Pay For Performance In Money ManagementBradford Cornell
On September 21, 2018, the Wall Street Journal ran an article entitled “Fund to Cut Pay for Market Laggards.” The articles described how Peter Kraus, the ex-CEO of successful investment research and money management firm, AllianceBernstein, was leaving to start a new firm, Aperture Investors. The paper reported that Aperture builds on the concept that Mr. Kraus championed in his last months at AllianceBernstein: Money managers should only charge higher fees than exchange traded funds when they beat the market. The idea sounds great in principle, but there are two hurdles that make it virtually impossible to operationalize in practice.
The first hurdle is defining what it means to be the market. Finance scholars have recognized for nearly a century that risk and return are related. An investor who holds a riskier portfolio, properly defined, can expect to beat the market but that is not really superior performance it is just added risk bearing. Therefore, beating the market must be defined on a risk adjusted basis. Easier said than done. For the last fifty years finance scholars have been arguing about how to properly measure risk. One solution is simply to ignore the problem and compare investment performance to market indexes without risk adjustment. But that produces incentives to bear added risk and to not diversify properly – hardly incentives that most asset owners would want a money manager to have.
But there is a bigger problem than risk adjustment. The volatility of stock prices implies that virtually all short-term performance reflects luck rather than skill. When asked what can one learn from one, three or even fiver years of past returns about a manager’s skill or the future performance of an investment, Nobel Prize winner Eugene Fama answered, “The short answer: usually almost nothing.” He went on to say, “This is especially true for active managers since active management almost by definition means low diversification and high volatility of unexpected returns (noise).” Despite this fact, when performance metrics are reported in the financial press and elsewhere they invariably involve horizons of five years or less and generally three years or less. Pay for performance over those time horizons is not really pay for performance. It is pay for serendipity.
Article by Brad Cornell's Economics Blog