In the following tables, the commodities which go to make up the Bureau of Labor Statistics index of wholesale prices are classified according to various characteristics; are grouped on the basis of frequency and magnitude of price change; and are presented, group by group, in the form of indexes, running on an annual basis from 1913 to 1937.

In table I various characteristics and classifications of the 784 individual items are listed. In the first column the code numbers are indicated, by which the prices are referred to in various Bureau of Labor Statistics publications. In a number of cases where several quotations were obtained by the Bureau of Labor Statistics for the same or very similar commodities, it was found desirable to combine such quotations into one index in order to avoid weighting such items too heavily in the unweighted averages used in computing various indexes. Where this was done the new composite price is denoted by the lowest code number among the prices entering into the composite with the letter c affixed, thus: 2c, 6c. The individual items which were combined to make up such composites may be identified by referring to the notes after column 16.

A brief description of each commodity price is given in column 2. More complete specifications of the commodities are available in the mimeographed pamphlet Wholesale Prices-Specifications of Commodities Entering Into the Composite Weighted Index, Bureau of Labor Statistics, Division of Wholesale Prices, September 1935.

In column 3 is given the period for which fairly complete data are available. In some cases scattered data are available outside the period given, but are too scanty for ready use in further analysis. On the other hand, the absence of a month or two from the series has usually not been considered sufficient to warrant rejecting the data; the missing month has usually been supplied by interpolation.

In columns 4, 5, and 6, the commodities priced are classified according to the degree of fabrication to which they have been subjected. In column 4 the classification into raw materials and manufactured goods follows that published by the National Bureau of Economic Research in appendix II of Economic Tendencies in the United States by Frederick C. Mills. Not all the prices are so classified by Mills, and in a few cases the precise quotation to which the classification made in Economic Tendencies refers is not quite certain. Column 5 gives the classification into raw, semifinished, and finished goods according to the clas

1 Appendix 2 was prepared by William S. Vickrey.

sification given by the Bureau of Labor Statistics, Division of Wholesale Prices, in Wholesale Prices for December 1937. Column 6 gives a somewhat more detailed classification prepared by the staff of the National Resources Committee. Commodities which pass through two or more stages of manufacture and so may at some time become a semifinished good are classified into raw, semifinished, and finished, designated R, S, and F, respectively. There are, however, some commodities which undergo no processing at all and are consumed substantially in the raw state. Such commodities are thus at the same time raw materials and finished goods, and are designated RF. There are also some commodities which undergo only one stage of manufacture and so never pass through the stage of being semifinished goods, at least in the open market. The raw material stage of such goods is designated RS, and the finished state SF.

Columns 7, 8, and 9 classify the items according to their durability. In general the criterion is durability under continuous use-i. e., normally rendering successive services-rather than mere absence of perishability. In column 7 the classification is that given to the corresponding Census industries (see column 16) in table 3 of Employment and Pay Rolls of the Bureau of Labor Statistics (e. g., the issue for October 1937). On this basis, the classification is somewhat indirect and occasionally yields anomalous results, as when matches are classed as durable. In column 8 the classification is taken from appendix I of Economic Tendencies in which various commodities are classified for use in production indexes rather than price indexes and the relation to the price indexes is somewhat indirect. In column 9 the classification by the staff of the National Resources Committee is given.

Column 10 classifies the items according to use, i. e., goods entering into capital equipment, producer's goods entering human consumption, and consumer goods. In column 11 the items are classified as food, clothing, and "other." Columns 12 and 13 classify the items as to source-in column 12 according to whether the source of the item is farm or nonfarm, and in column 13 according to whether the source is crop, animal, forest, or mineral. These classifications are taken from the National Bureau of Economic Research, except for the items classified as clothing and "other" which are by the staff of the National Resources Committee.

In column 14 the commodities are classified by the staff of the National Resources Committee according to source of activity, i. e., according to the industry that contributed the greatest value for that stage of

activity. Four classifications are given, i. e., agricultural, forestry and fishing, mining, and manufacturing. For the manufactured goods the classification was made according to value added, i. e., when valued added was 30 percent or greater the commodity was classed in manufacturing, otherwise it was classified in one of the other groups.

In column 15 commodities are classified according to the degree to which the product of a manufacturer is subject to the direct competition of similar products. Where the same commodity with substantially identical specifications is obtainable from several producers, the article is classified as a standard commodity, designated S; where there are differences between the product of one producer and that of another producing substantially the same commodity, or where the producer by advertising, packaging, use of trade-marks, or other means, has succeeded in creating the illusion in the mind of the public that such differences exist, the commodity is classified as differentiated, designated D; and where differentiation of the product is carried to the extent that the products of different producers are no longer directly comparable as to price without taking into consideration possible differences in the tastes of the consumer, the commodity is classed as unique, designated U.

In column 16 the Census classification number of the industry in which the commodity is produced is given. In the case of raw materials, the source is given, i. e., agriculture, mining, or import. An item is classed as an import if the preponderance of the supply is imported regardless of whether a considerable proportion of the supply may be produced domestically. Complete descriptions of the industries to which the industry code numbers refer are given in Industry Classifications for the Census of Manufactures, 1935 and 1937, Bureau of Census. The 1935 numbers were used for all industries but textiles, for which the 1937 numbers were used.

In column 17 is given the number of changes in price occurring during the period from January 1926 to December 1933, there being a total of 96 monthly quotations and therefore a possible maximum of 95 changes. In the case of composites, the total number of changes in all of the prices entering into the composite was divided by the number of components and the result evened off to the nearest integer. Prices were then classified into 10 frequency groups according to the frequency of change, the group intervals being adjusted so that the prices are distributed as evenly as possible among the ten groups. Prices changing least frequently were put in group I and those changing most frequently in group X. Group X was then further subdivided into Xa and Xb by putting into group Xb those prices which had shown a change every month, or a total of 95 changes, leaving in group Xa those

prices having shown 93 or 94 changes. These group numbers are listed in column 18.

2

In column 19 is given an index of depression sensitivity for each of the 617 items included in the frequency groups. The index is calculated by subtracting the 1932 price index from the arithmetic average of the 1929 and 1937 price indexes (all prices expressed as a percent relative of the 1929 price). This measure estimates in part the influence of the nondepression factors such as were at work between 1926 and 1929. Such a measure is superior to the drop in price from 1929-32 or a rise in price from 1932-37, because any change in the price not attributable to the depression is partly averaged out. On this basis, prices are grouped into 10 groups from the least to the most sensitive. The depression sensitivity group numbers are given in column 20 for each of the 617 items.

In columns 21, 22, 23 and 24 are shown the price indexes for each of the items used in the sensitivity groups for the years 1926, 1929, 1932, and 1937, in each case the base being 1929=100.

In table II the prices are listed by frequency groups. This table also shows the range of frequency corresponding to each group and gives the depression sensitivity group number for each price.

In table III the prices are listed according to depression sensitivity groups. The range of sensitivity for each group is given, and within each group the prices are given in order of increasing sensitivity.

Table IV gives annual indexes computed from these groupings. The first 10 lines are frequency group indexes, which are simple unweighted geometric averages of the items in each frequency group, shifted from a 1926=100 to a 1926-29=100 base. Similar indexes are also given for the 10 sensitivity groups. The frequency slope is the regression coefficient obtained when the frequency group indexes for a given year are related to the group number. To retain more nearly the normal distribution of the data and to avoid the influence of the general price level upon this coefficient, the calculation was carried out in terms of logarithms, the result being expressed in percent change in price per group. The Bureau of Labor Statistics Index of Wholesale Prices and the Federal Reserve Board Index of Industrial Production shifted to a 192629 100 base are also given for comparison.

=

Table IV-A gives data for the period 1913 to 1926 similar to those in table IV except that 11 items which showed extreme and disproportionate rises in price. during the war period when supplies of these commodities from Germany were cut off have been eliminated. These items are Nos. 598, 609, 611, (dyestuffs); 620, 659, 660, 661, 662, (potash salts); 632 (Continued on Page 200)

In those cases where the 1932 price is greater than the average of 1929 and 1937 prices, the index is negative. In such cases, the 1932 price is greater than the 1929 price or the 1937 price.