Jay D. Franklin

Just How Skilled of a Manager-Picker Do You Need to Be?

At Index Funds Advisors, Inc. we often refer to the proverbial monkey throwing darts at the stock pages of the Wall Street Journal and how active fund managers as a group have had about the same overall success as the monkey at beating the market. Case in point, the average alpha (the return of a portfolio beyond what can be explained by exposure to risk) of all the mutual fund managers in the CRSP (Center for Research in Security Prices) database was -1.5% for the decade ending 12/31/2010. It is no coincidence that this 1.5% shortfall is about equal to the average expense ratio  of all actively managed mutual funds based on William Sharpe’s paper, The Arithmetic of Active Management¹, which concludes that the expected value of active alpha is zero before costs and negative after costs by the amount of the costs.  Despite all the evidence and logic to the contrary, all too often we hear the refrain that goes along the lines of, “I’ll do better than average because I will find the best managers.” Hope springs eternal--especially since a whole fund rating industry exists just to assist investors in this fool’s errand.

Professor Jim Davis of Dimensional Fund Advisors asks the very interesting question of just how much skill must a manager-picker possess to have at least a 50% chance of achieving positive alpha? Davis sets up a hypothetical situation where 10% of managers have true skill (i.e., they are capable of delivering alpha of 1.5%), while the remaining 90% are unable to cover their costs and thus have an alpha of -1.0%. Given the CRSP result of -1.5% for all managers over the last ten years, these are generous assumptions. Now suppose we delegated the job of manager-picking to our monkey who would throw darts at a list of managers rather than a list of stocks, then our expected alpha would be -0.75% (assuming that 10% of the monkey’s picks are skilled). Alternatively, we can ask our would-be manager-picker to do the job, based on his in-depth analysis of Morningstar and Lipper data. Throwing in the additional assumption that that a manager’s realized alpha will vary randomly from his true alpha (i.e., truly skilled managers can be afflicted with bad luck and truly unskilled managers can be blessed with good luck), Davis calculates that our manager-picker must have at least a 40% accuracy rate in his identification of skilled managers  (four times better than what could be expected from chance) in order have a measly 50% chance of achieving positive alpha.

The bottom line is that if you are going to play the mug’s game of manager-picking, you better be sure that you are really good at it. The better option, or course, is not to play and instead invest in a risk-appropriate portfolio of globally diversified index funds.

¹Sharpe, William, 1991. The arithmetic of active management. Financial Analysts Journal, January-February, 1991, 7-9.