A Comparison of Relative Predictive Power for Financial Models of Rates of Return

Abstract

The need for an empirical measure of the relative value of financial theories becomes more important as we increase their number and sophistication. Although there are several methods of determining value, here we emphasize relative predictive power. According to this single criterion, a theory is insignificant if it fails to increase our present ability to predict future events. We recognize, however, that a theory can be useful if it merely increases our understanding of real events, or if it is based upon realistic assumptions.