Can Skewness Identify Future Outperforming Mutual Funds

The article “Can Skewness Identify Future Outperforming Mutual Funds” first appeared on Alpha Architect blog.

The annual SPIVA has documented that retail mutual funds underperform with great persistence, with any persistence of outperformance not significantly greater than would be randomly expected. The large body of research on the failure of active management led Charles Ellis to famously call it a “loser’s game”—one that is possible to win, though the odds of doing so are so poor the surest way to win is to not play. By that, he meant to avoid investing in funds that engage in individual security selection and/or market timing (active management).

Despite the large body of evidence against the use of active strategies, given the potential reward, there have been many attempts to find a methodology that would identify future outperformers ahead of time. Much of the effort has focused on active share as predictor. Unfortunately, the evidence, as presented in this article about the predictive power of active share, does not support the hypothesis that active share explains future outperformance. But hope springs eternal.

New Research

Jo Drienko, Chao Gao, and Yifei Liu, authors of the August 2024 study “A Skew is a Skill: Portfolio Skewness of Mutual Fund Holdings,” hypothesized: “Skilled managers are more likely to include winners and exclude losers in their portfolios, thus resulting in higher skewness in the cross-sectional return distribution of their holdings.” To investigate this hypothesis, they constructed portfolio skewness measures (the average monthly portfolio skewness of the past three months) from the cross-section of the holdings in each fund’s portfolio. Their data sample covered U.S. equity mutual funds over the period 1980-2023. Following is a summary of their key findings:

• The return cross-section of a mutual fund’s portfolio holdings is positively skewed on average.
• At the fund level, portfolio skewness varies substantially across funds yet is highly persistent over time.
• High skewness funds tend to be larger, less active, with lower turnover, and lower expense ratios—the last three being positive attributes.
• Actively managed mutual funds with high portfolio skewness outperformed funds with low portfolio skewness by 2.88% on an annualized basis. This association became stronger amid more investment opportunities (higher volatility, higher idiosyncratic volatility, and greater dispersion of returns) in the market.
• Shares added or tilted to by high skewness funds relative to low skewness funds significantly outperformed their counterparts, pointing to stock selection skill as an explanation for both the portfolio skewness and its predictability of fund performance.
• Using two measures of skewness, the highest decile of skewness funds generated statistically significant gross of fees annualized Fama-French-Carhart alphas of 1.238% (t-stat=2.88) and 1.633% (t-stat=2.98) in the nearest future month, while low skewness funds generated statistically negative annualized alphas of -1.148 (t-stat=-2.43) and -1.251% (t-stat=-2.27).

Unfortunately, as the table below demonstrates, the net of fee returns (what investors earn) of the highest skewness funds did not generate statistically significant alphas.

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained.  Indexes are unmanaged and do not reflect management or trading fees, and one cannot invest directly in an index.

It’s also worth noting that the authors found that active share did not show significant alpha predictability.

Their findings led Drienko, Gao, and Liu to conclude:

“We find that funds with high portfolio skewness significantly outperform funds with low portfolio skewness in the future. From a time-series perspective, the outperformance is larger when the investment opportunity is higher, meaning that skilled managers have better responses to investment opportunity innovations.” They added: “We infer that managers of high skewness funds achieve better performance through stock selection rather than active trading….  Portfolio skewness has implications for mutual fund investors as it only relies on limited historical holdings data to construct. It is particularly helpful in identifying funds consistently selecting inferior stocks as these funds significantly underperform, regardless of fund fees.”

Before drawing any conclusions, it is important to note that the authors use portfolio holdings, and the filings of holdings are delayed by 45 days. Thus, investors would not get data for the nearest future month until it had been over for at least 15 days. As an example, consider a fund with Q4 calendar end in December. You need the December holdings to determine whether the fund will do well in January. However, those holdings won’t be available until about Feb 15. Thus, their findings are not very useful for an investor. This issue could be addressed by addressed by lagging the data. Whether the results still hold remains to be seen, although the authors note that their measure of skewness doesn’t change too much from quarter to quarter.

Investor takeaways

Drienko, Gao, and Liu showed that the skewness of fund returns can help investors by identifying funds that are likely to persistently underperform. Unfortunately, being able to avoid persistent losers is not sufficient for successful management as skewness did not allow them to confidently select funds that could outperform. While the skewness metric did demonstrate that it could select funds with managers skilled a security selection, the fund’s expenses and implementation meant that the fund was just about able to cover its expenses, and that was before the negative impact of active management on after-tax returns—and the finding was not statistically significant at even the 10% level of confidence.   

Larry Swedroe is the author or co-author of 18 books on investing, including his latest Enrich Your Future.

Go Source