Can investors discriminate between lucky and skillful equity fund managers by examining the distribution of returns across fund holdings? In the September 2010 preliminary draft of their paper entitled “Home-Run Sluggers vs. Contact Hitters: Stock Performance Distribution inside Mutual Funds and Fund Managers’ Stock Picking Ability”, Peter Chung and Thomas Kim relate the skewness of the return distribution of equity mutual fund holdings to performance persistence. Specifically, they calculate the skewness of the distribution of four-factor (adjusted for market, size, book-to-market, momentum) alphas of individual fund holdings weighted according to position size. A fund manager who consistently picks outperforming stocks (gets lucky with one big winner) would have a negatively (positively) skewed distribution of alphas. Using reported holdings for 1,604 U.S. equity mutual funds and data to calculate the lagged six-month alphas for each of these holdings from the end of July 2002 through February 2006, they find that:
- Skewness of the distribution of value-weighted lagged alphas of mutual fund holdings relates negatively to fund future performance for up to two years.
- The predictive power of this metric concentrates in currently top-performing funds.
- Among all funds in the sample, those ranked in the bottom 20% of holdings return distribution skewness outperform those ranked in the top 20% by an average 9.6% over the year after funds file holdings (22.5% versus 12.9%).
- Among just those funds ranked in the top 20% of current performance, those ranked in the bottom 20% of holdings return distribution skewness outperform those ranked in the top 20% by an average 9.3% over the year after funds file holdings (27.6% versus 18.3%).
In summary, evidence suggests that mutual fund investors may be able to locate and exploit fund manager skillfulness by periodically: (1) screening for currently top-performing funds; and, (2) eliminating funds for which a “lottery winner” holding drives outperformance.
One caution regarding findings is that the sample period is short in terms of market conditions (roughly one bull market). The skewness-future return relationship could vary with market conditions.
Investors with diverse holding may want to apply this methodology (perhaps simplifying by adjusting for market returns but not other factors) to their own portfolios to assess whether they have been skillful or lucky.
