Distinguishing among Competing Theories of the Market Share Effect

Abstract

Empirical studies assessing the role of market share in influencing profitability have been challenged because of their inability to distinguish among competing hypotheses. The author addresses this concern by estimating a reduced form model of profitability. The estimated reduced form, producing market share coefficients indistinguishable from zero, is consistent with an underlying structural model with no direct market share effect. The bivariate correlation between market share and ROI arises primarily from a failure to control for the unobservable factors contemporaneously influencing both variables and inducing serial correlation in ROI models. An alternative structural model that both contains a substantial market share effect and is consistent with the estimated reduced form is not apparent.