Knowledge that Transforms

SOVIET Russia can serve as an extremely fascinating case study in the field of economic development-a field where scientific curiosity is certainly not matched by available statistical data, and where a good deal of ingenuity or foolhardiness is needed for the construction and evaluation of estimates from scattered pieces of information. In studies of economic development a method frequently used for checking estimates and establishing their comparative importance consists of applying the lessons of the past experience of some advanced country (such as the United States) to economically younger areas. Our present knowledge of economic development is insufficient for a final evaluation either of this method as a whole or of its applications, especially in the case of countries whose economic development has been unique. But attempts to draw such comparisons, whatever their direct validity in specific cases may be, appear promising and useful as a source of material from which a theory of economic development can be gradually constructed. This problem came up during the recent controversy regarding the validity of the Soviet machinery index which took place in this REVIEW between Professor Gerschenkron and Mr. Dobb.2 Gerschenkron maintains that the official Soviet index of machinery output based on I926-27 weights is overstated, essentially because new machinery, not yet produced in 192627, was subsequently included in it at inflated prices. He presented the following comparison between the machinery index and an index of ferrous metals in Russia: I As shown in column (3), the ratio of ferrous metals to machinery fell drastically between I928 and I932; and for the whole period I92838, output of ferrous metals increased 5.4 times, while that of machinery increased I6.4 times. While a part of this divergence could be due to an increased degree of fabrication involved in production of machinery as compared with other metal-using industries, Gerschenkron felt that the ratio of I6.4 to 5.4 was too large to be explained by actual experience.4 Hence, he con-

JT was something of a shock to many economists to learn from Secretary of Agriculture Brannan's testimony before a Congressional committee (April I95I) that in I950 had earned on average only 69 cents per hour for their labor compared with a national minimum wage of 75 cents an hour and rates of one to two dollars or more in other occupations. The 69-cent average (now revised to 70 cents) seemed contrary to several common notions. It did not square with general belief that farmers are getting too much. It seemed out of line with many undeniable cases of real farm prosperity. Since it was an official figure, friendly critics argued not that it was inaccurate but that it was unduly weighted downward by including many submarginal who produce very little for markets. Others raised additional questions, particularly as to validity of comparing farm returns per hour with returns per hour in other occupations, and as to why people living on farms are willing to take less in dollar income. The problem of comparing farm with nonfarm has been a baffling one for many years; in fact, ever since annual estimates of farm were started more than 25 years ago. Ir I92 0'S, equality for agriculture, parity, and the farmers' share of national income were phrases in common use by farm leaders, agricultural politicians, and agricultural economists. The specialists in farm comparisons in 1920's recognized that were both entrepreneurs and laborers; that farming was both a business and a way of life; and, if comparisons with nonfarm were to be made, it was necessary to assign some value for farmer's labor to get residual return on his capital, or to assign some value for use of his capital and management in order to compute residual return for his labor. This, of course, is still an arithmetical necessity. There are other debated issues, such as how to impute values to production used in farm home and how to deal with differences between other elements in farm and nonfarm living standards. Agricultural statistics have been expanded and improved a great deal over past twentyfive years, and a great deal more is known about farm and national income, about per capita farm and nonfarm income, about size and distribution of farm and nonfarm income, and about reduced labor and increased capital requirements in agricultural production, but for most part basic questions in comparative are still unsolved.

A NUMBER of the conceptual premises upon which statistical science and practice have long been based have recently been attacked and discarded as harmful to statistical development in the Union of Soviet Socialist Republics. It seems incumbent upon statisticians in other countries to study the Soviet theses, to determine the extent to which differences are real and not merely semantic, and to take issue with those appear to undermine the structure of upon which their own work is founded. The present paper is intended as a preliminary analysis of this kind and its conclusions should be regarded as tentative. The Soviet attack, in the writer's opinion, can best be interpreted as a new phase of the ancient conflict between dogma and science. The dogmas upon which it leans have the essential characteristics of those in a revealed religion. Pretending to be scientific, these dogmas are actually anti-scientific in spirit and in the consequences which would follow their general acceptance. They are part of a new orthodoxy which seeks to impose revelation and arbitrary theological dictates upon the reason and the scientific judgments of men. Freedom of and expression are incompatible with theocracy and they are debarred from Soviet statistical doctrine. In order to support this interpretation I must describe the recent promulgation of Soviet statistical doctrine in some detail. This occurred during a two day Conference on Methodology at the Central Statistical Administration of the USSR in Moscow on February 20-2I, I950. A summary record has been published and appears in translation in the official journals of some other countries.' Soviet leaders have previously pronounced orthodox doctrines in art and literature, economics, biology, and in other scientific fields. I am among those who felt there was reason to hope tllat as an important tool for scientific analysis, statistics could escape such a doctrinal imposition. This hope has been unfounded. The conference was opened by the Chief of the Central Statistical Administration, Mr. V. N. Starovskiy. He explained that harmful bourgeois influences and anti-Marxist distortion in Soviet statistical science and literature hamper its development. He identified the main obstacle to the development of statistical science as the formal mathematics school of thought which considers statistics (to be) a universal science for the study of nature and society based ultimately on the mathematical law of large numbers and not on MarxistLeninist theory. A report on the correct theoretical basis of statistics was then given by Mr. V. A. Sobol of Mr. Starovskiy's staff. Statistics, said Mr. Sobol, is a social science, whose tasks and theoretical foundation are formulated in the works of Lenin and Stalin. Its tasks are to aid in the building of a communist society. Its theoretical foundation rests upon historical materialism and (communist) political economy. Incorrect views, which insidiously lean upon the theory of probability, have been expressed by such writers as Academician V. S. Nemchinov, who supported the chromosome theory of heredity at the 6 August I948 session of the All-Union Academy of Agricultural Sciences. Among others tainted with heresy were Mr. I. Yu. Pisarev (also of Mr. Starovskiy's staff) who was the author of the article on Statistics in the Great Soviet Encyclopedia. Following Sobol's report, according to the record. there was lively discussion. Most of *A paper contributed to the Twenty-Seventh Session of the International Statistical Institute and published here with permission of the Institute and the Indian National Committee for the International Statistical Conferences of 195'. See, e.g., Vestnik Statistiki, c. I, 1950 (publication of the Central Statistical Office of the Ukrainian S.S.R.); and Statisticky Zpravodaj, Rocnik XIII, I5 rijna 1950, Cislo 809, pp. 253-69 (Statistical Bulletin published by the State Statistical Office of Czechoslovakia).

CCELERATED amortization, as it has been used both during World War II and in the present emergency, is a device to encourage expansion of essential industrial capacity by offering to private enterprise incentives which will have a minimum direct effect on market prices. While accelerated amortization has been widely viewed as primarily affecting the accounting for profits for tax purposes, it has in fact been more than this. A certificate permitting accelerated amortization has entitled the holder not only to the privilege of more rapid amortization of facilities for tax purposes, but also in many cases to larger gross revenues through higher prices for its products. This has resulted from the fact that many prices in our economy are cost-determined,' and that accelerated amortization has been recognized as an element of cost for some pricing purposes. This is especially true in our war or defense economy. The extent of this price or revenue concession is indeterminate, partly because of uncertainties concerning the volume of products likely to be sold and partly because of the complexities of the process by which military goods and their components are priced.2 At the present time, there are further uncertainties because of failure to clarify current policies. For this and other reasons the value to the recipient of a certificate of necessity permitting accelerated amortization, and the cost to the government of the privileges granted, are both highly conjectural. It is increasingly apparent that the program of accelerated amortization was entered into after Korea without any clear appreciation of what privileges it was intended to grant or what minimum concessions were necessary to induce the requisite expansion. There is, moreover, serious question whether, as a long run device in a society of high level mobilization, accelerated amortization will be a favorable device for strengthening the competitive forces in the economy. It is not surprising, therefore, that the program has been the subject of much criticism.3

to encourage international commerce make it important to have such knowledge currently. The writer, in Novemnber i95o, requested information on their foreign holdings from more than a hundred American manufacturing corporations.2 Seventy-two companies supplied the data which are summarized in Table i. As of May I943, the Treasury study showed the foreign assets of American corporations and individuals with a 25 per cent interest or more in companies abroad, as amounting to $7,365,000,ooo. About 6,ooo corporations and indi-