Prior work claims that the CAPM is mis‐specified based on evidence that beta and idiosyncratic variables such as size, book‐to‐market, and price‐earnings ratios combine to explain average cross‐sectional variation in stock returns. This paper set out to reexamine this research by employing more rigorous statistical methods to control for beta shifts.
TIMVAR program is employed to control for risk shift. Clean betas are used to reexamine the methods used in Fama and French (1992).
The paper shows that, even if the CAPM generates expected returns, the mentioned idiosyncratic variables may be correlated with expected returns. Moreover, if beta is measured with error, then it is possible for idiosyncratic variables to enter as explanatory variables. Evidence is provided indicating that, as measurement error in beta is reduced, then the role of beta in explaining cross‐sectional returns increases, while the role of idiosyncratic variables diminishes.
Beta continues to be a significant variable to measure the risk of securities.
This paper employs a novel approach (TIMVAR program) to detect and control for structural changes in linear regression. This approach has been widely used in quality control by industries. The audience of the paper includes both academics and practitioners.
