Where does efficient market hypothesis collide?

Eugene Fama named and formalised the efficient market hypothesis in his 1965 study of stock-price behaviour and his 1970 review in the Journal of Finance, defining an efficient market as one in which prices at every moment fully reflect all available information. The claim is that because any news is impounded into prices almost immediately by competing traders, no strategy using that information can earn returns above the market average on a risk-adjusted basis, so prices move only in response to genuinely new information and are otherwise unpredictable. Fama sorted the idea into three nested forms by the information set at issue: the weak form, where prices reflect the record of past prices; the semi-strong form, where they also reflect all public information; and the strong form, where they reflect private information as well. The hypothesis rests on antecedents in Louis Bachelier's 1900 mathematics of speculation, which modelled price changes as a random walk, and Paul Samuelson's 1965 proof that properly anticipated prices fluctuate randomly. Its reach is what makes it a landmark: it supplies the standard null hypothesis for tests of return predictability across stocks, bonds, currencies, and other assets, and it underpins the case for low-cost index investing over active stock selection.[1][2][3][4]

Unit

A theoretical claim about how fully asset prices incorporate information, stated in three forms graded by the breadth of the information set

Recovery means

Not applicable

Measured by

Tested through event studies, return-predictability regressions, and tests of whether active strategies beat risk-adjusted benchmarks after costs

The counter-camp, which holds that prices can drift far from fundamental value because the arbitrage that is supposed to correct mispricing is itself costly and risky. Shleifer and Vishny's 1997 The limits of arbitrage argued that real arbitrage uses scarce capital, is run by a few specialists on other people's money, and can fail exactly when prices are most wrong, so mispricing survives rather than being competed away. Shiller's 1981 finding that stock prices move far more than later dividend changes could justify gave the camp its founding anomaly of excess volatility. Where the efficient-markets camp reads a predictable return as compensation for risk, this camp reads it as an error that arbitrage cannot fully erase.[5][6]

Unit

A price's gap from fundamental value, and the cost of closing it

Recovery means

Not applicable

Measured by

Documented return anomalies, excess-volatility tests, and models of arbitrage risk and cost

The methodological point that market efficiency can never be tested on its own. Any test of whether prices are efficient is at the same time a test of the asset-pricing model used to say what returns should have been, so a rejection could mean the market is inefficient or that the benchmark model is wrong. Fama called this the joint-hypothesis problem and treated it as a permanent limit on cleanly falsifying efficiency. Grossman and Stiglitz's 1980 On the impossibility of informationally efficient markets sharpened the tension from the other side: if prices already reflected all information, no one would be paid to gather it, so a perfectly efficient market cannot exist in equilibrium. On this view efficiency is a limiting idea, not a state the data can simply confirm or refute.[7][8]

Unit

A test of efficiency bundled with a model of expected returns

Recovery means

Not applicable

Measured by

Not cleanly; only jointly, as return predictability read against an assumed pricing model