Social Costs of Oligopoly and the Value of Competition

TLDR: In this paper, the authors present a new methodology for estimating welfare losses caused by market power, which is based on an empirical model of oligopoly behavior and limit pricing, and provide estimates of: (a) actual social costs arising from existing market structures and (b) expected monopoly social costs that would occur if there were no competition actual or potential.

Abstract: In this study we present a new methodology for estimating welfare losses caused by market power. We depart from past studies by explicitly taking into account different levels of market power. We provide estimates of: (a) actual social costs arising from existing market structures and (b) expected monopoly social costs that would occur if there were no competition actual or potential. The difference between actual and monopoly welfare losses represents the value of competition in existing markets. We further estimate the separate contributions of actual and potential competition to this value. Our methodology is based upon an empirical model of oligopoly behaviour and limit pricing. From this model we estimate the markup which would occur were there no competition. We use this markup in turn to estimate industry demand elasticity at the monopoly price. With this elasticity and the assumption of linear demand we can characterise demand, cost, and welfare conditions at each equilibrium: monopoly, actual, and competitive. With this new methodology and some other modifications of earlier techniques we provide not only new estimates of welfare losses, but also estimates of the value of competition under existing conditions. We find that the actual deadweight loss triangle averages 2-9 % of value of shipments for a sample of 37 industries. We also estimate that were these industries to maximise joint profits with no threat of entry, the welfare loss would be I I *6 %. The difference, 8-7 %, we attribute to the beneficial effects of potential competition (4 9 %) and actual competition (3.8 %). Our monopoly benchmark thus yields additional understanding of the value of competition. We cannot represent our study as having solved all of the problems associated with social cost estimation of monopoly power. Indeed, given general equilibrium problems associated with horizontal, vertical, and cross-industry aggregation interacting with the 'second best' problem, we doubt that all the problems can be solved, although we can point out some of the potentials for bias. After presenting our estimates we discuss the implications of sampling methodology and aggregation problems, demonstrating that potentially strong, but possibly offsetting, biases exist in all studies, including ours.