Market Efficiency in an Irrational World

We discuss why investors are likely to be overconfident and how this behavioral bias affects investment decisions. Our analysis suggests that investor overconfidence can generate momentum in stock returns and that this momentum effect is likely to be strongest in those stocks whose valuations require the interpretation of ambiguous information. Consistent with this analysis, we found that momentum effects are stronger for growth stocks than for stable stocks. A portfolio strategy based on this hypothesis generated strong abnormal returns from U.S. equity portfolios that did not appear to be attributable to risk. Although these results violate the traditional efficient market hypothesis, they do not necessarily imply that rational but uninformed investors could have actually achieved the returns without the benefit of hindsight. To examine whether unexploited profit opportunities exist, we tested for a somewhat weak form of market efficiency, adaptive efficiency, that allows for the appearance of profit opportunities in historical data but requires these profit opportunities to dissipate when they become apparent. Our tests rejected the notion that the U.S. equity market is adaptive efficient. The psychology literature describes numerous behavioral biases. Potentially, these biases can explain almost any pricing anomaly imaginable. We believe the most important is overconfidence, and we concentrate on how it affects decisions and stock prices and how it results in markets that are not efficient.Numerous experimental studies have demonstrated the pervasiveness of overconfidence and its deleterious effects on decision making. That individuals tend to be overconfident should not be surprising; overconfidence endows people with clear advantages. Individuals who exude confidence are likely, for example, to more successfully woo mates—and clients. Thus, they should prosper and compose a large portion of the investor population. But overconfidence may hinder a person's ability to make good investment decisions. Because psychology studies suggest that individuals are likely to be the most overconfident in their abilities when they are evaluating information that is the most vague, we hypothesized that investors will be especially overconfident about their abilities to evaluate the stocks whose values depend more heavily on future (therefore, vague) growth options than about their ability to evaluate the stocks of stable companies. This behavior would open the door to pricing anomalies.To test this hypothesis, we examined the performance of a portfolio strategy for the July 1963 to December 1997 period that dictated buying the stocks of (U.S.-listed) companies with high book-to-market values that also had positive momentum and shorting the stocks of companies with low book-to-market values and negative momentum. The strategy produced strong abnormal returns that do not appear to be related to risk. We show that uncovering excess returns of this magnitude would be unlikely, even with considerable data mining, if markets were efficient.To test whether the returns would have been available to an investor at the time, one who did not enjoy the benefits of perfect hindsight, we tested for what we call “adaptive efficiency” in the market. Adaptive efficiency is a somewhat weak form of market efficiency that allows for the appearance of profit opportunities in historical data but posits that these opportunities will dissipate when they become apparent.To test this notion, we carried out a purely mechanical trading strategy in the sample period that each year followed investment styles that had performed well in the previous 10 years. This adaptive strategy did exceptionally well, and contrary to the predictions of adaptive efficiency, the excess returns it generated showed no sign of diminishing over time. Indeed, the strategy generated positive profits each year from 1982 through 1997.