The Literature on the Sales Tax Get access K. M. Williamson K. M. Williamson Wesleyan University Search for other works by this author on: Oxford Academic Google Scholar The Quarterly Journal of Economics, Volume 35, Issue 4, August 1921, Pages 618–633, https://doi.org/10.2307/1882429 Published: 01 August 1921

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

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

T HE census of manufactures of the United States classifies manufacturing industries into fourteen groups.' Of these groups and steel and their leads all others, whether the groups be ranked according to capital investment, amount of wages paid, or value added to materials by the manufacturing process.2 Moreover, the industrial importance of the iron and steel group is not fully revealed by the figures for that group considered by themselves. Among the other thirteen groups there are two, vehicles for land transportation and railroad repair shops, which are so intimately connected with iron and steel that the three might well be classified together, while the remaining eleven manufacturing groups are all dependent upon the iron and steel industry for tools, machinery and structural steel. Furthermore, the non-manufacturing activities, agriculture, mining, building, and transportation, are large consumers of iron and steel products. The manufacture of iron and steel, therefore, is not only important because of its magnitude but because it is more intimately related to every phase of our industrial life than is any other branch of manufacturing activity. Because of its importance and its unique interlocking with every branch of the modern industrial organization, the manufacture of iron and steel has been generally recognized to reflect the business situation as a whole. Direct evidence that this industry actually does offer good indices of business conditions may be found in various data published by the Harvard University Committee on Economic Research. We may remind the reader that in our original study, Indices of General Business Conditions, we found the monthly volume of production for I903-I4 to fluctuate concurrently with bank clearings outside New York City and general commodity prices. Consequently, production was one of the series selected for our curve B representing general business conditions.3 Comparison of the annual volume of production with indices of the output of manufacture as a whole led Professor Day to the conclusion that pig-iron production appears to give an amazingly accurate picture of the year-to-year fluctuation of physical production in manufacture. 4 Prices of iron and steel, as well as production, reflect the general business situation. The prices of pig iron and bar iron are included in our commodity price index of business cycles.5 Finally, the price indices of iron and steel stocks fluctuate in a manner very similar to that of the indices of other industrial stocks.6 It is the object of this article to bring together the significant material showing the fluctuations of the iron and steel industry during periods of business prosperity and depression. Prices. Chart I shows the monthly prices, I898I92I, in dollars per gross ton, of four iron and steel products in various stages of manufacture, namely, Bessemer pig iron, Bessemer billets, steel bars, and black steel sheets.7 The chart is constructed on the logarithmic scale, so that the equal vertical distances represent equal percentage changes. The important thing brought out by the chart is the manner in which fluctuations in the prices of pig iron are reflected in the prices of the semi-manufactured and manufactured products. During certain periods like I902-03, I90507 and I9I0, however, the prices of bars and sheets reveal a high degree of stability while pig iron and billets fluctuate considerably. That fluctuations in the prices of iron and steel products during periods of business prosperity and depression are representative of fluctuations in prices of other important commodities is shown by Chart II. The curves on this chart represent the prices (B) of pig iron and (C) bar iron compared with (A) our tencommodity price index of business cycles, based upon the prices of mess pork, cottonseed oil, hides, print cloths, sheetings, spelter, worsted yarns, and coke as well as pig iron and bar iron. This chart shows that the prices of pig iron and bar iron have similar cyclical movements to those of the ten-commodity index. The thr e curves have been adjusted for their respective long-time movements 8during the period of comparison, i898-i914, in order that the different long-time movem nts of the various price series might not interfere

Journal Article Beer's History of British Socialism Get access Herbert Feis Herbert Feis Harvard University Search for other works by this author on: Oxford Academic Google Scholar The Quarterly Journal of Economics, Volume 35, Issue 4, August 1921, Pages 596–602, https://doi.org/10.2307/1882427 Published: 01 August 1921

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.