Abstract

Market events of the past ten years have sparked an interest in tactical asset allocation, the dignified new title for market timing. The turbulence of October 1987 has only accelerated this interest. In the typical application, the portfolio manager seeks to assess the trade-offs between the expected return and risk of various asset classes, such as stocks, bonds, and cash. There are a myriad of techniques used, including valuation analysis, investor sentiment analysis, volatility analysis, and so on.

Here we examine the equity risk premium as a key component of the asset allocation decision. Our purpose is to test the predictive value of various equity market risk premium proxies. To do this we conduct statistical tests to determine if the risk premium proxy, measured in any month, is directly correlated with subsequent market returns in the following month. The results are encouraging, suggesting the following:

• Earnings yield and dividend yield are very effective tools for asset allocation. Indeed, they outperform the constant growth dividend discount model (DDM) approach, at least in this narrowly defined context.

• The earnings yield approach can be improved substantially with appropriate attention given to some of the quirks of the earnings stream.

• In all fairness, even the best of these disciplines can be enhanced by including additional information. Unfortunately, much of the “additional information” is not subject to quantification.

• By combining various techniques, we can produce an “equity excess return predictor,” which turned bearish three months before the October crash.