I read the review of the paper by Professors Cornell and Roll (Pensions & Investments, April 4). To readers who have managed funds, being told the capital asset pricing model is very limited is probably not news (as the traditional CAPM assumes that the entire market is made up of people who make their own investment decisions, or "principals," and ignores the bulk of your readership, or "agents"). The implications, however, are more profound than what has been highlighted by the academic research.
First, individuals in charge of managing pension assets are simultaneously principals to their external managers and agents of their boards. The academic research has yet to explore this complex facet, which impacts optimal portfolio structure and asset prices. Managers using the traditional CAPM to value assets or structure portfolios could be making a big mistake.
Second, the typical risk-adjusted performance measures, such as the Sharpe ratio or information ratio, derived from the simple CAPM are therefore not appropriate. Moreover, they do not tell clients enough about the skill of the manager/agent, something principals worry a lot about. I developed a risk-adjusted performance measure called the M-cube (an extension of Professor Franco Modigliani and Leah Modigliani's M-square) that attempts to correct for this problem, though we do not have a theoretical asset-pricing model to which to anchor this performance measure.
Finally, in all this talk about separating "alpha" from "beta," there is too much emphasis on targeting the "alpha." Few academics, with possibly the exception of Woody Brock, are telling clients that dynamically "managing the beta" may be the most important decision as the CAPM is wrong. As Professor Brock points out, we have to go away from a paradigm of an optimal investment portfolio to a paradigm of optimal dynamic investment strategies.
The CAPM may not be dead, but the challenge for academia is to capture the complex "principal-agent" relationships in this industry to give us a more general framework to help clients develop optimal investment strategies, risk-adjusted performance measures and asset pricing models.
Mcube Investment Technologies LLC