Investigating a new methodology for ranking international mutual funds

This paper tests the hypothesis that a composite forecast is, at times, more accurate than separate forecasts. The rationale for forecast superiority is that valuable information missing from one model may be captured by the other model. The expected asset returns from the traditional CAPM and the mean-lower partial moment CAPM have been combined to generate a composite systematic risk measure. The measure is then used in Treynor's performance index and tested on a sample of U.S.-based international mutual funds. The results show that the composite beta is a statistically significant and meaningful parameter. However due to sample size, the Wilcoxon signed rank test fails to provide strong evidence that there is a significant difference between Treynor's fund rankings when using the two different asset-pricing models.