Stock Return Predictability by Bayesian Model Averaging: Evidence from Stock Exchange of Thailand

This research paper examines the predictability power on future stock returns by employing the concept of Bayesian Model Averaging (BMA). The sample focuses on Stock Exchange of Thailand (SET) over 2001-2011. Predictors for return predictability contain financial information which are dividend yield, Book-to-Market, Earning yield, Default risk premium, Monthly rate of three-month Treasury bill, Term premium, Monthly inflation rate and Term spread. This paper also explores the predictability power over financial crisis, sub-period over 2008-2009. In addition, this paper compares expected returns from two models between BMA and traditional regression (Fama and Macbeth two steps procedure). Results indicated that BMA approach outperforms the traditional regression model.