Is Your Portfolio's High Active Share Really High?
18 Pages Posted: 29 Aug 2017
Date Written: June 9, 2017
Active Share ("AS") is increasingly used in the investment industry to measure the degree to which a manager makes active decisions. Some investors seem to fall into the logic trap of equating high AS with an automatic expectation of future outperformance. Being different from a benchmark certainly increases the probability that results will differ from that benchmark. But that just means the likelihood is greater that a portfolio will either materially outperform or underperform its index before fees.
But even in that context, there's an important first step that most investors seem to miss: what is high AS? Note that AS is a relative measure for any portfolio, not an absolute, because its calculation is dependent on the benchmark used. In this paper we show that the properties of the benchmark should be taken into account when assessing AS, otherwise there's a risk of misinterpreting this information. In this context there is no universal answer to whether AS is high or low. In particular, for a concentrated benchmark, it can be very hard to push AS to levels that are generally perceived as "high" without tilting the portfolio substantially to small cap stocks, which then may make the benchmark unsuitable. In this research, we provide a practical framework and method for investors to interpret AS correctly, and to compare different managers in order to make appropriate decisions in their portfolios.
The original developers of the Active Share concept, Cremers and Petajisto ("C&P"), showed a link between high AS and outperformance based on their research into a broad US mutual fund universe. This was documented in their 2009 paper "How Active is Your Fund Manager? A New Measure That Predicts Performance."
As the popularity of AS has grown, investors have typically taken the ranges of AS from that research and assumed they were universal. For example, C&P's research showed their top quintile of AS scores approximately corresponded to above 90%. So investors may believe that AS over 90% is a prerequisite for outperformance. Certainly an AS over 90% can be called high against almost any benchmark, but as we'll show, even a 90% AS can have very different meanings depending on the benchmark.
Subsequent research by C&P and various other entities has noted that AS varies significantly across differing universes, and varies most with capitalization. The argument has been made that the link between AS and outperformance can be explained by other factors such as small cap bias, given that small cap portfolios tend to have significantly higher AS than larger cap portfolios.
Our research shows that it is the specific benchmark that substantially defines the range of AS for portfolios managed against it. This leads us to provide a straightforward method for investors to evaluate the AS of any portfolio. Based on the construction of the benchmark, it is then possible to determine whether from a relative perspective AS can be viewed as high, low or somewhere in between, given the conditions imposed by that choice of benchmark index.
We also look at value for money in the context of what level of fees investors are paying for the level of AS. In theory one could argue that higher levels of AS should justify, or be associated with, higher fee levels. The original C&P work found little evidence to support this argument. With the ability to identify more appropriate peer groups to compare AS against fees, we can revisit the question of whether investors paying higher fees tend to get higher AS.
Keywords: active, share, AS, outperformance, underperformance, value, investing
Suggested Citation: Suggested Citation