Risk Attitudes and the Equity Premium Puzzle: empirical tests in a cross-country setting

Abstract: This study utilises panel data, Equity Home Bias measurements and a two-stage estimation process incorporating one version of the international CAPM to extract comparable input data and test country-scores for risk preferences, risk aversion and time discounting as well as country scores on broader cultural dimensions, on country-estimates of the Equity Risk Premia. The risk attitude scores, which just recently have been made available, are such that they may proxy for irrational behaviours which have been theorized to explain the Equity Premium Puzzle, and enable a rigorous way to empirically test such an effect. Rieger and Wang (2012) and Rieger and Ngo (2013) study the same parameters through cross-sectional regressions with macroeconomic variables as controls and find widespread evidence that they have an effect on the size of the Equity Risk Premia. Through highly differentiated methods and assumptions this study only finds this relationship in a limited number of subsamples for four out of the five parameters studied, of which only one has an augmenting effect on Equity Risk Premia levels. Additionally, regressions with interaction-terms are used to test the stationary assumption of cross-country risk preferences and cultural dimensions, rejecting the hypothesis of them always being constant over time.