Knowledge that Transforms

GENERAL business, after a nearly horizontal drift from June to July, resumed its downward course during the third quarter (Chart I, p. I65). Money rates, which had by June already canceled most of the violent advance incident to the I93I gold panic, declined in each month of the quarter. Speculation, following a very small drop during June, rose sharply in July and August and was only moderately reduced in September. VOLUME OF BUSINESS Adjusted data for bank debits in selected cities outside New York, our measure of the aggregate dollar volume of business, reached a new low level at the end of the quarter. The decline for July was negligible, but both August and September brought moderately large reductions. The check tax may have caused part of the third quarter drop in debits, but this effect is not measurable.

GENERAL business, after an important advance in April, declined during the remainder of the second quarter of I932, and stood in June at a new low level for the depression (Chart I, p. II3). Money rates, which had as early as February begun to ease from the peak following the I93I gold panic and bank liquidation crisis, declined in each month of the quarter. Speculation continued in its downward course during the quarter; and, largely in response to the Federal budget crisis and the renewed loss of gold, the declines in April and May were especially sharp.

T HOUGH generally accepted as indicative of business activity, figures on the production and use of electric power are seldom intensively utilized in analyses of the business s'ituation. The chief compilations are indeed faithfully reported, and as a rule the various composite indexes of business conditions make sure to include electric power as one constituent; but when it comes to current interpretation and analysis of business activity, not a great deal is made of the data. In part this comparative neglect may be ascribed to difficulties in the way of properly correcting the figures, among which the difficulty of making due allowance for normal growth is perhaps preeminent; but in part it seems to reflect uncertainty as to the precise significance and worth of the different figures available. In the present study it is proposed, therefore, to examine three outstanding series monthly production as reported by the United States Geological Survey, weekly production as reported by the National Electric Light Association, and monthly consumption as reported by Electrical Worldwith respect both to their statistical content and reliability and to their past performance as indicators of business or industrial activity.

THE weekly sensitive price index, to which there has been frequent reference in both the Weekly Letters and this REVIEW, was designed as an aid in forecasting intermediate fluctuations in business.' The series selected, therefore, were those which were found to be most consistent in anticipating such intermediate movements during the years upon which the choice of prices was based (I923, I924, and I925) when compared with indexes of wholesale commodity prices and of the physical volume of production of basic commodities. A list of the series is given in the table on page 44. The constituents of the index are mainly raw materials for use in manufacture, and their prices are therefore responsive to industrial prospects. Furtfhermore, several of them are byproducts; the supply of such goods is relatively inelastic, and the price consequently especially dependent upon demand. Finally, the commodities as a whole are subject to wider influences than their small number would suggest, since most are important in international trade, and many are used in numerous industries. The index is an unweighted geometric mean of actual prices, relative to the average for I926 as base.2 The considerations governing the choice of such an average are quite different from those that are important in the construction of an index to measure purchasing power or to be used in studying the quantity theory of money. In the present case, the purpose directs attention to the timing and direction of changes, not to the general level of the index. The type of average follows from this fact and from the several assumptions that follow concerning the relationship between commercial fluctuations and prices. These assumptions are: first, that any given change in the conditions of demand and supply for any one of the commodities has an influence that is relative to the price of the commodity at the moment; and, second, that the various commodities are of equal value for the purpose of the index. The index has been computed weekly, and is designated as applying to the week ending Wednesday; the prices included are those datedwithin the week either prices for a given day or averages of daily prices for a seven-day period. The index for any month is obtained by averaging the weekly indexes for those weeks whose major part falls in the month in question. The computation of the weekly index is, of course, performed with the aid of logarithms; the logarithm of the base is subtracted from the average of the logarithms of the actual prices to obtain the logarithm of the index.3 The number of commodities used in the index is relatively small, so that there is at all times possibility of distortion resulting from extraordinary fluctuations in one of the constituents. For example, the exceptionally wide movements in the price of rubber (arising from causes peculiar to this particular commodity) have at times had considerable undesirable effect on the index. But it is probable that such temporary disturbances have at least during the period covered merely necessitated caution in the interpretation of the index, and not impaired its usefulness. The period for which the sensitive index is now available (I92I-3I) is sufficiently long to make fruitful some review of the success or failure of the index in accomplishing the purpose for which it was designed.4 From I923 through 1925 (the test period) its performance was, as is

BETWEEN I 9 I 9 and I92I the REVIEW OF ECONOMIC STATISTICS published several studies of the balance of payments of the United States, which pointed to the conclusion that the War had made us a nation. We assumed that this term, which was well established in economic literature, had a definite and generally accepted meaning; and we did not undertake to define it. In recent years it has become clear that we should have defined the term with care, and that the United States has not become a in the sense that we intended at that time. Whatever their origin, and this is not important for present purposes, the terms and became well established in economic literature by the end of the nineteenth century. England was a nation. From interest on foreign investments, earnings of ships, and various other items in her annual accounts she had a large balance in her favor, which enabled her to import commodities greatly in excess of her exports and so to pay for a large unfavorable balance in such accounts. Upon the other hand, the United States, with a large balance against her on account of the invisible items, exported commodities greatly in excess of her commodity imports, and thereby paid her debts. Clearly the term had reference to a country's position on account of the invisible items in her balance of payments, and the visible proof of that position was the existence of an excess of commodity imports. If usage had been guided wholly by the idea of an annual reckoning of a country's international accounts, the terms might as well have been applied differently; and a country with a sufficient excess of commodity exports might just as well have been called a country. But interest on foreign investments bulked so large among the invisible items that a country, like England, that had long exported capital was in the old and familiar sense the of the countries to which her capital had flowed. Therefore it was natural that the country with a balance of invisible items should be called the nation, and the country showing a surplus of commodity exports should be termed the debtor. All this, of course, related to transactions on income account, and disregarded capital transactions which in any year might temporarily make a country a debtor if it exported an unusual amount of capital. Thus the richest country might be temporarily in debt, and export gold to balance her total accounts in any year; while at the same time, and for just the opposite reason, the poorest debtor country might temporarily become a and receive gold imports. Those who follow the exchanges from month to month, or try to determine a country's position at the end of every year, must of course consider capital as well as income transactions. But a country's normal position as debtor or is not determined by the capital transactions of any year, but by those on income account which are the ultimate determinants of a country's international position. If, therefore, economists are to use the terms and country, they should disregard capital transactions and consider only a country's normal position on income account. According to the older usage, therefore, a country was one which in any normal year had a balance in its favor on account of the invisible items entering into its foreign transactions on income account. Visible items might have been considered instead of invisible; but they were not, and this was probably because the countries having favorable balances of the invisible items were also the countries that made foreign loans or investments and therefore were creditors in the old-fashioned and legal sense of the word. This we believe to be a correct account of the term creditor nation as it was used prior to the War. Obviously a may be a debtor on account of the invisible items but so largely a on account of the visible items (commodity exports) that its total income account for a period of years shows a balance in its favor. Should such a be classified as a debtor? Obviously not, unless the term is going to be used in a technical sense not conducive to scientific insight and con-

I. THE STATISTICAL RECORD IT is the purpose of this article to present the quantitative evidence relative to total investment in new securities during the past twentysix years, to point out some of the important changes that have occurred, and to attempt an explanation of these changes. So far as new flotations cover the field, they provide a rough indicator of the volume of industrial investment. Attention will be focused on the broad changes in security issues as a whole, especially on the contrast furnished by the pre-war and the post-war decades. Records of corporate issues go back to I907. For the years through I9I8, the monthly listing provided in the Journal of Commerce has been used, and for the years since I9I9, the data presented each month in the Commercial and Financial Chronicle. The Chronicle series covers these issues rather thoroughly, and probably represents as complete an enumeration of publicly floated issues as is possible.' As for the Joutrnal of Commerce figures, the editor does not lay claim to complete statistical accuracy in their collection, but regards them as a large sample, prepared upon a basis which was as nearly comparable as possible. Neither series gives a complete picture of the facts; but both provide a satisfactory approximation thereto. The published figures cover flotations of securities issued by corporations, both foreign and domestic, and sold on the American investment market. Excluding refunding issues they thus cover, with minor exceptions, the capital raised by industry through the security market. Security issues of investment trusts, trading and holding companies are shown separately, as such issues represent duplication rather than any net increase in capital secured. That this increment of capital is not the same thing as total new investment must obviously be admitted, for it fails to include investment in government issues, private investment in partnerships, individual enterprises, houses and other durable consumers' goods, and that large element of capital growth Drovided through corDorate saving out of Drofits. There is also at least one element of overstatement involved, in that the figures include a small proportion of issues used to acquire plant already in existence, as well as the amount of underwriters' commissions. In spite of these flaws, however, it is believed that the series furnishes a reasonably serviceable index of investment in industry. The quarterly figures are presented in Chart i (p. I97) and Table 2 (p. I99).

THE MONTHLY RECORD NTIL November I930, the banks seemed to be weathering depression quite well. Though suspensions had increased somewhat in I929 even before the downturn of business and speculation, and thereafter increased more, yet as late as October I930 the record did not look bad against such earlier periods as the winter of I926 or I923-24; and the fact that suspensions during the spring and summer of I930 were exceptionally numerous for that time of year