Content area
Full Text
Finance theory tells us that no investment strategy can consistently earn abnormal returns. By abnormal returns, we mean returns in excess of those that should be earned given the level of risk borne (i.e., no "free lunch"). Yet, hedge funds have historically performed better, more or less, than theory would predict or permit. Many of these hedge fund strategies have been based on highly leveraging small arbitrage opportunities, opportunities that in theory should disappear quickly. These returns have often persisted for long periods, but the recent meltdown in global capital and credit markets has exposed even these perennial winners to the lessons of the market. Is there a strategy that appears to defy theory (and gravity) even during tumultuous times? This article examines just such a strategy, based on momentum trading.
With momentum trading, one buys the subset of assets with the highest returns in any period and simultaneously shorts the subset with the lowest returns in that period. The rationale of the strategy is that momentum will cause persistence in these relative returns. This strategy, if successful, is an obvious violation of market efficiency, so one must impose the utmost rigor in testing the strategy to give confidence in the results. We very carefully define abnormal returns with respect to conventional theory, and we conduct our tests over several economic scenarios, including the technology bubble and its bursting, and the housing bubble and its bursting. We conclude that the strategy is robust and does indeed present a pricing puzzle.
The two dominant paradigms in asset pricing are the consumption-based approach of Arrow [1964] and Debreu [1959] and the no-arbitrage approach of Harrison and Kreps [1979] and Harrison and Pliska [1981]. The two are equivalent under complete markets or the existence of a representative agent. Cochrane [2001] presents a unifying approach based on an asset pricing formula derived with fewer assumptions than the two classical models. We review some of the most important results of this theory, such as the bounds on asset return and volatility implied by linear factor models. When hedge funds are included in the sample, the bounds are severely violated, and the violation occurs over a very long period of time for an average of many funds within a particular...