Three claims, not one
Eugene Fama's 1970 review distinguished three forms. The weak form says past prices are already reflected, so chart-reading cannot work. The semi-strong form says all public information is priced, so reading the news cannot work. The strong form says even private information is priced, so insiders cannot profit — and that one is false, which is why insider trading is illegal and occasionally prosecuted.
The remaining two forms are supported well enough that index funds beat most active managers over long periods, a result replicated across countries and decades. The average active manager underperforms after fees by roughly the fee.
The Grossman–Stiglitz paradox sharpens the point: if prices were perfectly efficient, nobody would be paid to gather information, so nobody would gather it, so prices could not be efficient. Some inefficiency is the price of the information that removes it.
The anomalies and the joint hypothesis
Momentum — buying recent winners and selling losers — has survived decades of out-of-sample testing. Post-earnings-announcement drift persisted long after it was published. The value premium, small-cap effects and the low-volatility anomaly are all documented patterns that a strict reading does not predict.
Fama's answer is the joint hypothesis problem: any test of efficiency is also a test of the asset-pricing model used to define 'abnormal' returns. If your model is wrong, your anomaly is a modelling error rather than free money. That is not a dodge — it is a genuine identification problem, and it means efficiency is partly unfalsifiable as usually stated.
The practical resolution is that anomalies shrink after publication, which is exactly what efficiency predicts: the opportunity is real until enough people know about it.
Weak form: strongly supported. Simple technical rules do not produce reliable risk-adjusted returns.
Semi-strong form: broadly supported, with documented exceptions that are small, unstable and costly to trade.
Strong form: clearly false, and nobody serious defends it.
Behavioural finance accepts most of this and explains the exceptions through limits to arbitrage, not through investor stupidity — which is why the two camps have converged more than the headlines suggest.
The archive labels this live debate because the dispute is about the size and persistence of mispricing, not about whether markets are efficient in some absolute sense.
Why it is in the archive
Because 'the market is efficient' is used both as a serious scientific claim and as a slogan for deregulation, and the difference is what this page is for.
Related: bitcoin as digital gold, black swans.
Sources for this page
- Wikipedia, “Efficient-market hypothesis” — the three forms and the evidenceen.wikipedia.org
- Fama (1970), Journal of Finance — “Efficient Capital Markets: A Review of Theory and Empirical Work”doi.org
- Grossman & Stiglitz (1980), American Economic Review — the impossibility of informationally efficient marketswww.jstor.org
- Wikipedia, “Joint hypothesis problem” — why efficiency is hard to testen.wikipedia.org