Redefining Active Risk
Isolates Skill and Reveals Closet-indexers
Passive beta differences with a benchmark are a byproduct, typically unintentional, of any stock-selection process. Since consistent passive differences, once properly identified, can be freely obtained or offset, they are not part of active contribution.
Active contribution is not simply performance relative to a particular benchmark, it is instead only that portion of incremental return which could not have been obtained passively.
The true passive alternative to an active portfolio is rarely a single index, but rather a combination of index funds with different market, sector, and style exposures.
In fact, defining active contribution as performance relative to a single index, when that performance can be achieved with a combination of passive index funds and ETFs at a fraction of the cost, is economically unsound and increasingly risky from a legal and regulatory perspective.
Current methods of separating active from passive performance are insufficiently robust and provably unable to distinguish between portfolios taking genuine security selection risk and those merely taking high systematic market, sector, or macro risk.
Evidence that existing analytics fail to accurately measure risk exposures — and so cannot isolate active from passive performance — is their failure to predict future performance.
This is a serious limitation as passive differences obscure active skill and inflate active risk.
Statistical factor risk models, using only passively-investable factors, precisely measure all current risk exposures (region, market, sector, and style betas, and all idiosyncratic security risk exposures) that together fully predict future performance.
The validity of any model or approach is proven by the accuracy of its predictions. Our domestic and global risk models show 0.965 median correlation between returns predicted by passive exposures and subsequent fund returns.
We suggest users only trust analytics they can test and prove independently.
One-third of active mutual funds take too little active risk to ever overcome fees, even if highly skilled.1
Popular methods of measuring active risk (Active Share, tracking error, downside deviation, Sharpe, Sortino, and information ratios, etc.) fail to identify most closet indexers because they confuse a single-index benchmark with the portfolio’s passive alternative, leaving asset owners exposed to the risk of paying active fees for passive contributions.2
Statistically persistent skill
Properly isolated from passive market noise, security selection skill shows strong statistical evidence of persistence.
Managers with top decile security selection information ratios are twice as likely to outperform as not over the subsequent three years. Managers in the bottom decile are more than twice as likely to underperform as not.3
Portfolio structure and Qualitative manager assessment
Managers with complementary exposures can be combined, or passive ETFs can be incorporated, to offset unintended passive differences with the aggregate benchmark, reducing relative risk while retaining the active risk worth paying for.
Manager discussions can focus on those exposures that will drive future performance as well as any unexplained changes in risk exposures or potential conflicts with mandate.
If you’d like to take a look at some of your own managers, just email a few names to michele@peeranalytics. We’d appreciate any feedback.
* 0.97 median correlation between predicted and subsequent realized returns
1. mutual fund closet indexing
2. testing Active Share
3. performance persistence within style boxes and testing predictions of equity risk models
white paper: passive-factor risk models built for oversight