic concern with minutiae but from the firmly held belief that it was impossible for any manual to describe in detail all instances of a phenomenon that either did occur or were likely to occur over the next twenty years. Thus definition by enumeration was bound to raise problems sooner or later as an instance arose that was not covered in the enumeration list. On the other hand, it was felt that a definition by rationalisation should give enough indication to allow treatment of new phenomena to be determined that was consistent with the overall logic and conventions of the System.

6. Simplification While there was general agreement among the experts about what the goal of clarification entailed, the same unanimity was not always obvious as regards simplification, since the rejoinder was "simpler for whom?". Nor was the choice simply between user and compiler because different users and different compilers had different views of what constituted simplification.

176

A. Harrison/The SNA: 1968-1993 and Beyond

There is a strongly held body of opinion that suggests that compiling national accounts is simpler if they consist of direct aggregations of underlying basic data. A particular case in point concerns the commercial accounting returns prepared for enterprises. If the same accounting system is used for commercial and national accounting, data collection and data compilation becomes easier for both the provider of the data and the statistical office. The result, however, is a system that needs to reflect the real world in all its complexities. This in turn imposes burdens of complexity that may be regarded as being unnecessary and possibly undesirable in countries where commercial accounting is not so standardised or confirms to different norms. It can be argued that international accounting standards are desirable and are actively being developed so that with the passage of time this objection should disappear. But such standardisation is not the current position in many countries. National accounts should be compiled with a view to their use, and here two different perspectives are possible. For some microanalysts a system of national accounts that perfectly reflects in miniature the complexities of the real world is exactly what is needed. However, many macro-economists and planners look to the national accounts to provide a model of economic behaviour in the economy, a model that emphasises the key features in an easily grasped overview. A model of a complex world in microcosm does not provide this simple overview. There is another set of users, however, that is in favour of a more detailed approach. These are analysts concerned with living standards and the question of the adequacy of the basic provision of goods and services to households in different economic strata. These analysts would benefit from a greater integration between the macro-economic framework of the SNA and micro datasets coming, for example, from household budget surveys. As in the case of integration with enterprise accounting, the resulting system is likely to be more rather than less complex. On balance, therefore, the desire to provide integration with household and enterprise data tends to militate quite heavily against a goal of simplifying the System. Hence the adoption by the Expert Group of the understanding that simplification of the theoretical structure was impractical and the goal should be applied to the manner of exposition and to the advice on practical applications to be described in the handbooks.

7. Harmonisation The removal of minor discrepancies between alternative international statistical systems was one of the main aims of the review of the SNA. Three of the Expert Group meetings dealt specifically with the relationship between the SNA and the various specialised systems of the IMF. These were the meetings on the External Sector, on government accounts and on financial statistics. In the course of these meetings total harmonisation was achieved between each of these systems and the SNA. It was foreseen at the outset that this could only be achieved by making changes in one or the other system, and frequently it involved changes in both. During the balance of payments meeting, for example, it was agreed that balance of

A. Harrison/The SNA: 1968-1993 and Beyond

177

payments accounts would in future distinguish payments for what the SNA regarded as services from property income (factor and non-factor income in the old balance of payments terminology), a change that was conceptually simple but may give problems of implementation in a number of countries. Further, it was agreed that the balance of payments statistics should distinguish a category of capital transfers. There was some discussion about where these should be located. What the fourth edition of the Balance of Payments manual [8], calls a capital account corresponds to the SNA financial account, and capital transfers belong in the SNA capital account that so far has been missing in the balance of payments account. This change was suggested at the SNA Expert Group meeting and eventually agreed by a meeting of balance of payments compilers subsequently arranged by the IMF. This represented a major change to the balance of payments account. On the other hand, changes were also suggested in the SNA to be compatible with the balance of payments. One of these was the showing of reinvested earnings by foreign controlled enterprises as an outflow on the current account and a matching inflow in the capital account. Another change concerned the question of the classification of gold. In the balance of payments (monetary) gold may be held as a financial asset only by monetary authorities; all other (non-monetary) gold is treated as a commodity but not a financial asset. The 1968 SNA allows all sectors to hold gold as a financial asset. In the end, it was agreed to change the SNA to bring it into consistency with the balance of payments system Some changes were agreed to both systems. The reclassification of gold from monetary to non-monetary or vice versa, used in both systems to be treated as imputed imports and exports, but in future is to be recorded in the other changes in assets account. Similarly, it was agreed that in principle imports should be recorded on a fob rather that a cif basis, thus obviating the need to record domestic transport on imports as both an export and an import. The meetings on the government accounts and on financial statistics also resulted in agreements whereby the alternative systems would be clearly consistent with one another, with each derivable from the other. Another system where consistency was deemed to be of high importance was the relationship between the European System of Integrated Economic Accounts, the ESA [10], and the SNA. It was clear at the outset that the ESA would need to be rewritten the agreement is that the new version of the ESA to be published in 1994 will be fully consistent with the 1993 SNA. In some other areas attempts at harmonisation were regrettably rather less successful. Extensive discussions were held with the drafters of the introduction to ISIC to ensure that the definitions of establishment, enterprise and activity in ISIC and in the SNA are consistent. Although it was possible to agree on matters of substance, the goal of including exactly the same text in both publications was not achieved, principally because different points of emphasis were felt necessary by the two sets of drafters. Other areas where harmonisation is desirable but has not yet been achieved concerns other classifications which especially affect analyses of households. The classification by

178

A. Harrison/The SNA: 1968-1993 and Beyond

employment is important, but work by ILO on the definition of employment, self-employment and unemployed status is not yet finalised. The SNA experts believe that the separation between individual and collective consumption of government can best be done in terms of COFOG (the Classification Of Functions Of Government) but that some minor amendments to this classification would be needed principally to identify separately the administration of health and education type services. This issue is still pending. So also is the classification of individual consumption by purpose. At several points in the review process the desirability of resurrecting the classification of industrial inputs by purpose was raised in order to identify work on research and development, for example, or on environmental protection. Positive action to resume work on this classification, however, remains to be taken. Lastly, no attempt has yet been made to reconcile the classifications used in connection with industrial statistics with those proposed in the SNA, and until this is done the concept of "census value added", to name but one, is likely to be a continuing source of confusion.

8. Subject specific Expert Group meetings, 1986-1988 Six subject-specific Expert Groups took place between November 1986 and September 1988. The first concerned prices and quantity comparisons. In a sense the subject matter of this meeting was less material in the Blue Book itself as in M64 [4], The main preoccupation of the meeting was with clarifying points that were inadequately dealt with previously, and concern for how much of this subject matter should be included within the Blue Book. It was agreed, for example, that it was appropriate that some discussion of international comparisons as well as intertemporal comparisons was appropriate. It was also agreed that the main GDP income aggregates should be adjusted for terms of trade effects, though there was less consensus on how this should be done. In this area at least, flexibility is left with the countries implementing the proposal. Explicit recognition was given to the fact that many countries had adopted a series of chain indices in order to take account of changes that radically altered the structure of prices. Since the possibility of chaining had not been dealt with at any length in either the Blue Book or M64 the proposal of chaining at regular intervals could be regarded as the first major change in the System agreed to by the experts in their meetings, but one introduced in response to the established practice of users of the System. The second meeting on the external sector was mainly concerned with issues of harmonisation as described above. The third meeting, on the Household Sector, also introduced changes. Some of these were in the nature of clarification of the production boundary in respect of Household Production, again as described above. Further, there was discussion of the concept of the total consumption of the population as it appeared in the material product system, the MPS, [11], used in planned economies. The proposal, originally put forward by Petre, [12], of distinguishing between individual and collective consumption of government and between consumption expenditure and actual

A. Harrison/The SNA: 1968-1993 and Beyond

179

consumption of households was accepted. This was another major change bur again one where it was felt that there was general agreement that this was desirable. It was in the Household Sector meeting that the question of stressing the flexibility of the System through the use of social accounting matrices was also first proposed. The fourth specialist meeting concerned Government Accounts and most of the meeting was devoted to questions of harmonising classifications and concepts in the SNA and in Government Finance Statistics (GFS). It was at this meeting that a major concern of several developing countries emerged. They felt strongly that it was illogical and inconsistent that the SNA recommended that rent be imputed on governmentowned houses occupied by individuals but not on buildings owned and occupied by government in the conduct of its business. Further, misleading changes in government expenditure were recorded when government moved from owned to rented accommodation or vice versa. The proposal therefore, accepted by the Expert Group, was that imputed rent on government-owned buildings should be part of intermediate consumption of government, part of its output and part of its consumption expenditure. The fifth subject-specific Expert Group meeting on production accounts and input/output tables was held at the end of March 1988. In retrospect this seems to have been a particularly important meeting and it is interesting to try to identify why this was so. It was a long meeting, originally scheduled for ten days, and followed a meeting on input/output organised by the Austrian Statistical Office the preceding week. Since many of the experts attended this meeting also, together a period of about two and a half weeks was spent considering the SNA review. It had become apparent in earlier meetings that it was very difficult to give full consideration to items where only one aspect was considered. Many issues had been referred forward to this meeting. This was partly because of its scheduling towards the end of the specialised meetings and partly because of the co-ordinating function of input/output tables. The list of items for consideration on the agenda reads like a roll-call of the most difficult areas in national accounting. These include the separation of market and non-market production and the role of "departmental enterprises", valuation and the definition of basic prices, the treatment of value added tax, the valuation of imports, the measurement of bank service charges and the treatment of exchange rate differentials. In addition, a number of issues connected with the asset boundary also arose for the first substantial time at this meeting. These covered the appropriate treatment of military hardware and mineral exploration, research and development, and expenditure on the environment. As far as the external circumstances of the meeting, it is perhaps interesting to note a change of attitude among the participants. It was the sixth time these people had met together to discuss the SNA revision. The impact of the work, the need to attend meetings, and the necessary reading and preparation of papers for them, had had a major impact on almost everybody’s work patterns. The personal dynamics of the group became apparent in a very positive way; here was a group of people determined to achieve an overhaul of a major statistical system with the intention that this system should serve the statistical community well for the next twenty years. Indeed, it was at this meeting that the need to be forward-looking and to think about likely innovations

180

A. Harrison/The SNA: 1968-1993 and Beyond

in accounting practice and policy concerns was first articulated. This change of attitude, unfortunately, did not make the outstanding problems noticeably more tractable. An example of this might be the consideration of the imputed charge for bank services. In the summary and conclusions of the meeting there are eight points devoted to this topic, the last of which calls for a comprehensive study on a number of particular points. Despite the difficulties this implies, however, all commentators would surely agree with the first conclusion of the experts at that meeting, which again is quoted in full: "Banks produce genuine services; the problem is to measure their output". On this and on many of the other topics considered by the meeting the experts considered that further study was necessary before definitive conclusions could be reached. The last of the subject-specific Expert Group meetings was on Financial Statistics. Again, harmonisation with the IMF’s system of money and banking statistics was of high priority. For the most part, many of the subjects raised were new to the SNA rather than changes to it. There was much consideration of new financial instruments, of appropriate classifications of financial instruments and financial intermediaries, and the role of financial assets in the system, particularly in relation to balance sheets.

9. Drafting and co-ordination, 1988-1989 The original intention was that the entire draft of the new Blue Book should be prepared by Peter Hill, acting as "the scribe of the Expert Group", and he was relieved of his duties in OECD during 1988 so that he could proceed with the drafting on a full time basis. It was clear by the end of 1988 that one co-ordinating meeting to tie up all the loose ends and to review the draft would simply not be adequate. Instead, a series of three co-ordinating meetings was planned for 1989 alone. The first of these in fact, did not consider any draft but merely tried to reach consensus on outstanding issues. To look at the proposals made by the experts, especially when these are expressed in a terse and unelaborated form, can give the impression that the experts felt uninhibited about suggesting change and may even have sought to advocate change without pressing reasons. Despite this appearance, the spirit of the meetings was usually quite the opposite. Even though there may have been one or two enthusiasts advocating change, the body of experts as a whole tended to be rather conservative in their approach, and were reluctant to accept change except when the case for doing so seemed overwhelming. This is evidenced not only by the number of subjects that were continually referred forward to future meetings with more documentation to support the case, but the sheer number of meetings that were held from 1989 onwards in order to resolve such issues. Another change took place at about that time concerning the composition of the Group. The original proposal to have six core experts and six rotating specialists had worked fairly well, but by the beginning of 1989 the three core experts from developing countries had all changed their positions, and indeed one had had to retire from the

A. Harrison/The SNA: 1968-1993 and Beyond

181

revision process. It was therefore felt appropriate that some adjustment to the composition of the Expert Group was needed in order to ensure adequate representation of developing country concerns. Further, as the Group moved to a co-ordinating role it was felt that the original distinction between core and non-core experts should be dropped and that the original core members should be supplemented by one person from each of the subject-specific meetings with a speciality in that area. This revised Group first met in January 1989 and met in all on six occasions, the last in October 1992. As draft chapters became available these were circulated to members of the Expert Group for comment and subsequent revision. The co-ordinating meetings held in the middle of 1989 divided their time more or less equally between considerations of draft chapters and consideration of outstanding issues. Early in 1989 it had become clear that the work of redrafting the Blue Book was too much for a single individual to undertake, and it was agreed that Andre Vanoli would assist by working together with the United Nations Statistical Office to prepare detailed tables and classifications of transactions and assets, to provide a description of this framework of the System, and to describe how the basic System could be elaborated for specific circumstances by the development of satellite accounts and other complementary analyses.

10. Developments 1989-1991 At the third co-ordinating Expert Group meeting in September 1989 a difficult decision was made. It had become obvious both to members of the Expert Group and to the Inter-Secretariat Group that a draft of the Blue Book that could be ready for the 1991 Statistical Commission would be seriously deficient. In September 1989 not all of the first draft was complete. Those parts that had been circulated had received wide-spread comments would require further editing. Moreover, it was proposed to hold a round of UN regional commission meetings to enable national accountants from all countries to comment on the draft, and the direction the review was taking. This process in itself was envisaged to take a year. Part of the reason for this was logistical; several members closely associated with the revision process were required to attend each regional commission meeting in turn in order to make presentations, and slotting these meetings into their other workload meant a fairly extensive process. Further, there were difficulties seen if the drafting continued during this discussion process, since then the first regional commission would not be discussing exactly the same draft as the last one to meet. In effect, therefore, the effect of having a round of regional commission meetings was to suspend drafting on the document even though it was known there was still a very considerable amount of drafting to do. With these considerations in mind a recommendation was made to the Director of the Statistical Office, and accepted by him, that while a preliminary draft could be submitted to the Statistical Committee in 1991, it should be done on the understanding that there was still extensive work to be done and that a final draft would not be available before the 1993 Statistical Committee

182

A. Harrison/The SNA: 1968-1993 and Beyond

meeting. At the same time it was agreed that responsibility for the ultimate completion and publication of the new SNA should become the joint responsibility of all five organizations of the Inter-Secretariat Group and not of the UN alone. 1989 had also seen an unexpected development in national accounting. From the outset of the revision it was supposed that it would be desirable at some stage to consider the links between the SNA and the MPS. The ECE meeting in Geneva in 1989 showed for the first time that the two systems were not inevitably to be presented as separate but equal systems of the United Nations; rather, that representatives of planned economies wished to find an integrated system where features of both systems could be combined. This aspiration received further encouragement in an Expert Group meeting held in December of 1989 in Moscow to discuss points of detail between the two systems. It was agreed that there were points of substance where the systems differed because of political background, but that many of the differences between the two systems were incidental and due mainly to their separate development. The meeting therefore concentrated on eliminating these differences and, as with other meetings on harmonisation, changes were proposed for both systems. The boundary between intermediate consumption and compensation in kind received considerable attention for example, and it was noted that the MPS had an explicit treatment of losses which was missing in the SNA but should usefully be included. Thus suddenly and unexpectedly there was a requirement that the new Blue Book should adequately cover the conditions in centrally planned economies, as well as those in market economies. At the same time that the decision was made to postpone completion of the draft to 1993 it was also agreed that an end had to be made to the process of discussing difficult areas and firm decisions were to be made. Two further co-ordinating meetings were therefore scheduled, one to take place at the end of 1990 to review and tidy outstanding issues, and one in early 1991 to review the comments received from the regional commissions. One aspect of the fourth co-ordinating meeting at the end of 1990 was the concentration on assets. The first SNA in 1952 tended to deal very largely with income concepts. The 1968 revision was notable for the incorporation of input/output tables into the system, and a much greater elaboration of production. In the course of the latest revision, however, it had become clear that many of the areas not so far treated comprehensively in the SNA dealt with assets, including both the definition of the asset boundary and the suitable treatment of assets in different circumstances. A goal for this last meeting therefore, was to be able to take an overview of the whole question of assets and to reach an agreed definition of what exactly constituted an asset, and how this definition could be used to integrate the flow accounts and the balance sheets. The Statistical Committee meeting in 1991 had before it the preliminary draft of the new Blue Book and recommendations from the experts and the Inter-Secretariat Group that the revision process should be given another two years to reach conclusion. Reluctantly they agreed to this delay, but insisted that the process must come to completion in 1993, if necessary by reverting to treatment in 1968 SNA when agreement on a new treatment could not be reached. They also recommended that fewer meetings should be organised and greater reliance put on ad hoc consultation by telephone, fax

A. Harrison/The SNA: 1968-1993 and Beyond

183

and other electronic media, and greater responsibility to be taken by the Inter-Secretariat Group. The fifth co-ordinating meeting held in May 1991 was indeed the last to be held before the meeting scheduled for October 1992 in connection with a final inter-regional meeting. Peter Hill was seconded from the OECD to the Inter-Secretariat Group for another year in that month and an intensive period of drafting, commenting and redrafting began. As the year developed it became clear that a number of task forces should be set up dealing with specialised areas. These covered for example, the treatment of capital accounts, the developing of social accounting matrices, the treatment of input/output and supply and use tables, and the definitions of transactions and accounting roles. It also became apparent that the drafting needed to be distributed more widely than had originally been intended, and new authors were recruited, particularly in respect of the areas covered by task forces.

11. The final stages 1992-1993 By September 1992 a complete draft was available for consideration at an inter-regional meeting held in Mexico in the next month and for translation into the various United Nation languages for consideration by the Statistical Commission in 1993. The inter-regional meeting itself was planned as a quicker and cheaper means of gathering comments from national accountants around the world than another full set of regional commission meetings. In all representatives from about 60 countries met for a week to review the draft Blue Book and discuss how the proposals may be implemented. The report of that meeting shows that there was widespread enthusiasm for the draft Blue Book, for its range of coverage and clarity of exposition. Concerns were expressed over the continuing need for handbooks to explain compilation practices to complement the theoretical exposition in the Blue Book itself and that in few cases would national resources alone be sufficient to permit the full implementation of the new System on a timely basis. In the end, the draft revised Blue Book contained approximately 100 significant changes from the 1968 version. At the inter-regional meeting, although there were some participants who expressed reservations about one recommendation or another, in general it was recognised that a manual of this size could not possibly satisfy every reader on every point, and in a spirit of compromise it was accepted that there were only eight issues on which considerable resistance to the proposals was evident. In the light of this, the Inter-Secretariat Group made recommendations to the Statistical Commission on how these problems should be treated. On two, the treatment of gold as a financial asset and the treatment of the costs of mineral exploration as capital expenditure, the recommendations to change the SNA were upheld. On three issues there was regret expressed at the inter-regional meeting that changes had not been introduced into the System. These issues concerned environmental accounting, consumer subsidies and an articulation of a distinction between formal and informal activities. On

184

A. Harrison/The SNA: 1968-1993 and Beyond

all these there is widespread agreement that changes to the System are urgently needed but so far there is no concensus on alternative treatments. It was agreed it would be inappropriate to delay finalisation of the text until these problems were solved. The treatment in the 1968 Blue Book has therefore been maintained but a firm commitment made that the Inter-Secretariat Group, in consultation and cooperation with other interested parties, will pursue these issues as the highest priority areas on a research agenda leading to further changes in the SNA in the near future. On another two issues, it was decided to reverse the changes proposed in the revised Blue Book and keep to the 1968 treatment. Despite the conviction of the expert group, reaffirmed after a number of considerations, no estimates for the imputed rent on buildings owned and occupied by governments and non-profit institutions serving households will be introduced. In the dichotomy between market and non-market output, there had been differences of opinion about which category covered production for own-account consumption, which is generally marketable but not marketed. The final text therefore contains a three-way split, between market, own-account and other non-market production. The last issue of some contention was one that had been a source of difficulty every time it came up for discussion. There is universal acceptance that the provision of bank services is output but because no direct charge is made it must be measured indirectly. There is almost universal acceptance that that this output should be allocated to all users and not attributed by convention in total to intermediate consumption of a notional industry (the solution in the 1968 SNA). Unfortunately there is little concensus on how this allocation should be made and some even dispute that a satisfactory solution can be found. At the time of the inter-regional meeting, the draft simply recommended allocation to all users without specifying how this should be done. In response to comments received then and after, up until the Statistical Commission itself, a compromise was agreed whereby the text of the Blue Book allows two alternative treatments, one of allocation across all users and one maintaining the use of a notional unit but suggests bridging supplementary calculations should be made. Subject to these major revisions and to a multitude of uncontroversial but significant clarifications and refinements suggested by a wide readership of the September 1992 draft, the Statistical Commission in February 1993 recommended the approval of the new Blue Book and plans for the English version to be published in December 1993 with other language versions appearing soon afterwards. This endorsement of the draft by the Statistical Commission represented the culmination of nearly ten years work and represents a major milestone in the development of national accounting. At the same time the fact that some items have already been placed on a priority research agenda implies the recognition that it is not the end of the process.

A. Harrison/The SNA: 1968-1993 and Beyond

185

12. The achievements of the 1993 revision The English draft of the new Blue Book covers approximately 900 typescript pages. This compares with 150 pages of text in the 1968 Blue Book. Even allowing for the compression ratio between typescript and printed format this is still a very much extended publication. What exactly does this draft represent? In effect, it is a replacement for four existing manuals; the Blue Book itself and the manuals on Constant Prices, Financial Accounts and Balance Sheets, and the Classification of Tangible Assets. Further, there are three chapters devoted to the elaboration of National Accounts in a flexible manner, including the development of satellite accounts and social accounting matrices, which together could have constituted a separate manual. Although the original intention was to pay attention primarily to those areas of the SNA that were felt to be in need of review or updating, in the event it proved impossible to confine attention to a narrow range of areas. Thus the new Blue Book represents the result of a comprehensive review of all aspects of the national accounting framework; identification of transactors and transactions, classifications of these, times of recording of transactions, and detailed recommendations on valuation for a much more elaborate set of accounts than in the 1968 Blue Book. The 1952 version of the SNA was put together very quickly by an exceedingly small team. The 1968 revision took a little longer and involved rather more people. However, these were mainly drawn still from developed market economies, and this is reflected in the fact that the 1968 Blue Book contains considerations for developing countries as a separate chapter and not as an integrated part of the whole. By comparison the 1993 revision had called on a very wide range of experts from all types of economies, both poorer and richer developing countries and, most recently, countries from Central and Eastern Europe that previously operated entirely as planned economies. The active involvement of national accountants throughout the world has been sought throughout the revision process by the regional commissions of the United Nations so that, in principle at least, all countries were invited to comment on the revision process as it developed. Many, though not all, were able to attend regional commission meetings on the revision and the inter-regional meeting in Mexico in October 1992.

13. The shortcomings of the 1993 revision From the outset of the revision process, and clearly spelled out at the first meetings in 1986, was the strongly-felt need that a description of the System from a theoretical point of view should be accompanied by handbooks detailing practical recommendations for implementation of the System. So far, none of these is finalised, despite the original intention of simultaneous publication, and the prospects for seeing a wide range of these in print in the next few years is rather bleak. In part, this is a question of resources, not simply of money but also of personnel. Those people who might have

186

A. Harrison/The SNA: 1968-1993 and Beyond

worked on handbooks have in the main been diverted to working on the Blue Book itself, and whether such resources will continue to be available to the process of documenting national accounts is uncertain if not doubtful. Some exceptions should be noted. In September 1993 the IMF published the fifth edition of the Balance of Payments manual and an associated implementation guide is in course of publication. Eurostat is due to publish a revised version of the ESA in 1994. Handbooks on inflation accounting and the implementation of the SNA in countries in transition are in preparation, and draft proposals for environmental accounting are with the UN printer. Another shortcoming concerns language. One problem with the 1968 Blue Book is that in various places the text can be interpreted in different ways. This ambiguity leads to variations in implementation that was not intended. Recognising this, the Expert Group recommended that as the new text was prepared it should be simultaneously translated into at least French and Spanish in the belief that the translation process would reveal such ambiguities in time for them to be corrected. In the event, translation into other languages has followed the English drafting process rather than taking place simultaneously. This is due in part to the logistics of preparing such a large text in multiple languages and to the sheer cost of translation. In principle, it would have been possible that ambiguities revealed by the translation into other languages of the draft for the Statistical Commission could have been used to further refine the English text, but this would have meant finding the resources for someone familiar with each of the official languages and the SNA to check the translations, and agreeing to postpone the finalisation of the English version until feedback from all of the these cross-checks had been incorporated. An area not yet finalised is that of the glossary. It was recognised early on that an extensive and comprehensive glossary was essential if the Blue Book was to be used effectively as a reference document as well as an instruction manual. Initially it was thought that it might be possible to compile the glossary first and then elaborate the text around this, but this too proved to be overly idealistic. Until the various issues were discussed and a means of treating them agreed, it was impossible to write glossary-type definitions. Just as the text itself represents a distillation of the protracted discussion and discussion papers that have taken place, so the glossary also must emerge of a distillation even of the Blue Book text itself.

14. Rationalisation In SNA terms, the development of a new Blue Book is to be considered in the context of capital expenditures. Like many capital projects, though, it took longer and cost more than originally expected, and the immediate question is: why? One reason is the emphasis on consultation with a very wide range of users. By the 1980s it was considered unacceptable for a small group to determine the course of the SNA without wider consultation, especially when the small group might be seen to have excessive representation of the more developed market economies. Recognition of this

A. Harrison/The SNA: 1968-1993 and Beyond

187

need led to the involvement of the UN regional commissions, to the series of regional commission meetings during 1990, and the inter-regional seminar to be held in 1992. The regional commission meetings in 1990 took approximately one year. This year could have been saved from the overall schedule but only at the cost of reducing the inputs from other countries, and this was felt to be unacceptable. A second main reason for the delay was that the exercise was essentially under­ resourced. All of the member organisations of the Inter-Secretariat Group have had at least one person nominated to follow the SNA revision process, but in no case was this a full-time allocation of a person and, increasingly, as time proceeded the inputs provided by these people came from voluntary unpaid overtime. Such a comment applies equally to the members of the Expert Group. Those who criticise the time the process has taken should perhaps pause to consider the degree of effort this voluntary contribution represents and note the personal contribution and dedication of most of those individuals involved, which in many cases was carried through changes of positions, and in some cases organisations, during the revision process but with a continued commitment to the completion of the task in hand. Only one expert formally withdrew because of a change in responsibilities. It was originally thought that the new Blue Book could be drafted by a single person in a period of twelve months, and as already mentioned Peter Hill was made available by OECD for this purpose, beginning in 1988. In retrospect, it is clear that this was too early to begin the full drafting process because during that year a great deal of Peter Hill’s time was occupied in attending Expert Group meetings and preparing for them. Further, the task as it emerged simply became impossible for one person to handle alone, even in another twelve-month period. Quite apart from the pure drafting considerable resources are needed, both for secretarial work and for organisational support in order to circulate drafts for comments, to receive the comments, and then find a means of incorporating these into "completed" chapters as other chapters are still in the process of development. Realistically, a project of this magnitude needs a full-time secretariat of several members in addition to those with primary drafting responsibility.

15. An assessment of the changes Despite the original intention that there would be no major changes in the revised Blue Book the present draft does indeed incorporate a large number of changes, several of which may be described as major. This leads some people not immediately connected to the revision process to accuse the experts of having been innovative and to have ignored the basic mandate from the Statistical Commission. Looked at in another way, recommendations of the Expert Group may seem to be unduly conservative. There has been no major change to the production boundary; housework, for example, continues to be excluded; services derived from consumer durables are also excluded; no extensions have been proposed to deal with environmental costs directly; the asset

188

A. Harrison/The SNA: 1968-1993 and Beyond

boundary has been modified but not to the extent of including research and development, human capital, or consumer durables as assets in the SNA sense. An area which caused alarm in some quarters is the treatment of the imputed service charge for financial intermediation (the preferred new terminology for bank service charges) yet there are many outside the immediate world of national accountants who will say that the proposals do not go far enough along the lines of treating all interest payments as payments for a service. Many significant changes concern the format of the accounts and the use of terminology, but there has been little criticism of these. The annex shows an overview of all of these changes taking account of the last changes recommended to the Statistical Commission. As described above, the fact that on only eight of the almost 100 changes was reconsideration necessary, indicates the wide degree of acceptance of the new proposals among the national accounting community. Most of the disquiet leading to the reconsideration of the eight issues was voiced by people from countries who are members of the OECD, though the disquiet was by no means unanimous even on a single issue. In general, these reservations are expressed by individuals who are articulate and used to voicing their opinions in public, and who in general have fora in which to do this, the meetings at OECD, at Eurostat and at meetings of the International Association for Research on Income and Wealth. On the other hand, countries from other regions have expressed quite a different point of view. Is it that those countries with more experience of implementing a more or less complete set of accounts are more foresighted in anticipating the problems that the proposed changes may bring? Or is it that those who have implemented a system are more reluctant to consider change than those who have yet to extend their accounts in new directions? In many areas of human endeavour, progress is subject to a sort of pendulum effect. The pendulum swings towards innovation and change until it is halted by fear of going too far. It then swings back towards conservatism and a plea for continuity with past practices, though usually the swing back does not come quite to the last extreme in this direction. After a while the impetus for innovation again starts the pendulum swinging in the opposite direction. This pendulum effect can be seen throughout the process of reviewing the SNA. The Statistical Commission originally called for a review of the System driven by the need to update the System, incorporate new aspects and ensure consistency. By the time they were ready to launch the process, however, the mandate had been cut back to one of simplification and clarification. The Expert Group started from this position but inevitably moved forward along the path of innovation. At one point the list for innovations was much more extensive than that finally appearing, but the Expert Group too, in its last meetings moved back in the direction of conservatism and continuity with the past. Those immediately outside the revision process seem at present to be pressing for conservatism. It will be interesting to see whether further discussion may perhaps lead to a change of emphasis in favour of change.

A. Harrison/The SNA: 1968-1993 and Beyond

189

16. A wider perspective Consider the issues that national accountants regard as contentious, for example, the imputing of rent on buildings owned by government and non-profit institutions. There are arguments for imputing this rent (mainly theoretical) and arguments against (mainly practical). The overall effect on the level of GDP and more particularly its change over years is likely to be slight. The economists and policy analysts using the national accounts are unlikely to get very excited about the exact treatment of this item. On the other hand, there is a real risk that such users will complain effectively and vociferously about the changes that have been rejected by the experts. Measurement of unpaid household activity, accounting for the environment, the treatment of human capital and research and development are all areas where economists over many years have insisted that the SNA is deficient and major change should be incorporated. In discussing the delay to the SNA review at the 1991 Statistical Commission, the Director of the United Nations Statistical Office said that national accounts was too important to be left to the national accountants. When reviewing the 1993 SNA there is a risk that outside analysts will decide that national accounts is too important to be left to statisticians and will try to produce their own alternative systems. Statisticians complain that economists criticise the SNA because they do not understand the System, what it attempts to do and what it clearly stands aside from. This position is justified, but the message on the whole is not well put to the user community and does not help to answer the legitimate questions that the users are asking: What level of growth would result if account were taken for depletion of natural resources? How far is the level of well-being affected by unpaid household services? How far does growth in the future depend on the amount spent on education now? One of the problems of producing good national accounts is that one needs staff to work on the accounts who are meticulous in their attention to detail, who are content to assemble a lot of detail that piece by piece is not particularly interesting, but together gives a comprehensive view of the economy, and generally be content to work in an area where public appreciation and esteem is rather low-profile. The characteristics of such staff are likely to lead them more easily into a heated debate over the rights and wrongs of imputing rent on government buildings than on some of the more expansive issues listed above. The period of retrenchment in government statistics during the 1980s has begun to reverse in a number of countries at least. This provides a window of opportunity for national accountants and statistical offices to respond more positively to the challenges being placed before them. It is important that consensus be reached on changes to be incorporated in the SNA. Clearly, those outside the Expert Group and Inter-Secretariat Group have a right to have their voices heard, and if necessary for different decisions or changes to be made. However, above and beyond this, the important message to take to the world of the users is that the System has been comprehensively reviewed. It has seen major extensions, consolidation and integration in the area of balance sheets and the treatment of assets. Satellite accounts and other means of extending the accounts should be seen not as a refusal to face up to

A. Harrison/The SNA: 1968-1993 and Beyond

190

alternative areas, but as a positive means of developing new techniques in areas with difficult implications in ways that could be a forerunner to innovations to the System as a whole. In particular, such a treatment is appropriate for the development of environmental accounting, accounting for research and development, social protection, and measures of unpaid household activity. Only if such a positive message can be whole-heartedly endorsed by the statistical community will we be seen to be responding appropriately to the needs of our users, and to offer to new graduates the prospect that working in a statistics office could be as interesting, challenging and rewarding as the presently more high profile activity of economic forecasting and policy analysis.

References [1] [2]

United Nations (1968) A System of National Accounts, Series F, No. 2, Rev. 3, 1968. United Nations (1977) Provisional International Guidelines on the National and Sectoral Balance Sheet and Reconciliation Accounts of the System of National. Accounts, Series M, No. 60, 1977.

[3]

United Nations. Provisional Guidelines on Statistics of the Distribution of Income, Consumption and Accumulation of Households, Series M, No. 61.

[4]

United Nations (1979) Manual on National Accounts at Constant Prices, Series M, No. 64, 1979.

[5]

United Nations (1979) Guidelines on Statistics of Tangible Assets, Series M, No. 68, 1979.

[6]

United Nations (1980) R eport of the Expert Group Meeting on Future Directions for W ork on the United Nations System of National Accounts. E /C N .3/A C .9/5, May, 1980.

[7]

United Nations (1982) R eport of the Expert Group Meeting on the Review and Development of the

[8]

IM F (1977)Balance of Payments Manual, fourth edition.

[9]

Eurostat (1979) European System of Integrated Economic Accounts, Second edition.

UN SNA. C E S/W P.22/66 May 1982

[10] United Nations (1989) Basic Methodological Principles governing the compilation of the System of Statistical Balances of the National Economy, Series F, N o l7 /R ev .l, 1989. [11] Petre, Jean (1981) The Treatm ent in the National Accounts of Goods and Services for Individual Consumption Produced, Distributed or Paid for by Government. Paper prepared for 17th General Conference of the International Association for Research in Income and W ealth, August 1981.

A. Harrison/The SNA: 1968-1993 and Beyond

Annex I: Schedule of Expert Group meetings 1. SNA Structure

23-27 June, 1986

Geneva

2. Prices and quantity

10-14 November, 1986

Luxembourg

Comparisons 3. External sector

23 March-2 April, 1987

Washington

4. Household sector

30 August-4 September, 1987

5. Public sector

25-29 January, 1988

Florence W ashington

6. Production accounts and

21-30 March, 1988

Vienna

7. Financial flows and balances 8. Coordinating group

6-15 September, 1988

W ashington

23-27 January, 1989

Luxembourg

9. Coordinating group 10. Coordinating group

12-21 July, 1989

New York

13-22 September, 1989

New York

11. Reconciliation of SNA

4-9 December, 1989

Moscow

12. Coordinating group

3-7 December, 1990

Washington

13. Coordinating group

8-17 April, 1991

H arare

14. Coordinating group

1, 2, 10 October, 1992

Aquascalientes

input-output tables

192

A. Harrison/The SNA: 1968-1993 and Beyond

Annex II: Participants at Expert Group meetings Nam e

A ffiliation

H ans Adler

Canada

Peter Al

Netherlands

Odd Aukrust

Norway

Eneas Avondoglio Jack Bame Keith Blackburn A driaan Bloem Cornelius Bochove Carol Carson M. de Cartagena Shirly Carter U m a Roy Choudhury M argaret Costa Jean Courty Pierre Demangel Barbara Dunlop Bernardo Ferran Lourdes Ferran Alfred Franz Raul Garcia M.E. Comez-Luna Alan Heston Piroska Horvath Ralf Hussmans Yuri Ivanov P. Jayasundera Jagdish Kumar

Argentina USA Australia Netherlands Netherlands USA Ecuador UK India Brazil France Congo Australia Venezuala INSTRAW Austria ECLAC Mexico USA Hungary ILO CMEA Sri Lanka India ESCAP Japan Canada Colombia Germany Israel Mauritius Congo Netherlands Zimbabwe UK USA UK ECA Indonesia Hungary Canada Papua N.G. Finland Denm ark France Canada Mauritius

Y. Kurabayashi Kishori Lai M arian Libreros Heinrich Lutzel Pablo Mandler Devi Manraj Michel Mouyelo M. van Nieuwkerk Moffat Nyoni Colin Pettigrew Karin Polenske G raham Pyatt R en Rakatobe K. Salah Gyorgy Sandor Bogdan Sculz John Shadlow Erkki Tassia Bent Thage A ndre Vanoli Stewart Wells D. Zmanay

1 2 3 4 5 6 7 8 9 10 11 12 13 X X X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X

X

X

X

X

X X X

X X X X X

X X

X X X

X

X

X

X X X X X

X

X

X

X

X X

X

X

X

X

X X

X

X X X X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X

X

X

X

X X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X X X X X X

X

X

X X X

X

X

X

X

X

X

X

X

X X

X

X

A. Harrison/The SNA: 1968-1993 and Beyond

Piero E rba Alain Chantraine Brian Newson H ugon Krijnse Locker J. Rom an R. Salvat W erner Thon

Eurostat Eurostat E urostat E urostat Eurostat Eurostat Eurostat

A rie Bouter Mahinder Gill Chandrakant Patel Jonathan Levin Kevin O ’Connor

IM F IM F IM F IM F IM F

X

Peter Hill Derek Blades Erwin Veil Jeff Owens J.E. Chapron

O ECD O ECD O ECD O ECD O ECD

X

X

X

X

Laszlo Drechsler Jan van Tongeren Christina Hannig

UNSO UNSO UNSO

Ram esh Chandler John O ’Connor Ernst Lutz

W orld Bank W orld Bank W orld Bank

Anne Harrison

W orld Bank Consultant

X

X

X

X X

X

X

X

X

X

X

X

X

X X X X

X

X

X

X X

X

X X

X

X

X

X

X

X

X X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

x

x x x

X

x x x x x

194

A. Harrison/The SNA: 1968-1993 and Beyond

Annex III: Changes for the 1993 SNA 1. Accounts 1.1 The production account will be partitioned to show value added as an explicit balancing item leading to a generation of income account showing the various charges against value added. 1.2 The income and outlay account will be partitioned into a distribution of primary income account, a secondary distribution of income account, and a use of income account. 1.3 The distribution of primary income account will be further sub-divided to show an entrepreneurial income account. 1.4 Separate accounts will be introduced to show social transfers in kind and their impact on income and expenditure. 1.5 The capital finance account will be partitioned to show a capital account and financial account separately. 1.6 The reconciliation account is replaced by two other more comprehensive accounts, one showing other volume changes in assets, and one revaluation. 1.7 There will be a full articulation between the flow accounts and the balance sheets. 1.8 Production accounts will be introduced for institutional sectors, including the household sector, as well as for industries.

2. Sectors 2.1 Non-profit institutions serving households (NPISHs) will be kept as a full sector of the system. 2.2 Financial auxiliaries will be included in the financial corporate sector. 2.3 Quasi-corporations will be restricted to those unincorporated enterprises that have a full set of accounts including a balance sheet and withdrawal of income from quasi-corporations. 2.4 The system will allow the existence of unincorporated financial enterprises. 2.5 Seven sub-sectors will be suggested for the financial corporate sector. 2.6 Guidelines will be given on how to sub-sector the household sector. 2.7 For non-financial and financial corporations a distinction should be made between public, private, domestic and foreign controlled corporations. 2.8 Supplementary tables should show the non-financial public sector and the non-m onetary public sector.

3. Production 3.1 No major change will be m ade to the production boundary. Housework will still be excluded. 3.2 Intra-establishment deliveries will be excluded from the measurement of production. 3.3 Intra-enterprise deliveries will be included in the m easurement of production. 3.4 The measurement of growth of plants will be brought into line with m easurem ent of growth of animals. 3.5 The production of literary artistic works will be included within the production boundary. 3.6 Payments to use a patent will be treated as payments for a service and the action of making the patent available to another user treated as output. 3.7 Ancillary activities are described in detail with a recommendation that separate establishments should not be established in respect of these. 3.8 The action of storage is recognised as a productive activity representing transport over time, comparable to transport over space. This will mainly apply to seasonal goods, and in particular, agricultural products. 3.9 Illegal activities should not be excluded as a m atter of principle from the production boundary of the system.

A. Harrison/The SNA: 1968-1993 and Beyond

195

4. Consumption 4.1 A distinction will be m ade between individual and collective final consumption. 4.2 A corresponding distinction will be made between final consumption expenditure and actual final consumption. 4.3 Social insurance schemes will cover private funded schemes as well as social security and unfunded employee schemes. This implies, among other changes, that pensions in the SNA will be treated in a way similar to that presently found in the ESA. 4.4 By convention all expenditure by NPISHs will be treated as individual consumption. 4.5 The concept of consum er subsidy will not be introduced in the SNA. 4.6 Consum er durables will continue to be shown in the balance sheets only as a m em o item, and will not be treated as fixed capital.

5. Accumulation 5.1

Research and development will not be treated as fixed capital in the SNA.

5.2

Expenditure on mineral exploration will be treated as a fixed asset.

5.3

M ajor expenditure on computer software will be treated as a fixed asset.

5.4

Literary artistic work may be treated as a fixed asset.

5.5

A concept of valuables will be introduced as a third type of capital formation in addition to fixed

5.6

Gold may be held as a financial asset only by monetary authorities.

capital formation and changes in inventories. 5.7

No depletion allowance will be introduced in respect of the consumption of natural resources.

5.8

No degradation allowance will be introduced in respect of environmental pollution.

5.9

The scope of non-produced economic assets will be expanded to include non-cultivated biological resources and water resources (only those resources actively drawn into the production process).

5.10 Offensive weapons and their means of delivery will continue to be treated as current expenditure. 5.11 O ther expenditure on military durables will be treated as capital expenditure. 5.12 Hum an capital will not be introduced into the next SNA. 5.13 Structures such as roads, bridges and buildings constructed by the community will be recognised as fixed assets. 5.14 W here appropriate, consumption of fixed capital should be calculated on roads and other assets previously treated as having infinite service lives. 5.15 Capital stock information is to be shown cross-classified by owner and user. 5.16 Inventories held by Government will be extended to all inventories held, not just strategic materials. 5.17 The category of non-produced intangible assets will include transferable contracts. 5.18 The same category will include purchased good-will. 5.19 In principle, historical monuments should appear in the balance sheet in the same categories as structures serving a similar function. 5.20 W ork in progress may be recorded for the output of some service industries.

196

A. Harrison/The SNA: 1968-1993 and Beyond

6. Transactions 6.1

The item mixed income will be introduced to replace operating surplus as the balancing item for unincorporated enterprises on the generation of income account.

6.2

Reinvested earnings of foreign direct investment will be introduced for consistency with balance of payments data.

6.3

Government licenses and fees will be treated as taxes or fees for service in accordance with

6.4

Imports will be recorded in total fob.

6.5

Domestic transport provided in respect of im ported goods will no longer have to balanced by imputed

6.6

exports of services. The imputed service charge for financial intermediation should in principle be allocated by sector and

6.7

Observed interest flows will be adjusted by this imputed service charge in recording interest payments

6.8

The measurement of output of insurance companies will be increased by estimating a higher imputed

6.9

The value of this service charge will be estimated taking into account the investment income of

conventions in the IM F s Government Finance Statistics.

industries, but may not be. and receipts in the system. service charge to these companies. insurance companies and changes in their technical reserves. 6.10 This investment income will be treated as payable to policy holders and returned by them to insurance enterprises as part of the higher service charge. 6.11 The time of recording of insurance premiums and claims will be altered to coincide with the period covered by the premium and the time at which the eventuality occurs that gives rise to the claim, rather than the time of payment of premium and claims. 6.12 Acquisitions and disposals of capital formation will be shown separately. 6.13 Users will have the option of whether to show balancing items in the accounts on a gross or net of consumption of fixed capital basis. 6.14 In the Production account although the System gives preference to the use of basic prices, users will have the option of valuing output at producer prices. 6.15 Explicit recommendations will be given in the situation where multiple exchange rates exist. This will imply the introduction of imputed taxes and subsidies when multiple exchange rates are official. 6.16 The concept of financial leasing will be introduced explicitly. 6.17 Nominal holding gains and losses will be split into neutral and real holding gains and losses. 6.18 These will be shown only in the revaluation account and not in the accounts dealing with distribution and use of income. 6.19 The treatm ent of VAT will be made explicit and net recording recommended. 6.20 The conversion of commodity gold to monetary gold or vice-versa will no longer involve notional balance of payments transactions but will be recorded in the other changes in volume account. 6.21 A three-way table of financial transactions by creditor and debtor sector for each type of financial instrument will be introduced. 6.22 The system will recommend explicitly how losses are to be treated.

A. Harrison/The SNA: 1968-1993 and Beyond

197

7. Terminology 7.1

The distinction between commodities and other goods and services and between industries and

7.2

A distinction between m arket and non-m arket output and producers will be introduced.

7.3

Non-market output will include production on own account.

7.4

Changes in stocks will in future be referred to as changes in inventories. The word "stock" will be

7.5

The concept Gross National Product will be renam ed Gross National Income.

producers of other goods and services will be dropped.

kept for use in connection with capital stock. 7.6

Gross O utput will be replaced by Output.

7.7

The distinction between direct and indirect taxes will be dropped.

7.8

The term s "commodity taxes" and “other indirect taxes” will be replaced by the expressions "taxes on products” and "other taxes on production".

7.9

T rue and approximate basic prices will be replaced by a single concept called basic prices.

7.10 The term "purchasers

prices" will generally replace "market prices”.

7.11 Gross wages and salaries will be replaced by wages and salaries. 7.12 There will be no reference to factor incomes. 7.13 The term "corporation" will be used to denote corporate or quasi-corporate enterprise. 7.14 The term "banks" will be replaced by "financial intermediaries"and "banking” by "financial intermediation". 7.15 A distinction is to be m ade between consumption expenditure and actual consumption. 7.16 An aggregate for transactions over all resident sectors will be referred to as relating to the total economy rather than the national economy. 8. Constant prices 8.1 Chain-linking will be recommended as a standard practice for calculating flows at constant prices. 8.2 The concept of real net national income will be introduced. 8.3 In general, when the same goods are available at different prices, they are treated as different goods containing different amounts of service. Changes in the composition of the amount of goods purchased at different prices is therefore treated as a change in volume. In future, if price discrimination is practised such that the purchaser has no choice about the price he has to pay, a changing composition of the amount of goods purchased at different prices will show as a change in price rather than a change in volume.

9. Other 9.1 The Blue Book will recommend the introduction of satellite accounts as a means of extending the analyses available based on SNA concepts. 9.2 Flexibility of the system will be emphasised in a chapter on adaptations and applications. 9.3 The Blue Book will contain a chapter on social matrices explaining how these can be used to enhance the analyses available within a framework consistent with the SNA.

The Accounts o f Nations Z. Kenessey (Ed.) IOS Press, 1994

198

CONSTANCY AND CHANGE IN THE UNITED NATIONS MANUALS ON NATIONAL ACCOUNTING (1947, 1953, 1968 and 1993) Frits Bos National Accounts Unit Bl, Eurostat, Jean Monnet Building, L2920 Luxembourg

Abstract. In this paper, the successive United Nations Manuals on national accounting (1947, 1953, 1968 and 1993) are compared. The m ajor changes considered are: accounting structure, sector classifications and scope. Many basic concepts, like the production and the asset boundary, have hardly been changed since the 1947 report. Since the guidelines have just been revised, they will most probably constitute the core of national accounting conventions for more than 50 years.

Introduction In 1947, the United Nations published the report "Measurement of national income and the construction of social accounts" [9], The report was a landmark in the history of national accounting for two reasons. First, it contained the first fully elaborated and detailed national accounting system. Secondly, it contained for th e irs/ time international recommendations on compiling national accounts. Since that time, several international guidelines on national accounting have been published. In order to help planning the Marshall-aid, the OEEC developed a "Standardised System of National Accounts", which was published in 1952 [8]. In the next year, rather similar guidelines were published by the United Nations; it is commonly known as ’the first SNA’ (System of National Accounts) [9]. The similarity was not surprising, as Stone played a major role in constructing both guidelines. The OEEC guidelines and the first SNA can be dubbed as the second generation of international guidelines on national accounting. The third generation started by the publication of a new SNA in 1968 [10]. In revising the SNA, Stone again played a central role. In 1970, the EC issued their own guidelines: the European System of National Accounts (ESA) [4], Its basic concepts are rather similar to the second SNA. The ESA is intended to be in line with the UN guidelines, while clarifying some of the SNA-concepts and taking more explicit account of the needs of the EC-countries. At the moment of this writing, the revision of the SNA is nearly completed and the drafting of the ESA is reaching its final stages. This fourth generation of international guidelines is expected to be approved in 1993 and 1994. Since their first appearance, the international guidelines have been of great practical

F. Bos/On the UN Manuals

199

importance to national accounting because of their function as pedagogical instrument and international standard for harmonizing the national accounting concepts employed throughout the world. Their role as a pedagogical instrument was crucial in spreading the idea of national accounting. The guidelines made clear why national accounts figures should be compiled, explained the basic logic of national accounting and provided a clear example of how a national system could be set up. Furthermore, they employed concepts on which consensus was reached by leading international experts. The international organizations played a major role in spreading the use of the conventions adopted in the international guidelines. They stimulated their application by providing free statistical help and advice. And they enforced the concepts by asking member countries to submit figures following these concepts and by employing these figures also in taxing and subsidizing their member countries. A major purpose of the international guidelines was to improve the international comparability of national accounts figures by providing standard concepts and classifications. This is clearly expressed in the 1947 report: "The Sub-Committee hopes that the guiding principles and recommendations formulated ... will be applied to the widest possible extent in each country in the computation of national income and related accounts in order to secure greater international comparability than in the past" [9, Preface], In 1984, Richard Stone was awarded the Nobel-prize to a main extent for his role as principal architect of the international guidelines on national accounting. In this way, also the importance of the international guidelines to economic science was acknowledged. In this paper, the successive United Nations guidelines on national accounting (1947, 1953, 1968 and 1993) are compared. In section 2, the scope, global content and purposes of these guidelines are discussed. In section 3, the four reports are compared regarding some selected topics like the production boundary and the asset boundary. For a comparison on the other topics, we refer to [2], Conclusions are drawn in section 4. This paper does not discuss the historical context of the guidelines or the links between the national accounting concepts and economic theory. For these issues, we refer to [1] and [2], A much more detailed comparison of the 1968 and 1993 Manuals, including backgrounds of the revision process, can be found in the paper by Harrison in this volume [6]. The relationship between the international guidelines and the national accounts actually compiled by the various countries is rather complicated. A concrete example is given in [3], which discusses the implementation of the revised SNA for the Netherlands.

200

F. Bos/On the UN Manuals

1. Scope and global content

The reports have several differences in purpose, content, scope and amount of detail. We will limit ourselves in this section to some general remarks. No simple chronological sequence can be found in the reports, e.g. that the successive reports grow bigger and only add subjects and detail. In fact, the report of 1953 is in most respects simpler and less ambitious than the 1947 report. It is also the smallest in size. However, if we disregard the 1953 report, there is clear trend of increasing scope and detail which culminates in the 1993 report. In the 1947 report, relatively much attention is paid to the purposes and advantages of national accounting. In general it is stated that: "It is necessary to know in quantitative terms how the national income is related to its constituent transactions and to other totals of transactions. Thus modern enquiries which had their origin in an attempt to measure certain broad totals have changed their emphasis and now concen­ trate more on the structure of the constituent transactions and on the mutual interdependence of these transactions. It has come to be realized that for different purposes certain related but distinct aggregates are useful" [9, p. 24], Thus, the computation of national aggregates is just one of the main concerns of the 1947 report. In contrast to the 1947 report, the compilation of broad totals seems to be the only objective in the 1953 report; this is evidenced by the sector accounts: these should be registered in a very specific way in order to obtain national aggregates directly from the sector accounts. The 1968 and 1993 reports adopt an approach similar to the 1947 report. The 1968 report states that: "For many purposes of both analysis and policy, however, it is not sufficient to work with aggregates alone; it is also necessary to look at many aspects of the economy in greater detail. As a result, economic models are now built in which the main aggregates are subdivided, or disaggregated" [11, para 1.10]. In the 1993 report, in addition, also the multi-purpose character of the guidelines is stressed: "The System is primarily intended to provide disaggregated data to meet the needs of analysts and policy makers interested in the behaviour of markets and the factors responsible for major disequilibria such as inflation and unemployment. The design of the System and its coverage of economic activities has to be a compromise intended to yield the maximum benefits to users of all kinds and may not therefore be ideally suited for any purpose taken in isolation" [12, Ch. I, p. 22]. The limitations of GDP and national income as measures of welfare are discussed only in the 1993 report. For example, "[aggregate indicators such as GDP do not reflect any changes in distribution [of wealth and income] that may be taking place over time, so that such indicators need to be supplemented by micro-data for purposes of analysing changes in welfare" [12, Ch. I, p. 21]. Similarly, "a natural disaster may well lead to an increase in GDP by creating extra demands, even though the community may be no better off than in the previous period if the loss of welfare caused by the disaster exceeds the increase in welfare from the extra production and consumption" [12, Ch. I, p. 20],

F. Bos/On the UN Manuals

201

In the 1947 and 1993 reports, concepts are quite extensively discussed. In general, both the 1947 and the 1993 report can be regarded as introductions into the logic and basic principles of national accounting. The 1953 and 1968 reports are better characterized as systematic enumerations of recommended booking conventions. In these reports, the considerations underlying the choice of concepts are discussed much more sparingly. A major motivation for revising the 1968 report was that "its manner of exposition makes it inaccessable to many users especially those new to the System of National Accounts" [5, p. 336). So, in the 1993 report, "key concepts [are] defined by rationale and not by enumeration" [5, p. 336]. A problem with the discussion of concepts in the 1947 report is that sometimes more than one alternative is proposed, for example in the case of the concept of residency. Only in the 1947 report is some advice given on how to estimate the various items and aggregates. The accounting systems recommended in the reports, contain flows for which the values can readily be observed as well as flows for which values have to be imputed (e.g. income in kind or imputed services of owner-occupied dwellings). The fundamental problems with such imputations are already noted in the 1947 report: "The items in the tables may be divided, from one point of view, into cash terms and imputed items, the former being those elements which reflect market transactions, and the latter being those for which a calculation has to be made in the absence of market transactions. In view of the difficulty of finding a commonly accepted basis for the second type of estimate it is desirable that, as far as possible, items of this kind should be shown separately" [9, p. 18]. In the 1993 report, the distinction between imputed items and other items is explicitly incorporated in its supplementary classifications. In the 1947 report, guidelines on national accounting are regarded as a flexible instrument in achieving better and more comparable national accounts figures. According to the report, flexibility is required because national accounting conventions can not take account of all institutional differences in the world and all specific uses of national accounts figures. For example, it is clearly expressed that "in applying this [national accounting] system it will frequently be necessary to extend and adapt it to the particular circumstances of different countries" [9, p. 18, para 11], Furthermore, "[experience shows that unavoidable differences of opinion arise in the treatment of certain transactions due in large measure to institutional differences in different countries" [9, p. 18, para 10]. As a final case in point: "where income and expenditure components are used to work out behaviouristic or institutional coefficients such as the propensity to consume, the relation of imports to national income, etc., [it] may be argued that ... those definitions should be chosen which give rise to the simplest enduring regularities. Thus such questions as whether corporate taxes should or should not be included in national income may, if they cannot be settled on theoretical grounds, be decided by econometric analysis" [9, p. 24], In the 1953 and 1968 reports, such remarks on the interaction between national accounting concepts on the one hand, and particular circumstances and uses on the other are absent. The 1993 report returns to the philosopy of the 1947 report: "A prominent new

202

F. Bos/On the UN Manuals

feature of the revised SNA is that it emphasizes flexibility. The revised SNA includes a decription of how the central framework can be applied to policy and analytical requirements, data availability and other specific circumstances of different countries. Also, it describes how satellite accounts with alternative product and income concepts can be elaborated as an extension of the central framework" [12, Annex on changes]. The national accounting system outlined in the 1947 report is "based essentially on the model of an advanced industrial economy in which transactions in money are dominant" [9, p. 24). This contrasts with the approach in the later guidelines, which are also intended for countries in which barter transactions and own-account production play a more dominant role. In the 1968 report, there is even a separate chapter dedicated to supplementary classifications especially useful for the description and analysis of the economies of developing countries. The 1993 report intends to take care of the specific needs and data possibilities of developing countries (and countries in transition) by means of its flexible approach. A novel feature of the 1968 report was that the national accounting system was explained on the basis of a matrix. This matrix presents an overview of the accounts and their main classifications.0 It also integrates the input-output and supply and use tables with the sector accounts. Such an overall presentation of the system is absent in the 1947 and 1953 reports. In the 1993 report, the matrix is one of the alternative types of presenting national accounts (balancing statements, diagrammatic presentation, equations and matrix). The matrix plays of course the central role in the 1993 chapter on Social Accounting Matrices. In the 1947 and 1953 reports no attention is paid to the description of balance sheets, make and use-tables, input-output tables and figures in constant prices; these are all discussed for the first time in the 1968 report. Residency, the time of recording and the principles of valuation of transactions are only discussed in very general terms in 1947; the 1953 report contains a somewhat more elaborate treatment of these topics, but detailed recommendations are only presented in the 1968 report. The 1968 report pays only lip-service to the balance sheets and their link to the flow accounts. In fact, the 1993 report is the first report to discuss balance sheets and other changes in assets accounts. Only in the 1993 report is attention paid to the concepts of employment and unemployment. In its basic system, the focus is only on employment in terms of jobs, hours-worked and full-time equivalents [12, Ch. XVII], However, persons employed and unemployment are also incorporated in the SAM framework presented in the report [12, Ch. XX).

F. Bos/On the UN Manuals

203

2 . Selected topics

2.1 Production boundary In the reports, the definition of the production boundary is linked to transactions in money: in general, only the generation of goods and services which are exchanged for money is to be accounted for as production. However, the production boundaries are extended by accepting certain major exceptions to this rule: - Types of production that are not paid for separately at market prices: * production of government services; * banking services not explicitly charged for; * services by pension funds; * services by life and casualty insurance companies; * services of non-profit institutions. (A more extensive discussion of indirectly financed production can be found in section 3.2) - Own-account production of capital goods; - Imputed services of owner-occupied dwellings; - Output used for compensation of employees. The 1953 and 1968 reports have extended the production boundary even further. In the 1953 report, own account production by primary producers, namely those engaged in agriculture, forestry, hunting, fishing, mining and quarrying, is included [10, para 30]. The 1968 report draws an even wider production boundary by encompassing all production of primary products as well as the processing of primary products by the producers. Examples of such processing are the making of butter, cheese, flour, wine, oil, cloth and furniture [11, para 6.19]. The production boundary in the 1993 report differs in some respects from the one in the 1968 report: - "With regard to own-account production of goods by households, the revised SNA has removed the 1968 limitations which excluded [*] the production of goods not made from primary products, [*] the processing of primary products by those who do not produce them, [*] and the production of other goods by households who do not sell any part of them on the market" [12, Annex on Changes, p. 9]. - "The storage of agricultural goods produced by households is included as is supply of water (water carrying)" [12, Annex on Changes, p. 9], - "The revised SNA includes in gross output literary artistic works (i.e., the writing of books, composing music, etc.) which are produced for sale" [12, Annex on changes, p. 18], Despite these differences, the production boundaries in the successive guidelines correspond to each other in many important respects: they all include the services of owner-occupied dwellings and they all exclude unpaid household services, do-it-yourself activities, voluntary work and the services of consumer durables. These types of production are ignored despite the existence of paid counterparts that are counted as

204

F. Bos/On the UN Manuals

production. The services of labour, financial capital and land are often also exchanged for money. Nevertheless, such services in none of the three reports are regarded as being ’produced’. For example, payments of wages, interest and dividends are not registered as payments for output by households or enterprises. 2.2 Principles o f valuation In this section, we start with a description of the general principles of valuation employed. We then continue by describing some specific valuation conventions, e.g. the valuation of the output of pension funds, insurance companies and the government. At the end constant prices are discussed. In the four reports, market prices are the dominant principle of valuing transactions and transformations (i.e. production and consumption). Current market prices are generally preferred to a valuation at costs. In the 1953 and 1968 reports, this preference for market prices is made clear for e.g. barter trade, own-account production, changes in stocks, services of owner-occupied dwellings and income in kind [10, para 136; 11, para 6.21]; the 1947 report is only explicit about the services of owner-occupied dwellings [9, p. 66]. For practical reasons, the 1953, 1968 and 1993 reports allow for valuation at costs instead of at (producers’) prices on the market. The 1953 and the 1993 reports mention own-account construction [10, para 55] and the 1968 report mentions own-account capital formation and work-in-progress [11, para 6.24], Of course, for the collective services of the government valuation at costs is the only possible prin­ ciple of valuation. In the reports, nearly all invoiced values for the delivery of goods and services are accepted as ’true’ market values. Payments for goods and services of fiscal monopolies of the government (e.g. on the sale of alcoholic beverages), are regarded by all reports fully as payments at market prices. In that respect, the 1953, 1968 and 1993 reports recommend to record the operating surplus reduced by the normal margin of profits of business units, as indirect taxes. However, in some cases the reports do not accept the invoiced values (fully) as market values: - Nominal payments to the government and payments mainly meant to increase government revenues, e.g. payments for driving tests and licenses. - Transfer pricing. In the 1953 and 1993 reports only, transfer pricing is rejected for national accounts purposes. In the other reports, the occurrance of transfer pricing is not mentioned at all. - Charges in respect of delayed payment. In the 1968 and 1993 reports only, it is explicitly stated that market prices refer to "the price at which the commodities are sold against immediate cash payment" [11, para 6.16]. In the 1968 report, input-output analysis and supply and use tables are introduced. Valuation principles are advocated which are particularly suited to this type of analysis. In the accounts1, output is to be valued at producers’ prices and intermediate

F. Bos/On the UN Manuals

205

consumption at purchasers’ prices. The implication is that invoice values should not be used in case they include additional charges for trade and transport, i.e. when invoice values and producers’ values do not coincide. The 1993 report also incorporates input-output and supply and use tables. Like the 1968 report, intermediate consumption should be valued at purchasers’ prices. However, unlike its predecessor, valuing output at (approximate) basic prices is preferred to a valuation at producers’ prices [12, Ch. VI, p. 64], In the 1947 and 1953 reports, no valuation principles are mentioned for classifying goods and services produced and consumed. In the 1947, 1953 and 1968 reports, the concept of value added employed in the accounts and tables is at factor costs. In the 1993 report, value added at basic prices is recommended. The concepts of indirect taxes in the reports differ in some respects. The 1947, 1968 and 1993 reports define indirect taxes as as those taxes that are "levied on goods and services and chargeable as a cost against the proceeds of sale" [9, p. 31; 11, para 7.26; 12, Ch. VII, p. 17], The 1993 report therefore prefers the term "taxes on production". However, according to the 1953 report, indirect taxes encompasses also taxes on the possession or use of goods and services by households [10, para 255]. Examples given are motor vehicle duties and taxes on the operation of television sets. In the other reports, these are regarded as other current transfers paid by households to the government. Another difference occurs with respect to profits of fiscal and other public monopolies. In the 1953, 1968 and 1993 reports, profits of fiscal monopolies, like tobacco and alcohol monopolies, are to be recorded as indirect taxes [10, para 255; 11, paras 7.31 and 7.32; 12, Ch. VII, p. 24], The 1968 report also regards the monopoly profits of public enterprises (e.g. of public utilities) as indirect taxes. The 1947 report rejects this option "since private monopoly profit is usually treated as profit and not as a privately levied tax, there seems no advantage in choosing a different treatment in the case of public authorities" [9, p. 59], The 1947 report is silent about fiscal monopolies, while the 1953 report is silent about non-fiscal monopolies. The 1993 report also includes in indirect taxes (and taxes on products): - the profits of import and export monopolies; - implicit taxes resulting from the operation of an official system of multiple exchange rates. Both of these cases are not mentioned at all in the other reports. Subsidies are in all the reports defined to encompass only payments to producers and not payments for consumption purposes (e.g. ’subsidies’ for renting a dwellings). So, the 1953 report defines subsidies not fully as the mirror-image of indirect taxes. None of the reports mentions the wide-spread occurrance of implicit or indirect subsidies. Cases in point are the remission of income taxes for industrial Research and Development or allowing that capital formation is written down in one or two years in determining taxable income. Market prices may differ according to the type of purchaser. In some instances, e.g.

206

F. Bos/On the UN Manuals

compensation of employees in kind, a choice has to be made between imputing a market price from the point of view of the employer or the employee (e.g. wholesale trade prices or retail trade prices). In the 1953,1968 and 1993 reports, a choice is made in favour of the producer’s point of view regarding costs and revenues. The 1947 report does not discuss this issue. In the absence of market prices (e.g. for government services or for not directly charged banking services), other principles of valuation have to be recommended. In such cases, there is a clear need for specific conventions. Below, we will shortly describe the major specific conventions concerning the valuation of production (output). In all the reports, the value of trade production is not equal to the gross value of the sales, but is defined as the net value of the sales. So, the costs of purchasing the goods, services, etc. sold have to be deducted. All four reports agree that many banking services are not separately invoiced, but are financed indirectly, namely by the margin between, simply said, interest rates for lending and borrowing. More precisely, they recommend to value the production of such banking services as the excess of the property income they receive over the property income they pay out (excluding the income on investment of ’own resources’ and excluding the bank’s own distribution of property income, e.g. the dividends they pay to their shareholders). The reports also agree that premiums paid to casualty insurance companies, life insurance companies, pension funds and social security institutions are not (fully) to be recorded as a payment for services (in the period they are paid or due). The 1947 report defines the value of production of life insurance companies as the investment income plus the excess of premiums due over claims due [9, p. 42], According to the 1953 report, the value of production of life insurance companies should be equated to their expenditure on administration and control; in this case no operating surplus is mentioned. The 1968 report defines the value of production as: "the excess of premiums received over the total of claims paid and net additions to actuarial reserves, excluding the interest on these reserves which accrues to policy holders" [11, para 6.38]. The 1993 report recommends a similar, but more precise definition: * the excess of premiums earned over claims due and net additions to actuarial reserves and reserves for with-profits insurance * holding gains or losses on investments allocated to policy holders or claimants * net income from investments of the insurance technical reserves. [12, Ch. VI, p. 42], This implies that there is a major difference between, on the one hand, the convention in the 1953 report and, on the other hand, those in the 1947, 1968 and 1993 reports. The latter three reports have rather similar conventions for life insurance. In the 1947 and 1993 reports, the definitions of the gross output of casualty insurance companies are similar to their definitions for life insurance companies. The 1953 and 1968 reports employ different definitions for casualty insurance and life insurance companies. The 1953 report recommends to value the output of casualty insurance companies as the sum of administrative expenses plus operating surplus. A definition

F. Bos/On the UN Manuals

207

of operating surplus is not given. So it remains unclear, e.g., whether it should depend on accounting practice, or whether capital gains should be included or excluded in principle. A comparison to the conventions in the 1947 and 1993 reports is therefore difficult. The 1968 report regards the excess of premiums due over claims due as the value of production of casualty insurance companies. So, unlike in the 1947 and 1993 reports, net income from investments of the insurance technical reserves are not included. The 1947, 1953 and 1968 reports agree that the value of production of pension funds is equal to their expenditure for administration and control. However, the 1993 report recommends a valuation principle similar to that for insurance companies.2 Full agreement exists with respect to the value of production of social security institutions. The 1947, 1968 and 1993 reports advocate a valuation equal to the expenditure for adminstration and control. The 1953 report is not explicit on this issue. So, the 1947 and 1968 reports propose similar conventions for pension funds and social security institutions, while the 1993 report treats the pension funds on a par with life insurance companies. The reports agree that the value of production at ’market prices’ of the government and of non-profit organizations should be defined as equal to at least the sum of intermediate consumption, capital consumption, compensation of employees and indirect taxes paid [11, para 6.41], However, according to the 1947 report, the value of government production also includes the payment of interest on government debt as far as it arises "in connexion with government enterprises or the provision of ordinary peace-time service such as education, public sanitation, etc." [9, p. 72], Excluded are however interest payments related to public debt that "has been incurred in meeting such charges as the cost of war or temporary deficits on social security funds and as a consequence has little or no counterpart in the form of productive assets" [9, p. 72], Furthermore, only in the 1953 report, gross rent is imputed for buildings owned and occupied by the government3. However, an exception is made for historical buildings and museums. The imputation implies that a large part of interest on goverment debt as was to be included in the value of government output according to the 1947 report, is also covered in the 1953 report. In none of the reports, is a similar imputation recommended for the use of infrastructural works, like bridges and roads. Only the 1968 and 1993 reports pay attention to volume measures and constant prices. Both discuss in general terms the major problems, like changes in quality, new products and measuring the volume of non-market services. Both reports favour the use of Laspeyres indices. The 1993 report stresses the importance of regular rebasing of these indexes and recommends also the use of chain indices. The 1993 report also shortly mentions purchasing power parities.

208

F. Bos/On the UN Manuals

2.3 Sectors and units In this section, the treatment of sectors and units in the four reports are discussed and compared. In tables 1-4, the sector classifications in the four reports is represented. In the 1947 and 1953 reports, one sector classification is used for all the accounts. In the 1993 report, the Production Accounts are not only presented by sector, but also by economic activity. In a separate table (table XV.7), value added is broken down simultaneously by sector and by economic activity. All the accounts other than the Production accounts pertain only to the sector classification. The 1968 report distinguishes two sets of accounts: - For Production Accounts and Goods and Services Accounts4, an economic activity classification should be used; - For the Income and Outlay and Capital Finance Accounts the ’institutional’ sector classification is recommended. In this case, the economic activity classification is really presented as an alternative sector classification (dual sectoring). In the input-output and make and use tables of the 1968 and 1993 reports, the classification by economic activity should also be employed. In both reports, separate statistical units are used for the classification by economic activity (establishments, establishment-type units) and for the (other) sector classifi­ cation (enterprises). Only the 1993 report provides a clear definition of the statistical units to be used for the institutional sector classification. Enterprises in the corporate sector are defined as "the smallest legal entity for which complete accounts are available" [12, Annex on changes, p. 4], In the 1968 report, also the family of enterprise could be used. The 1947 and 1953 report do not discuss this statistical unit. An exception pertains to the delineation the national economy. In the 1947 and 1953 reports, the usefulness of additional classifications by economic activity is already noted. In the 1947 report, it is recommended to subdivide the sector Productive Enterprises into branches of activity. Statistical problems in employing a unit like the 1968 ’establishment’ for this subdivision are also noted: "enterprises frequently contribute to more than one branch of activity. In dividing out the different activities of a complex concern and in allocating costs, and still more profit, between them, cost accounting methods are unavoidably involved. In practice, these allocations are difficult to handle statistically. For certain purposes, it may be sufficient to use a classification of branches of activity coincident with the main activity of different groups of enterprises" [9, pp. 55, 56]. Rather similarly, in the 1953 report, the notion of a separate statistical unit for describing the production process is already present, although it is not deemed practically feasible [10, paras 90, 91]. The 1953 report also suggests to divide its sector Enterprises by legal forms of organization. In the 1968 report, establishments are characterized as units "which are relatively homogeneous in respects of the character, cost-structure and technology of production.

F. Bos/On the UN Manuals

209

If data are to be classified according to region, the establishment should in addition draw distinction in respect of the location of the activities" [11, p. 73, para 5.15]. "Fur­ ther it must be feasible to compile data on all of the gross output, employment, intermediate and direct inputs into the production, and capital formation of each establishment" [11, p. 74, para 5.17].

Table 1. Sectors and sub-sectors in the 1947 report.

1. Productive enterprises 1.1 Business enterprises. This includes, next to privately controlled enterprises, public authority enterprises, unincorporated enterprises and some non-profit institutions. 1.2 Persons (house-ownership) 2. Financial intermediaries 2.1 Banking system 2.2 O ther financial intermediaries 3. Insurance and social security agencies 3.1 Insurance companies and societies 3.2 Private pension funds. This consists of voluntary pension schemes 3.3 Social security funds. This includes all "compulsory funds organised by the government'' [8, p. 70], 4. Final consumers 4.1 Persons. This includes private non-profit institutions 4.2 Public collective providers. This includes pension funds of the government. 5. The rest of the world (all economic entities)

Table 2. Sectors and sub-sectors in the 1953 report

1. Enterprises 1.1 Unincorporated private enterprises 1.2 Households and private non-profit institutions in their capacity as landlords of dwellings whether or not they occupy their own properties 1.3 Private corporations other than private non-profit institutions mainly serving households 1.4 Non-profit institutions mainly serving enterprises 1.5 Public enterprises 1.5.1 Government enterprises (financially integrated with general government) 1.5.2 Public corporations 2. Households and private non-profit institutions, which are not established primarily with the aim of earning a profit and are not mainly rendering services to enterprises (included are pension funds) 3. G eneral government (included are social security funds, if their activities may be regarded as an instrument of the social policy of the government)

210

F. Bos/On the UN Manuals

Table 3. Sectors and sub-sectors in the 1968 report A. Sectors o f the Production, Consumption B. Sectors o f the Income and Outlay and Capital Expenditure and Capital Formation Accounts Finance Accounts A. 1 Industries. This sector includes: B. 1 Non-financial enterprises, corporate and - establishments the activities of which are quasi-corporate financed by producing goods and services 1.1 Private enterprises. This subsector for sale in the market at a price that is includes private non-profit institutions normally designed to cover the costs of serving business production; this includes such establishments 1.2 Public enterprises of the government - all the dwellings which households or private non-profit institutions own and use on their own account - the own-account construction o f dwellings, B.2 Financial institutions non-residential buildings and other projects 2.1 The central bank by households and private non-profit bodies, 2.2 Other monetary institutions (banks) as well as by government organs 2.3 Insurance companies and pension funds - the own-account production by households of 2.4 Other financial institutions (for example primary commodities, where the producer building and loan associations, hireactivities of the households had not already purchase companies and investment funds) been set off as industries by virtue of sale of the products on the market - private non-profit institutions primarily serving business A.2 Producers o f government services. This sector B.3 General government includes: 3.1 Central government - government units, like local and central 3.2 State and local government government (excluded are the establishments 3.3 Social security funds of the government classified in the sector industries) - social security schemes and non-profit bodies which, by virtue of the relations with the government, are clearly instruments of the social and economic policy of the government A. 3 Producers of private non-profit services to B.4 Private non-profit institutions serving households. Excluded are: households - their activities classified in the sector industries - fraternal groups, social clubs and similar bodies, which have the equivalent of less than two full­ time employees; these are to be included in the sector households A.4 Households. This sector includes: B.5 Households, including private non-financial - paid domestic services which one household renders unincorporated enterprises. Included are, to another i.a.: - all the dwellings which households - fraternal groups, social clubs and similar bodies, own and use on their own account which have the equivalent of less than two full­ - the own-account production by time employees households of primary commodities - the activities of individuals as consumers

F. B os/O n the U N Manuals

211

Table 4. Sectors and sub-sectors in the 1993 report.

1. Non-financial corporations 1.1 Public non-financial corporations 1.2 National private non-financial corporations (incl. non-profit institutes serving business) 1.3 Foreign controlled non-financial corporations 2. Financial corporations’^ 2.1 Central Bank 2.2 O ther depository institutions 2.2.1 Deposit money institutions 2.2.2 O ther (e.g. savings banks, mortgage banks, building societies) 2.3 O ther financial interm ediaries except insurance corporations and pension funds (e.g. investment Corp., hire purchase corp., corp. engaged in financial leasing) 2.4 Financial auxiliaries (e.g. securities brokers, loan brokers, insurance brokers) 2.5 Insurance corporations and pension funds 3. General government 3.1 Central government 3.1 Central gov.incl. related social security 3.2 State government 3.2 State gov.incl. related social security 3.3 Local government 3.3 Local gov.incl. related social security 3.4 Social security institutions 4. Households (by major source of income; in this sector also the services of owner-occupied dwellings are recorded) 4.1 Employers (owners of household unincorporated enterprises with paid employees) 4.2 Own-account workers (owners of household unincorporated enterprises without paid employees) 4.3 Employees 4.4 Recipients of property and transfer income 5. Non-profit institutions serving households (e.g. trade unions, political parties, churches, sports clubs, charities organisations) 6. Rest of the world

For the subsectors 2.2-2.5 a distinction between Public, National Private and Foreign Controlled should also be made.

The 1993 report employs a rather similar definition: "An establishment is defined as an enterprise, or part of an enterprise, that is situated in a single location and in which only a single (non-ancillary) productive activity is carried out or in which the principal productive activity accounts for most of the value added" [12, Ch. V, para 21]. The 1947 and 1953 reports do not contain a subsectoring of the sector Households over groups of persons or households. Therefore, such a distribution of income by types of persons/households is ignored de facto. In the 1968 report some suggestions are made for a subclassification of the sector Households for these purposes [11, table 5.1, p. 79] and one table is presented [11, table 23]. It is explicitly stated that the subclassi­ fication should be regarded as not more than tentative [11, p. 83, 5.75]. So, in fact, the 1993 report is the first to incorporate income and expenditure by household types in its basic accounting system.

2.4 Assets, capital formation and capital consumption A basic distinction can be made between financial and non-financial assets. This is also reflected in the accounting structures of the 1947, 1968 and 1993 reports. For example,

212

F. Bos/On the UN Manuals

in the 1993 report, financial assets are to be recorded on the Financial Account and non-financial assets on the Capital Account. All reports draw a rather similar boundary between financial and non-financial assets. The major exception pertains to the treatment of gold. The 1947 and 1953 reports regard gold as a commodity and not as a financial claim [9, p. 67; 10, paras 183, 184], The 1968 report distinguishes three kinds of gold: - monetary gold; - commodity gold for industrial use; - commodity gold held as a financial asset. Only commodity gold for industrial use is regarded as a non-financial asset. In the 1993 report all gold that is not held by the authorities as part of their international reserves is treated like a non-financial asset. There are substantial differences in the classifications, scope and amount of discussion of financial assets in the reports. In the 1947 report, various financial items are distinguished in the Reserve Account, like borrowing, net purchases of gold and silver bullion and coin, subscription to new issues, net purchase of existing securities, mortgage repaid and discounts and advances to business enterprises by banks. Financial assets are not discussed in the 1953 report. The 1968 report offers a classification of financial assets in 14 categories which stresses the distinction between short-term and long-term loans and other financial assets. This distinction is absent in the 1947 report. The 1993 report elaborates much more than all the earlier reports on financial assets. This partly reflects the rapid development of financial markets and the emergence of new financial instruments in the last decades. Examples of such financial instruments not discussed in the earlier reports are financial leasing, repurchase agreements, deep discounted bonds, options and other derivatives. In the 1993 report, the long­ term/short-term distinction is only present as a secondary distinction, because "of the improved facilities available in financial markets to re-finance short term financial assets" [12, Annex on changes, p. 23], In the 1968 and 1993 reports, the concept of capital formation does not encompass the accumulation of all non-financial assets. In both reports, the purchase of land, subsoil assets and some selected intangible assets are recorded as accumulation o f non-financial assets, but not as capital formation. In the 1968 report, these intangible assets are i.a. "exclusive rights to exploit mineral deposits and fishing grounds, other concessions and leases in respect of land, patents, copyrights and trade marks" [11, para 7.86]. Purchased goodwill is not at all to be recorded. In the 1993 report, intangible assets included in capital formation are expenditure on mineral exploration, computer software, and entertainment, literary or artistic originals. These are all called "produced intangible assets". The "intangible non-produced assets" are excluded from the 1993 concept of capital formation, but recorded at the 1993 Capital Account. This pertains to patents, leases and purchased good-will. We will now continue our description of the concepts of capital formation by focussing on specific items. In all reports, the purchase and production of machinery, other equipment, dwellings and other buildings are regarded as fixed capital formation. The usual expenditure on

F. Bos/On the UN Manuals

213

repair and maintenance are excluded from their concept of fixed capital formation. However, the expenditure on goods and services which significantly enlarge the length of life or productivity of capital goods are to be registered as fixed capital formation. In the reports, rather similar exceptions are made to the rule that the own-account production or purchase of tangible durables is fixed capital formation. This applies to the purchase of consumer durables as well as to military expenditure on weaponry and other specific military durables (e.g. tanks and military barracks). Durable goods to be used in the production process of which the value is relatively small are excluded from the 1953 and 1968 reports’ concepts of capital formation (if the customary business accounting procedure is to treat them as a current expense); this specific case is not mentioned in the 1947 report. A difference occurs with respect to the treatment of land. In the 1947 and 1953 reports, purchases and sales o f land are regarded as fixed capital formation, while according to the 1968 and 1993 reports they should not be incorporated. In all reports, large incidental expenditure for the improvement of land are regarded as fixed capital formation. Another difference pertains to military expenditure on durables that are not typically military. The 1947, 1968 and 1993 reports agree that military expenditure on family dwellings should be recorded as capital formation; the 1953 report is silent in this respect. In contrast to the 1953 and 1968 reports, the 1947 and 1993 reports recommend that military expenditure on infrastructure, schools and hospitals is recorded as fixed capital formation [9, p. 77; 9, para 139; 11, paras 6.66 and 6.122; 12, Annex on changes, p. 18]. In the 1953 and 1968 reports, fixed capital formation is limited to the production or purchase of the above mentioned tangible durables. In contrast, in the 1947 report, some intangible durables are accepted as fixed capital formation: "It is clear also that certain receipts from professional services, rendered for example by lawyers, accountants and surveyors, would also appear here [on the capital account] in so far as they are incurred, not in connexion with current output, but for the future benefit of the concern. In the same way, it would seem that other expenses of a non-recurrent character such as certain types of advertising expenditure should appear as sale to capital account in the first instance, which are then written off out of operating revenue over a period of years" [9, p. 54], Other types of intangible capital formation, like human capital, seem to be excluded in the 1947 report though. In the 1993 report, a more restricted amount of intangible assets is included in the concept of fixed capital formation. For example, expenditure on mineral exploration and computer software are included, but capitalizing the expenditure of a massive incidental advertisement campaign is not allowed. The 1953 and 1968 reports reject natural growth as capital formation, e.g. of standing timber or crops [10, para 182; 11, para 6.103]. Only changes in farm stocks such as grain or livestock are incorporated; the 1947 report neglects this issue. In the 1993 report, only natural growth under control of human activities are to be recorded as capital formation; other natural growth is regarded as an other change in assets. In fact, the differences between the 1953 and 1968 reports on the one hand, and the 1993 report

214

F. Bos/On the UN Manuals

on the other hand, mainly amount to a difference in the time of registration. In all the reports, capital formation is subdivided into fixed capital formation and changes in stocks/inventories. The 1993 report incorporates valuables in capital formation as a third category. It defines valuables as "goods of considerable value that are not used primarily for purposes of production or consumption but are held as stores of value over time" [12, Ch. X, p. 3], Examples of valuables are precious metals and stones, jewellery and works of art. The reports of 1947, 1953 and 1968 do not record changes in the stock of valuables at all. Only in the 1953 and 1993 reports, the fundamental distinction between stocks as inputs (materials and supplies) and stocks as outputs (work in progress, unsold output) is made. Exhaustion of natural resources is in neither of the reports regarded as changes in stocks or as consumption of fixed capital. In general, the four reports recommend to record capital consumption on fixed capital formation. However, the 1953 report proposes that no capital consumption should be recorded for all general government assets except buildings [10, para 208]. In the 1968 report, only infrastructure should not be written down [11, para 7.20]. In the 1947 and 1993 report, in principle all assets included in fixed capital formation should be depreciated (if necessary). The concept of capital consumption is in all the reports defined in terms of the replacement value of capital stock. The 1947 report gives a more precise description on what is to be maintained intact: "capital needed to yield a given amount of product" [9, p. 60]. The 1953, 1968 and 1993 reports agree that capital consumption should be based on expected economic lifetime, that unforeseen obsolescence should be excluded from the concept and that straightline depreciation is in most case a practical solution [10, para 440; 11, paras 7.19, 7.21; 12, Ch. VI, p. 57]. The 1947 report discusses the concept of capital consumption in relation to maintaining either real capital or money capital intact. However, a concrete choice is not made.

3. Conclusions For decades, all over the world, national accounts figures are being compiled on the basis of the concepts in the guidelines of the United Nations. The first international guideline was published in 1947 and the drafting of the fourth guideline has just been completed. The 1947 report was a pioneering effort as it showed for the first time a detailed and fully worked out national accounting system. Its successor, the 1953 report, contained a much simpler national accounting system with a very limited number of accounts and sectors. The 1968 report employs a number of sectors and accounts similar to the 1947 report. In the 1993 report, the amount of detail offered is drastically increased. For example, six accounts are employed to describe the distribution and use of income, while in the 1968 report only one account is used.

F. Bos/On the UN Manuals

215

A fundamental change in the 1993 sector classification is its institutional character. In contrast to the other reports, production and final consumption by households (e.g. of imputed services of owner-occupied dwellings) are not recorded in different sectors, but in the same sector. In addition to the drastic changes in the accounting structure and the sector classifications, major changes also occurred in the scope of the reports. In the 1968 report, for the first time recommendations concerning constant prices are presented; input-output and make and use-tables were also added to the standard national accounting framework. The 1993 report introduced balance sheets, purchasing power parities, Social Accounting Matrices, satellite accounts and the concepts of employment and the labour force. A change in concept, in particular important for developing countries, was the inclusion in the 1953, 1968 and 1993 reports of primary production for own-account. Other important changes were: - Only in the 1947 report, government production is valued at costs including part of the interest paid on government debt; - Only in the 1947 report, operating surplus excludes interest paid, realized holding gains are included in national income and compensation of employees does not include social security premiums paid by employers. - Only in the 1953 report, gross rent is imputed for buildings owned and occupied by the government; - Only in the 1947 report, some intangible durables are included in capital formation; in the 1947 report this may also include the costs of a massive advertisement campaign; - In contrast to the 1968 report, the imputed service charge for banking services is to be attributed to the various sectors in the 1947 and 1953 reports. - All reports recommend a different set of central aggregates of income and product. It is interesting to note that in the preliminary drafts of the 1993 report, some of the concepts in the 1968 report were abandoned and replaced by concepts from the earlier reports. Cases in point are the attribution of indirectly charged banking services and the imputation of rent on government buildings. However, in the final draft, the concepts of the 1968 report are retained. The 1947 report argues that it will frequently be necessary to adapt and extend its concepts for specific national circumstances. The 1953 report does not pay much attention to specific needs and circumstances. An exception pertains to the choice of the production boundary: own-account production of primary products is included because of its importance in developing countries. In the 1968 report, a separate chapter is dedicated to the data needs and possibilities of developing countries. The 1993 report advocates a flexible system which takes account of all kinds of specific purposes and national circumstances. This is achieved by incorporating chapters on "Applications of the integrated framework to various circumstances and needs", "Social Accounting Matrices" and "Satellite analysis and accounts". In the 1947 report, the concept of income in kind is already defined in terms of national norms and habits. In this respect, it is more radical than all the other reports,

216

F. Bos/On the UN Manuals

because they treat income in kind as a concept invariant to differences in time and place. Balance sheets, the concepts of employment and labour force and the distribution of income and expenditure by type of household are not discussed in the 1947, 1953 and 1968 reports. The 1993 report breaks with this tradition and drastically extends the scope of the standard national accounts by incorporating all these subjects. Substantial constancy in concepts is evidenced by e.g. the production boundaries. All the production boundaries exclude unpaid household services and do-it-yourselfactivities and they all encompass imputed services of owner-occupied dwellings. Even specific national accounting conventions like the imputed charge for banking services, the valuation of government production at costs and recording all government production as final consumption are all advocated since 1947. Considering that the guidelines have just been revised, they will most probably constitute an established core­ set of conventions for more than 50 years.

Endnotes 0. In fact, the matrix is more complete than the accounting system described in the rem ainder of the report: in the matrix, balance sheets and a revaluations account are shown, but they are not at all discussed in the 1968 report. 1. In some supporting tables, a valuation at approximate basic values, that is producers’ values exclusive of net commodity taxes (selected indirect taxes reduced by similar types of subsidies) is recommended. 2. In the draft report, the new SNA still agrees with the previous reports [11]. However, in a special meeting of experts, it was decided to change the convention [6], 3. The draft 1993 report adopted the point of view of 1953 report. It also recom mended the imputation for valuing the output of non-profit institutions. [11, Ch. VI, p. 28]. However, in a special meeting of experts, it was decided to side with the 1947 and 1968 reports in this respect [6]. 4. In the report, they are called the Production, Consumption Expenditure and Capital Formation Accounts.

References [1]

F. Bos (1992) The history of national accounting. National Accounts Occasional Paper nr. 48, Netherlands Central Bureau of Statistics, Voorburg. An earlier draft of this paper was presented at the History of Economic Thought Conference, Durham, 2-4 Septem ber 1991.

[2]

F. Bos (1993) Standard national accounting concepts, economic theory and data compilation issues; on constancy and change in the U nited Nations-Manuals on national accounting. National Accounts Occasional Paper nr. 61, Netherlands Central Bureau of Statistics, Voorburg.

[3]

F. Bos (1993) Implementing the revised SNA in the Dutch national accounts. National Accounts Occasional Paper nr. 63, Netherlands Central Bureau of Statistics, Voorburg. This paper was presented at the joint O E C D /U N -E C E meeting of national accounts experts, 21-25 June 1993,

[4]

Chateau de la Muette, Paris. Eurostat (1970) European System Integrated Economic Accounts (ESA). Office for official publications of the European Communities, Luxemburg.

F. Bos/On the UN Manuals

217

[5]

A. Harrison (1990) M ajor changes proposed for the next SNA: an overview, Review of Income and

[6]

A. Harrison, The SNA; 1968 to 1993 and beyond. Paper published in this volume. An earlier draft

W ealth, Series 36 (1990), pp. 335-352. was presented at the 22th G eneral Conference of the IARIW, Flims, Switzerland, August 30Septem ber 5, 1992. [7]

Inter-Secretariat W orking G roup on National Accounts (ISWGNA) (1992) National Accounts and Balances: System of National Accounts (SNA). R eport on decisions by the ISWGNA on specific issues

[8]

concerning the revised SNA at its 14-16 D ecem ber 1992 meeting held at OECD, Paris. O E E C (1952) A standardised system of national accounts. OEEC, Paris.

[9]

U nited Nations (1947) M easurem ent of national income and the construction of social accounts. Studies and Reports on Statistical M ethods no. 7, U nited Nations, Geneva.

[10] U nited Nations (1953) A System of National Accounts and Supporting Tables. Studies in Methods, Series F No. 2, Rev. 1, U nited Nations, New York, 1953. [11] U nited Nations (1968) A System of National Accounts. Studies in Methods, Series F No. 2, Rev. 3, U nited Nations, New York. [12] U nited Nations (1993) Revised system of national accounts; draft chapters and annexes, United Nations, New York.

218

The Accounts o f Nations Z. Kenessey (Ed.) IOS Press, 1994

THE MATERIAL PRODUCT SYSTEM (MPS): A RETROSPECTIVE Jlnos Arvay Hungarian Central Statistical Office, Budapest Abstract. After a conceptual interpretation of the Material Product System the study follows the career of the MPS from its first introduction in the USSR in the 1920s, adaptation as a common system for the CM EA countries in 1969, its development and extension during the following two decades; until its term ination in 1990. The MPS was a system of national accounting used in 16 countries and its history is presented in relations to its counterpart, the SNA used for macroeconomic statistics in the market economies. The study points to connections between socio-economic regimes and national accounting systems.

Introduction When this study was planned, the Material Products System (MPS) was one of the two internationally recommended systems of national accounts. However by the time when it was completed, the MPS was in limbo. Given this, what was the reason to include this system into considering the history of national accounts and the development of national accounting concepts? There are several reasons for it. Of course, when this study started, nobody was able to foresee the dissolution of the Council of Mutual Economic Assistance, the disintegration of the USSR and the dramatic political, economic and social transformation which have taken place in Central and Eastern European Countries (CEEC-s) during the last 2-3 years. Without these changes the MPS would still be in use today. The second important reason is the fact that the MPS was the official statistical standard used for measurement of economic performance and development for more than seven decades in the USSR and for 2-3 decades in 15 other Centrally Planned Economies (CPE-s)1/- This means that historical data on the economy of the former CPE-s available in national and international yearbooks or other publications conform to the definitions and classifications of MPS. Another reason for keeping in mind the basic principles of MPS is the role this system will play for several years after introducing the UN System of National Accounts because the new (SNA) aggregates will often be derived from the differences existing between these two systems. Last but not least, it is impossible to ignore the intellectual and practical efforts made during the last three decades in order to compare and bring closer the two international accounting systems. Beyond the UN the IARIW also has a long history of trying to harmonize the SNA and the MPS.

J. Arvay/The Material Product System

219

Besides and beyond these arguments the author of the present paper had a special personal reason for undertaking the description of the history of the MPS: he is one of the few witnesses still alive who participated in the first formulation of the MPS between 1965-1968; actively contributed to the comparison and linking of the MPS and SNA; was present in all significant developments and extensions of the system during the last two decades; and has an inside view on how the replacement of the MPS by the SNA proceeds today in all Central and Eastern European Countries.

1. Interpretation of the MPS There are two approaches of interpreting the national accounting system based on the concept of material production. One reflects the economic theory of productive activity expounded by Adam Smith more than two hundred years ago. In his famous work "The Wealth of Nations" [1] he restricted the scope of productive work and value creation activities to those carried out in the sphere of branches producing material goods. As it is known, more than for a century after Adam Smith, i.e. in the overwhelming part of the 19-th and a considerable part of the 20-th century the concept of material production was the dominating theoretical basis for the definition and estimation of national income in the developed world. Many outstanding economists, - among them Karl Marx - accepted and incorporated this conceptual basis of national income without any substantial changes into their works. The fundamental concept of national income was derived by both Adam Smith and Karl Marx from the theory of "value-creating labour". In contrast, the economic theory of Keynes and Tinbergen in the 1930s, the fundamental basis for the SNA-type national accounting system, relies on a broader concept of output. From another aspect the MPS can be interpreted as a statistical instrument consisting of a complete, consistent and interrelated set of accounts, balances and tables together with the definitions of the corresponding concepts and methodological explanations which describe the functioning and development of the economy. It is quite obvious that in the MPS the fundamental statistical concepts are based on the Smithian and Marxian economic theory. However, it should be pointed out that the MPS as a statistical system bears a certain degree of autonomy in respect of the coverage, scope, classification of economic flows and stocks, and the statistical interrelations among them. The initial version of the MPS-type comprehensive statistical system was elaborated in the USSR in the 1920s. The first relatively detailed data on the national income and other macro-economic aggregates were published for the fiscal year of 1923/1924. Basically the same system was applied in the latter periods of the economic developments in the USSR and this system was introduced in all countries which fell into the political and economic sphere of the USSR after World War II. It should be pointed out that the main features of the MPS were determined not only by the underlying economic theory, i.e. the narrow scope of productive activity, but also by the specific economic policy and economic management system prevailing in these

220

J. Arvay/The Material Product System

countries which can be called in short: the central planning system. This means that the aspects of economic views and analyses within the MPS differ significantly even for branches which are covered by both systems e.g. in industry or agriculture.

2. Adoption of the MPS as a common system of balances for the CMEA member countries 2.1. The national system o f the USSR A detailed description of the balances of the national economy was published in 1926 under the leadership of P.I. Popov, the chief of the Central Statistical Office of the USSR [2], This publication consists of two parts: in the first one, in 350 pages, the methodology and analysis of the data are described; in the second part the statistical materials are presented in 275 pages. It should be noted however, that this work has almost entirely been forgotten for many years even in the USSR and in the other countries in her sphere. The official compilations of the national balances were continued and stabilized in the second half of the 1920s and in the 1930s; they played an important role in the economic, financial and social planning. From the middle of the 1930s, however, the scope of the macrostatistical data was drastically restricted. Only some global aggregates of output expressed in constant prices were published, and therefore they were only used to a limited extent. This situation prevailed until the second half of the 1950s. After World War II when the "small socialist" countries in Central and Eastern Europe introduced the Soviet-type system of central planning (1948-1949), they also followed the practice of the USSR in respect of organizing the compilation of national balances. However, in this early period no detailed methodological description of the system of balances was available in the USSR. Therefore the statisticians of the "small" countries acquired the basic theoretical knowledge of fundamental concepts and practical methods in bilateral consultations held mainly in Moscow with Soviet statisticians. In this situation the compilation of national balances started in the new socialist countries without an overall description of concepts and methods; they relied frequently on the principle of "learning by doing": A long period passed before a short description of the practice of the Statistical Office in the compilation of national balances of the USSR was published in 1960. The author of this book was V.A. Sobol, head of the Division of Balances of the National Economy. [3] He presented the main balances and tables constructed at that time and outlined the main direction of future development. In this book the work of Popov published in 1926 is mentioned in a short paragraph only and no further reference to this or other publications is made. Nevertheless, Sobol’s book later on played an important role in the development of the national accounting system in the individual CPE-s and also in the elaboration of the common CMEA system because it presented a set of interrelated balances in a systematic way and extended the scope of macro-economic analyses.

J. Arvay/The Material Product System

221

2.2. Presentation o f the MPS to UN bodies. Discussions in Geneva in 1958-1964. It is interesting to note that a first, overall description of the MPS intended to cover the entire system and to be generally accepted by all member countries of the CMEA, was initiated in 1957 by the Statistical Commission of the United Nations. The background of this request was the desire to improve the international comparability of the main macroeconomic aggregates, but there was also a more specific goal: the different scope of national income had a considerable impact on the magnitude of membership fees to be paid by the individual countries to the UN. (There was a wide-spread view that the membership fee of countries using the MPS should be increased because the scope of national income in these countries is narrower than that of countries using the SNA.) The UN Statistical Commission requested the Conference of European Statisticians to carry out a conceptual and numerical comparison between the two systems. At its sixth plenary session in 1958 the Conference requested four Eastern and four Western countries invited to participate in this work task. The following countries agreed to work in the joint Group of Rapporteurs: [4] Countries using MPS USSR Czechoslovakia Poland Hungary

Countries using SNA United States United Kingdom France Italy

In the same year (1958), preceeding the sixth session of the Conference of European Statisticians, two documents were submitted by Czechoslovakia to the Secretariat of the Conference. One of them (CES/83) described the methodology of compilation of national income, the other (CES/84) presented the scope and structure of the system of balances of the national economy. These documents can be considered as the first official presentations of the MPS at the international level in three official languages of the United Nations on behalf of all member countries of the CMEA. It should be added that this description of the MPS was very concise; the two documents together did not exceed 25 pages. Relying on the description of the SNA existing at that time (1952 version) [5] and on the above mentioned Czechoslovakian documents and other studies prepared in the meantime by other Eastern countries, the Group of Rapporteurs identified the major conceptual differences between the two systems. This joint work of the representatives from the 4 + 4 countries was an excellent preparatory school for both parties for a later period when the overall revision of the two systems was launched in 1964. Looking back at this period from a sufficiently long historical perspective, it can definitely be stated that the comparison of the two systems exerted a mutual influence which was perceivable later when the development of the two systems proceeded. Two areas are mentioned below for illustrating the fruitful influence of the dialogue between the representatives of the SNA and the MPS on the discovery of shortcomings of both

222

J. Arvay/The Material Product System

systems. The first example relates to the definition of the consumption of the population. Following the conceptual discussion of this aggregate in the two systems, two countries from the Group of Rapporteurs, namely the United Kingdom and Hungary undertook a joint endeavour in 1962. The task to be met was to show the numerical differences between the SNA and MPS in respect of the consumption of the population. For that purpose Hungary defined this aggregate according to the methods used in the U.K. and vice-versa. Comparing the two results, the representatives of the two countries came to the conclusion that neither of the two systems properly reflects the "true" scope of the consumption of the population. The MPS is distorted by the omission of the most essential component of services (work of doctors, teachers, etc.), while the SNA is distorted by the omission of fundamental services (health, education) financed from the state budget. As a result of these observations, the representatives of the two countries proposed a third concept which avoids both deficiencies. This concept was the 'Total Consumption of the Population". It is worth noting that this proposal was made three decades ago. [6] As regards the implementation of this concept, it was incorporated in the MPS in 1968, whereas it will be introduced only in 1993 in the new SNA (under the term "actual consumption"). The acceptance of this wider concept of consumption will greatly improve international comparability of this important indicator not only between market and non-market economies but also within market economies. Another fundamental conceptual obstacle revealed by the Group of Rapporteurs related to the treatment of indirect taxes. In the comparison of national income between the SNA and MPS it turned out that the addition of value added originating in the non-material sphere is not the only major adjustment needed to bring the value of national income in Eastern countries to the level of national income as determined in the SNA It is also necessary to carry out another major adjustment in the opposite direction. As it is well known, in the 1952 version of the SNA national income is defined on the basis of factor cost, i.e. excluding net indirect taxes. On the other hand, in the MPS, national income is defined at prices paid by the final users, i.e. including indirect taxes. Therefore, if the CMEA countries followed the recommendations of the SNA in their compilation of national income, the final result would not significantly differ from the original official value because the net value added of non-material services would increase it by 14-17 per cent, however the deduction of the turnover tax would decrease the value of national income by 20-25 per cent.2 This was a telling argument against the intention to increase the membership fee of the CMEA countries and the corresponding committee of the UN withdrew its original proposal. it became obvious in the course of discussions of the 4 + 4 countries that the treatment of indirect taxes in respect of national income distorts the comparison not only between Eastern and Western countries but also among market economies because there are considerable differences among these countries as to the shares of direct and indirect taxes in their national incomes. The recognition of this fact was reflected later in the 1968 SNA when the concept of "disposable national income" was accepted. This term is defined at market prices and since that time this value is used as the basis for

J. Arvay/The Material Product System

223

UN membership fees. There was an essential turning point in the work of the Conference of European Statisticians in 1964 when the activity of the Group of Rapporteurs came to an end and a new body was established. At its twelfth plenary session the Conference founded a Working Party on National Accounts and Balances. Its primary task was to represent the European aspect and interests during the general revision of the SNA launched by the UN Statistical Commission in this year. However, in addition to this major program the Conference also included the following two points into the terms of reference of the Working Party: - description of the MPS: study of, and comments on, the generalized description of the MPS, if available; - general consideration of the means of establishing links between a revised and developed SNA and MPS and of drawing up a European statistical programme of national accounts and balances. [7] It should be noted that these points were formulated knowing that the Statistical Standing Commission of the CMEA was committed to the elaboration and publication of a system of national balances which was to be applied as a common standard for the CMEA member countries. It also became obvious to the Conference that the overall revision and development of the SNA and MPS and the establishment of links between them can not be confined to the considerations of the representatives of the 4 + 4 countries; this work was of concern to all member countries of the Economic Commission for Europe and therefore all countries should participate in the Working Party on National Accounts and Balances. 2.3. Activity o f the CMEA Statistical Standing Commission in the setting up o f the MPS between 1965 - 1969 In addition to the needs and requirements indicated at the meetings of the 4 + 4 countries in Geneva, a great number of unsolved problems hindering the statistical cooperation among CMEA countries necessitated the elaboration and documentation of a system of national accounts acceptable to all. It was expected that the new system should surpass the various documents presented so far at the national and international level. As with the SNA the MPS was to play an important role in the harmonization of economic statistics in order to remove the inconsistencies existing between the different branches statistics. This work started in 1965. Each member country assumed the elaboration of a given chapter of the MPS, e.g. the USSR was responsible for the chapter on production, the German Democratic Republic took on the chapter on accumulation, Hungary worked out the chapter on final consumption, etc. According to the original plans, the various chapters of the system should have been coordinated by the Division of Balances of the National Economy in the Central Statistical Office of the USSR. After the first round of work it became evident that a coherent accounting system can not be completed efficiently is this decentralized approach. Therefore, the work program was drastically

224

J. Arvay/The Material Product System

changed: the Statistical Office of the USSR alone prepared a new, relatively short draft of the MPS in which only certain parts of the country papers were taken into consideration. Finally, this description was approved by the CMEA Statistical Standing Commission at its eleventh session held in November 1968. It was declared as the common system of balances of the national economy recommended for international use. It was published in 1969 in Russian. A few years later the CMEA Secretariat submitted the newly adopted system to the UN Statistical Commission. In 1971 the UN Commission approved it and decided that this document was to be published and widely disseminated by the United Nations as one of the available international recommendations. As a result, the MPS was published as a United Nations publication.

[8] At the time of approval of the joint CMEA system of balances the practices followed by and the national requirements raised in the individual countries were rather different from each other and in most countries exceeded the scope of information envisaged by the new system. These differences mainly resulted from the various degree of advancement in introducing major reforms in the economic management. Toward the end of the 1960s many new ideas and proposals on economic reforms were under discussion in almost all countries and several of these reforms were introduced into the practice in some countries. These reforms, almost without exception, focussed on the increase of independence and self-management of enterprises and on the development of market conditions and financial instruments. All these required more and different statistical indicators from those offered in the CMEA accounting system. The need for a more developed and a reform-oriented system was mostly stressed in Hungary and Poland. The dissatisfaction with the MPS was also due to the influence of the experiences gained by the Eastern statisticians in Geneva when participating in the meetings of the Working Party on National Accounts and Balances dealing with the development of the SNA. Nevertheless, at that time it was not possible to reconcile these various views and aims in the common CMEA system and therefore, after many compromises, a rather restricted and simplified description of the MPS was adopted and published. Considering the links between international recommendations and actual national practices, L&szl6 Drechsler noted acidly in the middle of the 1970s that the difference between the SNA and the MPS is analogous to the difference between the smallest common denominator and the lowest common multiple in mathematics. When saying so he meant that in the CMEA countries the national practices of every country were broader and more detailed than the information requested by the MPS; while none of the market economies was capable to fully implement the 1968 version of SNA [9] in their national practices.

J. Arvay/The Material Product System

225

3. Scope, structure and main characteristics of the MPS 3.1. Structure o f the MPS The basic characteristic feature of the MPS (that new value is created in the sphere of material production only) is not mentioned repeatedly in what follows. It should however have kept in mind that this basic concept of production is reflected throughout the entire system. The original version of the MPS (1969) consisted of five chapters. The first chapter explained the most important concepts and definitions of the system and the next four chapters described the four basic balances, together with the complementary tables and the methodological explanations supplementing them. These four chapters define the scope and content of the economic processes including the measurement of stocks of production factors. The four basic balances are as follows: - The balance of production, consumption and accumulation of the global product (material balance). - The balance of production, distribution, redistribution and final disposition of the national income (financial balance). - The balance of manpower. - The balance of the national wealth and fixed assets. Two major indicators of output are of central importance in the MPS. One of them is global social product which is the sum of all material goods produced in the sphere of material production during the accounting year, including products used for production of other products and those used for final uses. According to the SNA terminology this value is the gross output o f branches producing material goods. The other major category of output in the MPS is the national income which is derived from global social product by deducting the intermediate consumption of goods and consumption of fused assets used for the production of other goods. In the international terminology this concept is referred to as Net Material Product (NMP) to avoid confusion with national income as defined by the SNA. It should be noted that the MPS does not mention the factor incomes transferred from or to abroad, because at the time of formulating this accounting system such types of incomes were negligible in the CMEA countries. As regards the final use of national income, the MPS distinguishes two major categories: one of them is final consumption which covers material goods consumed by the population as such (personal consumption) and materials, including depreciation of fixed assets, used for rendering services. The other major category of final use shows the value of accumulation which covers changes in stocks and gross capital formation net of depreciation. The end use of NMP also includes the difference between exports and imports. Turning to the primary distribution of incomes, the national income is subdivided again into two parts: to the primary income of the population and the primary income of the enterprises. The former category of income includes wages and other direct

226

J. Arvay/The Material Product System

incomes, earned by persons engaged in the material production, e.g. wages of workers employed in enterprises producing material goods and also the revenues of selfemployed craftsmen. The primary income of enterprises is determined as a residual. The redistribution of incomes covers all transfers and financial transactions before arriving at the final value used for obtaining material goods for consumption or for net accumulation. This latter category is indicated by the MPS as final income. It is important to point out that expenditures for services e.g. the purchase of a theatre ticket, or wages paid in the sphere of services e.g. the salary of a teacher fall in the category of secondary incomes according to the MPS, i.e. are treated as redistribution of primary income. The balance o f manpower identifies the number of the economically active population and the number of the available manpower resources deriving these figures from the total number of population classified by age, sex etc. In a complementary table the distribution of actually employed persons is shown according to sectors and branches of the economy. In the balance of manpower great emphasis is given to the separation of employees according to the material and non-material sphere; this makes it possible to measure the productivity of labour in the branches of material production. The balance of national wealth shows the accumulated stock of fixed assets and circulating material goods resulting from production. It is also stated in this balance that the surplus of foreign financial assets over foreign liabilities should be added to the reproducible stock of assets. Natural resources e.g. land and minerals are only indicated "below the line" and in physical units of measurements. It is explicitly stated that only goods created by labour are considered as components of the national wealth. In the balance of fixed assets the breakdown of the stock of fixed capital according to branches and sectors at the beginning and end of the year are also shown, as are the main types of changes between the initial and final stock of assets. Some specific features of the MPS are described below. 3.2. Production-oriented system If the MPS is judged exclusively on the basis of the recommendations as described in the document published by the CMEA and United Nations, it gives the impression that this system provides a well balanced and complete set of macro-economic information since all important aspects of the economic processes are reflected in the basic balances and supplementary tables. By studying it more carefully, however, one can discover that the role of financial and income flows are greatly subordinated compared to statistics reflecting the processes of material production. Although most of the transactions in income distribution are formally indicated, the categories and classifications are not in line with the channels and instruments actually used for the monetary flows. Different types of incomes are combined into a small number of broad categories, so it is difficult to understand their economic meaning. For example, income taxes and social security contributions are shown under one heading; government grants for current production and capital formation are also combined.

J. Arvay/The Material Product System

227

On the other hand, the presentation of the production processes in the MPS is clear and straightforward, of course, within the production boundary of the system. The so called material balance and the complementary tables provide understandable guidance for the measurement of the real flows of production. The central balance itself includes the core of a simplified input-output table. The production-centred character of the systems is much more obvious when turning from the recommendations to the actual implementation of the MPS in the individual countries. This is shown first of all in the composition and coverage of published data. However, there are several evidences that statistics on production compiled for internal purposes also had higher priority in comparison with statistics on income distribution. This tendency became more explicit from the time when the compilation of input-output tables was introduced. It should be mentioned that the production-oriented character of the MPS was in particularly sharp contrast with the 1952 version of the SNA. At that time the SNA was biased in the opposite direction giving the dominant role to the income flows and devoting limited attention to the production flows. The increased attention paid to the processes of production in the MPS was due to the system of economic management i.e. central planning prevailing in CMEA countries in the period when the MPS was introduced. At that time these countries found themselves in extremely severe economic situations, leading the planners to prescribe the quantities of output, imports, internal and external use of commodities in physical terms of measurement not only for theoretical but also for very practical considerations. The balances of the global social product and later the input-output tables were rather suitable tools for aggregating and harmonizing the detailed elements of plans and facts. In the given circumstances such economic phenomena as prices, profitability, saving, etc. played an entirely subordinated role in comparison with quantities of products which were in the focus of the plans and statistics. This way of looking at economic processes explains why the policy makers and scholars of the economic sciences did not feel the lack of a clear and detailed presentation of income distribution from the MPS. In the light of the above mentioned characteristics of the MPS, a good opportunity presents itself for drawing an important conclusion in respect of the superiority of the SNA over the MPS. The greatest advantage of the SNA is manifested not so much in the full coverage of economic activities since the inclusion of non-material services into the value of national income or gross domestic product increases their magnitude to a relatively small extent only (15-20 percent) and does not change the growth rate of the economy considerably. The superiority of SNA rests mainly in the complete and qualitative information it explicitly provides for understanding and analyzing the economic processes. It offers a series of indicators which can be used directly and realistically for evaluating the performances of the producing units and the financial equilibrium of the different sectors. Such concepts as operating surplus, disposable income, saving, net lending which are the main qualitative indicators of the SNA are completely missing from the MPS.

228

J. Arvay/The Material Product System

3.3. Uniform classification o f transactors throughout the system The MPS recommends applying one uniform classification of economic activities and ownership sectors irrespective of the types of transactors investigated in the system. This means that outputs, inputs, primary and secondary income distribution, capital formation and stock of assets are shown according to one type of classification by branches and sectors. This way of disaggregation of economic activity allows for monitoring the full spectrum of economic processes for every branch separately. This seems to be an extraordinarily great advantage of the MPS compared to the dichotomous structure of the SNA where production processes are primarily classified according to type of economic activity based on establishment-type units, while income distribution and financial flows and stocks are classified primarily by sectors based on institutional units. 3 However, this advantage of the MPS is rather illusory because it is based on the presumption that all economic units carry out homogeneous activity. This structure was imposed on the MPS by "theoreticians" who never compiled and analyzed actual figures and therefore, never met the actual conflicts between the various types of data. They did not take into account that even in the period of the most centralized organization of the economy it was not possible to arrange that the individual institutional units (enterprises, co-operatives) be engaged in one single type of activity only. And if an enterprise is engaged in several activities and therefore its products are to be classified into different branches (as strictly demanded in the MPS) then the attribution of profits, taxes and other types of incomes to the purely defined branches can only be carried out arbitrarily. No mention is made about this problem in the MPS. When explaining the classification and definition of the income distribution flows, the text of the MPS says hardly more than the theoretical Marxian reproduction schemes. Of course, if the balances of incomes are only approximate or are missing altogether, these conflicts of classifications are not apparent. During the revision of the MPS in 1986-1989 this problem was raised but no satisfactory solution was achieved in the revised version. (This is again a specific feature of the MPS which is not linked to its major underlying concept; it could easily be transformed into a similar dual classification system as proposed in the SNA.) 3.4. Aggregated presentation o f data There are considerable differences between the SNA and MPS in respect of presentation of data. It is only a matter of terminology rather than a real difference in content that accounts are consistently mentioned in the SNA and balances are referred to consistently in the MPS even when the content and the form of the scheme (e.g. a 'T ' type of table) is quite similar in the two systems. It is, however, an important feature of the MPS that it recommends highly aggregated indicators only. These guidelines were strictly followed in the practice of countries when publishing macro-statistical data in their yearbooks or specialized publications.

J. Arvay/The Material Product System

229

Aggregation relates equally to the classification of economic units according to branches (in general, global social product and national income is broken down into 5-6 major classes of branches only), and to the classification of transactions according to types of flows (e.g. different kinds of incomes are combined). It should be added that 2-3 years ago in some countries only a limited number of absolute figures were published. Instead, long time series of volume indexes and growth rates and percentage distributions were the preferred form of official publications. It should be noted, however, that both in disaggregation and in publication of macro-economic data there were very great differences between countries using the MPS; these differences have been diminishing recently.

4. Development of the MPS between 1969-1986 When the MPS was approved and published at the end of the 1960s the CMEA member countries had accumulated significant international and national experience in their own macro-statistical activities. However, in the first official version of the MPS these experiences were incorporated to a minor extent only. There were some concepts and rules which were included in the MPS already in the moment of its birth, although these concepts and methods deviate conceptually from the traditional scope of material production. Two such areas are mentioned below. 4.1. Inclusion o f total consumption o f the population into the MPS The history and role of this concept was described above. At the time of finalizing the MPS there was a general agreement between the CMEA member countries that this indicator should be included in the new system. A separate section is devoted in it to the "Total consumption of material goods and non-material services by the population". It is also worth quoting a sentence from this section: "It is advisable to calculate the indicator of the total consumption .... because it more completely describes the satisfaction of the personal needs of the population and the changes that have occured in the structure of the consumption by the population."4 It should be stressed that originally this concept represented an extraneous body within the MPS which has no organic relations with the other parts of the basic system. It gives similar weight and economic importance to services as given to material goods in the basic system. Nevertheless, in spite of theoretical inconsistencies brought into the MPS by this concepts, no real conflicts came up from its use in the statistical practice. 4.2. Recognition o f passenger transport as productive activity According to the classical interpretation of the concept of material production passenger transport and telecommunication provided for the population do not qualify as productive activities. In spite of this restriction, these services were included in the

230

J. Arvay/The Material Product System

sphere of productive activities in the MPS approved in 1968. There is, however, a reservation mentioned in the document when listing the branches of production. The followings are stated in a foot-note: "Only goods transport and communication serving branches of material production belong to the material sphere. However, for practical reasons, in the interest of the comparability of data, transport and communications are entirely included in the material sphere."5 4.3. Supplementing o f the MPS by input-output tables The harmonization of input-output tables among the CMEA member countries started in the middle of the 1960s; parallel with the elaboration of the MPS. A first version of international recommendations was approved by the Statistical Standing Commission of the CMEA at its 9th session in November 1967 under the title: "Basic methodological principles and indicators of the statistical input-output table". The scope of this system was in line with the production sphere as defined by the basic system, i.e. only material branches were included in the "inner" quadrant of the input-output table. Work on the input-output tables required the clarification of several methodological problems which could also be utilized in the finalizing of the MPS. It should be noted that the inputoutput tables were widely used in all countries of the region both for planning purposes and for economic analyses. A revised version of the recommendations on input-output tables was latter incorporated in the extended MPS. 4.4. Supplementing the MPS by balances for non-material services The narrow definition of the production concept of the MPS was more and more in contrast with the requirements of the actual economic and social policies and the needs of national and international comparison of economic performance of the CMEA member countries. In broadening the frame of the MPS a most extensive action was launched in the middle of the 1970s initiated by the Statistical Division of the CMEA Secretariat. The aim was to elaborate a set of indicators which describes as fully as possible the activity of the branches classified to the non-material sphere of the economy. The system o f the statistical indicators o f non-material services and the method for their calculation was the title of the recommendations approved by the CMEA Statistical Standing Commission in 1978. The adoption of this system was an extremely important breakthrough from a political and ideological aspect. In this document the outputs and inputs of services and their monetary flows are presented in the same way as the macro-statistical indicators of the material sphere. Consequently, they can easily be merged with similar indicators relating to material production. With the approval of these recommendations it was formally possible for the CMEA countries to calculate a number of important aggregates corresponding to the definitions of SNA. There remained only one restriction: gross output, GDP, components of value added, etc. are not shown in one single set of consistent tables; they are presented in two different sets of interrelated

J. Arvay/The Material Product System

231

balances. It must be admitted that from practical aspects this complementary system was not consequent enough: some rules deviated from SNA, etc., however it played an important role in the extension of the basic system. 4.5. Approval o f a "full" concept o f income o f the population A further important step was made in 1979, when the Statistical Standing Commission of CMEA approved a comprehensive statistical system covering the indicators of the income of the population. The full set of indicators still includes a category of income which is tied to the concept of material production. It is called final income o f the population and it is defined as the total value of material goods consumed by the population. It entirely coincides with the final consumption of the population shown as a component of the domestically used national income. Parallel with this narrow concept, however, another set of indicators was also accepted by the Commission; it is presented under the heading: personal and total income o f the population. These indicators are in harmony with similar indicators of the SNA and also with the system of statistical indicators on total consumption of the population approved by the CMEA Statistical Standing Commission in 1971. The basic components and structure of this system were worked out at the beginning of the 1960s in Hungary in order to provide overall information on the fundamental factors of living standards. Since that time these data were published regularly in the Hungarian Statistical Yearbook and in special publications. The first publication came out in 1964. [10] 4.6. Accounting system o f Hungary The national accounting system of Hungary deserves some attention because it represents one of the several initiatives aimed at developing and modernizing the CMEA statistical system. Hungary has always played an active role in broadening the scope and deepening the observation of economic activity, standard of living, external relations within the CMEA Statistical Commission. Wherever it was not possible to reach agreement with the other member-countries, the Hungarian Central Statistical Office followed its own way in developing its own statistical system, while also adhering consistently to the standards approved jointly. The introduction of a new national accounting system in 1970 illustrates clearly the Hungarian approach to the recommendations of the CMEA and those of the United Nations. Hungary actively participated in the elaboration of the MPS and accepted its recommendations for national and international comparisons. At the same time, however, Hungary considerably extended the scope of its macro-statistics and introduced a unique accounting system which was different from all other CMEA countries but also from all market economies. Basically the Hungarian system preserved all important indicators required by the MPS, but also incorporated all major macro-statistical aggregates proposed by the SNA. It means that Hungary integrated the SNA and MPS into one single system with certain compromises, of course. Integration of the two

232

J. Arvay/The Material Product System

systems can best be explained by an enlarged input-output table which is broken down into four parts along the vertical and horizontal borderline between the spheres of material and non-material activities. From the full matrix and from the four submatrices the main aggregates of the two international systems can be derived rather easily. Of course, for a meaningful integration, technical and incidental differences had to be removed. There was one major area of economic flows where the two systems could not be properly integrated because the underlying principles were completely different. This was statistics on primary and secondary income distribution. The dilemma of choice was solved on the basis of the actual statistical requirements. Accordingly the Hungarian system followed the recommendations of the SNA because the income categories of the SNA could be directly connected with the concepts and regulations pertaining to the new economic management system introduced in 1968. The detailed description of the Hungarian accounting system introduced in 1970 is given in [11].

5. Links between the SNA and MPS As it was mentioned above the discovery of the differences between the SNA and MPS started as early as the beginning of the 1960s by comparing the SNA on the basis of its 1952 version and the MPS which was not adequately documented at that time. It was the definite intention of the Conference of European Statisticians expressed at its plenary session in 1964 (as quoted before) that the revision of the two systems should be carried out in close cooperation and all possibilities should be used for bringing the two systems closer to each other. In addition it was also pointed out that after the completion of the revisions the links between the two systems should be identified in order to build bridges for transforming the main aggregates from one system into the other and vice-versa. This work was carried out in the first half of the 1970s and its results were published in two UN publications which are referred to in brief as document "F.20". [12] The study on conceptual relations offers a detailed description of the most important differences between SNA and MPS, separating differences resulting from the two distinct theoretical concepts of the two systems and those which occured incidentally. It presents three transformation tables showing the steps necessary to derive GDP from NMP and NMP from GDP. The first transformation table is used for deriving GDP (or NMP) from the aspect of production, the second table shows the transformation for the categories of primary income distribution and the third shows the transformation from the aspect of final use categories. In Part Two of this document ten countries: the United States, the United Kingdom, Japan, Austria, Finland, Peru, the Philippines, Zambia, (as SNA-using countries), Hungary and Yugoslavia (as MPS-using countries) show the steps and results of the transformation of a their own actual data into the other system. Even today it is worthy to evaluate from a historical perspective the work done in order to create links between the SNA and MPS during the 1960s and 1970s. For the

J. Arvay/The Material Product System

233

last three decades, up to 1989 the parallel existence and full autonomy of the two systems seemed to be destined to last for ever. Therefore, in that period of time the building of bridges between the two systems had invaluably great importance! The efforts made in the interest of transforming the data from one system into the other were received extremely positively not only from the point of view of statistics but also from the aspect of political and economic cooperation between the East and the West. Now the situation is completely different. There is a threat of belittling the role of the MPS and the work done to link the SNA and MPS in the previous decades and the exaggeration of the SNA in the present time and in the near future. Many economists and statisticians in Eastern and Central Europe consider the SNA as a magic tool and expect from its introduction significantly more than it will realistically provide in the first years. This is particularly valid in the transition period when the institutional framework of a market economy has not yet been established; prices do not properly reflect the relations between demand and supply and the structure of the economy is still greatly distorted, etc. No system of national accounts is able to provide a realistic picture of an economy which has distorted economic conditions and has no reliable instruments for measurement.

6. Revision of the MPS in 1968-1989 The Statistical Standing Commission of the CMEA launched a program in 1986 for updating the MPS. Three major objectives were focussed at. The first objective was to incorporate in one single document the revised MPS and the complementary systems approved in the twenty years following the adoption of the basic system. Secondly, it was the intention to make clearer and more extensive the existing methodological recommendations. The third objective was to remove several incidental differences between the SNA and MPS. The revision carried out during 1986-1989 brought about some remarkable results. All above mentioned complementary systems were added to the core system. Many important clarifications and additional methodological explanations were introduced, mainly in the chapters dealing with the description of the basic concepts and with the so called material balance which presents the concepts and definitions of the sources and uses of material goods. However, no substantional revision was made in the concepts and definitions of the income distribution processes; this was postponed for the years after 1989. Practically none of the technical and incidental differences between the SNA and MPS was removed. The improved and extended document was submitted to the UN Statistical Commission in 1987. The Commission agreed that this document should also be issued as a publication of the United Nations and be given wide circulation. Accordingly, the United Nations published the revised MPS in two volumes. [13] Volume I describes the basic methodological principles of the core and complementary system and Volume II contains the classification of the branches of the national economy and the system of

234

J. Arvay/The Material Product System

indicators of social statistics. This was the situation at the beginning of 1989 when the UN Statistical Commission discussed again the recent developments in the MPS and the work on linking of the SNA and MPS. During discussion of these items unexpectedly a sensitive question was raised which had been politely and carefully avoided for thirty years. The representative of Hungary proposed that instead of explaining the differences and creating shorter or longer bridges between the two systems the Statistical Commission should take definite measures for the integration of the two systems. This idea received general appreciation from all participants of the Commission. The report of the Commission noted that "The work on SNA/MPS links .... had entered a new phase. The main objective was now the achievement of substantial progress in convergence of SNA and MPS, with the ultimate goal of integrating the two systems. [14] At the same session the Commission endorsed the convening of an expert group meeting on SNA/MPS integration. The meeting was held in December of 1989 in Moscow with a broad participation of experts from both the East and the West. The meeting identified the most necessary and important task to achieve integration. This work was started and would have had extremely great significance if the political situation in the CMEA countries had remained unchanged. However, as a result of the radical changes in political events and the fundamental institutional restructuring following them in 1989 and 1990, the original plans for the integration of the SNA and the MPS became outdated.

7. Termination of the MPS in 1990 Some CMEA countries which were members of the World Bank and the International Monetary Fund already in the 1980s (Yugoslavia, Romania, Poland, Hungary), supplied data on their economic performance to these organizations according to the requirements of the SNA. In their own official publication, however, these countries (with the exception of Hungary) provided macro-statistical data exclusively according to the MPS. This was the practice until 1988. In 1988 and 1989 a breakthrough occurred practically in all Central and Eastern European countries. From one moment to the next most of the statistical offices of these countries started to publish data on the value of the GDP and on its major components as defined in the SNA. The Soviet Union began to publish data on GNP in 1988, and the other countries published such data in the following years. The roots of these unexpected changes are obviously found in those deep-reaching political, social and institutional transformations which took place in this region in the second half of the 1980s and culminated in 1989. The political transformation also had direct implications on the statistical system in order to serve the future needs of market type economies. Shortly after this breakthrough the Western countries offered their assistance for "countries in transition" to facilitate the transformation in all areas including the transformation of statistics. First the Conference of European Statisticians

J. Arvay/The Material Product System

235

organized a meeting in Geneva in Summer 1990, where the most urgent statistical tasks of the transition were identified. Then, a much wider conference was convened in Paris in September 1990 under the auspieces of the OECD. All the statistical offices of Central and Eastern European countries were represented at the highest level in both conferences. In respect of developing their national accounting systems the chief statisticians of these countries declared their intention in Paris to introduce the United Nations System of National Accounts or its European version implemented in the European Community [15] within a short period of time (in 2-3 years). These unanimous statements made in the Paris Conference empower the author of the present paper to declare 1990 as the year when the MPS ceased to function as an international guideline. This does not mean that all statistical data compiled according to the MPS disappeared in this year or will disappear in the near future. Such data will inevitably be used for several years in a number of countries, alone or parallel with data defined according to the SNA. After 1990, however, the MPS can no longer be considered as an active and mobilizing framework. It remains as a remnant of the past; it will be a point of departure for deriving SNA-type aggregates; it will serve as a documentation of macro-economic data compiled several decades before. As mentioned above, the resolution of the UN Statistical Commission on the integration of MPS and SNA has lost its force. For the time being there is no country in Europe which is interested in working on a version of the MPS which is suitable for integration with the SNA. On the contrary, they intend to replace the MPS by the SNA. Today all the statistical offices in Central and Eastern Europe are vigorously working on the introduction of the new SNA as quickly as possible. Finally, I wish to mention that in the discussion of the first version of this study two substantial comments were made. Frits Bos casted doubt on the assertion of the first chapter of the present work and also of the same idea in of the Studenski’s book [16] that the fundamental concept of production in MPS was based on the Marxian theory. Bos argued that in none of Marx’s writings are discussed any accounting procedures for measuring production. And Andre Vanoli expressed the view that the study in chapter 6, when dealing with the integration of the SNA and MPS, gives the impression that the idea of integrating the two systems was widely accepted and supported at the Moscow meeting in 1989. He pointed out that at this meeting the theoretical and practical approaches of integration were differently interpreted by the various participants. I acknowledge the statement of A. Vanoli, however in respect of F. Bos’ comment I wish to point out that several generations of Eastern and Western economists accepted the historical background of the MPS according to which the underlying principles of this system are based on the Marxian theory.

J. Arvay/The Material Product System

236

References [1] Adam Smith (1937) The W ealth of Nations, 1776. Published in M odern Library edition, N.Y. City. [2] Popov, P.I. (ed) (1926) Balance of the USSR National Economy in 1923-1924. (Balans narodnogo khoziaistva Soyuza SSSR 1923-1924 goda. SSSR), Trudi Tsentralnogo Statisticheskogo Upravlenia, Tom XXIX. Moskva.) [3] V.A. Sobol (1960) Review of the questions related to the balances of the national economy. (Ocherki povoprosam balansa narodnogo khoziaistva, Gosstadizdat) Moscow, p. 227. [4] R eport of the sixth plenary session. Conference of European Statistician/94. 25 June 1958. [5] U nited Nations (1952) System of National Accounts and Supporting Tables. Studies in M ethods. Series F. No.2. New York. [6] A. M6d and R. Beales (1963) The consumption of the population in the United Kingdom and Hungary. Statistical Standards and Studies. N o.l. United Nations, Geneva. [7] Report of the twelfth plenary session. Conference of European Statistician/221. 19-23 O ctober 1964. Para 69. [8] United Nations (1971) Basic Principles of the System of Balances of the National Economy. Studies in Methods. Series F. No.17. New York. [9] U nited Nations (1968) A System of National Accounts. Studies in Methods. Series F. No. 2. Rev. 3. New York. [10] Hungarian Cental Statistical Office (1964) National income and the living condition of the population. Periodical statistical publications, Volume 78. Budapest. [11] J4nos Arvay (1973) National production, national

income, national wealth. Publishing House of

Economy and Law. Budapest, 368. p. [12] United Nations, Comparison of the System of National Accounts and the System of Balances of the National Economy. Part one: Conceptual relations. Studies in methods. Series F. No.20. New York, 1977. Part two: The transform ation of SNA aggregates into MPS aggregates and vice-versa in selected countries. Studies in methods. Series F. No.20. New York, 1981. [13] United Nations (1989) Basic Methodological Principles Governing the Compilation of the System of Statistical Balances of the National Economy. Volume I and II. Studies in Methods. Series F. N o.l7/R ev.l. New York. [14] United Nations (1989) Official Records of the Economic and Social Council. Supplement No. 3. (E /1989/21 - E/C N .3/1989/25.) Para 61. [15] Statistical Office of the E uropean Communities (1979) European System of Integrated Economic Accounts (ESA). Second Edition. Luxembourg. [16] Paul Studenski (1958) The Income of Nations. New York University Press. Endnotes 1.

Countries using the MPS for some period included: Albania, Bulgaria, China, Cuba, Chechoslovakia, GD R, Hungary, Cambodia, Korean Dem ocratic Republic, Laos, Mongolia, Poland, Romania, Vietnam, Yugoslavia.

2.

In that time in all Centrally planned Economies turnover fax was the main channel of funnelling

3.

This dichotomy should be pointed out even if already the 1968 version of the SNA, and to greater

incomes into the State budget; the role of direct taxes on profit was negligible. extent the 1993 version of the SNA m ade considerable efforts to remove the barriers resulting from this dual classification. 4. 5.

See [8], paragraph 1, 119. See [8], para. 1.13. It needs to be added to this note that in their national system the Soviet Union and Czechoslovakia always excluded passenger transport and telecommunication provided for the population from m aterial production.

The A ccounts o f Nations Z. Kenessey (Ed.) IOS Press, 1994

237

A HISTORICAL NOTE ON QUADRUPLE-ENTRY BOOKKEEPING Harry H. Postner * Statistics Canada, Ottawa

Abstract. This note is concerned with the historical development of the concept of quadruple-entry bookkeeping. The concept serves to clarify the foundations of the integrated approach to national and sectoral economic accounting. A t the same time the concept is distinguished from and related to the double-entry bookkeeping of business accounting practices.

Introduction Two events have inspired the preparation of this note. First, we are now approaching the 500th anniversary of the publication of the classic treatise synthesizing the practice of "double-entry bookkeeping" [Pacioli, 1494], Some famous economic historians, such as Werner Sombart and Oswald Spengler, have characterized this event as one of monumental importance for understanding the development of the capitalistic business enterprise [see Ijiri, 1975]. The second event is the recent and long overdue recognition in a United Nations related document that a system of national economic accounts is ultimately based on, what is called, "quadruple-entry bookkeeping" [OECD, 1990]. Though the explicit notion of quadruple-entry bookkeeping was mentioned in the economic accounting literature over 40 years ago, the notion perse was not fully defined and its complete relationship to double-entry bookkeeping was typically not spelled out. Indeed, many discussions of what could be regarded as "quadruple-entry bookkeeping" during the past 40 years tended to avoid explicit use of that conceptual term. Moreover the recent recognition of quadruple-entry bookkeeping in the United Nations document is not concerned with the historical tracing of the subject matter nor with the historical development of its technical subtleties. These tasks are, indeed, the prime concern of the present note. The next section provides the general background discussion focusing on some key contributions to the literature of the 1940s to the early 1960s. Here it will be seen that the concept of "quadruple-entry bookkeeping" is often an implicit matter in the national accounting literature. For example, Sir Richard Stone, who many regard as the father of national economic accounting, never utilizes the term "quadruple-entry bookkeeping" and this practice has probably been the source of some misunderstanding in the post1945 development of national accounting. It also will be seen that our historical review * The views expressed are those of the author.

238

H.H. Postner/On Quadruple-Entry Bookkeeping

(both the background discussion in Section 2 and the technical considerations in Section 3) concentrates on the contributions to be found in the original Englishlanguage literature. For ease of exposition the following short-hand notation will be used throughout the note. Double-entry bookkeeping will be denoted as DEB and quadruple-entry bookkeeping will be denoted QEB. Some brief mention will also be made of single­ entry bookkeeping, SEB.

1. Historical Developments: Background Discussion It has long been acknowledged in the business accounting literature (e.g., De Morgan [1897]) that when a business firm, practising double-entry bookkeeping (DEB), engages in an exchange transaction with another firm, also practising DEB, then the two sets of double entries are the reciprocal of each other: the debit entry of the first firm with respect to a particular transaction category is matched by an equal credit entry of the second firm with respect to the same (or counterpart) transaction category. The firms’ balance sheet identity: assets = liabilities + net worth together with the notion of mutual exchange, then implies that the corresponding credit entry of the first firm with respect to some other transaction category must be matched by an equal debit entry of the second firm, again with respect to the same (or counterpart) transaction category. The four bookkeeping entries together could possibly be regarded as quadruple-entry bookkeeping (QEB), but this term is never used in the business accounting literature. Business accounting standards boards sometimes consider the two sets of reciprocal double entries and so does consolidation accounting of a divisionalized business enterprise. But by far, most often, the practical emphasis is on accounting for each business firm in isolation via DEB, without being concerned with mutual matching reciprocal double entries with other firms [Littleton, 1953]. It is noted, though, that DEB perse does not necessarily imply only two classification entries: one debit and one corresponding credit entry. A complex business transaction could easily involve more than two entries for a single firm, so long as the total of the debit entries equals the total of the corresponding credit entries. The essence of DEB is then not that there are only two accounting entries, one debit and one corresponding credit entry. The essence is rather that DEB is concerned with an offsetting dual-classification recording of the transactions of each accounting entity taken in isolation. The offsetting arrangement maintains the balance sheet identity. We now turn to our historical review of the national economic accounting literature. In the year 1949 two references to the concept of QEB appeared in the literature. The first edition of the well-known Richard Ruggles textbook [1949] pointed to the

H.H. Postner/On Quadruple-Entry Bookkeeping

239

presence of reciprocal double entries as an inherent characteristic of a sectoral breakdown of national income and product accounting. The reciprocal double entries arise as a result of the intersectoral articulation dimension of national accounting (following Stone, 1947) and the term QEB appears to be used here for the first time. (It might be noted, though, that the specific term QEB seems to be absent from the later edition of the Ruggles and Ruggles text.) In the same year Morris Copeland [1949] published an advance overview of his national flow-of-funds accounting system. This system was ideally characterized as one of QEB: (1) each sector distinguished in the system practices DEB, so that the summation of all debit entries (called uses of funds) across nonfinancial surplus and all financial flow categories equals the summation of all credit entries (sources of funds) across the same categories; and (2) each transaction flow category is in reciprocal national balance, so that the summation of all debit entries (uses) across the sectors equals the summation of all credit entries (sources) across the sectors, for each and every flow-of-fund category that is distinguished. One might refer to the first characteristic of this proposed QEB as one of sectoral internal balance; the second characteristic of the proposed QEB is then one of transaction flow external balance. All this is merely a generalization to the national economy-wide accounting level (and for an annual accounting time period) of the idea of "reciprocal double entries" that has already appeared in the business accounting literature. The essence of the proposed QEB follows from its two principal characteristics (but see further development to follow). We might also note that the general terminology of QEB has continued to dominate analysis of the flow-of-funds accounting system ever since its full conception in Copeland [1952], An important attempt at clarification with respect to DEB and QEB, in the context of national economic accounting systems, appeared in a key contribution by Stanley Sigel [1955]. There it was pointed out that both the national income and product accounting system and the input-output accounting system should not be regarded as satisfying the requirements of QEB. These systems do possess an element of double­ entry type recording in the sense that there are intersectoral and interindustrial articulations with respect to mutual transactions. But because both systems are restrained to exclude from their scope the monetary and financial aspects of their transactions, the particular sectors and industries cannot fully reflect the automatic balancing conditions of either DEB or QEB. In fact the sector accounts in national income and product systems and the industrial accounts in input-output systems are balanced only by the introduction of residual entries (e.g., savings and operating surplus) -- a possible characteristic of single-entry bookkeeping (SEB). Residual entries, however, are sometimes a preliminary step towards an acceptable DEB as explained in Ijiri [1989]. So Sigel [1955] in order to avoid conceptual confusion, would have preferred to characterize the actual structure of the national income and input-output accounting systems as one of SEB, but also noting the presence of double-entry type recording with respect to the nonfinancial aspects of mutual transactions. (Sigel’s position with respect to national income accounts is somewhat reminiscent of Kuznets [1948].) On the other

240

H.H. Postner/On Quadruple-Entry Bookkeeping

hand, Sigel agrees that the flow-of-fund accounting system can be regarded as one of QEB, though he prefers to utilize the language of DEB in this case. There does, however, seem to be an important inconsistency in Sigel’s arguments. This is because the nonfinancial surplus row of the flow-of-funds table is basically a residual entry (or at least a residual calculation). It is not surprising, then, that Stanley Sigel’s next major contribution [1962] dealt with the integration of national income and product and flow-of-funds accounting systems. The key step towards integration is the construction of a national (intersectoral) market transaction framework [Sigel, Table 1, p. 28] for an accounting time period. The national economy is described by a complete set of economic sectors, including the restof-world sector. Each sector is characterized by a column of debit transaction entries and a column of credit transaction entries. The total of all debit entries equals the total of all credit entries for each and every sector —a characteristic of DEB (there are no residual entries). The market transaction flow categories covered by the debit/credit entries are exhaustive, and include the full range of current/capital and nonfinancial/ financial transaction flows. At the same time, for each and every row signifying a transaction flow category, the summation of all debit entries across sectors equals the summation of all credit entries across sectors. (There are "zero" debit or credit entries when the sector/transaction category combination is not pertinent.) Sigel’s framework is reminiscent of the Canadian national transaction accounts as pointed out to the present writer by Thomas Rymes. So the two major requirements of QEB are satisfied under the market integration conditions. All would appear to be well. However, Ragnar Frisch [1953] may not approve because of his scepticism regarding the integration of real and financial flows in the same framework: the two sets of flows should be considered as moving on different "planes". Nevertheless, because of the restriction of the grand integration framework to market transaction flows, both real and financial, we have in effect "Hamlet without the Prince!" Sigel [1962] was well aware that a full integration of national income and product accounts with flow-of-fund accounts must additionally incorporate a wide range of nonmarket transaction entries. These would include: (1) income-originating entries such as business own-account construction operations, (2) cost-of-production entries such as capital consumption charges and inventory change allowances, (3) purchases of final product entries such as an imputed government consumption flow, (4) an imputed sales and purchase offset to in-kind income payments and the "free" provision of certain financial services, and (5) a range of strictly internal (interlocking) transaction entries in order to arrive at balancing subtotals of major analytic significance. Sigel [1962, Tables 5 and 6] shows that all these nonmarket transaction entries can be effected without disturbing what we could now state to be the two major requirements of QEB, namely that: (1) each sector (or transactor) practices DEB without residual entries, and (2) there is external debit/credit balance with respect to each and every transaction flow category including those flows derived from nonmarket considerations. Is this, then, the end of the story? The answer is: not quite! One development might

H.H. Postner/On Quadruple-Entry Bookkeeping

241

be stated briefly. (Another development will be raised in the following paragraphs.) The addition of nonmarket transaction entries to the "extended" integration accounting system obviously lacks the economic discipline provided by purely market transaction entries. In fact some of the nonmarket transaction entries are, in effect, very simple and subject to no economic discipline whatsoever even though the entries do not disturb the two accounting (debit/credit) requirements of QEB. This aspect has been raised, e.g., by Stone and Hansen [1953, p. 106] in a related context. So in order to preserve a QEB system with economic meaning, the nonmarket transaction entries must be made subject inter alia to a cross-sectoral (or cross-transactor) comparability principle along lines indicated in Stone and Hansen [1953]. The second development is that our description of QEB might be regarded as idealistic. We might accept the proposition that each sector, or transactor entity, practices DEB so long as the classification scope of transaction categories is "exhaustive". However, the second requirement of QEB implies that mutually transacting entities maintain their market accounting records in a perfectly consistent fashion. This assumption has been questioned by Thomas Schelling [1955 and 1958] who is also one of the very few writers of this period who actually used the term QEB. Is it really necessary for the construction of a comprehensive and integrated national accounting system to "force" such a consistency requirement? There are many examples where the transacting parties have different views concerning their mutual market transactions (e.g., Schelling [1955] and Ohlsson [1953]). The perspective of household sectoral accounts need not coincide in all respects with the perspective of business enterprise or government sectoral accounts - considering the asymmetric economic behavioral motivations and information-possessing environments of the different sectors. So a more flexible approach to QEB, along the lines suggested by Thomas Schelling, permits the existence of debit/credit transaction flow discrepancies across sectors, provided that the discrepancies are not all-pervasive and provided that the discrepancies are appropriately "tracked." Indeed when each sector practices DEB, then the arithmetic grand total of all transaction flow discrepancies, each measured by total debits minus total credits, sums to zero. Such a balanced set of discrepancies need not be limited to financial flows, but should cover nonflnancial flows as well. Sigel [1967] as expressed in Helen Tice [1967] would also agree. A related development might be added. Transaction flow discrepancies, for a particular accounting time period, have a tendency to be reversed in future accounting time periods. Consider the example raised by Raymond Goldsmith [1950], though in another context. A creditor might make a direct writeoff of an account receivable by debiting "bad debt expense" and correspondingly crediting "accounts receivable." The debtor, at that time, may record no change. So a "balanced" pair of transaction flow discrepancies appear in QEB accounts of the initial period. But in the next accounting period, when the debtor recognizes bankruptcy or similar proceedings, the debtor (or the executor) records a counterpart credit for "bad debt exemption" and a corresponding debit for "accounts payable." The creditor, at this time, remembers to record no further changes. Then the pair of transaction discrepancies is reversed in the second period.

242

H.H. Postner/On Quadruple-Entry Bookkeeping

This should be regarded as a supremely simplified example of what could turn out to be a highly complex sequence of adjusting entries affecting many transactors and affecting many other transactions categories (including nonmarket reserves and allowances) over several accounting time periods. Indeed, many such examples are characterized by the difficulty of "matching" the two or more parties involved in mutual transactions. There is often an anonymity element as well as an asymmetry element present that cannot be avoided. All this, however, has economic behavioral significance that is lacking when the national accounting statistician forces consistency via formal or informal adjustment and reconciliation procedures. A QEB with economic meaning, then, possesses appropriate flexibility and provisionality. This should not be surprising because traditional DEB with meaningful accounts is also characterized by flexibility and provisionality. Witness, for example, the discussion of "reserve accounting" [Yamey, 1987] and "suspense accounts" [Ijiri, 1989] in the context of DEB.

2. Historical Developments: Technical Considerations To recapitulate, the previous section began with an attempt to isolate the essence of DEB. It was also quickly discovered that the concept of reciprocal pairs of DEB, found in the business accounting literature, is merely a trivial extension of the traditional DEB. On the other hand we found that QEB, with its economy-wide accounting coverage, is a non-trivial extension of both DEB and the idea of reciprocal pairs of DEB. QEB leads to economic considerations beyond those, but not entirely unrelated to those, found in the DEB accounting literature. The reader should recall that QEB was analyzed in terms of a multisectoral/ multitransaction flow framework or system. The word "matrix" was deliberately avoided to describe that framework, because matrix economic accounting has a special meaning according to a tradition dating back to Richard Stone [1948]. Since the work of Stone has had so much influence on the historical development of national economic accounting, we could once more be accused of presenting "Hamlet without the Prince" unless some attempt is made to relate Richard Stone’s matrix apparatus to our historical review of QEB. Indeed it turns out that the nature of QEB and its historical development is clarified once the links to the Stone of [1948] and [1954] are exposed. First it should be noted that Richard Stone does not employ the debit/credit language originating in DEB business accounting. Instead, "debits" are called "accrued outgoings" or "payables"; "credits" are called "accrued incomings" or "receivables." However, for ease of exposition we will retain the debit/credit language so far used in this note. A second point is that Stone is concerned with an allocation of sectoral accounting records into the familiar subaccounts: production account, appropriation account, and, say, capital accumulation and finance account. For our purpose, such an allocation is not an essential aspect of QEB. So we will abstract from sectoral subaccounts and focus on each sectoral accounting record as a whole. A third aspect is

H.H. Postner/On Quadruple-Entry Bookkeeping

243

that the early work of Stone often refers to "transaction flow categories" as "different types of consideration." Once more we will retain our original terminology. It should be noted, though, that our conception of transaction flows, both nonfinancial and financial, is comprehensive and covers both market and nonmarket (imputed and internal) transactions. This agrees with the Stone framework except that Stone also embodies intrasectoral interlocking subaccounting entries that we presently exclude because of our abstraction from the allocation process. So the following analysis is essentially a faithful adaptation of the Richard Stone matrix accounting approach. The key problem of this technical section of our note, then, is to identify the two principal characteristics of QEB in the Stone matrix approach to national economic accounting. Suppose there are distinguished m number of economic sectors covering the national economy, and n number of types of transaction flows. Then it is possible to describe a Stone two-dimensional matrix, say 5*, of the order (mn x m)\ that is there are mn rows and m columns. This is essentially the set-up in the Stone of [1954]. However, it is considerably easier to visualize the analysis in terms of the Stone of [1948] where equivalently there is a three-dimensional matrix, say 5, of the order (m x m x n). There are m rows, m columns, and n layers (or two-dimensional planes). The typical element of such a three-dimensional matrix can be described by the notation 5^ that denotes a credit entry for sector i and a debit entry for sector j, both with respect to the k th type of transaction. Both the subscripts i and j run from l r..,m while the subscript k runs from For any given type of transaction flow (any particular layer or plane of the matrix), the row describes all the credit entries with respect to sector i; the ir* column then describes all the debit entries with respect to sector i. Now in order for an individual sector, say the ^ sector, to practice DEB (and QEB), we need: n m

Ek

j

n m

s ijk

=

Ek Yj,

s jik

(1

*

The left-hand side (LHS) of the above equation is the grand total credits (incomings) accruing to the ^ sector. The right-hand side (RHS) then indicates the grand total debits (outgoings) accruing from the same sector. Therefore in order for the first characteristic of QEB to be satisfied, this DEB-type requirement most hold for each and every sector that is distinguished with i= ly..,m. The second characteristic of QEB can now be identified as follows. Consider the proposed equation: m m

£E

i J

mm s ijk

= EE 1 j

s Jik

(2)

The LHS signifies the grand total of all credits (incomings) assigned to the k— transaction flow category. The RHS then signifies the grand total of all debits (outgoings) assigned to the same k— transaction category. Therefore in order for the

244

H.H. Postner/On Quadruple-Entry Bookkeeping

second characteristic of QEB to be satisfied, this condition must hold for each and every transaction type (or 2-dimensional plane) that is distinguished with k = l r..ji in the original 3-dimensional matrix. Of course, many of the planes of transaction category types may be only sparsely "populated." In view of Richard Stone’s orientation, it is not surprising that Stone explicitly highlights equation (1). But equation (2) also holds implicitly and, indeed, obviously in the Stone matrix approach. So the two required characteristics of QEB can be identified and established in a form consistent with the matrix approach to national economic accounting. Further technical analysis could show that the historical concerns of: Goldsmith [1950], Ohlsson [1953], Schelling [1955] and [1958], Sigel [1962] and [1967], and perhaps even the concerns of Frisch [1953], can be incorporated in the Stone matrix approach to national economic accounting. The key step towards the resolution of their concerns is the provision of an accounting memory mechanism combined with intersectoral transaction elements that can reflect the dual values of different accounting perspectives. These dual values can occur for at least some 2-dimensional planes of transaction categories. But other planes of transaction categories can remain uniquely valued. This also provides for a more flexible description of the QEB process in its historical evolution.

References [1] Copeland, M.A. (1952) A Study o f Moneyflows in the United States, National Bureau of Economic Research, New York. [2] Copeland, M.A. (1949) "Social Accounting for Moneyflows”, The Accounting Review, July 1949. [3] D e M organ (1897) A Common Sense Method o f Double Entry Bookkeeping, on First Principles, as suggested by De Morgan, Part 1, Theoretical, London. [4] Frisch, R. (1953) "Comments on Paper by R. Stone and J. Utting, in Input-Output Relations, Leiden. [5] Goldsmith, R.W. (1950) "Measuring National W ealth in a System of Social Accounting", in Studies in Income and Wealth, Volume 12, National Bureau of Economic Research, New York. [6] Ijiri, Y. (1975) Theory o f Accounting Measurement, American Accounting Association, Monograph No. 10, Sarasota, Florida. [7] Ijiri, Y. (1989) Momentum Accounting and Triple-Entry Bookkeeping, Am erican Accounting Association, Volume No. 31, Sarasota Florida. [8] Kuznets, S. (1948) "National Income: A New Version", Review o f Economics and Statistics, August 1948. [9] Littleton, A.C. (1953) Structure o f Accounting Theory, American Accounting Association, Monograph No. 5, 1953. [10] O ECD (1990) "Transactions, Imputations and W elfare”, Meeting of National Accounts Experts, Paris, July 1990. [11] Ohlsson, I. (1953) On National Accounting, (English Text), Stockholm. [12] Pacioli, L. (1494) Summa de Arithmetica, Geometria, Proportioni et Proportionalita: Distintio NonaTractatus XI, Particularis de computis et scripturis, Paganino de Paganini, Venice. [13] Ruggles, R. (1949) An Introduction to National Income and Income Analysis, M cGraw Hill. [14] Schelling, T.C. (1955) "National Income, 1954 Edition”, Review o f Economics and Statistics, November 1955.

H.H. Postner/On Quadruple-Entry Bookkeeping

245

[15] Schelling, T.C. (1958) "Design of the Accounts", in A Critique o f the United States Income and Product Accounts", Studies in Income and Wealth, Volume 24, National Bureau of Economic Research, New York. [16] Sigel, S. (1955) “A Com parison of the Structures of Three Social Accounting Systems, in Input-Output Analysis: An Appraisal, Studies in Income and Wealth, Volume 18, National Bureau of Economic Research, Princeton. [17] Sigel, S. (1967) "Comments on the Revision of the United Nations System of National Accounts," Review o f Income and Wealth, M arch 1967. [18] Sigel, S. (1962) "An A pproach to the Integration of Income and Product and Flow-of-Funds National Accounting Systems: A Progress Report", in The Flow of Funds Approach to Social Accounting: Appraisal, Analysis and Applications, Studies in Income and Wealth, Volume 26, National Bureau of Economic Research, Princeton. [19] Stone, R. (1948) "Social Accounting, Aggregation and Invariance", University of Cambridge, D epartm ent of Applied Economics, Reprint Series No. 11, 1948. [20] Stone, R. (1947) "Definition and M easurem ent of the National Income and Related Totals”, Appendix to Measurement o f National Income and Constniction o f Social Accounts, League of Nations, Geneva. [21] Stone, R. (1954) "Input-Output and the Social Accounts", in T. Barna, editor, The Structural Interdependence o f the Economy, John Wiley and Sons, New York. [22] Stone, R. and K. Hansen (1953) "Inter-Country Comparisons of the National Accounts and the Work of the National Accounts Research Unit of OEEC", in M. Gilbert, editor, Income and Wealth Series III, Cambridge. [23] Tice, H.S. (1967) "Report of a Conference on the Proposals for the Revision of the United Nations System of National Accounts”, Review o f Income and Wealth, March 1967. [24] Yamey, B.S. (1987) "Double-Entry Bookkeeping", in Palgrave Dictionary o f Economics, Macmillan.

246

The Accounts o f Nations Z. Kenessey (Ed.) IOS Press, 1994

INTELLECTUAL FOUNDATIONS FOR THE 1968 SNA Graham Pyatt Institute of Social Studies, The Hague

Abstract. This note identifies some of the most important innovations in the architecture of the 1968 SNA and seeks to identify their origin within the corpus of Richard Stone’ work. Particular attention is given to the central role of commodity balances in the system and the use of a matrix form at for presentation of the accounts.

Introduction 1 Few students of the history of national accounting would differ from the view expressed in Kenessey [1992] to the effect that ’Stone’s imprint on the UN work was most profound, and his seminal contribution to the 1968 revision of the S(ystem) of N(ational) A(ccounts) was probably the outstanding example of his effort’. Yet the context of this contribution and its roots in earlier intellectual efforts are not generally appreciated. Certainly, there is a contrast in this respect between these more recent developments and the series of earlier events which was triggered by the insight of Austin Robinson who, as a result of hearing a lecture by Keynes, was struck by the potential contribution which a quantified version of the national accounts could make to the formulation of policy. He accordingly set a young James Meade to the task, which was to result not only in the famous appendix to an official government paper [UK 1941] but also to two key texts for economists, Meade and Stone [1944] and Hicks [1942].2 The latter, in particular, was to make the link between theory and measurement which won for national income accounting a prominent position on the agenda of macro-economics for many years. The purpose of this note is to provide a point of reference for some of the new ideas which fed into the development of national accounting at a much later stage, and eventually found their place in the 1968 version of the SNA [UNSO 1968]. There are two topics in particular to consider, namely, the articulation of commodity balances as a central feature of the system, and the use of a matrix format for presentation of the accounts.3 Both were regarded as important by Stone and pre-date as an active part of his thinking the development of asset accounts as a complement to the accounting for income and outlays. The evolution of the assets accounts is a story in itself which might therefore justify a separate note. It is not a subject which is to be covered here.

G. Pyatt/Foundations fo r the 1968 SNA

247

1. The case for a new SNA

There is no need to look further than the 1968 SNA itself [UNSO 1968] for a clear indication of the reasons for thinking that a revision of the earlier system [UNSO 1953] was necessary: "The work of the last fifteen years which is relevant to this report has proceeded in two directions: the elaboration and extension of national accounting and the construction of disaggregated economic models. Each of these developments has helped to make possible the formulation of the new system and, at the same time, has made a new system necessary if international standards and international reporting are to keep pace with the work that is going on in a large and increasing number of countries'1. [UNSO 1968, para 1.4]. That Stone was himself a major contributor to the work on disaggregated economic models provides an important part of the explanation of how he was able to remain at the forefront of developments in national income accounting over a period of thirty years, and to bring a freshness of approach to the subject matter in doing so. It is this involvement in modelling which explains the existence of input-output system at the centre of the 1968 SNA and the cast of the complete structure within a matrix framework. These innovations emerged as most striking features of the 1968 system’s architecture because they were characteristic of the type of economic model which Stone had found to be most useful in analysing the structure of an economy. The development of Stone’s thinking about economic models and structures over the years following the publication of the 1953 SNA can be traced through a series of publications, with Stone and Croft-Murray [1959] and Stone [1960] paralleling the formation of the Cambridge Growth Project in 1959 with Alan Brown. These references and the subsequent publications of the Cambridge project [Cambridge 1962 et seq] document a clear line of intellectual development was to be the mainspring for the 1968 SNA. However, while the content of this contribution is important, these intellectual developments did not arise in a vacuum. They were motivated by practical concerns and facilitated by the quantum improvements in computing capacity which were becoming available at the time. Specifically, the Cambridge Growth Project was launched by Stone and Brown as an enquiry into the economic development of the UK economy and the reasons why it had failed to achieve a more impressive performance during the 1950’s. It was this agenda which motivated the modelling work and called for new developments in the national accounts in order that they could provide the data which proved necessary to support the model. It is in this sense that the development of social accounting matrices which is reported in Cambridge [1962 et seq] represents the intellectual foundations of the 1968 SNA as discussed below. These were motivated, in turn, by the need to support the policy models which interested Stone and Brown. Thus the policy concerns drove model developments, which in turn created a demand for new data systems and, therefore, a new SNA. The fact that similar policy concerns were being identified in other countries generalised the case for a revision of the architecture of the system.

248

G. Pyatt/Foundations fo r the 1968 SNA

2. Innovations in input-output analysis The titles of the volumes previously cited [Stone and Croft-Murray, 1959, and Stone, 1960] give a clear indication of their contents, with ’Social accounts and economic models’ describing the former and ’Input-output and national accounts’ the latter. Both contain early formulations of an input-output model which is driven by a final demand vector that is sensitive to relative prices through their influence on consumption demand. They represent, accordingly, and in parallel with the work of Johanson [1960], some early steps towards the development of what are now known as applied or computable general equilibrium models. Volume I of Cambridge [1962 et seq] provides a more developed formulation of Stone’s original conception which was subsequently implemented by ’Rocket’ (the name given by Alan Brown to the proto-type computable model). As part of this general effort, a distinction which receives a good deal of attention in UNSO [1968], namely, the distinction between industry and commodity technology models of the make matrix is also to be found in the earlier references cited above, but it is not until we reach the Growth Project literature that the need for a square make matrix in the latter model is clearly identified. It is at this point that the first empirical realisations of the distinction are generated, although the matrix representation of separate accounts for industries and commodities was built into the presentation from the outset.

3. The Use of SAMs The 1968 SNA is a particular example of a social accounting matrix (SAM), which is the name chosen by Stone for the matrix (single entry) presentation of accounts which he favoured. The name was suggested in the title of Hicks [1942]. Various authors had thought to propose the matrix presentation of accounts before it appeared in Stone [1948]. However, these various antecedents to the use of the framework within the Cambridge Growth Project did not include a flow of funds and, therefore, cannot be considered as a matrix characterisation of Walras’ Law. The innovation was important, since it is necessary for the introduction of asset accounts and therefore to the articulation of a total system. The use of a matrix format for presentation of the national accounts has been resisted by a significant number of practitioners, no doubt for a variety of reasons. The main argument, however, would seem to be that the older system of double entry accounts is possibly simpler to understand. For Stone there were a number of reasons for favouring the matrix format. Of these, I suspect that his dominant concern was that the matrix format is more elegant and, only incidentally, that it is also more efficient. This concern for elegance was at a premium in Stone’s view, not for its own sake, but because it invites a deeper understanding of the way in which an economy words. If the double entry approach

G. Pyatt/Foundations fo r the 1968 SNA

249

invites the understanding of an accountant, then the matrix form of representation was, and remains, a contribution to economics because it invites the psychological novelty of actually seeing, in its purest form, how one thing relates to another. Accordingly, it represents a quantum improvement on the insight which was originally provided by the work of Leontief.4 Thus, while Leontief revealed the structure of American industry, it fell to Stone, a generation later, to indicate how other aspects of structure could equally be represented within a matrix format. A particular aspect of this shift from double to single-entry format may serve to reinforce the general point. In Stone’s view, the use of double-entry was part and parcel of an excessive concern over economic aggregates and, especially, the gross domestic product. Aggregates had their uses, of course, but one of the virtues of the matrix approach was to move away from the previous emphasis on aggregates and refocus attention on the structure of an economy. It was, accordingly, a considered decision to omit row and column totals from Table 2.1 of UNSO [ 1968], which is the main blueprint of the 1968 SNA. The architecture of the 1968 SNA was designed to encourage an appreciation of structure and how it might evolve in future: the various macro-aggregates were no longer regarded as the only justification of the system, or even the main one.

Endnotes 1.

This note is based on introductory remarks m ade in presenting the Richard Stone Memorial Lecture

2.

I am indebted to Lady Stone for this history, which she solicited on my behalf from Jam es Meade.

3.

An alternative commentary on these m atters is Bos [1991] who provides a useful introduction to the history of national accounting.

4.

See Leontief [1941],

at the Tenth International Conference of the Input-Output Association, Seville, 1993. See Pyatt (1993)

References [1] Bos, F. (1991) T h e history of national accounting’ Occasional paper No. NA-048, Central Bureau of Statistics, The Netherlands. [2] Cambridge, D epartm ent of Applied Economics (1962 et seq) A Program me Growth

Vol I: a

computable model for economic growth. Vol II: a social accounting matrix for 1960. Vol I I I : Inputoutput relationships, 1954-66. Chapman & Hall, London. [3] Hicks, J.R. (1942) The social framework, Oxford University Press, Oxford. [4] Johanson, L. (1960), A multisectoral study of economic growth, North-Holland, Amsterdam. [5] Kenessey, Z. (1992) ’Postwar trends in national accounts in the perspective of earlier developments’, Twenty-second G eneral Conference of the International Association of Research in Income and W ealth, Flims, Switzerland. [6] Leontief, W. (1941) The structure of the American economy 1919-1929, Oxford University Press, Oxford. [7] M eade, J. and R. Stone (1944), National income and expenditure, Oxford University Press, Oxford. [8] Pyatt, G. (1993), ’Modelling commodity balances’ the Richard Stone Memorial Lecture presented to the Tenth International Conference of the Input-output Association, Seville.

G. Pyatt/Foundations fo r the 1968 SNA

250

[9] Stone, R. (1948), ’Social accounting, aggregation and invariance’ Cahiers du Congrfes International de Comptabilitd. French translation in Economic Applique, Vol. II. No. 1. [10] Stone, R. (1960), Input-output and national accounts OEEC, Paris. [11] Stone, R. and G. Croft-M urray (1959), Social accounts and economic models, Bowes & Bowes, London. [12] United Kingdom, H.M. Treasury (1941), ’An analysis of the sources war finance and an estim ate of the national income and expenditure in 1938 and 1940’. Command 6261, H.M. Stationery Office, London. [13] U nited Nations, Statistical Office (1953), A system of national accounts and supporting tables, United Nations, N.Y. [14] United Nations, Statistical Office (1968), A system of national accounts, Studies in Methods, Series F, No. 2. Rev. 3 U nited Nations, New York.

251

Author Index

Arvay, J. Aukrust, O. Bakker, G.P. den Bos, F. Harrison, A. Kenessey, Z.

218 16 66 198 169 1,109

Kurabayashi, Y. Postner, H.H. Pyatt, G. Reich, U.-P. Tiwari, S.G.

93 237 246 144 124