Analysts' Forecasts as Proxies for Investor Beliefs in Empirical Research
Posted: 23 Jul 1999
Abstract
We study the use of properties of analysts' forecasts as surrogates for unobservable constructs in empirical research. We provide an analytical framework with which to examine past empirical practice and contemplate future empirical research. Our model is used to interpret existing empirical research and suggest new testable hypotheses. Two conceptual links are examined: the link between the properties of forecasts and theoretical constructs of investor beliefs and the link between theoretical constructs of investor beliefs and price and volume reactions to earnings announcements. Among our reported results we show that forecast dispersion on its own will not measure the level of investor uncertainty that conditions market reaction. To control adequately for investor uncertainty in empirical tests, other properties of the forecasts including the number of analysts and the amount of common information in individual forecasts must also be included. Our results also highlight the fact that the relation between market reactions and investor beliefs depends critically on assumptions about the nature of private information acquisition.
JEL Classification: G14
Suggested Citation: Suggested Citation