In this paper the authors extend results by Harvey
and Zhou (1990, Journal of Financial Econometrics
26, 221–254) and Kandel, McCulloch, and Stambaugh
(1995, Review of Financial Studies 8(1), 1–53)
to derive the posterior distribution of a key parameter
in a Bayesian analysis of asset pricing models. It is shown
that this distribution depends upon the same terms that
constitute the standard asset pricing test of Jobson and
Korkie (1985, Canadian Journal of Administrative Science
12, 114–138). Contrary to the view held by other
authors, we find straightforward expressions for the posterior
distribution that can be calculated without resorting to
Monte Carlo methods.