What Is the Efficient Market Hypothesis?
A company has a great product, loyal customers, and profits that keep growing. Buying its stock seems like an obvious decision.
There is one complication: everyone else can see those things too. If investors have already bid up the shares in anticipation of success, the company may need to deliver extraordinary results just to justify what you paid. You can be right about the business but still have a disappointing return.
That is the problem at the center of the efficient market hypothesis, also called efficient market theory or EMH. The hypothesis says that prices fully reflect the information available to the market. Its practical implication is impactful: using information already reflected in prices should not give an investor a repeatable advantage after accounting for risk and costs.
The idea is closely associated with economist Eugene Fama, whose 1970 review helped establish it as a foundation of modern finance.1 It also presents a useful challenge to anyone considering a stock: what do you understand about this business that its price does not already allow for?
How Information Gets Into a Price
A market price emerges from people deciding what they are willing to pay or accept. They bring different forecasts, financial needs, and amounts of money to those decisions. Their trades determine the price at which shares change hands.
Suppose a company trades at $50 a share and announces that a major customer has signed a long-term contract. Investors who expect the contract to increase future profits may now be willing to pay more for the stock, and sellers of the stock can raise their asking prices as well. In today's high-frequency, high-data trading environment, the stock could move before someone reading the announcement has time to finish it.
The opportunity changes along with the price. An investment that looked attractive at $50 may be much less attractive at $60. The news still matters, but a buyer arriving later has to pay for the improvement it promises.
This process is called price discovery. Investors search for opportunities, and acting on what they find helps remove those opportunities. The hypothesis does not require every participant to be perfectly informed or sensible. It relies on competition doing enough of the work.
There is a subtle consequence here. A price reflects expectations about the future, so good news can still disappoint. Thus the old phrase of "buy the rumor, sell the news".
Imagine that the same company later reports profit growth of 20%. If investors had been expecting 30%, the stock might fall. The business is doing better than it did last year, but it has delivered less growth than investors were counting on when they bought the shares. The report's other details, especially management's outlook, can matter just as much as the headline number.
“Priced in” means an expectation has already influenced what investors will pay. It does not mean the expected event is certain to happen.
This helps explain the random walk idea. Once familiar information is reflected in the price, predicting the next move depends largely on predicting what will surprise investors next. That is a much harder task than explaining why a price moved yesterday.2
Market efficiency does not require prices to follow a literal random walk. Expected returns can change with economic conditions and risk. The question is whether a predictable pattern offers an advantage beyond compensation for taking those risks.
What Counts as Beating the Market?
When people say an investor has “beaten the market,” they often mean the investor earned more than an index such as the S&P 500, which tracks large U.S. companies. The comparison tells us who came out ahead, but we still need to understand why.
An investor who borrows money to buy more shares can gain more in a rising market and lose more in a falling one. A portfolio concentrated in a single industry can outperform spectacularly when that industry does well. Neither result, on its own, demonstrates superior stock selection.
A fair comparison has to account for those risks and deduct the costs of the strategy, including fund fees, trading expenses, and research. Taxes can further reduce what an individual investor keeps.
Researchers therefore ask whether an investment earned more than its risks would normally justify. They call the excess an abnormal return or excess return. Market efficiency concerns whether investors can repeatedly earn those returns using the information available to them, after accounting for the costs of doing so.
Luck matters too. Some investors will make winning bets even if none has a reliable advantage. A convincing case for skill needs more than a profitable trade or a good year.
Three Versions of the Hypothesis
The hypothesis depends on which information a price is supposed to reflect. Fama's framework distinguishes three forms:
| Form | Information reflected in prices | What should offer no advantage |
|---|---|---|
| Weak | Past prices and trading activity | Studying price charts and trading patterns alone |
| Semi-strong | All public information, including financial reports and news | Analyzing publicly available information |
| Strong | Public and private information | Having information unavailable to other investors |
Each version includes more information than the one before it. In each case, an “advantage” means the ability to earn abnormal returns, with costs considered when judging whether an opportunity is worth pursuing.
The strong form goes too far as a description of real markets: private information can give investors an advantage. For someone choosing stocks by reading financial reports and news, the semi-strong form is the most relevant. It says that even a better interpretation of public information should not provide a repeatable advantage after accounting for risk and costs.
That is a much stronger claim than saying prices react quickly to headlines.
What the Evidence Supports
Studies of earnings announcements, stock splits, and dividend changes found that prices often adjusted rapidly to new information. Those findings provided much of the support for the hypothesis, as Fama's 1991 reassessment describes.3
Fund performance offers another way to examine how difficult it is to outperform. Active fund managers choose investments in an effort to beat a benchmark, the index used to judge their results.
The SPIVA U.S. Year-End 2025 scorecard, published by S&P Dow Jones Indices, shows how difficult that has been. Over the 15 years ending December 31, 2025, only about one in ten active funds investing in large U.S. companies both survived and outperformed the S&P 500.4
The calculation counts funds that closed or merged during the period, so the results do not depend on examining only those still in business. Fund returns are measured after ongoing fees, though sales charges are excluded. The index itself has no fund expenses.
The result gives investors a reason to be cautious about their ability to choose a winning fund in advance. It tells us less about whether individual stocks were priced correctly. Fees can absorb a manager's advantage, and differences in risk can affect the comparison. The scorecard is more useful for understanding what investors experienced than for settling whether markets are efficient.
Why the Debate Persists
Suppose stocks in struggling companies earn unusually high average returns. Investors might have repeatedly underestimated those businesses. Or holding them might expose investors to losses at especially difficult times, with the higher expected return compensating for that risk.
To distinguish between those explanations, researchers need a model that estimates what return an investment's risk should justify. If a strategy earns more than the model predicts, the market might be inefficient, or the model might have overlooked a relevant risk. This is the joint-hypothesis problem: the test evaluates both the market and the model used to judge it.3
That makes the evidence harder to interpret. It does not mean a result can be dismissed simply by saying there must have been some hidden risk. Three challenges help explain why the debate remains open.
Past prices have sometimes helped predict returns. One example is momentum, a pattern in which recent winners keep outperforming recent losers for a time. In a 1993 study, Narasimhan Jegadeesh and Sheridan Titman found that buying recent winners and betting against recent losers generated positive returns over holding periods of three to twelve months. The risk measures they examined did not explain the results.5
A pattern based on past prices challenges weak-form efficiency. Whether an investor can use it profitably also depends on trading costs, exposure to losses, and whether the pattern holds beyond the historical data used to identify it.
Prices can move more than a simple valuation model explains. In his influential 1981 paper, Robert Shiller examined stock prices in relation to the dividends, or cash payments to shareholders, that companies subsequently paid. He found that prices fluctuated too much for those future payments to justify under the model he tested.6
Shiller's model assumed that investors always required the same rate of return. If that requirement changes, prices can change too: an investor demanding a higher return will pay less today for the same expected future payments. His finding challenged a particular explanation of stock prices and helped motivate research into investor psychology. It did not establish that every large price move is irrational.
Trading against a mistaken price can be risky. Suppose a trader thinks a $100 stock should be worth $60. Betting on a decline might eventually pay, but the stock could rise to $150 first. Losses on the bet could force the trader to close the position before the price falls.
Economists call obstacles like these limits to arbitrage. In this context, arbitrage means trading to exploit an apparent pricing error. The risks and cost of financing the trade can make it impossible to wait for a correction, even when the trader's judgment about value is sound.7
Prices can therefore be difficult to beat even when they do not make good use of information. Recognizing a mistake does not ensure that an investor can afford to act on it.
Who Pays for the Research?
The process that makes prices informative creates another difficulty for the theory.
Research costs money. Someone has to study the accounts, investigate a company's prospects, and decide whether its shares are attractive. If prices already revealed everything that work could uncover, investors could get the information free by observing prices. There would be little reason to pay for the research that helps make those prices informative.
In a 1980 paper, Sanford Grossman and Joseph Stiglitz showed why perfect informational efficiency cannot be sustained in their model when information is costly. For research to be worth paying for, investors need some prospect of profiting from what they learn. That requires information that is not yet fully reflected in prices.8
The argument explains why research and some degree of inefficiency can coexist. It does not tell us how large pricing errors are in real markets.
It also offers a practical way to think about efficiency. Competition can push prices toward better estimates until the remaining opportunity is too expensive to pursue. Investors' research makes prices more informative, while the cost of doing that work puts a limit on the process.
What This Changes for an Investor
“The company makes a great product” is a statement about the business. “Customers will buy much more of that product than the current price appears to assume” is the beginning of an investment argument. It still needs evidence, a reasonable purchase price, and an explanation of what could go wrong.
The same reasoning applies when choosing a fund manager. A strong past return deserves examination, along with the risks taken, the benchmark used, and the fees charged. To choose the manager on the strength of that record, an investor needs a reason to expect the advantage to continue.
This helps explain the appeal of low-cost index funds, which aim to track a chosen market instead of selecting its winners. Their case does not depend entirely on market efficiency.
In The Arithmetic of Active Management, William Sharpe explained that investors collectively earn their market's return before costs. Investors who hold the whole market in its existing proportions receive that return, so the remaining investors, taken together, must receive it too. The calculation counts each dollar equally, so larger portfolios contribute more to the average. When active investors pay higher costs, they keep less of that return.9
Some investors can still outperform. But a manager's fees need to be justified by more than the observation that markets sometimes get prices wrong.
Choosing an index still leaves decisions about which assets to own, how concentrated a portfolio should be, and how much loss an investor can tolerate. The article on passive investing examines those choices in more detail.10
The company with the great product may deliver years of growing profits and still disappoint its shareholders. The price they paid may already have assumed even more.
Sources
- Efficient Capital Markets: A Review of Theory and Empirical Work (Eugene Fama, 1970). The foundational review of market efficiency and its three forms.
- The Efficient Market Hypothesis and Its Critics (Burton Malkiel, 2003). Explains the relationship between random walks, return predictability, and market efficiency.
- Efficient Capital Markets: II (Eugene Fama, 1991). Reviews evidence on how prices respond to information and explains the joint-hypothesis problem.
- SPIVA U.S. Scorecard Year-End 2025 (S&P Dow Jones Indices). Provides the fund-performance results and explains how fees and funds that closed or merged are treated.
- Returns to Buying Winners and Selling Losers: Implications for Stock Market Efficiency (Narasimhan Jegadeesh and Sheridan Titman, 1993). Documents momentum in stock returns over holding periods of three to twelve months.
- Do Stock Prices Move Too Much to Be Justified by Subsequent Changes in Dividends? (Robert Shiller, 1981). Examines stock-price volatility using a dividend valuation model with a constant required return.
- The Limits of Arbitrage (Andrei Shleifer and Robert Vishny, NBER working paper, 1995; published in 1997). Explains how risk and financing constraints can prevent traders from correcting mispricing.
- On the Impossibility of Informationally Efficient Markets (Sanford Grossman and Joseph Stiglitz, 1980). Shows why perfect informational efficiency cannot be sustained in their model when gathering information is costly.
- The Arithmetic of Active Management (William Sharpe, 1991). Explains the relationship between active and passive investors' aggregate returns before and after costs.
- Passive Investing Is an Active Bet (James Warrick). Examines portfolio concentration and the choices that remain when investing through index funds.