Ambiguity, ambiguity aversion and foreign bias: New evidence from international panel data

Abstract

This paper provides novel empirical evidence on the relationship between ambiguity, ambiguity aversion and foreign bias in equities. We show that, after an increase in the level of domestic ambiguity relative to the level of foreign ambiguity, more ambiguity averse investors see a larger fall in foreign bias than less ambiguity investors. This finding is consistent with ambiguity operating through an information quality channel but at odds with interpretations based on familiarity. We proxy for the level of ambiguity using the prediction errors around variance forecasts based on state-of-the-art volatility models for 23 developing and emerging markets. Our measure of ambiguity aversion comes from an international survey and is based on an Ellsberg-type urn experiment. Cross-sectional regressions show that home country ambiguity aversion correlates positively with foreign bias. The results are robust to the inclusion of a broad set of control variables, changes in sample size and different ways of computing variance forecasts and the level of ambiguity. However, the results suggest that this fall in foreign bias is not associated with an improvement of international diversification.