This section develops an econometric framework that allows an investor to combine in- formation in the data with prior beliefs about both pricing and skill. Nonbenchmark assets allow us to distinguish between pricing and skill, and they supply additional information about funds' expected returns. In addition, nonbenchmark assets help account for common variation in funds' returns, making the investment problem feasible using a large universe of funds. The Bayesian econometric framework here is very similar to that in P¶ astor and Stambaugh (2001), who address performance estimation rather than investment decision making. Consequently, they specify noninformative prior beliefs about the degree of skill a fund manager might possess
