For almost five decades, the literature on the investment performance of mutual funds has found that very few managers possess sufficient stock-picking or market-timing talent to allow them to consistently and reliably produce positive risk-adjusted performance after considering their fees. In other words, there’s little to no evidence of outperformance beyond the randomly expected.
As my co-author Andrew Berkin and I discuss in our book, “The Incredible Shrinking Alpha,” while perhaps disheartening, this result shouldn’t be surprising given the very high skill level of active managers competing fiercely in a zero-sum game, even before expenses. Thus, investors shouldn’t expect there to be many opportunities for a free lunch.
In addition, because we should expect the scarce resource to earn any “excess returns” that occur (and the ability to generate alpha is far more scarce than investment capital), it is naive to expect that mutual fund managers won’t charge sufficient fees or attract a sufficiently large amount of assets to effectively capture any alpha they generate. Said another way, investors should not expect to be the beneficiaries of the manager’s skill.
Despite the large body of evidence demonstrating that it’s a loser’s game (one that, while possible to win, has odds so poor that it’s not prudent to try), the most common strategy used by both institutional (such as pension plans and endowments) and individual investors to select a fund manager involves hiring outperforming managers and firing underperforming ones.
Buying The Winners
Bradford Cornell, Jason Hsu and David Nanigian contribute to the literature on this strategy with their February 2016 study, “The Harm in Selecting Funds That Have Recently Outperformed.”
The authors note that research on investor behavior has found that, in defiance of the evidence, fund flows “are positively correlated with past performance. Anecdotally, investment consultants and fiduciaries acknowledge that past outperformance is ‘a’ if not ‘the’ dominant manager selection criterion, because it is intuitive and thus defensible to investors.”
To test the strategy, Cornell, Hsu and Nanigian examined whether selecting managers based on recent performance leads to outperformance for investors. Given that three years is the typical investment horizon used by institutional investors in making hiring and firing decisions, they used that horizon in their study.
To simulate the impact of actual decision-making based on track record, they compared the performance of investment policies that involve investing in a “winner strategy” (an equal-weighted investment in the top decile of funds based on benchmark-adjusted returns) to results from a “median strategy” (an equal-weighted portfolio of funds ranked between the 45th and 55th%iles) and a “loser strategy” (an equal-weighted portfolio of the bottom decile of funds).
The authors note: “Funds in the ‘Winner Strategy’ bucket would generally be the funds that are selected by wealth management platforms as part of their buy or recommended list and recommended by financial advisors to clients for consideration. Funds in the ‘Loser Strategy’ bucket would generally be funds that are not on any recommended list and are actively being eliminated from client portfolios by financial advisors.”
They also examined the investment performance produced by the strategy of investing in funds that underperformed their benchmarks by more than 1% per year, and the even more extreme case of investing in funds that underperformed their benchmarks by more than 3% per year. They wondered if doing so would eliminate future bad performance from the portfolio (or, perversely, if it would lead to future outperformance due to mean reversion).
The Winners Lose
Based on the historical evidence that there’s persistence of underperformance among the highest-cost funds, the authors eliminated the funds in the top decile of funds ranked by expense ratio from their sample. The dataset covered the period from 1994 through 2015. Portfolios were formed based on funds’ most recent 36-month performance and rebalanced monthly to maintain equal weighting across funds. At the end of the three-year period, the process was repeated. The following is a summary of their findings:
- The winner strategy produced by far the lowest return (8.0%) and the worst Sharpe ratio (0.29). The returns and Sharpe ratios for the median and loser strategies were 9.8% and 0.48%, and 10.4% and 0.51%, respectively.
- The Carhart four-factor model (beta, size, value and momentum) alphas were -2.7% for the winner strategy (with a t-stat of -3.0), -0.4% for the median strategy (with a t-stat of -0.9) and 0.2% for the loser strategy (with a t-stat of 0.2).
Shockingly—at least for those who believe in using past performance to make hiring and firing decisions—switching from the winner strategy to the loser strategy would have resulted in almost doubling the Sharpe ratio of the portfolio (from 0.29 to 0.51). In addition, the four-factor alphas of the winning strategy were highly negative, at -2.7%, and statistically significant at the 1% confidence level (t-stat of 3.0). In addition, it was 2.9 percentage points below the four-factor alpha of the loser strategy.
This certainly calls into question the belief that past outperformance was the result of any skill. What’s more, paying fees to consultants to help you make these hiring and firing decisions would only add to the negative impact of such a strategy.
Mind The Gap
These findings help explain the well-documented “returns gap” (what my colleague and fellow author Carl Richards called “the behavior gap”) experienced by individual investors. Due to performance chasing, the returns they earn are below the returns of the very funds in which they invest.
Providing more fuel for the fire, Cornell, Hsu and Nanigian found that the strategy of investing in the funds that underperformed their benchmarks by more than 1% actually outperformed the strategy of buying funds that beat their benchmarks by more than 1% (9.8% with a Sharpe ratio of 0.48 versus 8.7% with a Sharpe ratio of 0.37). The losers produced a four-factor alpha of -0.4% (with a t-stat of -0.7), while the winners produced an alpha of -1.7 (with a t-stat of -3.4).
The findings were the same for the strategy of investing in the funds that underperformed by at least 3%. They managed to outperform the funds that had outperformed by at least 3% (10.0% with a Sharpe ratio of 0.48 versus 8.9% with a Sharpe ratio of 0.39). The losers produced a four-factor alpha of -0.2% (with a t-stat of -0.4) while the winners produced an alpha of -1.4 (with t-stat -3.3). It doesn’t get much uglier than that.
To test the robustness of their findings, the authors also examined a 24-month horizon instead of a 36-month one. The results changed little.
Looking At Expense Ratios
In another interesting test, the authors started with a universe of the top 90% of managers ranked by lowest expense ratio and computed their annualized future outperformance from the selection date to the end of the performance reporting sample—in other words, perfect foresight. They then examined the impact of starting with this fund universe and additionally screening funds based on recent performance. Once again, they found the same results.
Equal-weighting the top 25% of winners would have returned 12.3% (with a Sharpe ratio of 0.61). However, then screening for the most recent winners would have lowered returns to 10.6% (with a Sharpe ratio of 0.43) while screening for the most recent losers would have returned 13.5% (with a Sharpe ratio of 0.68).
The authors write: “Even if investors start with a select list of good managers, using recent outperformance to further screen managers is a harmful practice.” And their findings are consistent with those of prior research on the hiring and firing decisions of pension plans and other institutional investors. Consider the following …
Hiring And Firing
Amit Goyal and Sunil Wahal, authors of a May 2005 study, “The Selection and Termination of Investment Management Firms by Plan Sponsors,” evaluated the selection and termination of investment management firms by plan sponsors (public pension plans, corporate pension plans, union pension plans, foundations and endowments).
They built a dataset with the hiring and firing choices from approximately 3,700 plan sponsors from 1994 to 2003. The data represented the allocation of more than $737 billion in mandates to hired investment managers and the withdrawal of about $117 billion from fired investment managers. The following is a summary of their findings:
- Plan sponsors hire investment managers after large positive excess returns up to three years prior to hiring.
- Return-chasing behavior does not deliver positive excess returns thereafter, with post-hiring excess returns indistinguishable from zero.
- Plan sponsors terminate investment managers after underperformance, but the excess returns of these managers after being fired are frequently positive.
- If plan sponsors had stayed with the fired investment managers, their returns would have been larger than those actually delivered by the newly hired managers.
It is important to note that the above results did not include any of the trading costs that would have accompanied transitioning a portfolio from one manager’s holdings to the holdings preferred by the new manager. The bottom line: All of the activity was counterproductive.
Cornell, Hsu and Nanigian drew two main conclusions: “First, a heuristic of hiring recent outperforming managers and firing recent underperforming managers turns out to be 180 degrees wrong …. Second, consistent with previous research, it appears superior investment performance is more a function of the systematic exposures (a persistent investment style) that managers embed into the portfolio, not some nebulous talent—elusive and unique alpha skill.”
Their findings clearly present a tremendous challenge and problem to investors who base decisions on past performance. The practical implication is that investors should clearly focus on factors other than past performance when selecting fund managers.
Thus, the logical conclusion should be that the strategy most likely to allow you to achieve the best results is to focus instead on the selection of passively managed funds that provide you with the desired amount of exposure to the well-documented factors that explain the differences in returns of diversified portfolios—factors such as beta, size, value, momentum and profitability/quality for equities and term and default for bonds—and do so in a low-cost and, for taxable investors, tax-efficient manner.
Unfortunately, the belief in active management as the winning strategy is so deeply embedded for so many investors that, faced with such evidence, the likely outcome is that they will experience cognitive dissonance. As a result, the evidence is likely to be ignored because facing up to it would be too painful an admission of belief in a false theory.
Larry Swedroe is the director of research for The BAM Alliance, a community of more than 140 independent registered investment advisors throughout the country.