What was the LTES Project?

by Konosuke Odaka

In this essay I shall provide a brief explanation of the LTES (long-term economic statistics) research project, predecessor of the Asian Historical Statistics Project.


The term LTES refers to a collection of historical economic statistics which were compiled and published as a fourteen-volume series under the joint editorship of the late Professor Kazushi Ohkawa and Professors Miyohei Shinohara and Mataji Umemura. The editors were all faculty members at the Institute of Economic Research at Hitotsubashi University where the LTES project was conducted. The series offers the best available set of quantitative records of Japan's modern economic growth for the period between 1868 and 1939.(1)

The system of national income accounting was used as the analytical basis for compiling and editing the statistics. The adoption of and adherence to such a conceptual framework is the dominant characteristic of the LTES series, and one may well view it to be both the major strength and most pronounced weakness of the series. The analytical framework was established by Simon Kuznets and later refined and enhanced as the System of National Accounts (abridged as SNA). SNA has provided the basis for the world-wide compilation and publication of national income statistics, conducted, especially since 1953,(2) under the leadership of the UN Statistical Office. Consequently, the statistics are essentially macro in nature; that is to say, they are formulated with the nation-state as the basic unit of observation.

The fourteen volumes of the LTES were released between 1967 and 1987 by Toyo Keizai Shimposha (The Oriental Express). Each volume comprised (1) a description, in Japanese, of the estimating procedures and an interpretation of the results, and (2) the supporting statistical tables, all of which came with English captions and a brief English summary of the statistical footnotes. The volume titles, translated into English, are:

Of these, volumes 1, 4 through 10, and 14 deal directly with the components of national income, whereas the balance of the series supplies economic statistics which are of particular importance in tracing the modern economic history of the country.

The English version of the historical national income statistics of Japan was published separately as Patterns of Japanese Economic Development, co-edited by Kazushi Ohkawa and Miyohei Shinohara (New Haven: Yale University Press, 1979). The original authors contributed to the English volume, drawing on the estimates and statistics compiled in the original Japanese 14-volume set. However, there are some notable differences in the figures reported in each version. The gap is attributable mainly to the time that elapsed between the two separate dates of publication, as the statistical estimates were subject to a process of continuous revision and improvement.

It should be noted that the major time-series in the LTES volumes have since been converted to computer-readable form and are available from the Documentation Centre of Statistical Information on the Japanese Economy, Institute of Economic Research, Hitotsubashi University. Manuals for this version of the LTES have been released as LTES deetabeesu kaisetsu [The LTES database: a user's handbook] and LTES deetabeesu kensaku shisutemu kaisetsu [A manual for the LTES database retrieval system] (1991).

The reader may also find it useful to know that there are several additional time series statistics that are closely related to the LTES series, compiled by Shozaburo Fujino, Fukuo Igarashi, and Juro Teranishi. (The statistical figures in the LTES are confined to the regions comprising Japan proper and exclude former Japanese colonies.) Subsequent to the publication of the fourteen volume series, a companion volume was prepared on the ex-Japanese colonies and released from the same publisher, The Oriental Express, with Professors Mataji Umemura and Toshiyuki Mizoguchi as co-editors (1988).(3)

A Brief Historical Account of the LTES Project

The initial stimulus for preparation of the LTES came from Yuzo Yamada's Nihon kokumin shotoku suikei shiryo [Statistical source materials for estimating Japanese national income] (Tokyo: Toyo Keizai Shimposha, 1951). This work is seldom used now that the LTES series is complete, but it still contains some useful statistics which are not readily reproducible in the LTES. The figures for corporate savings are a notable example.

Compilation of the series, particularly the identification and collection of basic statistics, was an arduous task. The method of compilation, accuracy of the statistics, and characteristics of the statistics were all scrutinized and, when necessary, corrected and/or revised. The outcome of this work was assembled into a number of worksheets, the preparation of which in those days was largely dependent on the abacus, which was still used (concurrently) even after the introduction of desk calculating machines. The manual calculators were then replaced by machine-operated calculators, which were in turn replaced by electronically controlled machines. The statistics were released periodically as provisional output series under the heading of "processed economic statistics."(4) The series was then revised, if necessary, before being fully integrated to form part of the GDP (or GNP) time series.

In any event, the compilation of the LTES was a heavily labor intensive operation whose success was made possible by the cooperation not only of all the Institute staff but also of temporarily hired young assistants. Financial contributions from the Rockefeller Foundation to the project, which was headed by Professor Ohkawa, was instrumental in pushing the work ahead, especially in its early stages.

In the late 1950s, Kazushi Ohkawa, in association with Miyohei Shinohara, Mataji Umemura, Masakichi Ito, and Tsutomu Noda, produced the first official (although still provisional) publication of the LTES estimation project. This was entitled Nihon keizai no seicho ritsu (Tokyo: Iwanami Shoten Publishers, 1956). Its English version was The Growth Rate of the Japanese Economy since 1878 (Tokyo: Kinokuniya Bookstore, 1957).

A number of prominent economic studies have made extensive use of the outcomes of the LTES series. Foremost have been the publications of Professor Takafusa Nakamura, including Senzenki Nihon no keizai seicho [An analysis of the pre-war Japanese economy] (Tokyo: Iwanami Shoten Publishers, 1971) and Nihon keizai [Post-war Japanese economy] (Tokyo: University of Tokyo Press, 1978; second edition 1980).(5) Other contributions of a similar nature include Shozaburo Fujino's Nihon no bijinesu saikuru [Business cycles in Japan] (Tokyo: Keiso Shobo, 1965) and Shotoku riron [National income theory] (Tokyo: Toyo Keizai Shimposha, 1973; revised edition 1984); Ryoshin Minami's Nihon no keizai hatten [Economic development of Japan] (Tokyo: Toyo Keizai Shimposha, 1981; second edition 1992); and Shunsaku Nishikawa's Nihon no seicho shi [A history of Japan's growth] (Tokyo: Toyo Keizai Shimposha, 1985).(6) (JA further notable work is that of Professor Toshiyuki Mizoguchi, who has proposed a few important hypotheses for furthering the understanding of the macro-economic significance of the ex-Japanese colonies in his Taiwan Chosen no keizai seicho (The Economic Growth of Taiwan and Korea) (Tokyo: Iwanami Shoten Publishers, 1975).

Important Characteristics of the LTES

The LTES is the first set of economic historical statistics of Japan to be based on the modern national income accounting system. As has been mentioned earlier the data was subjected to rigorous checks in the process of compilation.

Modern statistical information is relatively abundant in Japan, even by contemporary European standards. This is a result of efforts of central and local government bureaucrats to gather such statistics and to their accord with the political rulers that such information was necessary. Naturally, some of the sources and methods of compilation go back to the Edo era.

These statistical sources were first utilized extensively by Marxist-oriented economic historians who were searching for factual evidence that the Japanese economy was facing political-economic catastrophe (according to the prediction of Karl Marx) in the early twentieth century.(7) LTES not only used the same data sources from a totally different perspective but identified and utilized numerous entirely new data sources (such as the chohatsu bukken [survey on mobilizing resource capabilities]).

Some Deficiencies of the LTES

In this writer's view, there are at least three problems with Japanese LTES.(8) First, due to the length of time over which the research was conducted and the large number of scholars who participated, some inconsistencies have emerged in the basic concepts. In some cases, two or more figures could be derived depending on the selection of source materials. For example, two separate sets of time series data for domestic capital formation (investment) can be obtained either by (1) carefully adding the current output values of producers' durable equipment, or by (2) taking differences between the previous and current year's capital stock values.

Second, a consistency check has not yet been completed between different time-series statistics contained in the LTES. For instance, we do not as yet possess much knowledge about inter-industry transactions of goods and services between different sectors of the economy. H.B. Chenery, Shuntaro Shishido, and Tsunehiko Watanabe have made an important contribution in this respect by attempting an estimate of an input-output table of the Japanese economy in 1914 and 1954. This was followed by a monograph by Masahiko Shintani, Senzenki sangyo renkan kozo no henka ni kansuru suryoteki kenkyu, sangyo renkan bunseki ni yoru sekkin [A quantitative study of the changing industrial structure of prewar Japan, an input-output table analysis] (Fukuoka: Seinan Gakuin University, 1988).

Third, the coverage of the statistics is not comprehensive enough. All the statistics are confined to annual data due to the nature of the source materials. Moreover, certain indispensable constituents of national income - such as product inventory and corporate savings - have not been incorporated into the final figures, mainly because of technical difficulties in estimating them. Professor Mataji Umemura has long argued that agricultural output, especially in the earlier decades, is quite likely to be under-reported due to omissions in the record of miscellaneous crops, which comprised a very important source of calorific intake for the farmers in the period. The data pertaining to the production and consumption of energy-generating industries are also a possible source of under-representation. Of course, one of the most serious sources of deficiencies in the statistical information stems from the difficulty of correctly assessing the activities of the tertiary sector (commerce and other service industries).

Future Tasks of the LTES Japan

The following three items should be considered as priorities in any possible revisions of the LTES Japan in the future.

Admittedly these tasks are not easy to conduct, since the statistical source materials were collected by entirely different principles before the Meiji era and diverse institutions were involved in data collection (if it was collected at all). Regional diversity was in all likelihood much higher then. Nonetheless, the construction of a continuous series should be started as soon as possible as it would be infinitely rewarding not only to the students of Japanology but to all those who are concerned with matters pertaining to comparative economic history.


(1) "Modern economic growth'' refers to a process of economic change with the following characteristics: (1) a sustained rise in per capita real income, (2) continuous population growth, and (3) conscious application of scientific and technological advances to economic development, causing rapid structural change in the economy. The term was first coined by Simon Kuznets in his Six Lectures on Economic Growth (Glencoe, Illinois: Free Press of Glencoe, 1959).

(2) A description of national income accounting, as applied to Japan, is available in J.R. Hicks and Nobuko Nosse, The Social Framework of the Japanese Economy (Tokyo, London: Oxford University Press, 1974). The first handbook of SNA was issued in 1953, and was amended and expanded in 1968. See also Yoshimasa Kurabayashi, Studies in National Economic Accounting, (Tokyo: Kinokuniya Bookstore, 1977) and United Nations, A System of National Accounts and Supporting Tables (New York 1953).

(3) Preceding publication of the companion volume, an interim report containing estimates of the national income statistics of Taiwan and Korea under Japanese rule was made public in Shigeru Ishikawa and Miyohei Shinohara, eds., Taiwan no keizai seicho [The Economic Growth of Taiwan] (The Institute of Developing Economies, 1972).

(4) A set of the processed provisional output series are kept in the library of the Documentation Centre of Statistical Information of the Japanese Economy.

(5) The English versions of the Nakamura volumes have been published as The Postwar Japanese Economy: Its Development and Structure (Tokyo: Tokyo University Press, 1981) and Economic Growth in Prewar Japan (New Haven: Yale University Press, 1983).

(6) Minami's volume has also been published in English as The Economic Development of Japan: A Quantitative Study (Basingstoke: Macmillan and Company, 1986; second edition, 1994).

(7) See, for instance, Moritaro Yamada, Nihon shihonshugi bunseki [An analysis of Japanese capitalism (Tokyo: Iwanami Shoten Publishers, 1934) as well as a later handbook version.

(8) It is worth noting here that an overall review of the LTES series was once attempted by an empirically-minded theoretical economist, Shunsaku Nishikawa, in Kikan riron keizaigaku [Economic Studies Quarterly] (Vol. XXVII, No. 2), pp. 126-34.

(9) Some preparatory research, especially in connection with the Japanese economy during World War II, has already been conducted by Professors Takafusa Nakamura, Yoshiyuki Mizoguchi, Akira Hara, and others.