An important but still partially unanswered question in the investment field is why different
assets earn substantially different returns on average. Financial economists have typically
addressed this question in the context of theoretically or empirically motivated asset pricing
models. Since many of the proposed “risk” theories are plausible, a common practice in the
literature is to take the models to the data and perform “horse races” among competing asset
pricing specifications. A “good” asset pricing model should produce small pricing (expected
return) errors on a set of test assets and should deliver reasonable estimates of the underlying
market and economic risk premia. This chapter provides an up-to-date review of the statistical
methods that are typically used to estimate, evaluate, and compare competing asset pricing
models. The analysis also highlights several pitfalls in the current econometric practice and
offers suggestions for improving empirical tests.