This article explains the efficient market hypothesis as a concept. It is educational and general, not personalised investment advice, and nothing here is a recommendation to buy or sell any security. Where it references anomalies or research results, those describe widely observed patterns and an ongoing academic debate; past patterns do not predict future results.
The efficient market hypothesis is the idea that you cannot beat the market by studying it, because whatever you could learn is already in the price. It is one of the most influential claims in finance, the intellectual foundation of index funds, and also one of the most argued-over, because markets keep displaying patterns the strict version says should not exist. Here is what the hypothesis actually says, the three forms it comes in, the deep problem that makes it hard to test, the anomalies that strain it, and the “efficiently inefficient” middle ground most working practitioners actually occupy.
What the efficient market hypothesis is
The efficient market hypothesis (EMH), formalised by Eugene Fama in 1970, is the claim that asset prices already reflect all available information, so consistently earning excess returns from public information alone is not possible; prices move only on genuinely new, unpredictable information. That is the whole idea. If a piece of information is known, the reasoning goes, some trader has already acted on it and moved the price, so by the time you read it the edge is gone. What is left to move prices is only the next surprise, and surprises, by definition, cannot be forecast. Under strict efficiency, a chart is a record of past surprises, not a map of future ones.
Fama’s 1970 review is what turned the idea into a testable framework: it organised the scattered evidence into a single structure and defined the forms the field still tests. The intuition is older, running back to Louis Bachelier’s 1900 work on price movements and Paul Samuelson’s argument that properly anticipated prices fluctuate randomly, but it is Fama’s three-form taxonomy that the field still uses. Note what EMH does not say: it makes no claim that prices are always “right” or that bubbles never happen. Its claim is narrower and harder to dodge, that publicly available information cannot be reliably turned into excess return.
The three forms: weak, semi-strong, strong
Fama split the hypothesis into three nested versions, each defined by which information set is already priced in. They are worth keeping straight, because most real arguments about market efficiency are really arguments about one specific form.
| Form | Information already in the price | If true, what fails to give an edge |
|---|---|---|
| Weak | All past prices and volume | Charting past price and volume alone |
| Semi-strong | All public information (prices, filings, news) | Charting and fundamental analysis of public data |
| Strong | All information, public and private | Everything, including material non-public information |
The weak form says prices already reflect all past price and volume data, so studying historical charts alone cannot produce excess returns. This is the version that, if true, would sink the simplest kind of technical analysis. The semi-strong form goes further: prices reflect all public information, financial statements, news, analyst reports, so neither charting nor fundamental analysis of public data should give a durable edge, and prices should jump to new public information almost instantly. The strong form is the most demanding: prices reflect all information, public and private, so even an insider with material non-public information could not profit. The strong form is widely regarded as false: insiders are documented to earn abnormal returns on private information, which is why the strong form is rejected and why such trading is regulated. The interesting debate lives almost entirely in the weak and semi-strong forms.
Why it matters
EMH is the load-bearing argument behind a trillion-dollar industry. If markets are efficient, then the average actively managed fund cannot beat a cheap index after fees, because any edge it might find is already priced away and its costs are pure drag. A closely related result, William Sharpe’s “arithmetic of active management,” reinforces the point: it holds even without assuming efficiency (as an accounting identity, the average actively managed dollar must trail the average passive dollar after costs), and efficiency only sharpens it. Together they form the intellectual foundation of passive index investing and rank among the most consequential ideas in the history of finance. The empirical record is largely kind to it: over long horizons, the majority of active managers underperform their benchmarks, roughly what a mostly efficient market would predict.
It also reframes what a would-be outperformer is actually claiming. To say “I can beat the market” is to say “the market is inefficient in some specific, exploitable, repeatable way, and I have found it.” EMH forces that claim into the open. It does not prove the claim impossible, but it sets a high bar: the burden is on the active manager to show a real, persistent inefficiency, not on the market to prove it has none. That burden is the reason a serious systematic strategy should be tested against a survivorship-bias-free record rather than a story, a discipline discussed in why backtests lie.
The joint-hypothesis problem: why efficiency is so hard to test
Here is the deep difficulty, and it is the single most important idea in this article. You can never test market efficiency on its own, because to decide whether a return was “excess” you first need a model of what a fair, normal return should have been. Every test of efficiency is therefore also a test of the pricing model you used to define normal. Fama himself named this the joint-hypothesis problem.
The consequence is genuinely awkward. Suppose you find a pattern that appears to earn excess returns. There are two explanations, and the data alone cannot always tell them apart. Either the market is inefficient (a real anomaly), or the pattern is fair compensation for some risk your pricing model failed to capture (the model is incomplete). A believer in efficiency can always argue that any apparent anomaly is just a missing risk factor; a skeptic can always argue the risk story is invented after the fact to rescue the theory. This is why the efficiency debate has never been cleanly settled by evidence: the question is entangled with a second question that is itself unsolved. It is also why the honest framing of any edge is that it is part mispricing and part risk, in unknown proportion, the same conclusion reached about momentum in momentum investing explained.
The anomalies that strain strict efficiency
The strong case against strict efficiency is a stack of persistent, documented patterns, “anomalies,” that a perfectly efficient market should not display, or should have arbitraged away once published. The best known:
- Momentum. Recent winners keep outperforming recent losers over intermediate horizons, documented by Jegadeesh and Titman in 1993 and replicated across decades, countries, and asset classes. It is the anomaly efficiency struggles most to explain, because it survives out of sample and has no clean risk story, more on this below.
- Value. Cheap stocks (low price relative to book value or earnings) have historically outperformed expensive ones, the effect Fama and French built into their three-factor model. Whether this is mispricing or risk compensation is a textbook case of the joint-hypothesis problem.
- The small-cap effect. Small companies have historically earned returns beyond what their market risk alone predicts, first documented by Rolf Banz. Part risk premium, part possible mispricing, and partly weakened since publication.
- Post-earnings-announcement drift. After a genuine earnings surprise, prices keep drifting in the direction of the surprise for weeks rather than jumping instantly to the new level, a direct violation of the semi-strong form’s claim that public information is priced in at once.
The efficiency-side rebuttal is disciplined and not to be dismissed: many anomalies shrink or vanish after they are published (arbitrage crowds them out), some were data-mining artefacts that never truly existed, and the survivors may simply be compensation for risks the standard models miss. All three rebuttals are sometimes correct. The systematic factors that do survive scrutiny, and the honest way to tell a real one from a fluke, are the subject of factor investing.
The middle ground: “efficiently inefficient”
The most defensible position sits between “markets are perfectly efficient” and “markets are easily beaten”: a paradox first made rigorous by Sanford Grossman and Joseph Stiglitz in 1980, and the frame most thoughtful practitioners actually hold.
Their argument is elegant. Gathering and analysing information is costly. If markets were perfectly efficient, prices would already reflect everything, so no one could earn a return on that costly research, and therefore no one would bother to do it. But if no one does the research, prices could not reflect the information in the first place, so the market could not be efficient. Perfect efficiency is self-defeating: it destroys the incentive that creates it. The resolution is an equilibrium in which markets are almost efficient, inefficient by just enough to pay the people whose trading removes the inefficiency. Information gatherers earn a return that compensates their costs, no more. The market is, in the memorable phrase later adopted as a book title by Lasse Pedersen, efficiently inefficient.
This middle ground matters because it makes both extremes wrong in a useful way. It says beating the market is possible but hard and expensive, that the edge available is roughly the size of the cost required to capture it, and that easy, free, obvious edges are precisely the ones competition erases. It reframes outperformance as a fair wage for genuinely costly, genuinely skilled effort. That is a far more honest picture than either “you can’t beat the market” or the get-rich pitch on the other side.
Where a systematic momentum system fits
This is the one place the theory connects naturally to a working system. Momentum is the anomaly the efficient market hypothesis has the hardest time absorbing: it is robust, it has been replicated in samples and asset classes its discoverers never touched, it persisted for years after publication, and it lacks the clean risk story that lets efficiency comfortably reclassify value or small-cap as “just risk.” A systematic momentum strategy is, in the language of this article, a concrete bet on one specific crack in efficiency, the wager that the momentum anomaly is real, is driven by documented behavioural frictions, and is persistent enough to be worth the cost of harvesting it.
Shishin is that kind of bet, made in the open. It is a systematic, rules-based US-equity momentum system, and it does not claim to have repealed market efficiency, only to exploit the best-documented place where strict efficiency leaks. Consistent with the Grossman-Stiglitz view, it treats any edge as small, costly to capture, and worth verifying rather than asserting: the record is meant to be inspected across a survivorship-bias-free backtest and a live, independently attested track, not taken on faith, which is the entire point of the published track-record analysis and the attestation log at /verify. A rules-based process can, at most, show its work on the one crack it is built to trade.
The limits of the theory
EMH is a lens, not a law, and its own limits are as instructive as its claims. It is unfalsifiable in isolation because of the joint-hypothesis problem, so it can never be cleanly disproved, which is itself a scientific weakness. It sits uneasily with the historical record of bubbles and crashes, where prices detached from any plausible information for extended stretches, the challenge behavioural finance presses hardest, explored in behavioral biases. And the strong form is simply false. What survives, and what is worth carrying, is the humbled core: markets are mostly efficient most of the time, edges are real but small, costly, and competed against, and anyone claiming an easy, obvious, repeatable way to beat the market is almost certainly wrong. Strict efficiency overstates the case; dismissing efficiency entirely understates how hard the game really is.
So, are markets efficient?
Mostly, and imperfectly, which is the only answer the evidence actually supports. Prices absorb public information fast enough that the average active manager cannot beat a cheap index after costs, which is the practical victory of the hypothesis and the reason index funds exist. But the market is not perfectly efficient, it cannot be, by the Grossman-Stiglitz logic, and it leaves persistent anomalies, momentum chief among them, that a disciplined, costly, systematic process can attempt to harvest. The efficient market hypothesis is best read not as “you can’t win” but as “here is exactly how hard winning is, and why the free lunches you can see are already gone.”
Sources & further reading
- Fama, E. F. (1970). “Efficient Capital Markets: A Review of Theory and Empirical Work.” Journal of Finance, 25(2), 383 to 417. Introduces the three-form taxonomy and the joint-hypothesis problem.
- Grossman, S. J. & Stiglitz, J. E. (1980). “On the Impossibility of Informationally Efficient Markets.” American Economic Review, 70(3), 393 to 408. Sets out the “efficiently inefficient” equilibrium.
- Jegadeesh, N. & Titman, S. (1993). “Returns to Buying Winners and Selling Losers.” Journal of Finance, 48(1), 65 to 91. Documents the momentum anomaly.
- Related reading: momentum investing explained, factor investing, and behavioral biases.