Knowledge that Transforms

In an important recent article,' Modigliani has renewed his analysis of the relationship between the real and monetary parts of the Keynesian system. One of Modigliani's major points in his latest article is that his famous 1944 model 2 embodied an error, for, in his opinion, it incorrectly specified the forms of the real aggregate consumption and investment demand functions.3 While we find much to agree with in this new model, we do not agree with him as to the fundamental source of error in the 1944 model. Consequently, we are in disagreement with some important aspects of the latest formulation and, as a result, with some of his more important conclusions. basic flaw in Modigliani's 1944 model is not the fact that the aggregate real spending functions were not homogeneous of degree zero with respect to prices.4 A quite different source of error was involved, and this error has been partly perpetuated in the 1963 model. While Modigliani's 1944 system contains a direct relationship between the price level and the money-wage rate,5 this relationship is omitted entirely from the monetary subset of equations. It is this relationship, however, which provides an essential link between the monetary and real sectors. In 1944, Modigliani claimed that the four equations which defined . . the system of monetary equilibrium were completely independent of the other equations in the system and formed a determinate subset.6 Consequently, Modigliani concluded that Since the money income is determined exclusively by the monetary part of the system, the price level depends only on the amount of and therefore a change in the money-wage rate could affect the price level only if it resulted in a change in the level of output.7 This surprising result is not only in conflict with Keynes' analysis of the effects of money wage changes,8 but it is also incompatible with Modigliani's own conclusion in a later section of that same 1944 paper. Towards the end of his 1944 article, Modigliani argued that a reduction in the moneywage rate will . . depress the interest rate, the money income, and money wages without affecting the real variables of the system, employment, output, real wage rate. 9 But how then can a change in the money-wage rate affect the rate of interest? If the money-wage rate change does not affect real output, then, by Modigliani's earlier assertion, it cannot affect the price level. If both the price level and the level of real output are unaffected, there should be no change in the transactions demand for cash balances and consequently no change in the rate of interest! Of course, a change in the moneywage rate does affect prices and output and, therefore, the rate of interest. Accordingly, Modigliani's belief that the monetary subset is independent of the money-wage rate must be in error. In his latest formulation, Modigliani explicitly introduces the price level variable into his demand for money equation in order to demonstrate that, given nominal money balances, changes in the money-wage rate will shift the entire money market function relating the rate of interest and the level of output.10 This implies that contrary to the 1944 model, at any given level of output, a decrease in the money-wage rate involves a reduction in prices and money incomes and therefore a decrease in the demand for transactions balances. Modigliani can now demonstrate that the moneywage rate has an impact on the demand for money equation, but, in his system, a change in the money* authors are associate professor of economics at the University of Pennsylvania and Lilly Faculty Fellow, University of Chicago, respectively. 'F. Modigliani, The Monetary Mechanism and Its Interaction With Real Phenomena, Review of Economics and Statistics, xxxxv (Feb., 1963), 79-101. 2 F. Modigliani, Liquidity Preference and the Theory of Interest and Econometrica, xiI (Jan., 1944), 4588; reprinted in Readings in Monetary Theory (New York: Blakiston, 1951), 186-239. All references are to the Readings in Monetary Theory reprint. 'F. Modigliani, The Monetary Mechanism and Its Interaction With Real Phenomena, 82. 'In fact, we shall argue that the real aggregate consumption function is incorrectly specified in the 1963 model. 'F. Modigliani, Liquidity Preference and the Theory of Interest and 188, equation 7. Ibid., 190. 7Ibid., 210. 'J. M. Keynes, General Theory of Employment, Interest, and Money, (New York: Harcourt Brace, 1936), Chap. 19. 'F. Modigliani, Liquidity Preference and the Theory of Interest and 220. 0F. Modigliani, The Monetary Mechanism and Its Interaction with Real Phenomena, 89-90.

T HREE years ago, Seymour Harris wrote that we must solve the problem as a condition for appropriate policies for growth, employment, aid, and defense; and several contributors to his book The Dollar in Crisis called for devaluation of the as the proper and indeed the only appropriate solution to the dollar problem. But economic growth in the postwar period has been universally high by historical standards. It has been especially high in the industrial countries of Continental Europe and Japan, both in comparison with the United States and in comparison with their own past. Real GNP in the six members of the European Economic Community, for example, rose 70 per cent between 1950 and 1960 -an average annual increase of no less than 512 per cent. Japan's growth was even higher. And even postwar United States growth was substantially higher than historical trends until the late fifties. In a year in which the international payments system is under close collective scrutiny by the major industrial countries, it is worth asking what role, if any, international monetary arrangements played in stimulating this growth. Dollar deficits may in fact have made a substantial contribution, and devaluation of the dollar, far from furthering free world growth, could well slow it significantly. Undoubtedly, no single factor can claim credit for postwar growth. The war-induced disturbance to old patterns of behavior and disillusionment with the prewar, static, competitive view of the world clearly fostered receptivity to rapid change in Europe and Japan. Psychological and economic momentum, gathered during the quick restoration of output to prewar levels, contributed to the sense that rapid growth was both possible and desirable. To these intangible factors were added powerful government incentives to business investment and extensive public investment. But surelv one imDortant element in the rapid postwar expansion was the presence of high export demand for European and Japanese goodsan export demand so high that additional output could always be sold profitably as soon as it became available. High demand was due in part to the continuing growth in total world demand. But it was also due to the possibility of substituting European and Japanese manufactured goods for American goods which had sometimes been the only goods available right after the war -in home markets, in third country markets, and even in the United States market. This substitution (in the context of growing markets) proceeded on a grand scale. The share of world exports of manufactures supplied by the six members of the European Economic Community, for example, rose steadily from 33 per cent in 1951 to 46 per cent in 1961 (Japan's rise was from 4 to 7 per cent), while the United States share fell sharply during the same period. United States imports of finished manufactures rose 250 per cent from 1950 to 1960, compared with a rise in United States industrial production of only 45 per cent. European and Japanese goods could be sold easily abroad because the currency devaluations of 1949 and other postwar currency changes made them cheap relative to goods, or very profitable to export at going market prices. Crude comparisons suggest the goods and services of Western Europe had declined in price by 24 per cent relative to the prices of American goods and services between 1938 and 1953.1 In 1950, the total purchasing power of the at official exchange rates was over 30 per cent higher in many European countries than it was in the United States.2 As might be expected, the of the in Europe was considerably greater for local services than for goods which move in international trade. Some products, notably produc-

Evsey D. Domar, Scott M. Eddie, Bruce H. Herrick, Paul M. Hohenberg, Michael D. Intriligator, Ichizo Miyamoto, Economic Growth and Productivity in the United States, Canada, United Kingdom, Germany and Japan in the Post-War Period, The Review of Economics and Statistics, Vol. 46, No. 1 (Feb., 1964), pp. 33-40

good deal of emphasis has also been placed, however, on the claimed benefits of such a shift to the United States balance-of-payments position. For example, the CED statement goes on to say that, A major advantage of a general excise tax is that it would tend to improve the ability of the United States to compete with others in world markets. This view assumes that the direct tax that is cut is the corporate profits tax and that a cut in it reduces prices of corporate output. The present paper calls to attention some influences that are frequently neglected in appraising the effects of such a shift on the balance of payments when these assumptions are valid. It then considers the effects of a cut in corporate taxes that does not reduce prices. Finally, it considers the balance-of-payments effects of making the cut in direct taxes in the individual income tax.

ECONOMISTS have been interested in the accuracy of anticipations for two reasons. First, effective countercyclical policy depends on accurate forecasts of future investment, inventories, sales, etc. Second, the degree of accuracy of anticipations has implications about the way anticipations are formed. This paper deals with the accuracy of sales anticipations. The source of the data on anticipations is the Annual Survey of Business Anticipations of Sales collected by the Securities and Exchange Commission and the Department of Commerce. The only systematic study of these data has been conducted by Modigliani and Weingartner.2 In the first section, we test whether sales anticipations are more accurate than the forecasts of several naive models and whether anticipations correctly predict the direction of change of sales. Next, we test Theil's hypothesis that predictions refer to a shorter period than they are supposed to. In the last section of the paper, an accuracy measure proposed by Theil is adopted to determine (1) whether an accurate forecaster is less likely to commit systematic forecasting errors and (2) whether anticipations conform to the rational expectations hypothesis. Nature of the Data

T HE determination of interest rates in rural areas throughout the underdeveloped world is best explained in micro-economic terms. The typical village moneylender will either be an outright monopolist, or he will be an imperfect competitor.1 The market for loans will center around the village itself. The farmer will normally only borrow from the one or more moneylenders that the village can support. He will not often have access to a bank or other lending institution. In these circumstances, the moneylender will face a demand curve for his loans which will slope downwards from left to right. The rate of interest will be on the vertical axis and the volume of loans on the horizontal axis.2 He will also have a schedule of costs for lending. This will be compounded of the administration and risk charges on each unit which he lends, together with the opportunity cost of his raw material money. It is this last cost component which corresponds to the pure rate of interest of existing theory. The average cost-of-lending curve will describe the arc familiar to the student of the principles of economics. There will be a certain volume of loans which will maximize the moneylender's net returns. This volume will be at his equilibrium level of lending, and it will determine his most profitable interest charge. It is the opportunity cost of each dollar or rupee which he advances at this equilibrium point which we will analyze here. Questions of average administration and risk charges, as well as of monopoly profit, must be left to other discussions. But it should not be supposed that these are relatively unimportant considerations. Risk and administration costs in particular probably play the major role in forcing high interest rates upon farmers in poor countries.3 If we view the opportunity cost of the lender's money as one of the determinants of his costs, then we could draw a curve representing these charges. It would probably run parallel to the volume of lending axis to begin with and then rise quite sharply as the moneylender adds to his loans. The unit opportunity cost is thus registered on the vertical axis and it forms one of the components of the interest rate which the farmer must ultimately pay. The reasons why the opportunity cost of the money used in a lender's loans will describe such a curve can best be explained under two separate headings. They are: (1) the returns on alternative investments, and (2) liquidity preference.

Introduction JN spite of the voluminous studies that have been made on the theory of the consumption function, one important question remains unanswered. Does the Keynesian theorem of consumer behavior operate in any modern community as Keynes claimed it would? ' The first section of this paper will be devoted to testing the Keynesian hypotheses: (1) the level of current income is the main determinant of the level of current consumption in the short run, and (2) the marginal propensity to consume is less than unity. In the second section, we shall examine factors affecting the differences in aggregate consumption ratios of various nations. While much effort has been spent on study of the aggregate consumption function of the United States, our knowledge of the consumption patterns of countries in the rest of the free world, particularly of less advanced countries, continues to lag.2 The relative scarcity of research in this area has been due primarily to the absence of reliable data. Until the introduction of a uniform national account system by the United Nations in 1947, national income data were virtually non-existent except for the highly developed countries.3 As reports of the member nations have been published for a number of years, sufficient data are now available to calculate and compare the aggregate consumption functions of various nations. The following criteria were used in selecting countries for this study:

CONSIDERATIONS of the impact of changes in the volume and composition of expenditures by the defense and space agencies will be misleading if they ignore the regional component. In considering disarmament, traditional monetary and fiscal policy responses produce an effect which is nationwide in scope. Such policies may not be of much help to states and communities, such as California or Wichita, whose economies are heavily dependent upon defense expenditures. In addition, shifts in the regional pattern of these expenditures can produce similar stresses in local economies. A major problem in this connection has been the measurement of the defense-space expenditure impact in a region. Aside from an induced impact operating through regional consumption and business investment functions, the impact on income and employment can be divided into two components: (1) the direct impact through prime contract awards to firms, and (2) the indirect or inter-industry impact through subcontracts and purchases of supplies by prime contractors. While there have been some attempts to measure the direct impact, little work has been done which also accounts for the indirect impact. This paper reports on an effort to measure both the direct and indirect impact of defense-space expenditures on the manufacturing sector of the Los Angeles-Long Beach Standard Metropolitan Statistical Area (SMSA). The task is an empirical one. Hence, we shall briefly review some of the various techniques of measurement and present the results of a short-cut method to measure the impact on Los Angeles manufacturers. Empirical Difficulties Almost all approaches to the measurement of the regional impact of defense-space expenditure involve variations of an input-output framework.1 Unfortunately, given the present state of data availability, they are not operational, at least without extended research effort. National data from the 1947 table, while useful, are somewhat out of date.2 Regional data are all but nonexistent. Pending the development of more adequate data, some short cuts need to be examined. In the search for short cuts, it is useful to keep in mind what the gross flows data of an interregional input-output table reveal. Row information reflects where sales are made in terms of industries, final demand sectors, and regions. Column information indicates the source of inputs from other industries both inside and outside of the region. Short cuts, essentially, involve something less than the complete cross-check of independent estimates of the row and column entries. Most regional input-output studies, in fact, get these estimates sometimes from row information and sometimes from column information, but rarely from independent estimates of both. A column-oriented approach, which has a good deal of appeal, simply traces down the subcontractors. There is some evidence to suggest, as an order of magnitude, that half of a specific defense or space program prime contract is subcontracted.3 It would seem that tracing down a few layers of subcontractors would account for most of the impact. Unfortunately, this is not the case. When a prime