December 2019 SPIVA Persistence Scorecard
Since 2002, S&P Dow Jones Indices has published its biannual Indices Versus Active (SPIVA) reports, which compare the performance of actively managed equity funds to their appropriate index benchmarks. It also puts out a pair of scorecards each year that focus on persistence of performance. Persistence is an important issue because if it is not significantly greater than should be expected at random, investors cannot separate skill-based performance (which might be able to persist) from luck-based performance (which eventually runs out).
Following are some of the highlights from the just-released December 2019 Persistence Scorecard, with data through September 2019:
- Irrespective of asset class or style focus, few fund managers consistently outperformed their peers.
- An inverse relationship exists between the length of the time horizon and the ability of top-performing funds to maintain their success. Less than 1% of equity funds maintained their top-quartile status at the end of the five-year measurement period. And no large-cap fund was able to consistently deliver top quartile performance by the end of the fifth year.
- Over long horizons, most fixed income categories showed no persistence: Of the 13 fixed income categories, seven showed no managers remaining in the top quartile over the five-year horizon.
- Of 545 domestic equity funds that were in the top quartile as of September 2017, 8.1% managed to stay in the top quartile through September 2018 and again through September 2019, slightly more than the 6.3% we would randomly expect to do so.
- Over the two non-overlapping five-year periods, it was almost as likely for a bottom-quartile large-cap fund to rise to a top-quartile performer (9.6%) as it was for a top-quartile performer to remain a top-quartile performer (10.6%). The results were similar for small-cap funds, with 18.8% of top-quartile performers becoming bottom-quartile performers while 16.1% of bottom-quartile performers became top-quartile performers. The results were better for mid-cap and multi-cap funds.
- Over the five-year horizon, the results show a lack of persistence among nearly all the top-quartile fixed income categories, with a few exceptions.
While the evidence demonstrates that historical performance is only randomly associated with future performance, many institutional investors (such as endowments and pension plans) still engage in the practice of selling funds or firing managers once they have underperformed the market over the previous three years, typically replacing them with funds or managers that have recently outperformed. They do this while ignoring the evidence that past performance has almost no predictive value and the research that found the hired managers go on to underperform the fired managers.
An explanation for this strange outcome involves evidence of mean reversion in mutual fund performance. Robert Arnott, Vitali Kalesnik and Lillian Wu, the authors of the January 2018 study The Folly of Hiring Winners and Firing Losers, found it can be explained by the tendency of underperforming managers to hold cheaper assets with cheaper factor loadings, setting them up for good subsequent performance, whereas recently winning managers tend to hold more expensive assets.
The SPIVA scorecards provide powerful evidence regarding the persistent failure of active managers to outperform. They also provide compelling support for Charles Ellis’ observation that while it’s possible to win the game of active management, the odds of doing so are so poor it’s imprudent to try – which is why he called it “the loser’s game.”
Larry Swedroe is the director of research for The BAM Alliance, a community of more than 140 independent registered investment advisors throughout the country.
 When measuring the persistence of performance over a five-year period, S&P looks at each year sequentially. For a five-year period, a “persistent” fund manager must be in the top quartile in the first year, and remain there in the second, third, fourth and fifth years. The odds of that happening randomly are 25% in each year; the odds of that happening randomly for all five years sequentially is approximately 0.4% (25% to the fourth power).