How Do Alphas and Betas Move? Uncertainty, Learning and Time Variation in Risk Loadings

I employ a parsimonious model with learning but without conditioning information to extract time-varying measures of market-risk sensitivities, pricing errors and pricing uncertainty. Parameters estimated for U.S. equity portfolios show significant fluctuations, along patterns that change across size and book-to-market categories of stocks. Time-varying betas display superior predictive accuracy for portfolio returns against constant and rolling-window OLS estimates. I also study the relationship of betas with business-cycle variables, finding that those of high BE/ME stocks move pro-cyclically, unlike those of low BE/ME stocks. Investment growth, rather than consumption, predicts the betas of high BE/ME and small-firm portfolios.