T HE whole problem of capital requirements in economic growth has received increasing attention in recent years under double impetus of post-Keynesian preoccupation with full employment in more highly developed countries and new interest in theory and strategy of economic development for underdeveloped areas. Growth models such as those of Harrod, Domar, and Fellner explicitly focus on productivity of investment, i.e., productivity of incremental additions to capital stock, as a determinant conditioning full-employment rate of growth.' At same time, most of projections for underdeveloped areas are based on models in which capital-output ratio figures as one of principal growth variables.2 In this context, there have been two schools of thought as to comparative levels of marginal capital-output ratios in developed versus underdeveloped economies. The first school, following good classical doctrine, maintains that on an a priori basis one would expect a comparatively low marginal ratio in underdeveloped countries, where capital is relatively much scarcer than labor. On other hand, second school argues that in early stages of development a large share of investment resources will need to be allocated to social overhead capital, productivity of which is low. Therefore on this account, as well as owing to shifts in industrial structure, while initial marginal capital-output ratio tends to be high, it will decline at more advanced stages of economic growth. For instance, Colin Clark maintains that the figure of capital requirements per unit of output rises . . . in early stages of industrialization, but that later effect of modern technology is to bring this figure down. I With this background in mind, it seemed to us that an investigation of incremental capital-output ratio in Soviet Union might throw some additional light on factors impinging upon problem of capital requirements during periods of rapid growth in early stages of development. The Soviet process of economic growth with its marked emphasis upon development of producers' goods and defense industries has been generally considered a highly capital-intensive one. However, as we will attempt to show, this conclusion seems to be belied by realities of Soviet development experience for economy as a whole, although it does have some validity for industrial sector alone. Specificially, in this paper we propose (i) to develop measurements of aggregate and industrial incremental capital-output ratios in Soviet Union, (2) to indicate some of factors which may account for comparatively low ratios observed, and (3) to point to some of implications of our results for measurements in general.

NATIONAL income accounting provides a broad avenue for systematically surveying the state of a country's statistical development, for identifying the most serious statistical gaps, and for appraising the reliability of the data. This function may be particularly important in areas where statistics are not regularly published, but are presented in scattered and unsystematic fashion. While this statistical mobilization and survey function is methodologically quite important from the standpoint of further studies of China's mainland economy, it is a subsidiary rather than a primary purpose of this study. Thus, the major focus is upon an analysis of the structure and flows in the economy, upon composition rather than just size of national product. Our work on the Chinese Communist economy, its performance, its capacity for growth, the relative importance of different sectors, the rate of saving and investment, the sources of saving, etc., has been greatly handicapped by the absence of national income estimates for some recent period. All of the previous estimates are for the I93I to I936 period.' Owing to the profound structural and institutional changes in the Chinese mainland economy, these would be inapplicable to the I950's even if they were methodologically and statistically unassailable. For recent years, the Chinese Communists have published some highly aggregated estimates.2 However, it is difficult to work with these, since they are not buttressed by a detailed exposition of sources and methods used. The national income concept used in this context is apparently the same as in the Soviet Union, that is, it excludes passenger transport, government, and many other types of services. In recent years, particularly since I955, we have begun to obtain much more information on various aspects of the Chinese economy, e.g., agricultural and industrial production, government budget, agricultural taxation and marketing, profits of state enterprises, levels of investment expenditure, etc. However, it has been very difficult to interpret and appraise these data since they could not be fitted into a broader framework and could not be assessed against the background of performance and intersectoral flows in the economy as a whole. In effect, -all of our investigations have suffered from the lack of an analytically meaningful yardstick. It is my hope that this study, confined to an analysis of national income accounts in one year, will at least go part of the way toward filling this gap. It should be viewed as an attempt to explore a new field of inquiry rather than to furnish definitive estimates. One of the purposes here is to apply the national accounting framework to an underdeveloped sovietized economy and to work out a method of approach to the available data which can then also be used for future estimates. There is no doubt that as the work on Chinese national income proceeds, these methods will be improved and new information will become available so that our present estimates will need to be revised. Space forbids a detailed exposition of methods and sources. Neither is it possible to discuss at length the pricing and valuation problem. These tasks will necessarily be left to a forthcoming monograph on national income and product in Communist China. Therefore, the primary purpose of this paper is to summarize the findings, to give some general indications of the methods of estimation pursued, and to bring out some of the analytical implications.