The Profitability-Concentration Relation: Market Power or Efficiency?

INDUSTRIAL economists have been studying the relationship between profitability and concentration using cross-industry samples for many years now. However, as sometimes happens, once it appeared that some consensus had been achieved: With few exceptions, market concentration and industry profitability are positively correlated (Peltzman, [I977, p. 229]), people began to question the interpretation of those results. In particular, Demsetz (e.g. [I973], [I974]) and others suggested that, rather than concentration/ market power leading to higher prices and profits, the causal direction in fact ran from greater efficiency to both higher profits and, incidentally, to higher concentration. Our purpose here is to subject these conflicting hypotheses to cross-industry empirical test using U.K. data. In order to do this, we first develop a model which is sufficiently general to allow for both alternative explanations of a positive concentration-profitability correlation. In the light of this model we then examine the implications for the within-industry relation between market share and firm-level profitability. More specifically we explore Demsetz's claim that the empirical form of this relationship should signal which of the alternative hypotheses is consistent with the between-industry profitability concentration correlation. Our model suggests that, in certain special circumstances, this claim is partly justified, in that it should be possible to assess the relative strengths of the market power and efficiency effects from the within industry relation. More generally, however, factors such as product differentiation and scale economies may seriously impair the information content of the relation, with at least a strong suspicion that the estimates will tend to exaggerate the evidence in favour of the efficiency hypothesis. In the event, our results tend to confirm that, for most industries, the within-industry data offer little conclusive support for either hypothesis. But, for a not insignificant minority of industries, the evidence suggests some support for the view that both efficiency and market power are at work simultaneously. Our model is detailed in section II and estimated on U.K. data in section III, after which we offer a few comments in section IV. First, however, we Drovide a backcloth to the later analysis.