EXECUTIVE SUMMARY
The predictive power of Active Share in respect of future performance has been debated in the investment industry for over a decade. Active Share is a holdings-based measure of a portfolio’s difference from benchmark. It’s a necessary condition for outperformance because if a portfolio is not materially different from benchmark, it cannot perform differently. But different does not mean better: high Active Share can result in underperformance or outperformance.
While there is little evidence that Active Share on its own can predict future performance, combining it with other measures has shown some promise. In a 2014 paper1 , Gillman, Khusainova and Mier (“GKM”), combined Active Share with portfolio concentration (measured by Concentration Coefficient2 ) using the Fundamental Law of Active Management (Grinold, 19893 ). They found this approach had some predictive power for US and international equity portfolios during 2009-2013, although not for the financial crisis years of 2007-8. More recently, our 2019 paper introduced Active Share Risk Profiles4 . As well as providing granular data on the components of Active Share, the data in this research used a different portfolio universe (global equity) and a longer time period (4Q 2009-3Q 2019) compared to the 2014 study. Based on this new data, we address two questions:
1. Does the latest data support similar predictive power to the 2014 findings using the Fundamental Law (combining Active Share and portfolio concentration)?
2. Can we improve the predictive power of Active Share when we include other measures, including its various components as researched in the Active Share Risk Profile paper?
CONCEPT
For the impatient reader, the short answers to those two questions are yes and no!
We did find that using Active Share within the context of the Fundamental Law provided some predictive power for subsequent performance, but just as in the 2014 study, the statistical link was modest.
We then investigated possible links between the components of Active Share (e.g. country, sector and stock-specific exposures) and future performance. As well as testing these individually, we looked at combinations including portfolio turnover and concentration.
We included these two metrics to test the hypothesis that high concentration and long-horizon (low turnover) would provide predictive power if combined with high Active Share. However, none of these tests found any predictive power for subsequent outperformance.
So the “magic bullet” that links Active Share (and related metrics) to future performance is still elusive. We must acknowledge the possibility that it does not exist.
While high Active Share is a necessary condition for outperformance, we have no significant proof that it is sufficient…..yet! As such, there seems no definitive support for marketing claims that high Active Share is associated with future outperformance.
We admit that our data sets are limited, both in number of equity portfolios and in length of period. The 2014 study included 174 U.S. and international portfolios,5 and the 2019 paper included 46 global portfolios. It’s possible that a larger number of portfolios might produce a different conclusion, although we would be surprised by this.
In our opinion, it’s more likely that a different conclusion might come from examining different and longer time periods. The 2009-2018 period shows only limited predictive power. A much longer set of data might show that the modest apparent link in the last decade was “noise.”
Alternatively, it might reveal that a stronger link exists in other periods than was apparent in the period we tested. We would be happy to share our data with others who might want to address this topic.
Continue reading here.
