[1] The proposition that financial markets are informationally efficient—that prices at any moment fully reflect all available information—dominated academic economics for much of the later twentieth century. [2] In its strongest form the hypothesis implies that no investor, however diligent, can consistently outperform the market by analyzing public information, since any exploitable pattern would already have been competed away. [3] The theory drew much of its appeal from a compelling logical foundation: if predictable profits existed, rational actors would move to capture them, and in doing so would erase the very predictability that made them possible.
[4] By the 1980s, however, an accumulation of empirical anomalies had begun to strain this elegant framework. [5] Researchers documented patterns—the tendency of stocks with low price-to-earnings ratios to outperform, or the persistence of momentum over intermediate horizons, or the tendency of small companies to yield higher returns than their risk seemed to warrant—that appeared difficult to reconcile with a market that had already priced in all relevant information. [6] A parallel body of work, drawing on experimental psychology, contended that investors are subject to systematic biases: they overreact to salient news, extrapolate recent trends unduly, weigh losses more heavily than equivalent gains, and cling to prior beliefs in the face of contrary evidence. [7] From this behavioral perspective, prices could depart substantially and persistently from underlying values.
[8] Defenders of efficiency were not without replies. [9] Some contended that the documented anomalies were artifacts of flawed statistical method, likely to evaporate once discovered and traded upon—as several indeed appeared to do. [10] Others maintained that apparent excess returns were merely compensation for risks the researchers had failed to measure, so that what looked like a free lunch was nothing of the kind. [11] The dispute proved difficult to adjudicate precisely because the efficient-market hypothesis, absent an agreed model of what compensation various risks ought to command, is nearly impossible to test in isolation; a finding that appears to refute efficiency can always be reinterpreted as evidence that the assumed model of risk was wrong.
[12] The result has been less a decisive victory for either camp than a wary accommodation. [13] Few economists now defend the strong claim that prices are always correct, yet fewer still deny that markets impose a genuine discipline on those who would trade against them. [14] The prevailing view holds that markets are neither perfectly efficient nor reliably foolish, but rather difficult to beat—an assessment that, while less tidy than its predecessors, may prove more durable.
