Asset-pricing anomalies at the firm level

We introduce a hierarchical Bayes approach to model conditional firm-level alphas as a function of firm characteristics. Our empirical framework is motivated by growing concerns in the literature regarding the reliability of inferences from portfolio-based methods. In our initial tests, we confirm the existence of several CAPM anomalies at the firm level. Prominent multifactor models deliver only a modest improvement, however, as they often resolve only those anomalies which are directly linked to their additional factors. Further results suggest that the economic importance of CAPM anomalies is overstated. We find that anomalies are primarily confined to small stocks, few characteristics are associated with CAPM alphas out of sample, and many firm characteristics do not contain unique information about abnormal returns.