“Measuring the Level and Persistence of Active Fund Management” (pro) and “Fund Activeness Predicts Performance?” (con) summarize debate on the ability of Active Share, how much portfolio holdings differ from a benchmark index, to predict mutual fund performance. The authors of the con paper summarized in the latter (principals of AQR Capital Management) assert that “neither theory nor data justify the expectation that Active Share might help investors improve their returns.” In his June 2015 paper entitled “AQR in Wonderland: Down the Rabbit Hole of ‘Deactivating Active Share’ (and Back Out Again?)”, Martijn Cremers rejoins the debate by examining the methodology and motives of the con paper. Using data on active U.S. equity mutual funds from the original research, and holdings/performance data for seven AQR Capital Management funds offered to retail investors that concentrate in U.S. stocks as available through December 2014, he finds that:
- The assertion in the con paper relies on sorting funds by benchmark and assessing the predictive power of Active Share separately by benchmark, finding predictive power for only eight of 17 benchmarks. However:
- The eight affirming benchmarks are among the most widely used, comprising 72% of the data.
- The non-affirming benchmarks comprise only 28% of the data. Parsing data into many small samples: (1) changes which funds are designated high-Active share; and, (2) makes statistical reliability inherently elusive, especially because the relationship between Active Share and future fund performance is non‐linear.
- Weighting results across benchmarks by strength of findings strongly supports belief in the predictive power of Active Share. (Even findings from the con methodology support some belief, but not at conventional academic confidence levels.)
- Regarding AQR Capital Management U.S. equity mutual funds:
- Average Active Shares are the range 48%-67%, among the lowest 20% of similar funds either overall or by benchmark.
- The ratio of average expense ratio to Active Share is fairly high compared to similar funds.
- As a group across different ways of assessing performance, these funds do not meaningfully beat their benchmarks.
In summary, evidence indicates that Active Share, especially applied as originally specified, can help investors choose good mutual funds.
A possible framing of this debate is as a contest between research and marketing.
Cautions regarding findings include:
- In general, use of indexes as benchmarks ignores the costs of maintaining liquid tracking assets and so overstates achievable benchmark returns.
- As implied, investors should take care in assessing arguments offered by those selling something.
- Both investment managers and researchers are selling something.
