Working Paper: NBER ID: w18050
Authors: Martijn Cremers; Antti Petajisto; Eric Zitzewitz
Abstract: Standard Fama-French and Carhart models produce economically and statistically significant nonzero alphas, even for passive benchmark indices such as the S&P 500 and Russell 2000. We find that these alphas arise primarily from the disproportionate weight the Fama-French factors place on small value stocks, which have performed well, and from the CRSP value-weighted market index, which is historically a downward-biased benchmark for U.S. stocks. We propose small methodological changes to the Fama-French factors to eliminate the nonzero alphas, and we also propose factor models based on common and tradable benchmark indices. Both kinds of alternative models improve performance evaluation of actively managed portfolios, with the index-based models exhibiting the best performance.
Keywords: Performance Evaluation; Benchmark Indices; Alpha; Fama-French Model; Carhart Model
JEL Codes: G11; G14; G23
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Fama-French and Carhart models (C58) | nonzero alphas for passive benchmark indices (G12) |
| methodological choices (C90) | positive alphas for S&P 500 and Russell 1000 (G12) |
| overweighting of small value stocks (G41) | positive alphas for S&P 500 and Russell 1000 (G12) |
| inclusion of non-US stocks (G15) | positive alphas for S&P 500 and Russell 1000 (G12) |
| adjusting factor models (C51) | significantly lower index alphas (C43) |
| seven-factor index model (C38) | more accurate reflection of fund performance (G23) |
| model construction choices (C52) | inflated or deflated performance metrics (E31) |
| negative alphas for small-cap indices (G12) | biases in performance evaluation models (C52) |