This article hypothesizes that resources employed in concentrated industries are more specialized and durable and therefore less mobile than resources employed in atomistic industries. The hypothesis is tested by estimating the relationship between changes in the market value of a firm's securities and changes in its earnings levels and comparing this relation under different market structures. The response of market values to earnings changes is found to be greater for concentrated industries than for atomistic industries. This difference is interpreted as evidence that abnormal earnings persist longer in concentrated industries because of the lower degree of resource mobility.

This article hypothesizes that resources employed in concentrated industries are more specialized and durable and therefore less mobile than resources employed in atomistic industries. The hypothesis is tested by estimating the relationship between changes in the market value of a firm's securities and changes in its earnings levels and comparing this relation under different market structures. The response of market values to earnings changes is found to be greater for concentrated industries than for atomistic industries. This difference is interpreted as evidence that abnormal earnings persist longer in concentrated industries because of the lower degree of resource mobility.

The mean-variance capital-asset-pricing model forms the basis for much of the theoretical and empirical work in modern financial economics. While this model defines the relevant measure of the risk of a security 0 in a general equilibrium context. the relationship between this measure and the microeconomic variables of a firm has not been studied in the literature. This paper develops a model of the firm under uncer-tainty and derives the relationship between systematic risk and such firm variables as monopoly power, demand elasticity, and the labor-capital ratio. The general con-clusions are surprisingly robust and point to several interesting empirically testable hypotheses. I.