Content area
Abstract
This study is an investigation of analyst forecast dispersion as a risk measure. The study discusses the arguments that have made dispersion an intuitively appealing risk measure, and, based on these arguments, conducts empirical tests aimed at determining whether dispersion does represent risk.
The latest way of studying the earnings-returns relationship, the earnings response coefficient (ERC), makes finding a good proxy for future uncertainty an important goal. ERC's are hypothesized to be a function of noise and risk. If accounting researchers are to use ERC's to test the effects of accounting policy and choice on information system noise, an empirical proxy to control for risk differentials is needed.
The structure of the empirical tests flows from the argument that earnings uncertainty represents risk in the underlying equity. This argument rests on the assumption that earnings play a central role in valuation. Recent work in earnings-based valuation theory is utilized to construct the setting in which forecast dispersion is tested for its ability to represent earnings uncertainty and therefore risk. The empirical model constructed is a multiple regression equation with market value of equity as the dependent variable, and book value of equity, current earnings, earnings growth expectations, and forecast dispersion as independent variables.
The empirical results show that forecast dispersion behaves much like the arguments would suggest. Although earnings is the dominant factor in explaining cross sectional variations in equity value, earnings uncertainty, represented by forecast dispersion, at times plays a role in explaining why some firms are valued less by the market.
