Trends of Concentration in American Manufacturing Industries, 1947-1958

FOR all its many shortcomings, the homely concentration ratio is a direct and fairly clear indicator of industry structure, and it is available on a comprehensive basis for manufacturing industries.' Although cross section comparisons of concentration are hazardous, these ratios are relatively reliable and useful as indicators of changes in industry structure over time-2 Recently, the Census Bureau's set of concentration ratios for 1958 has become available.3 Taken together with the 1947 and 1954 ratios, the 1958 compilation provides a nearly complete set of comparable ratios for 4-digit industries covering a span of 11 peacetime years, with an intermediate year to help distinguish trends from erratic movements. present study uses these new ratios to analyze changes in concentration over time, and their relations to some other industry conditions, including growth and oligopoly. first section considers broad trends, and begins with the traditional and natural question: has concentration generally increased or decreased over the 1947-1958 period? Changes in concentration in the basic metals and engineering industries will also be reviewed. second section offers tests of several hypotheses which have been put forward relating industry concentration (as cause or effect) with industry growth or contraction. Some evidence on the role of entry is also presented. findings also bear on discussions of the size distribution of firms. In the third section, trends of concentration in industries which had structures in 1947 will be surveyed. This will give some indication of how vulnerable the dominant firms have been, and will thereby throw some light on two theories of oligopolistic behavior which have been advanced by J. S. Bain and G. J. Stigler. * I am indebted to Leonard Schifrin, Charles H. Berry, George Stigler, William J. Fellner, and Paul MacAvoy for a number of helpful comments on an earlier draft of this paper. Leonard Schifrin also suggested the inclusion of coverage ratios in the regressions of the second section. multiple regression program was borrowed from James Friedman. 1Among the obvious weaknesses are the following: Industry definitions according to the Standard Industrial Classification (SIC) system do not accurately delineate true markets, nor do they allow for potential entrance by firms across industry lines. Based as they are on national figures, they do not allow for higher concentration in regional submarkets. In some cases they neglect the role of imports in domestic markets, and of export markets for domestic producers. ratios do not describe the entire firm distribution, but only one slice of it; and they give no information about the relative positions among the top group of firms. ratios also fail to reflect turnover among firms over time. Finally, there is the obvious but important point that concentration is purely a structural indicator, and says nothing explicitly about behavior or performance. For a thorough treatment of these and other problems of concentration ratios, see National Bureau of Economic Research, Business and Price Policy (Princeton, 1955). 2 Recent examples of studies involving concentration ratios for individual industries include G. Rosenbluth, in Canadian Manufacturing National Bureau of Economic Research (Princeton, 1958); A. L. Phillips, Scale and Technological Change in Selected Manufacturing 1899-1939, Journal of Industrial Economics (1956), 179-193; R. T. Selden, Accelerated Amortization and Industrial Concentration, Review of Economics and Statistics, xxxvii (Aug., 1955), 282-291; W. G. Shepherd, Comparison of Industrial in the United States and Britain, Review of Economics and Statistics, XLIII (Feb., 1961), 70-75; and V. R. Fuchs, Integration, and Profits in Manufacturing Industries, Quarterly Journal of Economics, LXXV (May, 1961), 278-292. A different approach to concentration (based on size inequality of firms) has been taken by P. E. Hart and S. J. Prais. See especially S. J. Prais, The Analysis of Business Concentration: A Statistical Approach, Journal of the Royal Statistical Society, Series A (1956), 150-191, and P. E. Hart, Concentration in Selected Industries, Scottish Journal of Political Economy, (1958), 185-201. This approach is discussed in the second section of the present paper. 3 Ratios in Manufacturing Industry: 1958, 87th Congress, 2nd Session, No. 78696, Government Printing Office (Washington, D. C., 1962), Part I. volume containing ratios for 1954 and 1947 is in American Industry, 85th Congress, 1st Session, No. 46358 0 (Washington, D. C., 1957). In addition, there is expanded coverage of 5-digit product classes, and for many products there are ratios for both 1954 and 1958. Also, a smaller companion volume (Part II) presents some estimates of concentration within regional markets for a number of 4digit industries.