Key Principle
Market efficiency is not produced by rational individuals but by diverse individuals. When investors use heterogeneous strategies and information sources, their errors cancel in aggregate and prices converge on fundamentals. Efficiency is the default state. The dangerous exception is diversity breakdown: when imitation causes strategies to correlate, error-cancellation ceases, positive feedback overwhelms negative feedback, and markets depart violently from value. The question is never "are agents rational?" but "is the agent population diverse enough?"
Why This Matters
- Reframes the entire efficient market debate from individual rationality to collective diversity — a structural, potentially measurable criterion.
- Contrarian investing must target correlated irrationality (diversity collapse), not isolated individual error. A few irrational actors cannot move markets.
- The same mechanism that makes markets usually efficient is the one that, when it breaks, produces bubbles and crashes. Understanding the mechanism reveals both sides.
- Stock markets are uniquely fragile among decentralized systems because they lack terminal values, exhibit reflexivity, and use price as a behavior trigger — all of which incentivize imitation over independent analysis.
Good Examples
- Treynor's jellybean jar: Average guess of finance students consistently approximates the actual count and beats almost all individual guesses — aggregation with diversity produces emergent accuracy. (Ch. 14)
- Lost nuclear bomb (1966): John Craven assembled diverse experts who placed bets on the bomb's Mediterranean location. No single expert was correct, but the bet-generated probability distribution located it. (Ch. 14)
- Iowa Electronic Markets: Beat nearly 600 polls in predicting presidential elections 75% of the time — decision markets harness diversity plus financial incentives. (Ch. 29)
- Norman Johnson's maze experiment: Collective solution quality improves rapidly with ~5 contributors, each bringing unique exploration data. Aggregation converts varied partial information into "super information." (Ch. 28)
- Army-ant circular mill: 1,200-foot circumference, two-day duration, dissolved only when a few workers broke away — recovery from herding requires reintroduction of behavioral diversity. (Ch. 13)
- Acquisition scenarios: When a defined terminal value appears, stock prices converge accurately on ultimate value because independent analysis is again rewarded over imitation. (Ch. 30)
Counterpoints
- No reliable diversity barometer exists. Detecting when diversity has collapsed requires evaluating both public and private opinion — there is no single metric. (Ch. 14)
- Bee colonies lack reflexivity. The honey bee analogy for decentralized intelligence breaks down because bee dances inform without inducing imitation, whereas stock prices trigger imitative behavior. (Ch. 29)
- Convergence requires long horizons. Arthur's El Farol model needed 20,000+ iterations to converge. Real markets, where agents have short horizons and limited diversity, depart from efficiency frequently. (Ch. 15)
Key Quotes
- "Sufficient investor diversity is the essential feature in efficient price formation. Provided the decision rules of investors are diverse — even if they are suboptimal — errors tend to cancel out and markets arrive at appropriate prices." (Ch. 14)
- "So the issue is not whether individuals are irrational (they are) but whether they are irrational in the same way at the same time." (Ch. 14)
- "The key to successful contrarian investing is to focus on the folly of the many, not the few." (Ch. 14)
- "A full ecology of strategies is sufficient to generate efficient markets. But when diversity is jeopardized — which it frequently is — markets depart significantly from the underlying fundamentals." (Ch. 15)
- "Stock prices tend to be efficient when investors are heterogeneous. But when heterogeneity does not prevail and investor errors become nonindependent, markets become subject to excesses." (Ch. 29)
- "This perspective shifts the onus of rationality away from individual investors and suggests that allocative efficiency arises from the structure of the market itself." (Ch. 29)
- "Investors who appreciate how and why markets are efficient will have better insight into how and why markets are inefficient." (Ch. 30)
- "When people are free to do as they please, they usually imitate each other." — Eric Hoffer (Ch. 13 epigraph)
Rules of Thumb
- Monitor the crowd, not the individual. Cataloguing individual biases is less useful than detecting when the population's strategies have become correlated.
- Diversity collapse is the buy/sell signal. Extreme consensus — measured through media uniformity, positioning data, strategy crowding — flags the transition from efficiency to excess.
- Diversify your own information diet. Deliberately allocate time to unfamiliar disciplines and unconventional sources. This is the human equivalent of the ant colony's "wild hair" mechanism that prevents lock-in. (Ch. 28)
- Agency costs amplify herding. Managers minimize tracking error to protect business value, not portfolio returns. The more managers herd, the more deviation becomes career-threatening, narrowing diversity further. (Ch. 13)
- Beware open-ended horizons. Unlike decision markets with finite endpoints, stocks lack terminal values — this rewards selling to someone willing to pay more over independent valuation, incentivizing imitation. (Ch. 30)
- Recovery is slow and individual. Markets go mad in herds but recover their senses "slowly, and one by one." Position for the asymmetry. (Ch. 13, quoting MacKay)
Related References
- Process vs. outcome thinking (decision journals as structural countermeasure to hindsight bias)
- Complex adaptive systems and emergent behavior
- Fat tails and power laws (diversity breakdown is the mechanism that generates extreme return events)
- Reflexivity and feedback loops (Soros's framework as practitioner version of herding mechanics)