A Bayesian model averaging approach to enhance value investment

Abstract

Simple financial ratios such as book-to-market are often used to identify value stocks. This paper examines the extent to which fundamental accounting information can be used to better identify truly undervalued value stocks to enhance profit in a simple value strategy. Gibbs sampling and model averaging are used in a logistic regression setting, employing fundamental accounting information as explanatory variables, in the design of an implementable investment strategy applied to markets in the US, the UK and Australia.