i. GeneralRichard Ruggles
iii. Mainland ChinaTa-Chung Liu
During the first half of the twentieth century there was considerable evolution in the scope and content of economic data. At the turn of the century, statistical abstracts of economic data generally contained only information relating to the production and prices of specific agricultural and mineral commodities, the value of imports and exports, and, in some cases, information on tax revenues and outlays by central governments. The financial crises of the late nineteenth century and early twentieth century stimulated an increase of interest in banking and financial statistics, and as central banks were set up and systems of reporting regularized, data on the operation of the banking system emerged as another type of basic economic data.
Economists as such did not play an important role in fashioning economic data prior to World War i. In the 1920s, however, the situation changed rather dramatically. One of the first areas of data in which the activities of economists became important was that of index number construction. Irving Fisher and Warren Persons did much to advance index number theory and to foster the development of price and output measurements. Fisher not only developed the measurement of index number formula (1922) but he also pioneered in the development of what later became known as econometrics —the advancement of economic theory in its relation to statistics and mathematics. Warren Persons’ work and, later, that of Wesley C. Mitchell gave impetus to the analysis of economic time series; they concentrated on the task of separating trends and seasonal and cyclical factors from one another. Mitchell was the founder of the National Bureau of Economic Research, an institution in the United States that has continued to be in the forefront in the development and analysis of economic data. In England during this period, work on economic data was concerned somewhat more with the development of statistical theory. Bowley and Stamp were particularly influential in the development of statistical techniques and techniques for the analysis of economic data.
In spite of the rapid development during the 1920s, however, it is fair to state that prior to the great depression of the 1930s economic data remained highly fragmented, and there was no unifying framework to integrate the large number of independent series. Economists had been preoccupied with the search for a few key economic indicators that would solve all their problems, revealing the true economic cycle and showing the level of prices and industrial activity.
A completely different view of the functioning of the economic system was introduced by John Maynard Keynes in The General Theory of Employment, Interest and Money (1936). By emphasizing the basic idea that the functioning of the economy depended upon the interaction of its parts, Lord Keynes fostered the concept of economic data as a systematic body of interrelated information. For the first time a comprehensive set of economic data covering all production, consumption, saving, and investment for the economy as a whole was introduced in the context of an operational economic model. The pioneering work in the field of national income measurement that had been done by Will-ford I. King (1930), Simon Kuznets (1941), and Erik Lindahl (see Stockholm, Universitetet 1933–1937) was developed into a social accounting framework by Richard Stone (1951). Simultaneously, Wassily Leontief (1941) was developing another operational economic model involving the interrelation of inputs and outputs of specific industrial sectors of the economy.
The development of national income accounting and input-output analysis thus introduced the idea that economic data are a set of systematically interrelated pieces of information that can be utilized to show the functioning of the various parts of the economy. However, this concept of a framework for economic data did not supplant interest in the growing body of partial data in almost every area of economic analysis. Data in such fields as agriculture, labor, manufacturing, trade, etc., continued to expand. The significance of the more systematic approach to economic data lay partly in its ability to digest, integrate, and put order into the massive increase in information about specific sectors of the economy.
World War ii provided a further impetus to the expansion of economic data. The mobilization of the war economy required vast amounts of information about the productive capabilities of various industries and their supply requirements. The suppression of inflation through taxation and price control required the further development of national income accounts. In almost all countries considerable advances were made in the development of systematic bodies of economic data, and in the immediate postwar period these advances continued. Programs of economic recovery were instituted, and international organizations such as the Organization for European Economic Cooperation and the United Nations stimulated countries to make basic improvements in their statistical reporting systems.
The trend toward the utilization of more comprehensive bodies of economic data in econometric models is continuing. Prior to World War ii econometric models were generally quite aggregative, but in recent years these models have become considerably less so. For example, the model Lawrence Klein developed for the United States economy in 1950 contained 16 equations; the 1959 Klein-Goldberger model contained 32 equations; and the 1963 Social Science Research Council model contains over two hundred equations. Similarly, the research studies by Jan Tinbergen in the Netherlands and by Richard Stone in England have been moving toward disaggregated econometric models involving complex interrelations among the variables. [SeeEconometric models, aggregate.]
Recently a somewhat different approach has been developed, utilizing computer simulation techniques in order to project economic behavior over a future period or to analyze how an economic system behaves under specific conditions. Basically, simulation utilizes economic models in such a way that major pieces or segments of the model are treated as probability samples for which behavioral characteristics are taken from sample surveys or masses of detailed data observations at the micro-economic level. By utilizing a building-block approach, simulation techniques can handle many different levels of aggregation simultaneously and therefore are capable of utilizing both highly detailed microeconomic data and more aggregated macroeconomic data. The recent advances in computer technology have thus made feasible the utilization of economic data in a manner that was previously not thought possible. [SeeSimulation.]
Types of economic data available
It is obvious that the content of economic data is undergoing profound change and that at the present time we are in the middle of these changes. Nevertheless, it will be useful to identify what are the main types of economic data currently available and to discuss the ways these bodies of data are changing.
National income accounts. Within the past decade national income accounts have come to be recognized as one of the most basic types of information about a nation’s economic activity. In a number of countries the government is required to present, along with its budget, a forecast of the national income accounts for the coming period, showing how the proposed government budget relates to anticipated economic developments.
National income accounts today generally cover five sectors of the economy. (1) The gross national income and product account shows who purchases the final output of the economy and how the income generated by such economic activity flows to individuals as income, to governments as taxes, or to corporations as profits. (2) The personal income account shows the income received by individuals, whether in their role as participants in the productive process or as recipients of social security benefits or other transfers from government. The personal income account also shows how people dispose of their income, i.e., how much of their income goes for consumer expenditures and personal taxes and how much is saved. (3) The public sector account, which has become considerably more important in recent years, shows, on the receipts side, what the government receives from taxes on individuals or on businesses and how much it receives as property income, such as the surplus of government enterprises. The expenditure side of the government account shows the outlays of the government in terms of the functions which it undertakes, such as education, health, defense, general administration, and economic development. (4) The balance of payments account shows the foreign trade of an economy in terms of imports and exports and transfers to and from abroad. This account shows whether a nation is running down its reserves of foreign exchange or acquiring a surplus. (5) The final account, the saving and investment account, on one side brings together the excess of income over expenditure in each of the other sectors, and on the other side shows the capital formation by government and by public and private enterprises.
These five sector accounts are so set up that an expenditure in one account corresponds to a receipt in some other account. Thus, for example, the amount of consumer expenditures in the personal income account corresponds exactly to the sale of goods to consumers appearing in the gross national product account. [SeeNational income and product accounts.]
Flow-of-funds data. The national income accounts cover only current income and expenditure transactions; they exclude those financial transactions that merely alter the assets and liabilities of the various sectors of the economy. Because this type of financial information is important in any analysis of fiscal and monetary policy, sets of information on the sources and uses of funds by the different sectors of the economy have been developed in many countries. In the United States information on the flow of funds is published by the Federal Reserve Board. This information is generally consistent with the national income accounting data, but the two sets of data are not fully integrated. In some other countries—for example, Canada, the Netherlands, and France—the flow-of-funds data have been integrated with national income data to produce one consistent set of data interrelating the income accounts with the financial transactions of each sector.
Input-output data. Input-output data showing the purchase of inputs and sale of outputs by industrial sectors of the economy are also becoming available for many countries, but unlike national income and flow-of-funds data, this information rarely appears on an annual basis. The statistical requirements of a complete input-output system are very considerable, and it is usually possible to obtain such a body of data only for periods for which economic censuses are available. Again, however, there is an increasing tendency toward integration of national income accounting data and input-output data. For example, the gross product originating by industrial sector has come to be a common breakdown in the national income accounts, and this often matches the industrial breakdown provided in the input-output matrix. [SeeInput-output analysis.]
National wealth. The development of flow-of-funds accounts and input-output matrices has fostered increased interest in balance sheets for various sectors of the economy. Data on the financial transactions of each sector and on capital formation make possible, via a perpetual-inventory approach, the development of national wealth and sector balance sheets. This work, however, is still in its early stages, and few countries at present publish official current data on national wealth. [SeeNational wealth.]
Data coverage, collection, and processing. One of the major characteristics of economic data systems at the present time is increasing detail, both in terms of time coverage and in terms of disaggregation. For example, today most of the industrial economies publish quarterly data on national income accounts; and for some components, such as personal income, they may even publish monthly data. There is also increasing interest in regional breakdowns of information within nations. In large part this is due to the fact that the different regions of a nation develop differently, and economic development must recognize differences in regional patterns, for the consideration of problems of location of new industries and for dealing explicitly with problems of regional poverty. [SeeRegional science.]
In developing these new bodies of data, governments have made use of two new tools which have made possible the development of much more reliable data at lower cost: (1) the use of sample surveys and (2) the development of electronic data processing. Since World War ii there has been a considerable increase in the use of sample surveys that are specifically designed to obtain desired information. Many of these sample surveys have been very successful in providing data which heretofore were completely unavailable. Private groups as well as governments have been able to develop important bodies of data. For example, in the United States the Survey Research Center at the University of Michigan has developed an important body of information on consumer expenditures, and the McGraw-Hill Publishing Company has developed an interesting and useful survey of investment plans of industrial firms.
The development of electronic data processing techniques has permitted governments to make use of existing bodies of information in ways which were previously impossible because of the extremely high cost of processing. The increased use of high-speed computers to process and edit data has meant that major bodies of data that are collected for other purposes, e.g., income taxes and social security, can be utilized for statistical purposes. Furthermore, bodies of information, such as census data, can be made available in a space of a few months rather than the several years which used to be required before final tabulations were available.
Problems of measurement and reliability
There can be little doubt that the quality as well as the quantity of economic data has shown substantial improvement in recent years. In part this improvement has been due to increasing knowledge of sample design with resulting improvements in the ability to measure biases and to take into account errors in reporting. It is generally true that careful sampling in any given area can produce much-higher-quality data at very much lower cost than can complete reporting. The flexibility of sampling techniques permits the analysis of a large number of different factors in a way that would not be possible if complete and uniform reporting were required of each responding unit.
Furthermore, the reliability of data is not determined purely at the point of collection. Careful editing is very important if the final results are to be meaningful and consistent, and great strides have been made in recent years in this area. The computer is especially valuable in this context. It is now possible to carry out many internal checks within a body of data to reveal inconsistencies and to resolve such inconsistencies in an optimal manner. Perhaps even more important, however, is the increasing possibility of bringing together information from a number of different sources for comparisons and consistency checking. For example, matching census data with income tax data provides some overlapping information which, when compared on a case by-case basis, may reveal unexpected weaknesses in one or both sets of data. Awareness of problems is the first step toward the improvement of the quality of data.
The problem of reliability of data and the measurement of error in empirical work in economics is complex. In the simplest case the concept of sampling error is not too difficult to comprehend or measure. One is essentially asking how much variance can be expected to occur on repeated sampling of a given type of information. In point of fact, however, a given piece of economic data is never used in isolation; it is always related to other pieces of economic data, which also have sampling or reporting errors. The problem of measurement of error, therefore, comes down to the joint error distribution of the different pieces of data that are used in conjunction with each other. For example, data on the volume of consumer expenditures in a particular period may have a certain sampling error, but it does not follow that the error in the change in consumer expenditures from one period to another can be calculated on the basis of sampling errors. The reason for this is rather simple. The sampling errors in the consumer expenditure data in different periods may be expected to be somewhat correlated with each other, and to the extent that this is true the error in the change in consumer expenditures will be affected. It would be quite possible, for example, for the error in the level of consumer expenditures to be quite substantial; but as long as the error is relatively constant, the change in consumer expenditures may be quite accurate.
Validity of measurement and operationalism. For economic data as a whole, the problem of the validity of measurement is often considerably more important than that of reliability. It is, of course, quite proper that economists approach empirical information with a framework that embodies preconceptions as to the nature of the data. To do otherwise would result in a chaotic mass of information without any principles for organizing the data. Nevertheless, the economists’ preconceptions often represent an oversimplification of the actual situation.
In the field of price measurement, for example, economists reasoning from a static model envisage an economy where all transactions can be expressed as the product of a price and a quantity, and the quantity of output of the system as a whole can be measured in terms of a properly constructed index. This basic concept of price and output underlies our measurement of the gross national product in constant prices. Unfortunately, there are impediments to making this simple theoretical model operational. Few products remain unchanged over time, and there is no satisfactory way to evaluate whether changes in price do or do not correspond to changes in quality. Furthermore, one of the most common features of economic development is that new products are developed to replace old products. Unfortunately, the newly developed products seldom bear a one-to-one relationship with the old products. Thus, central heating may replace heavy wool clothes, and telephone communications may replace travel and mail correspondence. These substitutions are very subtle and cannot be perfectly measured.
The statistician must make tremendously oversimplifying assumptions in order to make data about the real world fit into the theoretical economic models. For example, price statisticians usually assume implicitly that new products are exactly equivalent in utility and cost to the products they replace and that old products improve in quality only in a few, rigidly specified ways. The problem with such a solution, however, is that the results of measurement in a given situation are often dictated by the assumptions that are made rather than by what has actually happened. In measuring output in such areas as construction and services, for instance, the statistician is often forced to substitute the measurement of inputs for the measurement of the output, thus, by definition, assuming that there is no change in productivity in these sectors.
Similar problems have arisen in many other areas of economics where the statistician has attempted to implement concepts developed by the theorist. For example, in the development of the concept of capital, theorists have been vague as to the definition and meaning of real capital over time. Statisticians have dealt with the problem by adopting certain conventions, such as formal rates of obsolescence and depreciation, but in doing so they have in part determined by assumption the answer to the question of how much the growth in real capital contributes to the expansion of output.
In view of these limitations, economics is still far from being able to answer such interesting questions as how much better off we are now than at some previous time or how much productivity in various sectors has advanced. But such questions, interesting as they are, may not be the most important or productive questions that might be asked. Perhaps it is more important to find out how the economy functions, what parts are growing relative to other parts, and how the change comes about. Questions of this sort are much more likely to be answered by the growing body of economic data now becoming available. It is much easier to describe and analyze economic change than to evaluate the welfare or productivity effects of such change.
Copeland, Morris A. 1952 A Study of Money flows in the United States. Publication No. 54. New York: National Bureau of Economic Research.
Fisher, Irving (1922) 1927 The Making of Index Numbers: A Study of Their Varieties, Tests, and Reliability. 3d ed. Boston: Houghton Mifflin.
Goldsmith, Raymond W. 1955–1956 A Study of Saving in the United States. 3 vols. Princeton Univ. Press.
Goldsmith, Raymond W. 1962 The National Wealth of the United States in the Post War Period. National Bureau of Economic Research Studies in Capital Formation and Financing, No. 10. Princeton Univ. Press.
Keynes, John Maynard 1936 The General Theory of Employment, Interest and Money. London: Macmillan. → A paperback edition was published in 1965 by Harcourt.
King, Willford I. 1930 The National Income and Its Purchasing Power. Publication No. 15. New York: National Bureau of Economic Research.
Klein, Lawrence R.; and Goldherger, A. S. 1955 An Econometric Model of the United States: 1929-1952. Amsterdam: North-Holland Publishing.
Kuznets, Simon 1941 National Income and Its Composition: 1919–1938. Publication No. 40. New York: National Bureau of Economic Research.
Leontief, Wassily W. (1941) 1951 The Structure of the American Economy; 1919-1929: An Empirical Application of Equilibrium Analysis. 2d ed. New York: Oxford Univ. Press.
McGraw-Hill Publishing Company, Incorporated, Department of EconomicsBusiness Plans for New Plants and Equipment. → Published annually.
Michigan, University of, Survey Research CenterSurvey of Consumer Finances. → Published since 1945/1946.
Mitchell, W. C. (1915) 1938 The Making and Using of Index Numbers. 3d ed. U.S. Bureau of Labor Statistics, Bulletin No. 656. Washington: Government Printing Office.
Ohlsson, Ingvar 1953 On National Accounting. Stock-holm: Konjunkturinstitutet.
Orcutt, Guy H. et al. 1960 Simulation: A Symposium. American Economic Review 50:893-932.
Orcutt, Guy H. et al. 1961 Microanalysis of Socioeconomic Systems: A Simulation Study. New York: Harper.
Ruggles, Richard; and Ruggles, Nancy (1949) 1956 National Income Accounts and Income Analysis. 2d ed. New York: McGraw-Hill. → First published as Introduction to National Income and Income Analysis.
Stockholm, Universitetet, Socialvetenskapliga Institutet 1933–1937 Wages, Cost of Living and National Income in Sweden: 1860–1930. Volume 3: National Income of Sweden: 1861-1930, by Erik Lindhal, Einar Dahlgren, and Karin Kock. London: King.
Stone, Richard 1951 The Role of Measurement in Economics. Department of Applied Economics Monographs, No. 3. Cambridge Univ. Press.
Tinbergen, Jan 1938–1939 Statistical Testing of Business-cycle Theories. 2 vols. Geneva: League of Nations, Economic Intelligence Service. → Volume 1: A Method and Its Application to Investment Activity. Volume 2: Business Cycles in the United States of America: 1919–1932.
United Nations, Statistical Office (1953) 1964 A System of National Accounts and Supporting Tables. Studies in Methods, Series F, No. 2, Rev. 2. New York: United Nations,
Adequate collection of economic information, rapid transmission and processing of the data, and proper utilization of scientific procedures for drawing conclusions from the data are vital for any modern society—a fortiori, for centrally planned economies. The Soviet Union and the communist-ruled east European countries are socialist planned economies, that is, economies whose “commanding heights” (banking, industry, transport, etc.) are in the hands of the state. From these commanding positions the communist leadership attempts to guide the main economic processes of each country within the framework of comprehensive plans that set specific goals for the economy as a whole and for its main sectors and branches. Resources are channeled by a variety of means toward fulfilling these goals, and performance of each production unit is systematically followed up as plan implementation unfolds.
Industrial and nonindustrial data
For planning purposes a vast and comprehensive amount of information is needed. The statistical organization must reach deeply into the economic and social system and must have access to all the relevant data. Systematic collection of data on many facets of economic life, adequate condensation of information at various levels, feeding of the appropriate information to planning and management organs, methodic follow-up of the operational aspects of plan implementation, and manifold analyses of the developmental aspects of the economy are permanent tasks in continuous planning. These tasks raise complicated problems with respect to methods of aggregation and disaggregation of the data, accuracy in record keeping, comprehensiveness and timeliness in reporting, effectiveness and speed in auditing and processing and, above all, boldness and freedom in analyzing the results obtained.
In the Soviet Union and in each east European country, the backbone of statistical apparatus is the Central Statistical Administration (CSA)—in Yugoslavia, the Federal Institute of Statistics. The CSA handles all the statistical data, that collected by itself as well as that collected by all other agencies involved in statistical operations (banks, ministries, etc.). The main stream of information flowing to the statistical offices comes from the producing enterprises. The producers report daily, weekly, monthly, etc., so-called operating-technical data, in which inputs and outputs are expressed in physical terms. A second basic stream of information, furnished in value terms, flows to the CSA from the producers and from the banking and financial system, particularly the national and investment banks and, in Yugoslavia, the Social Auditing and Accounting Agency. On the basis of the data processed by the CSA and the information collected by their own organs and consolidated at local, regional, or ministerial levels, the planning organs elaborate the operational (short-term) and “perspective” (long-term) plans within the framework laid down by the directives of the policy makers.
There are avowedly numerous deficiencies in the producers’ input-output data used in preparing the planning statistics. The deficiencies are due primarily to the following: (a) performance is judged on the basis of reported data; (b) operating-technical and value data are not well coordinated; (c) far too heavy demands are continuously placed on the producers for data on all aspects of their activity.
Since plan fulfillment is evaluated and managerial rewards are assigned on the basis of reported data, managers are tempted to manipulate and misreport statistics in a variety of ways.
Further, the Soviet and east European enterprises (excepting the Yugoslav) obey numerous, often conflicting, commands and have to keep a variety of overlapping and poorly integrated accounts: operating—technical records of physical flows; accounting records in value terms; controlling records of plan fulfillment. The flow of materials organized on the basis of specified norms of uses of inputs and outputs is not well coordinated with the value aspects of production either at the level of the plant or at higher levels. There are no unified balance sheets at the level of the enterprise on the basis of which the over-all situation could be easily and rapidly appraised, and there is lack of integration between recorded material flows and value indicators at the level of the economy as a whole.
Finally, the enterprises are called upon to produce an enormous volume of statistical reports, often overlapping and superfluous, in response to the urgent requests of statistical, banking, or planning organs at local or regional levels. Swamped by innumerable forms which they must complete for a variety of purposes, the enterprises deliver a vast but low-quality output of data. The fight against paperwork (the Kampf gegen die Papierflut as the East Germans call it) has long been a continuous but losing battle in all the east European economies except Yugoslavia.
At the heart of almost all of these problems is the limited opportunity for decision making granted to the managers of producing enterprises, that is, the centralization of economic management and decision making. The greater the volume of commands from above, the greater the volume of norms to be complied with from below; the broader the decentralization in decision making, the better the incentive system—as in Yugoslavia since the early 1950s—and the smaller the needs for specifications, norms, and continuous checking from above.
The deficiencies of the nonindustrial statistics are even more glaring than the deficiencies of the industrial statistics. Agriculture, for instance, furnishes notoriously poor information because of its over-all low level of development, its type of organization, its system of statistical reporting, and the poorly trained personnel used for record keeping, particularly on the collective farms.
The principal responsibility of the CSA is to compile output indexes by branches, sectors, or regions, in physical or value terms, and feed them to the appropriate planning and management organs. (The efforts of these organs depend in effect upon the completeness, rapidity, and accuracy of the data collected by the central statistical offices.) Moreover, the CSA is charged with the elaboration of social accounts. The accounts deal with production, consumption, and investment, with transfers and “redistributions” of the social product, and with labor resources. Integration and aggregation of data are achieved by the CSA following the Marxian definitions of production. According to these definitions, production includes services directly connected with the processing and distribution of goods and excludes all other services.
Only a relatively small portion of the information available to the CSA is published. What is published is still in some cases annoyingly incomplete, of questionable quality, or of little relevance for assessing the true performance of these economies. During the Stalin era—1928 to 1952— concealment, bizarre manipulations, and outright distortions became the trademarks of Soviet statistics. A cloak of secrecy was cast over most of the key domestic economic data. This secrecy has been appreciably lifted since the mid-1950s, but even now many economic aspects of the Stalin period for the Soviet Union, and of the early postwar decade—1945 to 1955—for all the area, remain unclear; further, some crucial data are still withheld.
The Soviet and east European policy makers have a high stake in the performance of their planned economies. Hence, they are always sensitive about some troublesome indicators, such as the rate of growth of agricultural output or the rate of growth of the economy as a whole in certain years. Generally, however, ample statistical materials are currently released on industrial outputs, manpower, investment, state budgets, etc. Much of this data is not readily comparable to the Western data. The differences are due not only to dissimilarities in underlying statistical methodologies and frames of reference but also to underlying discrepancies between East and West concerning taxation and pricing systems. The Soviet-type tax system blurs the line between requited and unrequited factor returns, and administrative price manipulation raises complex problems concerning the valuation of output. The severe limitations in the nature of the underlying ruble prices and of the definitional frameworks used, along with the various other deficiencies mentioned, hamper the usefulness of Soviet social account and of Soviet income indexes. A systematic effort has been made in the West to eliminate the distortions of the available Soviet statistical materials and to recompute the Soviet national income series for the whole of the planning era opened in 1928. A similar effort of recomputation has been made for various east European countries.
During the late 1950s and early 1960s, Soviet and east European statistical recording, processing, and publishing have increasingly improved. Early in 1945, the fledgling east European statistical administrations achieved the necessary shift in focus from the prewar type of social statistics to the type of economic statistics needed for planning purposes. After centering their attention on the compilation of industrial or agricultural inventories, statistical administrations made a serious effort to progress from the collection of “photographic” inventory-type data to the collection of “cinematic” production-type data that would be valuable for planning and forecasting. During the 1950s the CSAs solved in part the problems of following up the plan performance and of elaborating output measurements by branches, sectors, planning controls, etc. During that period the first complete postwar population censuses were published in eastern Europe. Since then, the energies of the statistical machines have been turned toward the complex problems of social accounting. In some countries—for example, the Soviet Union, Yugoslavia, Hungary, Poland, and East Germany—relatively successful steps have been taken toward the elaboration of input-output tables. The volume and quality of statistical information published have notably increased, and the publication of detailed and informative statistical yearbooks has been undertaken by all these countries. Statistical and economic journals are published regularly.
Concepts and methods
All this activity has been highly dependent in the area as a whole, except Yugoslavia, on statistical concepts and methods elaborated during the Stalin era. Severe doctrinaire limitations have existed in regard to the use of probability theory and to the use of mathematical methods in general in the social sciences. The key contention that dominated Soviet statistical thought under Stalin was that socioeconomic statistics form a special “social science” based on Marxian dialectics; the body of methods known as mathematical statistics based on probability and the law of large numbers—to which, incidentally, many great Soviet scientists contributed—was considered applicable only to the natural sciences.
Since 1960, however, sustained efforts have been made in all these countries to break away from the specious Stalinist concepts and methods. Mathematical methods are increasingly being introduced into economics and statistics. Throughout the area, there is growing interest in Western theories of decision making under uncertainty, Western techniques of sampling, planning, social accounting, and activity analysis, and Western concepts of efficiency and optimality. These influences will play a crucial role in the late 1960s and beyond, when Soviet and east European statistics will have to tackle new and more complex tasks. Hopefully, they will result in better training of statisticians, new points of view in statistical elaboration, and less high-handedness in the publication of data.
Improvements in the general quality of the data will require several reforms and innovations. Reporting demands on producing enterprises must be reduced, and the output indicators now compiled should be supplemented with quality indicators. Automating the collection and transmission of data would significantly increase speed and efficiency in the performance of these tasks.
Developing an optimal system of unified accounting for economic planning and control is an ambitious task. Further progress toward this goal will necessitate a vast amount of research, not only in mathematics and statistics but also in economics, cybernetics, and sociology. Complex multi-staged programming models must be formulated, and methods must be developed to solve short-term and long-term planning control problems at various levels. To implement these models and methods, an integrated network of computational centers is indispensable.
The prospects for improvement and progress also depend very much on reorganizations of the economies themselves. The volume of centralized commands will probably decrease as the market mechanism is increasingly relied on to determine resource allocation. This should lead to an overhauling of incentive systems and should produce price systems that indicate relative scarcities and production costs of goods more accurately than present price systems. If such reorganizations take place, serious progress in the elaboration and dissemination of economic data may be expected. Statistical theory and experimentation have a long tradition in these countries; however, the basic challenge now is not so much renewing ties with the past but successfully meeting the new tasks and problems that a developing economy poses to economists and statisticians.
Statistical yearbooks have been published regularly by all the communist-ruled countries of the area since the mid-1950s. After World War ii, only Yugoslavia and Poland published yearbooks before the mid-1950s. Nearly all the countries publish a bulletin (usually monthly) of current statistics and a journal covering theoretical or methodological problems. The bibliographical items below discuss the concepts and problems of information in historical perspective.
Bergson, Abram 1961 The Real National Income of Soviet Russia Since 1928. Cambridge, Mass.: Harvard Univ. Press.
Campbell, Robert W. 1963 Accounting in Soviet Planning and Management. Russian Research Center Studies, No. 45. Cambridge, Mass.: Harvard Univ. Press.
Egermayer, Frantisek 1959 The Evolution of Soviet Views on Statistics. Czechoslovak Economic Papers 1:229-265. → Also published in 1959 in Czech.
Fedorenko, N. P. 1965 Devising Methods of Managing the National Economy. Washington: Joint Publications Research Service. → Also published in 1965 in Russian.
Feshbach, Murray 1960 The Soviet Statistical System: Labor Force Recordkeeping and Reporting. U.S. Bureau of the Census, International Population Statistics Reports Series P-90, No. 12. Washington: U.S. Bureau of the Census.
Grossman, Gregory 1960 Soviet Statistics of Physical Output of Industrial Commodities: Their Compilation and Quality. National Bureau of Economic Research, General Series, No. 69. Princeton Univ. Press.
Jasny, Naum 1962 Essays on the Soviet Economy. New York: Praeger.
Kaser, Michael 1960 Economic Development in the Soviet Union and Eastern Europe: With Special Reference to Statistics. Statsokønomisk tiddskrift> 74:284-295.
Kotz, Samuel 1965 Statistics in the USSR. Survey: A Journal of Soviet and East European Studies , no. 57:132-141.
Nemchinov, V. S. 1963 The Use of Statistical and Mathematical Methods in Soviet Planning. Pages 171-188 in International Conference on Input-Output Techniques, Geneva, 1961, Structural Interdependence and Economic Development: Proceedings. New York: St. Martins.
Scientific Conference on Statistical Problems, Budapest, 1961 1962 Input-Output Tables: Their Compilation and Use. Edited by O. Lukács et al. Budapest: Akadémiai Kiadó.
The communist regime was established on the Chinese mainland in October 1949. A substantial amount of economic statistics was published by the regime during the period 1949–1960. While poor in quality, the statistics for 1952–1957 could, with proper caution and careful evaluation, serve as a reference for understanding the performance of the economy. The data for the earlier years, 1949–1951, are less reliable and those for the “great leap forward” years, 1958–1959, are extremely untrustworthy. No significant economic statistics have been released by the regime since 1960.
There is no clear evidence of intentional falsification of statistics at the level of the central government prior to 1958. The poor quality of the data Was primarily due to the underdeveloped nature of a large sector of the economy. The collection of reliable statistics from hundreds of millions of old-fashioned producing units was inherently difficult; and the numerous drives to fulfill targets, often unreasonable even before the “great leap forward” in 1958, induced and compelled local units to file exaggerated reports of achievement.
Development of a statistical network
There was no effective national statistical reporting system until the establishment of the State Statistical Bureau on August 8, 1952. By the end of 1953 roughly 160 local statistical offices had been established at the provincial level for the relatively large cities. During 1955 and 1956 efforts were made to extend this network to smaller towns and rural units, but even as late as the end of 1957 the difficult task of establishing perhaps a quarter of a million local statistical offices had not been completed. Without an effective organization to collect them, the data cannot be expected to be reliable (Li 1962).
Major statistical undertakings completed from 1952 to 1957 include the population census of 1953, the unification of statistical reports on state-managed enterprises and construction in 1954, the 1954 census of private industry and individual handicraftsmen, the 1955 census of private trade and restaurants, and the 1955 census of workers and employees in the socialist sectors. Reporting on agriculture had to be entrusted to the regular local government units. “Model surveys” of questionable quality were carried out mainly for food crops and cotton in 1953 and 1954, the choice of the sample being left to the judgment of the individual investigators. A survey of the income distribution of agricultural producer cooperatives was carried out in 1955 and 1956. Sample surveys of the incomes of farm households and of urban workers and employees were conducted in 1956. A study of the age and sex distribution and the educational and political background of workers and employees of all industrial, construction, and transportation enterprises, as of February 1957, was reported to have been made.
The effort toward improving the statistical system was abruptly shattered by the “great leap forward” in 1958. The collection and reporting of statistics were in effect taken over by the Communist party mechanism and were integrated into emulative drives to increase production. This led to the unprecedented statistical fiasco of 1958, from which the regime learned a sober lesson. A serious effort was made in 1960 to rehabilitate and strengthen the statistical system. With this effort, however, the publication of national economic statistics practically stopped.
Important data for 1952–1957
For 1952–1957, a period of relatively good statistics, the best data are those on government budgets and such homogeneous products as steel and electric power, produced by large industrial enterprises. Poor as it is, the population count of 1953 must be considered better than the reports on agricultural and handicraft production. A substantial amount of retail trade was transacted by small private traders, for whom no reliable information was available. Estimates of national income had to be made with these basic data of varying degrees of unreliability. Only a brief review of some of the available data will be given.
Population.Table 1 presents the U.S. Census Bureau’s summary of the regime’s 1953 population census. This census has been criticized by scholars both inside and outside mainland China. The census was not obtained by house-to-house canvassing. Instead, heads of households were required
|Table 1 — Population of mainland China, by age and sex, June 30, 1953|
|(In per cent)|
|Source: Aird 1961, pp. 81, 83.|
|45 years and over||19.0||21.3||20.2|
to report to registration stations set up for this purpose. Since the census was taken primarily for registering voters for the national election of 1953-1954, it is doubtful that political undesirables would have come to register. Inadequate preparation and insufficient pilot surveys, the prolonged period of enumeration, the highly inaccurate “pencil stroke system of tallying,” the troublesome problem of temporary residents (supposedly reported by families or neighbors at place of permanent residence), and the tradition of underreporting of females must have resulted in substantial inaccuracy. The high degree of social control in Communist China, the one factor favorable to an accurate count, was not sufficient to counterbalance all the unfavorable factors.
Further doubts were cast on the accuracy of the census on the basis of demographic analysis. Aird (1961) found a marked deficiency of persons 5 to 14 years of age and an excess of persons 45 and over. The former, however, could have been the result of the Sino-Japanese and the nationalist-communist conflicts from 1937 to 1949, which were unprecedentedly violent even for China. The latter could have been the result of the relative absence of large-scale wars during the last two decades of the Ching dynasty. Aird also constructed theoretical models of population dynamics for mainland China and found the vital rates implicit in the age distribution of the reported 1953 population to be incompatible with the demographic history of China. Conversely, using the vital rates reported for the years near 1953, Aird could not derive the reported 1953 population. These analyses, while interesting, are inconclusive. It is doubtful that vital rates were constant in China during any reasonably long period of time, an assumption made in these analyses. The vital rates reported for the years near 1953 were not a part of the 1953 census and are even less reliable than the 1953 population count. Thus, they should not be used to invalidate the census itself.
The 1953 census is certainly not comparable in accuracy to censuses of the more developed countries. It was, nevertheless, collected by a systematic and comprehensive effort, however poorly designed. Comparable efforts have never been made to obtain statistics for the output of agriculture and the small producing units in the underdeveloped sectors of industry and transportation or for the incomes and expenditures in these sectors. These other statistics must be considered inferior to the population data for 1953.
Industrial production. Modeled on the Soviet concept of gross value of industrial production, the official data on this important item are aggregates of the gross value of production of the reporting enterprises. With some important exceptions, double counting was not permitted within an enterprise; raw materials produced and used by the same enterprise in making its own output were excluded from its value of production. Purchases from other enterprises, however, were not excluded. It is known that managers of individual enterprises went as far as purchasing one another’s intermediate products in an attempt to increase their respective gross values of production and to fulfill or overfulfill quotas.
Prices for the third quarter of 1952 were used as weights in computing the official gross values of production for years prior to 1957. Prices of consumer goods were depressed in the fall of 1952 because of the “Five Anti” campaign against private enterprises (a movement outwardly directed against (1) bribery; (2) tax evasion; (3) theft of public assets; (4) fraud in fulfilling government contracts; and (5) theft of government information), but prices of producer goods were little affected. Thus, producer goods were overvalued relative to consumer goods. Since output of producer goods increased faster than output of consumer goods after 1952, the official rates of growth of industrial production were upward biased. The valuation of new products at trial-manufacturing expenses (deflated by 1952 prices) also exaggerated the increasing trend.
|Table 2 — Indexes of industrial production of mainland China|
|Year||Official la||Official llb||Chaoc||Liu-Yehb|
|a. Gross value of factory output.|
|b. Net value added in factory production.|
|c. Factory production.|
|Sources: Official I index from China 1960, p. 16; Official II index from Liu & Yeh 1965, p. 150; Chao index from Chao 1965, p. 88; Liu-Yeh index from Liu & Yeh 1965, p. 66.|
Attempts were made by scholars in the United States to correct some of the biases. Since the concepts used in the different studies are not the same, only the rates of change are roughly comparable. The different estimates are therefore presented as index numbers in Table 2.
Agricultural production. There is a consensus among Western scholars that the official data on production of food crops underestimated output in earlier years, so the rate of growth for the 1952-1957 period was overstated (Eckstein 1962, p. 32; Hollister 1958, pp. 19, 29; Li 1959, p. 65; Liu & Yeh 1965, pp. 43-46; Wu et al. 1963, p. 185). The official data and the other estimates are given in Table 3.
|Table 3 — Food crops production of mainland China (billions of metric tons)|
|Sources: Official data from China 1960, p. 119 Liu-Yeh data from Liu & Yeh 1965, p. 132; Wu data from Wu et al. 1963, p. 185.|
National income. The concept used by the regime in computing national income is the net material product, that is, the net value produced by agriculture, industry, construction, freight transportation, the part of communications serving other productive sectors, trade, and restaurants. Income originating in government administration (including armed forces), cultural and educational organizations, and personal services is excluded.
|Table 4 — National income of mainland China (billions of 7952 yuan*)|
|*Official rate of exchange: 1 U.S. dollar equals 2,343 yuan in 1952. For well-known reasons, the yuan figures should not be converted into U.S. dollars at the official exchange rate for the purpose of comparing them with the national income of the United States or of other countries.|
|Sources; Official data from Liu & Yeh 1965, p. 220; Eckstein data from Eckstein 1962, p. 56; Hollister data from Hollister 1958, p. 2; Li data from Li 1959, p. 106; Liu-Yeh data from Liu & Yeh 1965, p. 66; Wu data from Wuetal. 1963, p. 241.|
Scholars in the United States have attempted to construct independent estimates following standard Western concepts. Eckstein (1962) estimated the gross national product of 1952 by using the value-added approach for agriculture and the factor shares method for other sectors. Hollister (1958) applied the final-expenditure method to estimate gross national product. Li (1959) adjusted the official estimate of net domestic product to derive an estimate of net national product but cautioned that the implied rate of growth in his estimates was excessively high. Liu and Yeh (1965) used the value-added approach to estimate net domestic product. Wu and associates (1963) used the LiuYeh framework but modified the estimates in several respects.
The official data and the independent estimates are presented in Table 4. Because of the different concepts of national income used, only the rates of growth can be even roughly compared. The average annual rates of growth for the 1952-1957 period are as follows: official, 9.0 per cent; Hollister, 8.6 per cent; Li, 8.8 per cent (an overestimate, according to Li); Liu and Yeh, 6.0 per cent; and Wu, 5.6 per cent.
The 1958 fiasco and its aftermath
The production targets set for 1958 during the “great leap forward” were extremely ambitious and completely unrealistic. The statistical reporting system gradually developed since 1952 was cast aside, and the local cadres dutifully reported unprecedented but impossible achievements. In April 1959 the regime announced 100 per cent increases in the output of food crops, cotton, and steel in 1958 alone. A drastic downward revision of the claims, however, was announced in August 1959, reducing the estimated production of food crops and cotton by one-third. The claimed increase in iron and steel production was also scaled down substantially but in a more subtle way. Official economic statistics published for 1958 and 1959 (China 1960) are of much lower quality than those of 1952–1957, and the several independent Western estimates made for 1958 and later years can be considered only as conjectures.
The disruptive economic policies undertaken during the “great leap forward,” together with severe floods and droughts in 1959 and 1960, resulted in a serious agricultural crisis and a sharp reduction in output in the ensuing years. The rest of the economy also suffered setbacks, resulting from the shrinking supply of raw materials from the agricultural sector, the withdrawal of Soviet technicians in 1960, and the generally lower morale and efficiency of workers because of food shortages. The almost complete blackout of economic statistics from the Chinese mainland since 1960 makes it exceedingly difficult if not impossible to give any quantitative evaluation of the performance of the economy from 1960 to 1964. There are some indications that by 1964 the 1957 level of production may have been regained in most sectors of the economy.
Aird, John S, 1961 The Size, Composition and Growth of the Population of Mainland China. Washington: U.S. Department of Commerce, Bureau of the Census.
Chao, Rang 1965 The Rate and Pattern of Industrial Growth in Communist China. Ann Arbor: Univ. of Michigan Press.
Chen, Nai-Ruenn 1963 The Economy of Mainland China, 1949-1963: A Bibliography of Materials in English. Berkeley: Social Science Research Council, Committee on the Economy of China.
Chen, Nai-Ruenn 1966 Chinese Economic Statistics: A Handbook. Chicago: Aldine.
China (People’s Republic), Kuo chia t’ung chi chŪ (State Statistical Bureau) 1960 Ten Great Years: Statistics of the Economic and Cultural Achievements of the People’s Republic of China. Peking: Foreign Languages Press.
Eckstein, A. 1962 The National Income of Communist China. New York: Free Press.
Emerson, John P. 1965 Nonagricultural Employment in Mainland China: 1949-1958. Washington: U.S. Department of Commerce, Bureau of the Census.
Hollister, William W. 1958 China’s Gross National Product and Social Accounts: 1950–1957. Glencoe, III.: Free Press.
Li, Choh-ming 1959 Economic Development of Communist China: An Appraisal of the First Five Years of Industrialization. Berkeley and Los Angeles: Univ. of California Press.
Li, Choh-ming 1962 The Statistical System of Communist China. Berkeley: Univ. of California Press.
Li, Choh-ming (editor) 1964 Industrial Development in Communist China. New York: Praeger.
Liu, Ta-Chung; and Yeh, K. C. 1965 The Economy of the Chinese Mainland: National Income and Economic Development, 1933-1959. Princeton Univ. Press. → Data in tables, reproduced by permission. Copyright © 1965 by The RAND Corporation.
Orleans, Leo A. 1961 Professional Manpower and Education in Communist China. Washington: Government Printing Office.
Wu, Y. L.; Hoeber, F. P.; and Rockwell, M. M. 1963 The Economic Potential of Communist China. Unpublished manuscript, Stanford Research Institute.