Download This PaperWhat Difference Fat Tails Make: A Bayesian MCMC Estimation of Empirical Asset Pricing Models
Bayesian Inference in the Social Sciences, edited by Ivan Jeliazkov and Xin-She Yang, John Wiley & Sons, published in October, 2014
28 Pages Posted: 22 Aug 2017
Date Written: July 11, 2013
Abstract
This chapter presents an empirical application of Bayesian MCMC estimation to the three main asset pricing models in use in the financial econometrics literature, namely, the Capital Asset Pricing Model (CAPM), the Fama-French (1992) three-factor model, and the Carhart (1997) four-factor model for decile portfolios sorted on market capitalization and book-to-market ratios. I use a SUR model similar in spirit to Zellner (1976) with a Student t-link to allow for fat tails in the distribution of the return innovations. The empirical findings are encouraging in that fat tails lead to posterior distributions of the asset pricing errors that encompass zero and indicate that previous findings of statistically and economically significant intercepts might be an artifact of using a Gaussian error distribution.
Keywords: asset pricing errors, fat-tailed error distributions, size, value, momentum, capital asset pricing model, Fama-French three-factor model, Carhart four-factor model, Bayes credibility interval, Bayes factor, model comparison, Student-t link
JEL Classification: G11, G12
Suggested Citation: Suggested Citation