Knowledge that Transforms

T HE curves of our Index Chart for I903-I4 were obtained by sorting twelve statistical series into three groups of four each, and then averaging the series of each group. The criterion for sorting the series was statistical; that is, when corrected for seasonal influence and normal growth the series of each group moved in the same general direction (upward or downward) at the same time. The three groups, however, did not undulate simultaneously but moved in a well defined sequence. After the statistical analysis of the series had been completed, it was noticed that the series of each of the three groups possessed economic similarity as well as

I NDICES of the cost of constructing a standard reinforced concrete factory building, supplied by the Aberthaw Construction Company of Boston to the REVIEW, are shown in the accompanying table. The basic figure is the actual cost of construction of a sevenstory reinforced concrete factory building erected by the Aberthaw Company in I9I4. Estimates have been made of the cost of putting up the same building if construction were begun on January I, I920, July I, I920, and on the first of the month from January to October I92I. The index number for any given date is secured by expressing the estimated cost of construction, beginning on that date, as a percentage of the actual cost in I914. Although the indices are based on the estimated costs of constructing a selected type of building they show the fluctuation in costs of reinforced concrete construction in general. The building chosen as a base was actually constructed between August 22 and December 31 I9I4; building costs were so uniform during the year that it is accurate to consider the base to be the entire year. This b,uilding was chosen for two reasons. The building itself is typical of the reinforced concrete structures erected by the company, as is also the proportion of costs involved for the various classes of labor and various kinds of materials. In the second place, its construction was an almost ideally handled job ; that is, a job which involved no waste and no unusual costs. Consequently, actual and estimated costs correspond closely, both for the job as a whole, and for its component parts.

T HE year following the armistice was devoted mainly to readjusting productive industry from a war to a peace basis. This required many changes in the direction of production and the return of several million soldiers to the pursuits of civil life. The process was not easy; but, after an initial period of depression, it was accomplished without serious disturbance, and a year ago there was every reason to be satisfied with the achievement. Normal conditions had by no means been restored, and it was clear that years must elapse before the world could liquidate all the accounts left by the great war. But it was a great gain and the first indispensable step in the return to normal life that the army had been demobilized and production readjusted from a war to a peace basis. The work accomplished in i919 had been greatly facilitated by several developments which were bound to occasion difficulty in I920. The export trade of the country, which had greatly increased during the war, instead of declining proceeded to rise to new high levels, thereby continuing the stimulus which American business had been receiving ever since I9I5. Added to this was the stimulus to active trade caused by the extravagant expenditures, both public and private, which had developed to a considerable degree during the war and acquired increased impetus with the return of peace. To finance foreign and domestic business and provide for the continuing abnormal requirements of the government, credit expansion had continued during i919, and prices were mounting to a higher level than had been reached during the war. It was clear that the United States, like the rest of the world, was caught in the toils of a new inflationary movement, which would ultimately lead to a reaction, the time and severity of which were impossible to estimate. To the year I920, therefore, the year i919 bequeathed a high and unstable price level, an unbalanced foreign trade situation, and growing money strain, which were bound to prove serious liabilities. Although readjustment of discount rates had been inaugurated by the federal reserve banks in November, i919 and liquidation in security markets had already set in, the year I920 opened with very active trade and increasing commodity prices. The early months were marked by great apparent prosperity, and many new records were scored in various lines of production and trade. But the prosperity was largely artificial and was approaching its end. Our foreign trade was unbalanced, and the burden of financing it was falling more and more upon the banks -a situation which could not continue indefinitely. Money rates were rising, and there was every evidence of increasing credit stringency. Liquidation in security markets continued apace, but was inadequate, as it always had been, to relieve the situation; all the conditions spelled liquidation in commodity markets as forecasted in December, i919 by our Index of Business Conditions. But, as always, only the discerning few realized what was impending; and even when liquidation finally began, many failed to realize its import. In the late spring the tide turned, and by summer liquidation of commodity markets was fairly under way. As it continued, the movement gathered force, and finally proceeded at an unprecedented rate. Although the REVIEW OF ECONOMIC STATISTICS had for months forecasted a reversal of business conditions, the editors had not expected a reaction of such acute severity. We had looked for a return to some such level as had prevailed in the few months following the armistice, and as late as July expected nothing so drastic as the events of the last half year. There were various reasons, not worth recounting now, which had seemed to justify the belief that in most lines of industry readjustment would come more gradually. But, as so often happens, when the turn came one untoward development followed another with the result that reaction was swift, intense, and world-wide. Surveying the year in retrospect, which is so much simpler than prospect, it is easy to trace the general course of events. By the spring of I920 consumers began to react somewhat violently against the continued increase of prices which had reached inordinate levels in many lines. It now appears that at the peak of wholesale prices retailers were, in many cases, unable to pass the entire load along to their customers. Some of them, foreseeing this, had already begun to buy very conservatively; and others, learning by experience, pursued thenceforth a conservative buying policy. The revolt of the consumer was natural and even inevitable, but it was intensified by the fact that he had been looking forward to relief from the burden of war prices and, therefore, resented more keenly the great increase which occurred between the spring of i919 and the spring of I920. Since purchasing power was not yet impaired, retail trade continued upon a fairly high level, but buying had become conservative; and this reacted upon the wholesaler, the manufacturer, and the producer of materials. Just when these developments were becoming serious the country ran into critical difficulties in transportation. The movement of freight was slowed down greatly during the early summer, and this retarded the circulation of goods and prevented liquidation of credits. In the agricultural sections a considerable part of the crops

Edmund E. Day, An Index of the Physical Volume of Production: IV. Agriculture, Mining, and Manufacturing Combined, 1899-1919, The Review of Economics and Statistics, Vol. 3, No. 1 (Jan., 1921), pp. 19-22

PROFESSOR IRVING FISHER, in a paper read at last annual meeting of American Statistical Association,' proposed a formula for computation of index numbers of prices and of quantities which, he maintained, was the best for all purposes. 2 The formula which he recommended for indices of prices and quantities as ideal is mean between two ratios of aggregates. The aggregates depend upon a combination of prices and quantities of commodities in a given year with those in a base year.3 The data necessary for computation of Fisher's formula for a series of years consist of prices and quantities in every year of all commodities selected for index. For an exact comparison of average prices or quantities in two years, by Fisher's method, it is not sufficient to compute indices for those two years relative to a third year as a fixed base; it is necessary to use one of two years compared as base. Extensive data and laborious computations are required for construction of Fisher's indices. Professor Fisher has designated several types of averages as biased and hence unreliable. One of averages he condemns, because it will have a downward bias, is geometric with weights dependent on prices at or near base year. 4 Since this type of average was used by Professor E. E. Day in computing

John H. Williams, The Balance of International Payments of the United States for the year 1920: With a Statement of the Aggregate Balance, July I, 1914-December 31, 1920, The Review of Economics and Statistics, Vol. 3, Supplement 1 (Jun., 1921), pp. 169-212

T HE object of this study is to construct an index of wholesale commodity prices for a special purpose. That purpose is to measure changes in general business conditions during alternating periods of prosperity and depression. In other words, our problem is to select and combine seriesof wholesale prices of commodities in order to secure an index of business cycles. This is not the problem which, heretofore, makers of commodity price index numbers have had before them; their problem has been, rather, to measure changes in the cost of living or in the general exchange value of money. The price index of business cycles, which we present here, is a new type of index number constructed by methods adapted to the special object in view.