A Bayesian view of temporary components in asset prices

This paper studies models in which the a stock price contains a random walk and a stationary component, as in Fama and French [Fama, Eugene F., and Kenneth R. French, 1988, Permanent and Temporary Components of Stock Returns, Journal of Political Economy 96, 246-273.] and Poterba and Summers [Poterba, James, and Lawrence Summers, 1988, Mean Reversion in Stock Prices: Evidence and Implications, Journal of Financial Economics 22, 27-59.]. We extend this model to allow for two latent factors which generate short term and long term autocorrelations, respectively. To facilitate econometric identification, we assume that these factors are common across multiple asset returns, and we estimate the factor loadings. In an application to size and book/market sorted portfolios, we find the short term factor economically and statistically insignificant. Estimates of parameters relating to the long range component suggest that portfolios of small firm stock display about three times the amount of mean reversion than for large firm stocks. Overall, the evidence suggests that mean reversion is largely a small firm phenomenon. The evidence is consistent with dynamic equilibrium models in which asset prices co-integrate with aggregate consumption or dividends.