Market efficiency

Market efficiency is the proposition that financial-market prices can embody a better forecast than any individual or competing system. In the book’s account, this does not mean that every investor is wise, or that prices are always right. It means that if markets forecast well, their success cannot be credited solely to a tiny cadre of arbitrageurs correcting a foolish crowd: the crowd’s aggregated judgment must itself contain substantial information. The argument’s force comes from holding two ideas together—markets can be collectively intelligent, yet the mechanisms that produce that intelligence remain vulnerable to imitation, constrained capital, and distorted information.

Definition and Scope

The book adopts a practical definition of market efficiency from Robert Shiller: a market is efficient if it consistently provides a better forecast of companies’ future discounted free cash flow than any other individual or system. This is a comparative standard, not a claim of perfect foresight. A market can be wrong about a future outcome and still, in principle, be the best available forecasting mechanism; the relevant question is how its performance compares with alternatives.

That distinction matters because the common account of efficiency assigns intelligence to specialized traders rather than to the market as a whole. In the familiar metaphor, ordinary investors are sheep whose systematic mistakes create opportunities, while arbitrageurs are wolves who spot mispricing and force prices back toward true value. The book challenges the assumption that this division of labor can explain market accuracy by itself. If prices are efficient, the aggregate judgments contributing to them cannot be dismissed as irrational noise while the work of rational investors alone bears the explanatory weight.

Arbitrage and Its Limits

The “wolves and sheep” account relies on a fragile expectation: after an arbitrageur trades against a perceived mistake, other investors will recognize the correction and stop repeating it. But investors who are mistaken do not necessarily know that they are mistaken. A trade signaling that they are wrong is not guaranteed to change their minds. If the arbitrageur’s capital is exhausted, they may resume the same behavior, leaving prices exposed to the very misjudgment the trade was supposed to correct.

There is also a problem of what makes arbitrage profitable. A rational trader can earn money from a position only if prices eventually move toward rationality. If the arbitrageur’s own trading alone makes the price correct, the resulting profit will be small. As short seller Jim Chanos put it, “We can only hope that eventually the market will realize 2+2=4.” The hope is not merely that the arbitrageur can identify a discrepancy, but that the market itself will ultimately converge toward a more accurate price.

Capital and market depth impose another constraint. Even intelligent investors may not have enough resources to outweigh the combined buying or selling power of others. The book uses the 1998 Long-Term Capital Management crisis to expose the danger in assuming that a dominant rational trader can always stabilize prices. Long-Term Capital Management was, in many of its markets, the largest participant and should have been the marginal investor—the trader whose actions set the price. Yet when conditions turned against it, the absence of buyers drove down the value of its assets. Its positions could be rational and its capital immense; neither could protect it when other market participants were running in the opposite direction.

The Crowd as an Aggregator

The implication is not that every market judgment is accurate. It is that price, whatever its accuracy, represents the weighted-average judgment of investors participating in the market. If market prices provide reliable forecasts, then the collective contribution of investors must be doing real informational work. The “sheep” cannot be wholly irrational in the aggregate while the market is efficient: collectively, the sheep are wolves.

That conclusion connects market efficiency to a wider account of collective intelligence. Groups can produce better judgments when many people contribute different information, but aggregation depends on how judgments are elicited and combined. In classroom experiments, information cascades occurred when people chose sequentially and paid attention to those who had already acted. When Angela Hung and Charles Plott instead made participants’ reward depend on whether the group’s majority answer was correct, students paid more attention to their private information; the collective judgment became significantly more accurate than that of cascading groups. The example clarifies a condition relevant to market claims: aggregation is not magic. The structure of participation and incentives affects whether separate information remains available to the group.

Nor does a collective forecast settle what ought to be done. The book distinguishes the accuracy of a decision from its legitimacy and public purpose. A market may gather dispersed assessments of what is likely to happen; that does not establish that its output serves the common good or that everyone affected had a fair voice. In the discussion of the proposed Policy Analysis Market, public access was defended partly because people outside the intelligence establishment might possess information the agencies would not otherwise hear. But the proposal also provoked public outrage at the prospect of ordinary people profiting from terrible events. Predictive value and ethical acceptability are separate questions.

Information, Institutions, and Evidence

The broader institutional argument helps explain why the reliability of aggregation cannot be detached from the movement of information. Decentralized participants may possess specialized or local knowledge, yet useful information can remain isolated rather than reaching the rest of the system. The book’s example of intelligence agencies emphasizes that fragmented holdings and weak information sharing can prevent institutions from recognizing the combined significance of data they already possess. For markets, the parallel is analytical rather than a claim that the systems are identical: information must be contributed and made consequential before aggregation can improve a forecast.

The references supporting the discussion place market efficiency alongside behavioral finance, investor psychology, and limits to arbitrage. They also point to real-time research on market efficiency and to Jack Treynor’s bean-jar experiment as relevant evidence and experimental inquiry. These citations frame the proposition as something to investigate against competing accounts, not as a slogan that prices must be right. The central test remains comparative forecasting performance, while the surrounding analysis asks whether participants can act on information, whether their resources are sufficient, and whether the market’s composition allows separate judgments to affect the result.

Market efficiency therefore names a collective achievement with boundaries. It does not promise that a rational specialist can always defeat an irrational crowd, that every price is correct, or that accurate prediction is the same as a just decision. Its strongest claim is more exacting: when markets forecast well, the explanation must take seriously the intelligence embedded in the judgments of the many, as well as the conditions that let those judgments accumulate into a price.