I This paper attempts to provide a framework for clarifying and testing a version of the "income-smoothing" hypothesis. This hypothesis has been variously expressed in terms of income levels, rates of change in income, and rates of return, inter alia.1 In general, each version of the hypothesis is concerned with the extent to which managers may attempt to affect the volatility of a series of reported accounting numbers (or, in the case of rates of return, a series of relationships among accounting numbers) via selections and applications of accounting procedures. The alleged motivation for this behavior is a desire to reduce the extent to which "bad times" and-at the other extreme-"good times" are revealed by reported accounting numbers. It is suggested by some that a "smoothed" series of accounting numbers, particularly income numbers, will enhance the value of a firm. A typical statement of this argument was provided by Hepworth: "Certainly the owners and creditors of an enterprise will feel more confident toward a corporate management which is able to report stable earnings than if considerable fluctuation of reported earnings exists."2