ENTRY plays a crucial role in microecoIJi nomic models of market structure and performance. However there has been very little direct empirical investigation of entry and its determinants over a broad cross section of industries. In this paper we construct and estimate a model which assumes entry is a function of the incentives to enter relative to the level of entry barriers. The subject of the analysis is the cross-section differences in entry between the three-digit industries of the Canadian manufacturing sector. Previous studies of entry have either concentrated on only several industries or have attempted to make conclusions about entry conditions by regressing profit rates on variables representing entry barriers. Bain (1956) examined 20 (mostly four-digit) United States manufacturing industries and concluded that the most significant barriers to entry were product differentiation, economies of scale in plant or firm and control of patents or scarce resources, respectively. Mann's study (1966), which was limited to 30 of the United States manufacturing industries, did not examine the relative importance of the various barriers to entry. Mansfield's (1962) sample was limited to four industries. Capital requirements was the only barrier he considered. Most econometric investigations of entry barriers have been indirect tests. They have regressed the profit rate, rather than entry, on those structural characteristics considered to be barriers to entry (Comanor and Wilson 1967, Miller 1969). Unfortunately this specification does not permit reliable conclusions regarding the effectiveness of these variables in deterring entry. There are theoretical reasons for questioning the often assumed strong positive relationship between entry barriers and the true profit rate. Additionally of course, there is the infamous gap between true and measured profits. This paper's treatment of entry barriers has several important advantages over previous work. Those variables considered to be entry barriers are introduced directly as determinants of entry rather than the profit rate. This is a direct rather than indirect test of the propensity of these factors to deter entrants. A most important result is that our conclusions are less sensitive to those unavoidable measurement errors in the profit rate. Our estimating equations consider a more extensive list of entry barriers over a larger (71) sample, covering all types of manufacturing.

The determinants of entry: a study of the canadian manufacturing industries

Concentration ratios and the degree of monopoly

Factor market distortions, production, and trade: a survey

ECONOMISTS have been alternately fascinated and frustrated with industry or market concentration ratios ever since they were first calculated from census data for the Temporary National Economic Committee for 1937. They are fascinated because concentration ratios are the single best available index of the degree of oligopoly. The frustration stems from the absence of precise coincidence between the Standard Industrial Classification System (SIC) used by the Bureau of the Census and economically relevant markets. Yet, when all is said and done, most industrial organization economists agree that concentration ratios based on SIC industries not only are the best available, but provide useful measures of one dimension of the extent of oligopoly in American industry.1 This is not to imply, of course, that market concentration is the only index of oligopoly or market power. Economic theory suggests and empirical studies verify that entry barriers, product differentiation, firm conglomeration, among others, also may influence firm conduct and industrial performance. But changes in industrial concentration are uniquely significant because often they reflect, at least partially, changes in other structural variables as well. For example, if entry barriers are declining, because of growing markets or whatever, this tends to become reflected in lower concentration ratios. Hence changes in market concentration may also reflect what is happening to other structural variables affecting the discretionary power of sellers. We shall not here argue the question of whether or not concentration ratios are meaningful indices of market structure or whether they are causally related to industrial performance. Those assuming that the answer to both questions is yes, gain aid and comfort from the most comprehensive review of the empirical evidence on the subject.2

Trends in Industrial Market Concentration, 1947 to 1970

Dynamic market models in industrial organization

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. The 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. The 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. The 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. The findings also bear on discussions of the size distribution of firms. In the third section, trends of concentration in industries which had "oligopolistic" 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. The 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. The 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. The 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 Concentration and Price Policy (Princeton, 1955). 2 Recent examples of studies involving concentration ratios for individual industries include G. Rosenbluth, Concentration in Canadian Manufacturing Industries, National Bureau of Economic Research (Princeton, 1958); A. L. Phillips, "Concentration, Scale and Technological Change in Selected Manufacturing Industries, 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, "A Comparison of Industrial Concentration in the United States and Britain, Review of Economics and Statistics, XLIII (Feb., 1961), 70-75; and V. R. Fuchs, "Integration, Concentration 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 Concentration Ratios in Manufacturing Industry: 1958, 87th Congress, 2nd Session, No. 78696, Government Printing Office (Washington, D. C., 1962), Part I. The volume containing ratios for 1954 and 1947 is Concentration 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.

William G. Shephard

Papers

Trends of Concentration in American Manufacturing Industries, 1947-1958