Puzzles in international portfolio investment.

Abstract

This dissertation investigates two important puzzles in international finance: the home bias puzzle and the contagion puzzle. In the first chapter of the dissertation, a rational forward-looking model of portfolio choice is used to study home bias, the tendency to overweight home securities in investment portfolios. The model is applied to monthly returns on equity indices for Canada, France, Germany, Italy, Japan, U.K., U.S. and the world in the 1970--1998 period. Three main findings emerge. First, if a U.S. investor uses a forecasting model and all of the currently available information to form and update the forecasts about the mean and variance of future returns, it is optimal for the investor to hold an internationally diversified portfolio. Second, the conditional approach generates much more stable and sensible portfolio weights than the unconditional approach. Third, taking into account the predictability of the variance can reduce the standard error of optimal allocation weights substantially compared with the traditional method of using the historical sample mean and variance. These findings imply that a well-diversified portfolio is statistically significantly different from the home-biased portfolio. The second chapter of the dissertation focuses on the contagion puzzle: the simultaneous withdrawal of capital from emerging markets apparently without differentiation during crises. The theoretical model of rational contagion generates a co-movement of capital outflow from both crisis and non-crisis countries through the transmission of uncertainty. The model describes the portfolio allocation problem facing an investor from a developed country who invests in the home market and in emerging markets. Returns in emerging markets are assumed to be predictable by local dividend yields to some degree. The model is used to demonstrate that when uncertainty about the forecasting parameters in one emerging market increases (as is likely during periods of crisis), it may be optimal for the investor to withdraw from all other emerging markets, even if fundamentals in those markets are unchanged. Data on U.S. equity holdings show that recent capital outflows from Latin America during the Asian crisis and from Asia during the Mexican crisis are consistent with the theoretical model of rational contagion.