Measures of equity home bias puzzle

The paper develops measures of home bias for 42 countries over the period 2001 to 2011 by employing various models: international capital asset pricing model (ICAPM), classical mean-variance, minimum-variance, Bayes-Stein, Bayesian, and multi-prior correction to Bayesian. ICAPM country portfolio weights are computed relative to world market capitalization. Bayesian model allows for various degrees of mis-trust in the ICAPM and multi-prior model's investors' ambiguity aversion. Mean-variance computes optimal weights by sample estimates of mean and covariance matrix of sample return and Bayes-Stein improves precision associated with estimating the expected return of each asset. The paper finds that, for few countries, there is not much change in home bias measures using various models. Foreign listing, idiosyncratic risk, beta, natural resources rents, size, global financial crisis and institutional quality have significant impact on home bias. There are policy implications associated with home bias.