Collusion and the Pattern of Rates of Return

Abstract

This paper considers methods of separating firms and industries into collusive and competitive subsets, and the differences in patterns of rates of returns which are to be expected among these groups. The approach adopted here is quite distinct, both in structure and in policy implications, from the customary technique of estimating a linear relation between concentration and profitability or margins, which is then used to indicate a level of concentration associated with an excessive exercise of monopoly power.' Elsewhere [6] I have indicated that the low levels of significance which typify estimates of the concentration-profits relationship, their observed nonlinearity, and the heteroskedasticity commonly visible in residuals computed from such estimates all indicate that concentration, while an important determinant of the feasibility of collusion, is far from the only factor conditioning its success. As an alternative this paper outlines a proposed classification scheme for the identification of collusion which is based on price-cost margins in U.S. manufacturing for 1967. Margin data derived from Census industry tabulations are corrected for a variety of inter-industry influences compatible with competition. Following these corrections, the distribution of residuals at high concentration levels is shown to exhibit both positive skewness and more dispersion than is observed for low concentration industries, suggesting the presence of contamination by monopoly elements. The residual distribution is separated into competitive and monopoly components with large positive outliers being regarded as monopolized. One important result of applying this technique to U.S. data is a finding that many high-concentration industries appear to achieve only competitive margins. This result indicates that across-theboard deconcentration would involve restructuring a number of competitive industries. This classification scheme is outlined in more detail in Section II. The second stage of the process of identifying collusive industries involves presentation