Every time someone comes up with a “new and improved” way to invest or predict which active managers will shine, it seems that subsequent research finds that it’s really just about focusing on different betas. The latest example is a new study (“Deactivating Active Share”) from AQR, a quantitative money manager, which documents that it’s (still) all about beta when it comes to looking for clues about which active managers are more likely to deliver statistically significant alpha.
Under scrutiny is a methodology known as Active Share, which was outlined in the widely discussed 2009 paper “How Active is Your Fund Manager? A New Measure That Predicts Performance”, by Martin Cremers and Antti Petajisto (C&P 2009). The basic idea is that funds that strayed further from a benchmark, the higher the potential to deliver market-beating results (alpha). According to C&P 2009, mutual funds with high Active Share rankings have a history out of outperforming their benchmarks. The key message: avoid the closet indexers and load up on strategies with high Active Share ranks.
For some managers, a high Active Share measure is a badge of honor, as it seems to show a fund’s dedication to raising the odds that it will beat beta. But as the AQR analysis shows, reality is more nuanced. “Overall, our conclusions do not support an emphasis on Active Share,” the AQR authors write. “Predicting investment performance is difficult and there do not seem to be any silver bullets.”
Consider, for instance, how Active Share statistics compare when sorted by benchmark, as per AQR’s analysis. As the chart below shows, high Active Share ranks among equity funds are concentrated in portfolios with a small-cap bias.
“This presents a clear problem,” the AQR authors explain.
[Research] papers that sort funds on Active Share (as do C&P) end up sorting funds on their benchmarks. In practice, few investors would evaluate all managers on a particular dimension and then accept whichever benchmark falls out. Instead, they would start with a benchmark and select a manager from within that benchmark.
