International Portfolio Allocation under Model Uncertainty

This paper revisits an old argument, hedging real exchange rate risk, as an explanation of the international home bias in equity. In a dynamic model, the relevant risk to be hedged is the long-run risk as opposed to the short-run risk. Domestic equity is indeed a good hedge with respect to long-run real-exchange-rate risk. Two new frameworks are able to explain a large share of the observed US home bias: a model with Hansen-Sargent preferences in which agents fear model misspecification and a model with Epstein-Zin preferences. These two models are also immune to the risk-free rate puzzle. (JEL C58, F31, G11, G15)