Working Paper: NBER ID: w26707
Authors: James J. Choi; Kevin Zhao
Abstract: A seminal study of persistence in mutual fund performance is Carhart (1997), who found that U.S. equity mutual funds’ past-year returns positively predict their raw excess return and one-factor alpha over the next year. Based on these results, an investor may believe that she can earn higher returns by buying mutual funds with high past-year returns. We are able to replicate Carhart’s results in his 1963-1993 sample period, but we find that significant performance persistence does not exist in the 1994-2018 period. Even during the 1963-1993 period, performance persistence weakened in later years. The disappearance of significant performance persistence is due to lower returns to favorable styles, as well as less favorable style tilts and increased style-adjusted underperformance by past winning funds.
Keywords: No keywords provided
JEL Codes: G11; G12; G23
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Past-year returns of U.S. equity mutual funds (G23) | Raw excess return over the subsequent year (G17) |
| Top decile of one-year returns (G11) | One-factor alpha over the subsequent year (C29) |
| Bottom decile of one-year returns (G12) | One-factor alpha over the subsequent year (C29) |
| Top decile of one-year returns (1963-1993) (G12) | Significant difference in raw excess returns vs bottom decile (G19) |
| Performance persistence disappears (1994-2018) (D29) | Top decile no longer earns significantly positive one-factor alphas (D33) |
| Difference in one-factor alphas between top and bottom deciles (1994-2018) (D33) | Positive but not significant (C29) |
| Decline in performance persistence (D29) | Changes in fund style and performance of those styles (G23) |