An alternative approach to market structure and the markup ratio

Abstract

Hall's studies of market power emphasize that the markup ratio of price over marginal cost is an important indicator for assessing the role of market structure in resource allocation (without using concentration ratios) and is also a good measure of allocative efficiency. The basic idea was that price is equal to marginal cost if there is perfect competition and constant markup ratio. A statistrical model was developed to estimate the markup ratio. If the markup ratio estimated for an industry is unitary, then the null-hypothesis of competition is not rejected. If, however, it is greater than one, the alternative hypothesis of non-competition is not rejected. The theory mandates that monopoly power is dynamically constant despite a nonconstant productivity residual. Hall assumes, in etimation, that technical progress (or solow residual) has a randomness. This assumption implies that the procyclicity of technical progress is not a true phenomenon. The spurious procyclicity is revealed in the estimated markup ratio as greater than one, and provides evidence against competition. Although Hall's method is very attractive for studying market power and the insight into the behaviour of technical progress in the business cycle is quite striking, the estimation method suffers from observational equivalence, not to mention some degree of arbitrariness in the selection of instrumental variables. That is, the satistical equation estimated can be expressed in two different forms. In the method of instrumental estimation, the tow different equations look exactly the same, because they cannot identify their own instrument. It is im possible to determine in actual estimation what is estimation what is estimated, and, thus, any arguments made on the basis of the estimation results are not reliable. Similarly, we conclude that all of DHP's conclusions based on the ocmparision with Hall are suspicious if not invalidated. This may also explain why, for instance, Shapiro applies a variant of Hall's approach and finds contrary to Hall and DHP a strong relation between concentration and price-cost margins. An alternative method for the study of market power is suggested, avoiding the observational equivalence problem. Under the assumption of profit-maximization, Hall's markup ratio of price to marginal cost is replaced by the alternative markup ratio of price to margianl revenue. The hypothesis testing is quite similar to that in Hall. If price is greater than the estimated marginal revenue, the null-hypothesis of competition is rejected. The alternative method assumes that the price movements have a substantial random element. This presumption is based on the idea that the price movements are spuriously procyclical. The spurious procyclicity of price change is evidnece against competiton, and is expected to be revealed in the estimate of the alternative markup ratio as greater than one. In this alternative method, the statistical model is expressed as in Hall's method in two different forms. However, unlike Hall's method, the two different equations can identify their own instrument, som which equation is being estimated is determinable. Besides this identifiability of the instruments, there are two other advantages of this alternative method. First, unlike Hall's method, it is not constrained by the assumption of constant returns to scale. Second, the number of variables needed for the test is much smaller, and the data are easier to obtain.