Dynamic asset allocation modeling for international investment : a comparison of information-based active strategies versus passive strategies for the EAFE and S&P 500 portfolios

Tactical asset allocation has become popular in asset management since the stock market crash in October 1987. Researchers and practitioners have always promoted the benefits of international diversification. Much research has been done in domestic asset allocation and global asset allocation. However, a portfolio mix between the S&P 500 Index and the MSCI EAFE Index is a novel combination for tactical asset allocation. The objective of this study is to develop a dynamic asset allocation strategy dealing with such an asset mix. A rolling binary logit model is built using the preceding sixty months of data and is used to forecast the next month's movements of these two indices. Forty-eight trading strategies are implemented to validate the forecastability of the prediction model using the out-of-sample data from January 1978 to September 1999. This study affirms that a dynamic asset allocation strategy can be established to time the market and generate a superior abnormal return on a portfolio investing in these two assets. A prediction model can be built on public information variables to successfully forecast the upcoming movements of these two indices. Even with transaction costs, an investor can rely on the signals to make asset allocation between these two indices and produce a terminal wealth significantly larger than the passive portfolios invested in either one of the indices alon