By LARRY SWEDROE
One of the advantages active managers tout is that they have the opportunity to adjust factor exposures, taking advantage of regime shifts. Is that ability an advantage, or just one that is fraught with opportunity?
Manuel Ammann, Sebastian Fischer and Florian Weigert contribute to the literature on factor investing with their study Factor Exposure Variation and Mutual Fund Performance, published in the Fourth Quarter 2020 issue of the Financial Analysts Journal.
Ammann et al. investigated the relationship between a mutual fund’s variation in factor exposures and its future performance. They estimated a fund’s dynamic exposures to the factors of the Carhart four-factor model — the market (MKT), size (SMB), book-to-market (HML) and momentum (UMD) factors.
They then measured a fund’s factor exposure variation by the volatility of the factor and created a factor exposure variation indicator (FEV). Using this measure, they investigated whether performance differences exist between funds with high FEV and funds with low FEV in a large sample of U.S. equity mutual funds in the period from late 2000 to 2016. Following is a summary of their findings:
— Factor timing is particularly prevalent among funds with long management tenure, high turnover and high total expense ratios. — “Mid Cap,” “Small Cap” and “Micro Cap” funds tend to have less stable factor exposures than “Growth,” “Growth and Income” and “Income” funds. — FEV seems to be prevalent in different market situations and periods of economic booms and recessions. — Funds with volatile factor exposures underperformed funds with stable factor exposures by a statistically significant (at the 1 percent confidence level) 147 basis points per annum.
— Sorting funds on individual MKT, HML or UMD factor exposure variation resulted in underperformance of the most volatile funds by 102, 82 and 120 basis points per annum, respectively, with statistical significance at least at the 5 percent level. — The abnormal returns monotonically decreased in market, value and momentum exposure variation as well as in the overall FEV Indicator. — A one-standard-deviation increase in factor exposure variation reduced abnormal future returns by 71 basis points per annum. — Differences between high FEV and low FEV funds remained statistically and economically significant when using the Fama and French five-factor (beta, size, value, investment and profitability) model plus the momentum factor for the computation of FEV. The results were similar using other factor specifications — the Frazzini and Pedersen (2014) betting against beta factor, the Baker and Wurgler (2006) sentiment factor, or the Pástor and Stambaugh (2003) liquidity factor.
— The underperformance is neither explained by volatile factor loadings of a fund’s equity holdings nor driven by a fund’s forced trading through investor flows.
— The results were confirmed by various tests of robustness.
Their findings led Ammann, Fischer and Weigert to conclude: “Fund managers voluntarily attempt to time factors, but they are unsuccessful at doing so.”
They added: “Our findings do not support the hypothesis that deviations in factor exposures are a signal of skill and we recommend that investors should carefully take our results into account before investing in funds with high FEV.”
The bottom line is that another myth (along with timing the market in general) about active investing has been exposed.