Optimal Asian Multi-Currency Strategy Portfolios with Exact Risk Attribution

Abstract

In an earlier paper (Los, 1998a), the exact and complete return attribution framework of Singer and Karnosky (1995) was extended to include market risk measurements for n countries. Exploiting a selection matrix based on the cash accounting identities, the resulting degenerate portfolio choice problem is solved as a lower dimensional, non-degenerate problem of fundamental investment choices between stock market premiums and currency swap returns. The original n2 multi-currency strategic investment allocations are uniquely retrieved from the resulting 2n optimal fundamental choice allocations. This new optimal return-risk attribution accounting framework is applied to monthly return data of Hong Kong, Indonesia, Japan, Malaysia, the Philippines, Singapore, Thailand, the USA and Germany from June 1992 through December 1997. This includes the illustrative period of the Asian currency crisis of July - December 1997. The USA and Germany are included as alternative low risk strategic investment allocations in the Asian portfolio for further diversification. Throughout this five and a half year period, Asian risk levels, as measured by the GMV standard deviations of return, were about five times the corresponding average returns. The evidence shows that most of the strategic investment risk in Asian countries is attributable to the risk amplitudes of the stock markets, followed by those of the currency markets and, least, the cash markets. The Thai stock market was the most volatile market to invest in throughout the period. The currency swaps caused the spreading contagion. For a U.S. dollar based Asian investor, GMV portfolio risk could have been reduced by half and his return could have been doubled, when the USA would have been included in his portfolio. In contrast, diversification to Germany (Europe) would only marginally have contributed to portfolio risk reduction. Risk management in Asia was hazardous. Spectral analysis of the covariance risks shows that during the investigated period the Asian portfolio efficiency frontier was non-stationary and that the Asian diversification risk and systematic risk changed over time in a few specific countries.