ESTIMATES OF the magnitudes of monopoly welfare losses (MWL) appearing in the economics literature have tended to differ quite widely. At one end of the spectrum we find figures of the order of o. I % of GNP quoted by Harberger [I 954] in his pioneering study, while at the other extreme Cowling and Mueller [ I 978] have derived estimates of around I O % of gross corporate product from some of their measures. Two of the major sources of such differences lie in the treatment of (a) the general equilibrium aspects of deviations between prices and marginal costs, and (b) the impact of market power on the cost levels of firms; but, unfortunately, the tendency of writers to adopt rather different analytical frameworks often makes it difficult to accurately assign the variations in the estimates to specific factors. In the present paper the issue of monopoly welfare loss is considered in the context of a differentiated goods model based upon work on monopolistic competition by Spence [I976] and by Dixit and Stiglitz [I977]. Within a set of common assumptions about demand, the effects of varying cost conditions and differing equilibria on the level of MWL are examined. On this basis, four major factors connected with suboptimal resource allocation can be identified: (a) supernormal profits, (b) product differentiation, (c) suboptimal scale, and (d) factor bias. The first of these needs little explanation: its association with welfare loss has been well explored in structure-conduct-performance theory and is used extensively in applied anti-trust economics. Attention will therefore be focused upon the remaining three, all of which can produce MWL in zeroprofit equilibria. The paper is organised as follows. Section II sets out and discusses the differentiated goods framework which will be used throughout the rest of the analysis. It is also shown how a simple transformation of variables reduces the model to the homogeneous goods case and thereby makes immediately available a number of familiar results from the structure-conduct-performance literature. In Section III the focus is on welfare analysis of monopolistic competition. The results here are essentially restatements of earlier work, although the transformation of variables helps to clarify their foundations. Section IV extends the analysis to incorporate additional decision variables such as advertising, research and development, etc., while section V considers the case where market conditions will not sustain a large number of firms in equilibrium and suboptimal scale effects become important. Factor bias is analysed in
Read full abstract