Mutual Fund Performance Evaluation: A Comparison of Benchmarks and Benchmark Comparisons

Our primary goal in this paper is to ascertain whether the absolute and relative rankings of managed funds are sensitive to the benchmark chosen to measure normal performance. We employ the standard CAPM benchmarks and a variety of APT benchmarks to investigate this question. We found that there is little similarity between the absolute and relative mutual fund rankings obtained from alternative benchmarks which suggests the importance of knowing the appropriate model for risk and expected return in this context. In addition, the rankings are quite sensitive to the method used to construct the APT benchmark. One would reach very different conclusions about the funds' performance using smaller numbers of securities in the analysis or the less efficient methods for estimating the necessary factor models than one would arrive at using the maximum likelihood procedures with 750 securities. We did, however, find the rankings of the funds are not very sensitive to the exact number of common sources of systematic risk that are assumed to impinge on security returns. Finally, we found statistically significant measured abnormal performance using all the benchmarks. The economic explanation of this phenomenon appears to be an open question.