The Efficient Market Hypothesis — What It Says and What It Doesn't

The Three Forms and What Each Claims

Eugene Fama formalized the Efficient Market Hypothesis in 1970. The hypothesis has three versions, and most popular discussions conflate them in ways that make the hypothesis seem either obviously true or obviously false.

The weak form says that current prices reflect all past price and trading volume information. You cannot consistently profit by looking at price charts, moving averages, or historical patterns — because those patterns, to the extent they have predictive value, have already been incorporated into prices by the traders who noticed them first. Technical analysis, in this view, is the financial equivalent of finding money on the ground and expecting it to stay there.

The semi-strong form says prices reflect all publicly available information — not just price history but earnings announcements, news coverage, analyst reports, macroeconomic data. You cannot consistently profit by trading on public information because that information is already priced in by the time you access it. Fundamental analysis of public financials, in this view, is also futile as a consistent edge.

The strong form says prices reflect all information, including private and insider information. Even insiders can’t consistently trade profitably. This form is widely rejected by evidence — insider trading prosecution exists because insiders demonstrably profit from private information.

The Mechanism That Makes It Work

The EMH is not an assumption — it’s a consequence. If prices consistently underreacted to public information, sophisticated traders would exploit the predictable drift, and their trading would push prices to fully reflect the information faster. If prices consistently overreacted, traders would short the overreaction and profit from the reversion. Competition among informed traders drives prices toward full information incorporation.

This is the key insight: the EMH is an equilibrium condition, not a description of how any individual market participant behaves. Individual traders can be irrational, overconfident, and wrong. As long as their errors are uncorrelated — as long as they don’t all make the same mistake in the same direction at the same time — their noise cancels out and sophisticated arbitrageurs correct the remaining systematic errors.

The EMH breaks down when: (a) irrational traders make correlated errors (herding), (b) sophisticated traders cannot easily exploit the mispricing due to limits on arbitrage (cost of shorting, capital constraints, risk of the mispricing getting worse before it corrects), or (c) the market lacks sufficient participation by sophisticated traders to enforce efficiency. All three happen regularly in real markets.

What the Evidence Shows

The evidence on the weak form is largely supportive. Academic studies of technical trading rules consistently find that they don’t generate abnormal returns after accounting for transaction costs and risk. The patterns that appear in backtests tend to be overfitted to historical data and don’t hold out-of-sample.

The evidence on the semi-strong form is mixed and event-specific. Prices do react rapidly to public information — earnings announcements, merger news, and macroeconomic data releases are largely incorporated within minutes. But several anomalies persist: the value premium (cheap stocks by various metrics outperform expensive ones over long periods), the momentum effect (recent winners continue to outperform recent losers over medium horizons), and the size effect (small-cap stocks have historically outperformed large-cap). Whether these are compensation for risk, artifacts of data mining, or genuine mispricings exploited by informed investors is debated.

The most damning evidence against the EMH’s practical implications is the frequency of asset price bubbles that in retrospect are obvious. Prices that rise far above fundamental value and then collapse are inconsistent with efficient pricing if the fundamentals were knowable. The dot-com bubble, the housing bubble, and numerous smaller episodes suggest that correlated irrational behavior can sustain mispricings for long enough to matter, and that sophisticated arbitrage fails to correct them in time.

The Joint Hypothesis Problem

Testing the EMH runs into what Fama himself called the joint hypothesis problem. To determine whether a market is efficient, you need a model of what the correct price should be. Any test of efficiency is simultaneously a test of efficiency and a test of the asset pricing model used as the benchmark. If you find abnormal returns, it could mean the market is inefficient, or it could mean your asset pricing model is wrong.

The value premium, for example, could mean that value stocks are systematically underpriced (market inefficiency) or that value stocks carry risks not captured by the standard models and their higher returns are compensation for those risks. Distinguishing the two is hard and depends on which risks you consider relevant — which circles back to needing an asset pricing theory.

This is not a dodge. It genuinely means that “is this market efficient?” cannot be answered in isolation from “what is the correct model of asset pricing?” Both questions are open.

What You Should Take From It

The EMH is most useful not as a claim about whether markets are exactly right, but as a prior about how hard it is to be systematically wrong. If a simple, obvious strategy could consistently beat the market, enough people would use it that the opportunity would disappear. Durable, consistent edges in markets are rare; most apparent edges are noise, risk compensation, or survivorship bias.

The practical implication that flows from this — even for those who don’t believe in strong-form efficiency — is that beating the market through active management is harder than it looks, the costs of trying are certain even when the success is not, and most investors would do better indexing. This conclusion doesn’t require the EMH to be true in any strong form. It just requires accepting that markets are competitive enough that consistently outperforming them is exceptional rather than routine.

The behavioral economists who attacked the EMH were not arguing that you can easily profit from market irrationality. They were arguing that markets have systematic failures, that prices can deviate from fundamentals for long periods, and that understanding those deviations matters for understanding market dynamics — even when arbitraging them is difficult. Both observations can be true.