Keywords

JEL Classifications

National accounting is a product of the 20th century, more precisely of the Great Depression, the Second World War and the subsequent period of recovery and economic growth. However, two and a half centuries earlier, estimates of national income had started with William Petty and Gregory King in England, and Vauban and Boisguilbert in France. This innovation in England, by the end of the 17th century, has been attributed to ‘the spirit of the age’ (Phyllis Deane 1955), ‘an age of great intellectual vigour, scientific curiosity and inventiveness’ (Richard Stone 1986). This early work had two main purposes: on the one hand, taxation and fiscal reforms, and on the other the assessment of the nations’ comparative economic strength in an age when England, France and the Netherlands were frequently at war. Exceptionally, King, an outstanding pioneer, made consistent estimates of various economic magnitudes (income, expenses, increase or decrease in wealth, and so on) for a series of years. However, as a rule, national income was estimated as an isolated magnitude using various methods. Estimates were intermittent and extended slowly (according to Studenski 1958, national income had been estimated at least once for only eight countries by the end of the 19th century, and for some 20 by 1929. From 1850, earlier in England, evaluations of fortune or wealth, more numerous, were disconnected from national income estimates.

From National Income Estimate to National Accounting

The influence of the First World War was limited, with some exceptions (for example, an NBER 1909–19 series in current and constant dollars published by Wesley Mitchell et al. in 1921–22). The 1929 crisis was a turning point. Official demand appeared (US Senate 1932; Carson 1975, p. 156) leading to a 1934 report prepared by Simon Kuznets and his assistants (National Income 1929–1932, in current prices, by type of economic activity and distributed income). Estimates were then extended to expenditures (final consumption and capital formation) by Clark Warburton. In a number of countries – the Netherlands (Jan Tinbergen), Sweden, Denmark (Viggo Kampmann) – large programs were developed, such as the one resulting in National Income in Sweden 1861–1930 published in 1937 by Erik Lindahl, Einar Dahlgren and Karin Koch. Working on his own, Colin Clark in the United Kingdom extended his previous 1932 estimates to a quite comprehensive coverage (National Income and Outlay1937).

The 1930s were a period of maturation in economics, apart from the conceptual and methodological deepening directly involved in this stream of quantitative estimates. The stimulus to quantitative macroeconomics given by Keynes’s General Theory (1936) provided the theoretical basis for the estimation of interdependent economic aggregates, for the relationships between income and expenditure and between saving and investment were central to his argument. Such interrelationships had not previously been absent from economic theories (think of Quesnay’s Tableau économique, Marx’s reproduction schemes or Walras’s general equilibrium analysis). However, after the Great Depression, such concepts and their statistical representations became central to macroeconomic concerns and policies. Keynes’s works were focused on macroeconomic relations, but others sought representations of the economic system as a whole in different ways. Ferdinand Grüning in Germany (1933) analysed the economic circuit at a level later called ‘mesoeconomic’, halfway between the macro and micro levels. Wassily Leontief’s research (1941) introduced input–output analysis at the level of homogeneous industrial groups, with a much broader view, in terms of general equilibrium, than the descriptive detailed balances of relations between branches (industries) prepared by P.I. Popov (1926) in the Soviet Union. The idea of an accounting approach for the economy as a whole, similar to the business accounting approach, was introduced either as a tool for improving national income estimates (as by Morris A. Copeland, following an intuition of Irving Fisher) or as part of a new proposed economic organization (André Vincent in France, Ed Van Cleeff in the Netherlands). The idea of micro/ macro relationships was present in much of this work. Coming from a very different perspective, Ragnar Frisch developed an axiomatic, bottom-up representation of economic circulation.

The Second World War was the second, decisive, turning point. National accounting, often called at the beginning social accounting, crystallized in a direct response to the problem of war finance in the UK, as explicitly stated in the April 1941 White Paper (UK Treasury, An Analysis of the Sources of War Finance and Estimate of the National Income and Expenditure in 1938 and 1940). This was backed up by a technical paper by James Meade and Richard Stone in 1941. A more elaborated ‘social accounting’ system was soon proposed by Stone in an appendix to Measurement of National Income and The Construction of Social Accounts (published by the United Nations in 1947). Inspired by business accounting, it included sector accounts grouping accounting entities and their transactions organized according to a sequence of sub-accounts, with a set of detailed definitions and the discussion of many unsettled issues. Although it covered neither balance sheets nor a detailed analysis of the productive system, this accounting system was well ahead of its time. Actually, before and during the war, the United States was in advance in both national income and related aggregates estimates and their use, as for instance in the 1942 feasibility study of the Victory Program led by Kuznets or the analysis of the inflationary gap (Carson 1975, p. 174–7). However, the National Income Division of the Commerce Department, with Milton Gilbert, evolved towards a simple accounting framework rather than a developed accounting system.

Though they encountered many difficulties and though it was a very uneven development, mostly due to deficiencies in statistical information and staffing, national accounting experienced a kind of golden age in the three decades following the war. Economic reconstruction and growth policies, the large increase in the economic role of government and the welfare state, the extension of international cooperation (for example, the Marshall Plan and, later, the Common Market in Europe), with the consequent emphasis on measuring of the rate of growth, led to a great demand for national accounts. This comprised the requirements of Keynesian macroeconomic demand management for short-term economic budget forecasts and longer-term projections needed for various types of indicative planning (the latter being particularly important in France). The development of econometric techniques and national accounts estimates reinforced each other. This trend towards greater use of national accounting data was general, even though the economies involved ranged from basically liberal economies such as the United States to more controlled economies such as France, the Netherlands and Norway.

International Harmonization and Extensions

Country experiences interacted with the process of international harmonization very early. Discussion between Canada, the UK and the USA took place in September 1944. There was a meeting of a League of Nations Committee, for which Stone prepares a memorandum, in December 1945. Stone played a prominent role in the first generation of standardized systems (OEEC 1950, 1952; United Nations 1952). This first attempt at standardization across the Western world as a whole, however, was too limited in scope, and was very far from the ambitions of the 1945 accounting scheme. Conceived as a simplified model for countries that were only beginning to develop their national accounts, it could not meet the needs of countries that were already more advanced, such as Scandinavian countries (Odd Aukrust in Norway, Ingvar Ohlsson in Sweden) or even a country like France. Under the impulse of Claude Gruson, France was, in the 1950s, in order to implement far- reaching economic policies, beginning the process of building a comprehensive and ambitious system of its own, integrating accounts for economic agents, input–output tables and financial transactions in a way that was more integrated than the Copeland’s money-flows accounts in the United States.

Until the end of the 1960s the Western stage was characterized by the existence of a variety of national systems that were difficult to reconcile, even among those countries that adopted, in principle, the same comprehensive concept of production, including non-market government services. The new French system adopted a narrower concept of production, limited to market goods and services. The Soviet Union and its satellites used the even more restricted concept of material production, limited to goods and the so-called material services (mostly the transport of goods), following the old tradition of Smith and Marx. However, during the 1960s intense international discussions took place, on the basis of the wide range of national experiences in Europe and North America and the demands of international organizations. The result was the adoption of a second generation of standardized systems, the 1968 System of National Accounts (SNA) and the new European System of Accounts (ESA 1970), prepared on the basis of a report by Stone for the UN (the OECD deleting its system) and a French expert for the European Community. The European Community, thinking the 1952 system was too narrow and unsuited to harmonizing the accounts of its original six members and to meeting the needs of Community policies, had decided in 1964 to establish its own system.

The new system (they can be described as a single system, for SNA and ESA were very close) was closer to Stone’s 1945 inspiration and to the French, Scandinavian and British systems than to the 1952 standardized system, in terms of coverage (in particular of input–output tables and financial accounts), integration and institutional orientation. The main weakness remained the absence of balance sheets, despite the pioneering work of Raymond Goldsmith in the United States at the beginning of the 1960s. Fixed capital formation was limited to tangible assets and the relation between income and changes in wealth was not fully shown.

The System of Balances of the National Economy, built around the material product concept, was also standardized, though little innovation was involved, through the framework of the Council of Mutual Economic Assistance, and then published by the United Nations (1971). Careful comparisons between the SNA and the Material Product System (MPS) were carried out in the UN European Economic Commission in Geneva.

France decided to leave its own peculiar system and join, via ESA 1970, the international system, this being achieved by 1976. The USA was not actively involved in the elaboration of the 1968 SNA, keeping its National Income and Product Accounts, whose accounting and conceptual framework had evolved little since 1947.

A quarter of a century later, a third generation of normalized systems has taken the trend towards a universal system a step further. The 1993 SNA/ESA 1995 closed the accounting framework by including balance sheets and completing the accumulation accounts with the introduction of a revaluation account (holding gains and losses) and an account for other types of capital gains and losses. Intangible capital formation was partly accounted for. In the current accounts, the analysis of income distribution was deepened (primary income distribution, secondary distribution, and redistribution in kind), actual final consumption was differentiated from final consumption expenditures, via the re-routing of social transfers in kind from government to households. This clarification of the accounting relation between income and changes in wealth (net worth) has deep implications (see below).

Nearly full integration was achieved between the SNA and the International Monetary Fund manuals (Balance of Payments, Government Finance Statistics, Monetary and Financial Statistics). The MPS disappeared at the beginning of the 1990s with the collapse of the Soviet Union and the fast transition of China towards a market economy. Paradoxically, the USA followed a slower path towards adopting the SNA framework.

During this long process of extension and harmonization of the accounting framework, the substance of the accounts changed dramatically in comparison with what was involved when the focus was on estimating national income. The product aggregate soon became the most important one, on a par with the expenditure aggregate. The income aggregate not only lost its position of being the single aggregate, but was often given a secondary position. From that, a series of consequences resulted.

The factor cost method of valuation, when still in use, was reduced to a lower rank than the market price valuation (in spite of the recurrent objection of ‘double counting’). The latter was much more convenient for the valuation of expenditure and the analysis of consumer behaviour. In an integrated framework, the market price valuation was then applied also to the product aggregate (domestic product takes progressively the first place) and much later on to the income aggregate. In the 1993 SNA, full recognition was given to the concept of national income at market prices, which is in fact the new name given to the earlier concept of national product (which was not actually a product but an income concept).

Partly for similar reasons, gross concepts have generally come to be preferred in practice, even though net concepts, that is, after deduction of consumption of fixed capital (depreciation in the usual business terminology), were considered closer to what was generally understood by the idea of national income. Both gross and net concepts of product, income and expenditure are finally considered part of the SNA/ ESA.

The analysis and measurement of production and flows of products (goods and services), both in current value and in volume, have been given an increasing importance in relation to the integration of supply and use or input–output tables (a characteristic feature of the 1968 SNA/ESA 1970). This is increasingly done using the framework of annual tables. The integration with income estimates is less clear in practice, though the concept of value added, a significant improvement, and not only in words, on the old expression ‘net output’ or ‘net product’, provides the necessary link.

In this context, thanks to Stone’s contribution, significant improvements in valuation concepts were made in the 1968 SNA. This widens and differentiates the usual notion of market prices. Basic prices, excluding net taxes on products, were introduced on the output side, resulting in the measurement of value added at basic prices. All taxes, minus subsidies, on products are then introduced. On the use side, acquisition prices are defined as purchasers’ prices including only non-deductible taxes.

Measures in constant prices (described as volume measures), combining quantity and quality changes, also changed significantly. The trend was from globally deflating national income using a single price index in the 1930s, to deflating each of the main items in the balance of products (output, final consumption, and so on) using specific indices, and finally to an integrated system of price and volume measures, at a detailed level, using an input–output framework when annual tables were available (with Denmark, France, the Netherlands and Norway leading here). Double deflation, of output and inputs respectively, was used for value added in this context. International manuals by Stone (1956, 1968 SNA, ch. 4) and Peter Hill (1972; United Nations 1979) recommended such an approach. Later on the 1993 SNA/ESA 1995 recommended replacing the traditional fixed base indices with chain indices, preferably Fisher volume and price indices or acceptable alternatives.

Much more complex, both conceptually and practically, international comparisons of volume levels of aggregates were the object of an International Comparison Project (ICP), launched in 1968, after the pioneering research of Colin Clark (1940) and Gilbert and Irving Kravis (1954) at the OEEC. Purchasing power parities, more significant than exchange rates, were calculated. The results of the ICP, however, were not as widely implemented or as widely accepted as national volume measures, something that is unfortunate in a globalized world.

Beyond the progressive completion of its integrated framework, attempts were made to broaden the scope of national accounting by developing semi-integrated additional constructs, such as the satellite accounts whose idea was introduced (by Vanoli) by the end of the 1960 (for example, accounts for social protection, health, education, and environmental protection). In such an approach, the fully integrated system itself becomes the central framework (the expression often used, ‘core accounts’, is ambiguous).

Social accounting matrices (SAMs) were designed by Stone and Alan Brown in 1962, in order to achieve more flexibility than was possible using the usual account presentation. Though the word ‘social’ here means only ‘for the whole economy’, it gave rise to a certain ambiguity. SAMs are sometimes presented as a kind of alternative framework.

In the late 1980s, the Dutch proposed an ambitious ‘system of economy-related statistics’ as a way of organizing a vast array of statistics. A ‘core system’, narrower than the SNA central framework, was linked with ‘system modules’, such as social and environmental modules. This proposal had some similarity with the unsuccessful attempt by Stone, in the first half of the 1970s, to design for the United Nations a system of social and demographic statistics. It echoes the growing importance given to the micro–macro linkages (for example, Richard and Nancy Ruggles 1986), in parallel with the increased availability of micro-databases.

Concern for statistical coordination had, of course, been present in national accounting from the very beginning.

New Challenges Since the Mid-1970s

The achievements of national accounting, in the face of an enormous development of statistics, have been impressive. However, many countries are still far from fully implementing the international system (for example, few countries prepare integrated balance sheets), and economic and social conditions have changed drastically, especially since the mid-1970s. As a result national accounting, often questioned, sometimes radically, has had to face new challenges.

Since around 1980, after the supply shocks of the 1970s and the decreasing role played by macroeconometric models, national accounting has no longer been supported by the Keynesian paradigm. Some people even think it is obsolete. However, the demand for national accounts continues to grow, even if it also changes. Predominantly short-term concerns have led to a pressing demand for quarterly accounts, and even sometimes for a monthly GDP, resulting in conflicts between timeliness (early estimates are required) and accuracy. Though more accurate, through successive revisions, annual accounts seem less used and their results are less commented upon.

In the opposite direction, computable general equilibrium models have multiplied since the mid-1970s as a means of studying policies aimed at structural change. Without any concern for the setting up of time series, they are based on the accounts of a single year supplemented, as required, by other data dictated by the models’ specificities and purposes. Although they use the somewhat misleading SAM terminology, they actually need national accounts bases.

It remains true, however, that for the study of structural and social policies economists and social researchers, since the last two decades of the 20th century, have generally preferred to make use of micro-simulation models. The role of national accounts data is relatively reduced in this context.

In contrast, a considerable extension of the institutional and political role of national accounting took place during the 1990s, mostly in Europe. Certain aggregates (GDP or GNP) had been used fairly early for administrative purposes such as country contributions to international organizations, eligibility thresholds to preferential World Bank loans, regional allocation of European structural funds, and the ‘Fourth own budgetary resource’ of the Community budget. However, the debate over accession criteria to the European Economic and Monetary Union (the creation of the euro) marked a qualitative jump in the consideration of national accounting by policymakers and public opinion. Most Maastricht criteria were defined in reference to the ESA (ratios of public deficit and public debt to GDP). The ESA became compulsory for member states of the European Union. This marked the culmination of the European statistical strategy adopted in the 1960s. Closely related to the international statistical systems, like the SNA, European statistical tools are in effect very often legally based.

The policy uses of the ESA necessitate effective harmonization of the content of the accounts. A procedure of verification and evaluation of the comparability and representativeness of GDP is established. Full harmonization is, however, difficult. Because conceptual and statistical issues and political considerations intervene, especially in the procedure for identifying excessive deficits, specific cases have to be studied, sometimes through a rather difficult process. Here, and in issues such as the ratio between compulsory levies and GDP, national accounts appear at the forefront of sensitive political concerns. While it clearly shows their importance, this situation may also have less positive aspects for the national accounts. There is the possibility of political pressures, though this is rare; there may be lack of flexibility; official obligations and procedures can be very time-consuming and, as a result of limited human resources, European national accountants may become insufficiently involved in research work.

No similar policy-led process is taking place at the world level. However the need for regulation on a global scale is increasingly felt. Monitoring and intervention aimed at remedying local and regional crises and at preventing systemic crises falls to the International Monetary Fund, in agreement with the principal economic powers. Hence the growing role of the IMF in the supply, by member states, of timely and well-documented harmonized information. In the last decade of the 20th century, the Fund set up a system of standards to guide countries in data dissemination, including meta-information concerning various characteristics of the data. The structuring role of national accounts has been particularly highlighted. The Fund has conducted assessment missions in order to evaluate the quality of countries’ national accounts and data systems.

The impressive increase in the demand for and use of national accounts statistics has taken place against the background of economies which have become much more complex, and hence more difficult to describe and measure, than was the case in the three decades following the Second World War. The number and sophistication of available products have grown; changes in product quality have become more rapid; the share of services, generally more difficult to measure, especially in volume, has increased. The effects of technical change, opening the global economy, the transformation of enterprises and groups, refinements of price policies and consumer behaviour, continuing financial innovations, frequent extension of informal activities, and so on have caused a tendency for economic information systems to maladjust. Hence many controversies arise, notably on price and volume measurements of capital goods – quality change based on performances (Robert Gordon) or on resource cost (the traditional solution championed by Edward Denison) – or measurement of consumption goods and services, where the Boskin Report in the United States (Boskin et al. 1996) argued that the price increase was overestimated.

Significant methodological progress has been in areas such as the measurement of quality change of durable goods based on the change in their performance, the US having taken the lead. However the field is huge, and research is mostly concentrated on information and communication technology products. The measurement of financial and insurance services is in progress. Intangible assets are increasingly investigated. For non-market services, the necessary focusing on direct output–volume measurement instead of the traditional input–volume approach opens, at the start of the 21st century, another wide field of research. It soon appears that the relationship between the concepts of output and outcome must be clarified. On the other hand, some very important issues, like interest and inflation, the treatment of R&D expenditures and the extraction of subsoil resources, have remained outstanding for a long time, defying consensus, though relevant solutions do exist.

After a long emphasis on the relationship between production, income and expenditure, national accounting concerns have in recent decades been extended to the full set of relations between production, income, accumulation and wealth. This raises complex issues concerning the analysis and measurement of capital, particularly intangible assets, and consequently income. By the end of the 20th century business accountants faced similar difficulties with the emerging international accounting standards, moving from historical cost, which national accounting always rejected, to fair value valuation of assets.

Thus, national accounting is fighting for a better coverage of its traditional object at the same time that, at least since the early 1970s, new social concerns have given rise to requests for aggregate monetary indicators synthesizing broader sets of phenomena. There remain things that national accountants cannot do. One is the provision of a welfare indicator, a function that Kuznets assigned to national income, and which gave rise, in the 1940s, to an intense debate involving John Hicks and Paul Samuelson that reached negative conclusions (William Nordhaus and James Tobin 1973, later tried to provide such a measure with their ‘measure of economic welfare’). Another is the measurement of an environmentally adjusted domestic product. The suggestions in this direction included in the 1993 United Nations Handbook, Integrated Environmental and Economic Accounting, do not reach a consensus and are not implemented. There was then a move towards wanting a sustainable product or income measure, but this does not make any answer easier, though Hicks’s concept of income (the maximum amount that can be consumed in a period while expecting total wealth to be unchanged at the end of it) has increasingly been advocated in recent decades.

Most difficulties relate to the observation and measurement of non-market nonmonetary flows and stocks. Economists propose at least partial measurement solutions, within the framework of standard economic theory, using, for instance, contingent valuation methods (which raises problems of combination with actual exchange values, transfer of results and aggregation), or theoretical constructs with idealized conditions, seeking to justify a possible interpretation of net domestic product in terms of both welfare and sustainability. Other approaches, however, lean towards synthetic indicators combining both monetary and non-monetary variables.

Tensions between social concerns, theoretical issues and observation constraints of actual economies are increasingly at stake.

See Also