THE PROBLEM

Increasingly recognized as an industry shortcoming is that little is actually known about skill – what it looks like, how it should be measured, or who has it.1 Sure, there are mountains of academic and practitioner papers discussing skill. And every manager search or review includes some commentary about skill. These assessments rely upon well-established metrics such as relative return, style, active share, information ratio, batting average/slugging ratio, and multi-factor alpha. These and other conventional analytics deliver tremendous insights into a fund’s results and the risk/return tradeoffs used in generating those results. Unfortunately, conventional analytics offer scant information about skill itself. They hint at its presence or absence – pretty much stopping there. The underlying issue is that conventional analytics use as inputs either a fund’s return series or its history of daily holdings. These data are themselves fund outcomes. They do not contain the information necessary to identify and quantify specific skills. A frequently used simile is that attempting to identify manager skill using fund returns is like trying to gauge a tennis player’s ability to serve balls knowing only how many games the player won and lost. Intuition may nudge you toward formulating a judgment but it is based more on guessing and/or desire than analysis.