In addition to index numbers of production in major branches of industry or of total production for the whole country, index numbers have been prepared for the more important single lines of production. For practical business purposes, these individual indexes of production are probably of greater value than the more general indexes. They make possible a comparison between the actual current rate of production at any time with the estimated normal rate for that period. This normal rate is estimated from the average rate of growth over a period of years and from the average month to month seasonal variation. If the rate of current production in any given line is above normal, knowledge of that fact enables business to avoid over-extension of equipment and over-expansion of production. This type of knowledge is a primary necessity if we are ever to moderate that general unbalancing and maladjustment of production which has always been associated with the breakdown of prosperity. By use of such knowledge, it is possible to sustain a more even pace in the different lines of industry. It is possible to know whether we are producing very much more pig iron or copper or rubber than the normal need, and therefore, to know in what lines there is too little expansion and in what lines there is too rapid expansion.

A form of differentiation of data between industries which is particularly important is differentiation between production of consumers' goods and production of producers' capital goods. In general, the fluctuations in manufacture of basic materials, machinery, and industrial equipment are more severe than those in manufacture of consumers' goods. The more violent fluctuations of industries making fixed capital goods are a main feature of the unbalancing of production which is characteristic of the breakdown of prosperity. The charts on page 600 show the wide fluctuations in pig iron production, a major producers' good, and a comparison between producers' basic materials and consumers' goods. Pig iron fluctuations are similar in kind to, but more extreme in degree than, the fluctuations of producers' goods in general. (7) The Volume of Consumption. The available data on retail trade indicate that the common notion of a consumers' strike" is a popular error. In the depression of 1920-1921, the sales of department stores, drug stores, grocery stores, and cigar stores were well sustained until the latter half of 1920.12 This was after wholesale prices had broken and after physical production in leading lines had begun its downward course. The depression was well under way before the main branches of retail trade fell off.13 The consumers' strike followed, but did not. initiate the downfall of prosperity.

The term "consumers' strike" is loose and misleading. In the sense of a supposed definite stoppage of retail buying, the implication of the

12 Mail-order sales began a decline in February, 1920, but even this decline came after the fall in production of both consumers' and producers' goods had begun.

13 H. B. Vanderblue, Problems in Business Economics, pp. 63-68, 614; L. B. Mann, Journal of the American Statistical Association, June, 1922, pp. 255-258.

term is contrary to the facts. What really happens appears to be substantially as follows: 14 Manufacture proceeds more rapidly during a period of activity than physical volume of consumption. Not that consumption fails to increase. It does increase, but not so rapidly as factory

COMPARISON OF VOLUME OF MANUFACTURE OF CONSUMERS' GOODS AND PRODUCERS' BASIC MATERIALS WITH ESTIMATED NORMAL

Adjusted indexes as compiled by the Harvard University Committee on Economie Research, in Harvard Economic Service, Volume III, No. 52, Dec. 27, 1924, p. 377.

Adjusted indexes, ibid., Volume III, No. 35, Aug. 30, 1924, p. 250.

output. The maladjustment then is a matter of maladjustment between two unequal rates of increase, the swift increase of manufacture and the slow increase of consumption. Supply outruns demand at current prices, with the resulting period of liquidation of stocks and readjustment of prices.

14 The Problem of Business Forecasting, edited by Warren M. Persons, Chapter II, by Willford I. King, Chapter III, by W. Randolph Burgess; Federal Reserve Bulletin, Jan., 1923, p. 104.

This lack of coördination between consumer buying and factory production has received a number of explanations. Some authorities have attributed it to the lag of wage increases behind price increases, with the consequent lag of purchasing power which wages represent. According to this view, a more liberal wage policy pursued during prosperity would put labor in possession of the wherewithal to make retail demand effective, and would insure that goods would be taken off the markets as fast as manufacture puts them on the market. Other authorities have attributed the lag of consumption to the failure of corporations to disburse profits as rapidly as they are earned. Various policies of reserves and surpluses mean a withholding of earnings from the hands of stockholders and hence a diminution of their purchasing power in consumers' markets. Other authorities have attributed the consumption lag to various theories of over-production due to the inherent superefficiency of machine methods of manufacture; of over-saving due to excessive capital accumulation during prosperity; and of under-consumption due to an unjust distribution of income. None of these explanations fully accounts for the maladjustment. Analysis of the business cycle at this point becomes somewhat vague and indecisive, for the reason that statistical data have not been carried far enough to yield definite conclusions. The fact that consumer buying is outrun by factory output at current prices appears to be established, but the explanation of the fact and the means of remedy require further research.

In addition to the lag of consumption behind manufacture, there is a marked difference in the amplitude of fluctuation of the two factors. A striking feature of retail trade is its relative stability. In general, retail trade shows a steady volume, with slight interruption due to business movements. The amplitude of fluctuations in physical volume of consumption is much less than that in physical volume of manufacture.15 In consequence of this difference, there appear alternate periods of excess and deficit in the stocks of goods on hand. At times of activity and forward buying among dealers, inventories accumulate in the hands of manufacturers, wholesalers and retailers. At times of dullness and hand-to-mouth buying among dealers, stocks become exhausted and shortages appear. The total accumulation of stocks of each line of product thus varies from phase to phase of the business cycle. When manufacture overfloods with new production an already overstocked market, the inevitable result is costly and wasteful liquidation. Control over stocks and inventories is, therefore, essential to a better adjustment between. the relatively mild fluctuations of consumption and the relatively violent fluctuations of manufacture.

To summarize, two significant factors in volume of consumption are noted: first, a lag of consumption behind manufacture in leading lines;

15 A significant bit of evidence in this connection is the fact that the maximum cyclical decline in the number of employees in retail establishments in 1920-1921 was under 3 per cent as compared with 14 per cent for all industry and 26 per cent for factories. See W. I. King, Employment, Hours and Earnings in Prosperity and Depression, p. 30.

second, a discrepancy in amplitude of fluctuation between retail trade and manufacture, accompanied by alternate over-stocks and shortages of goods, and consequent price liquidations and readjustments.

(8) The Volume of Unemployment. An index of employment is calculated to show changes in the relative intensity of employment between two points of time, without necessarily measuring the total volume of employment or unemployment. What is measured directly is simply the fluctuations of employment. From these fluctuations, it is easy to infer the fluctuations in unemployment. But the absolute volume of unemployment is not indicated. When the President's Conference on Unemployment set out to ascertain the absolute volume of unemployment in 1921, the estimates were so uncertain that they ranged all the way from two to six millions. It would be useful to have exact knowledge on the absolute volume of unemployment, but such exactitude has not as yet been attained. The main step in this direction has been the measurement of "fluctuations in employment" between various periods of time. The accompanying diagram shows such an index of employment in factories for the period from 1914 to 1924. Relative changes in employment are the factors here measured.

GENERAL INDEX OF EMPLOYMENT IN FACTORIES
(Average Monthly Employment 1923 100)

=

From data compiled by the United States Bureau of Labor Statistics, representing weighted indexes based upon the number of wage earners in the respective industries in 1919. See Monthly Labor Review, April, 1924, pp. 129-132.

The cyclical decline of employment varies widely from industry to industry. This variation, as shown in the table on page 603, has been measured by W. I. King for the 1920-1921 period, taking as the form of measurement the "total hours actually worked by all employees." 1

The range of decline was least for retail trade and greatest for metal manufactures. Variation in decline of employment varied with the size of establishments as well as with the type of industry. The greatest stability of employment existed among small establishments, and the greatest fluctuation among large establishments.

The indexes of employment are valuable for a number of purposes. They reflect the general course of production and thus supplement the 16 W. I. King, Employment, Hours, and Earnings in Prosperity and Depression,

pp. 55-59.

figures on industrial activity. They forecast changes in the buying power of a large body of consumers, since the amount of purchasing power flowing into the hands of wage earners is conditioned directly by the state of employment. They are a useful guide to banking policy and the credit situation, by showing to what extent the physical resources of production are being employed by business men seeking bank loans. They indicate whether the labor market is over-supplied or under-supplied with potential workers, and so help in formulating wage and employment policies in industry. Finally, they indicate the fluctuations in social well-being associated with the business cycle, such as poverty, crime, suicide, and migration.

Particularly because of the heavy social cost and distress associated with cyclical unemployment, much attention has been given to the means for prevention of unemployment. The concrete proposals toward this end are discussed later in the present chapter. The necessity for stabilizing employment is urgent. In the past, the insecurity of the job has been a primary source of demoralization and suffering in the ranks of labor. It has been a main cause of discontent, ill-will and bitterness in industrial relations. It has been a constant stimulus to fear, worry, humiliation, and resentment among workers' families. To establish a reasonable degree of security in the job, which represents the laborers' only chance to earn a living, is a problem pressing for solution.

Some Theories of Business Cycles.-A wide variety of theories have appeared in explanation of business oscillations. The following list is by no means exhaustive, and the theories mentioned represent considerable overlapping. However, the enumeration does separate the main types of theories and the points of view which they specially emphasize.

(a) Crop cycles due to periodic variations of rainfall and temperature have been assigned as the cause of variations in all other lines of enterprise, such as manufacture, trade, and finance. Statisticians have made out a good case for climatic variation, but even if this were