Confidently Wrong: A Reading of Official Expert Consensus That History Eventually Corrected

Let us begin with a stipulation that is both obvious and frequently ignored: expertise is real, it is valuable, and the alternative to consulting it is generally worse. This article is not a brief for ignoring credentialed authority. It is a brief for reading the historical record of credentialed authority carefully before deciding what weight to assign it — because that record contains a pattern that anyone making consequential decisions ought to understand.

The pattern is this: institutions composed of intelligent, trained, well-credentialed people periodically arrive at a consensus that is not merely incomplete but actively wrong in ways that cause measurable harm. The consensus is defended with confidence. Dissent is marginalized. The correction, when it comes, arrives later than it should and at greater cost than necessary. Then the institution moves on, and the next consensus forms.

This is not a modern phenomenon. It is not a product of politicization, social media, or regulatory capture, though all of those things can accelerate it. It is a feature of how human beings, organized into credentialed hierarchies, process information about uncertain questions. The psychology of deference to authority looks identical whether the authority is a Roman physician or a contemporary research institution. History is the longest-running study of this dynamic ever conducted.

The Mercury Physicians and the Confidence of the Trained

For a substantial portion of Western medical history, mercury was not merely an accepted therapeutic agent — it was a prestigious one. Calomel, a mercury chloride compound, was prescribed by trained physicians for conditions ranging from syphilis to yellow fever to constipation. It was not prescribed by charlatans operating outside the medical establishment. It was the medical establishment's treatment of choice, defended in the professional literature and administered with the full authority of credentialed practice.

The harm was extensive and documentable. Mercury poisoning produces a recognizable syndrome: tooth loss, neurological damage, excessive salivation, and, at sufficient doses, death. Patients who survived their illnesses frequently did not survive their treatments. The physicians administering these treatments were not ignorant men. Many were among the most educated practitioners of their era. Their confidence in calomel was not a failure of intelligence. It was a failure of the feedback mechanisms that should have connected observed outcomes to revised practice.

The lesson is not that 19th-century doctors were foolish. The lesson is that a consensus built within a credentialed community can persist long past the point where the evidence against it has become substantial, because the social cost of revising a consensus is high and the individual cost of maintaining it is low — at least for the people maintaining it.

The Postwar Depression That Never Arrived

At the conclusion of World War II, a significant portion of the American economics profession was confident that the end of wartime production and the demobilization of millions of servicemen would produce a severe economic depression. The logic was coherent: remove the artificial stimulus of military spending, flood the labor market with returning veterans, and the result should be a contraction. Several prominent economists put versions of this prediction on the record.

What followed instead was the longest sustained economic expansion in American history to that point. The mechanisms that the consensus had not adequately weighted — pent-up consumer demand, veteran education subsidies, suburban housing construction, the Marshall Plan's effect on export markets — drove growth rather than collapse.

The interesting question is not why the prediction was wrong. Macroeconomic forecasting is genuinely difficult, and error is inherent in the enterprise. The interesting question is why the prediction was held with such confidence, and what the professional consequences were for those who made it. The answer to the second question is: very few. The consensus was wrong, the economy boomed, and the economists who had predicted disaster moved on to the next consensus. The accountability mechanisms that would, in a well-functioning information system, attach consequences to confident error were largely absent.

Ulcers, Bacteria, and the Cost of Defending a Wrong Answer

For most of the 20th century, the medical consensus on peptic ulcers held that they were caused by stress, diet, and excess stomach acid. The treatment was accordingly oriented toward managing these factors: dietary changes, antacids, stress reduction, and in severe cases, surgery. The consensus was not a fringe position. It was the position of the field's leading practitioners, reflected in textbooks and clinical training programs.

In the early 1980s, an Australian physician named Barry Marshall proposed that a bacterium, Helicobacter pylori, was the primary cause of most peptic ulcers and that the condition was therefore treatable with antibiotics. The response from the established consensus was not cautious skepticism. It was active resistance. Marshall's papers were rejected. His findings were attributed to contamination. The professional machinery of the field mobilized, with considerable energy, against a correct answer.

Marshall eventually resorted to drinking a solution containing H. pylori, developing gastritis, treating himself with antibiotics, and documenting the outcome — a demonstration of personal conviction rarely required of researchers working within a consensus their field accepts. He and his collaborator Robin Warren received the Nobel Prize in Physiology or Medicine in 2005.

The ulcer case is useful precisely because it is not a story of a corrupt institution or a politicized consensus. The physicians who resisted Marshall's findings were not acting in bad faith. They were acting in accordance with the well-documented human tendency to defend prior commitments against disconfirming evidence — a tendency that does not diminish with education and may, in some respects, increase with it.

The Efficient Market and the Crisis That Wasn't Supposed to Happen

The financial crisis of 2008 arrived in a professional environment that had, for decades, developed increasingly sophisticated mathematical frameworks for understanding and managing risk in financial markets. The efficient market hypothesis, in its various forms, had become the organizing consensus of academic finance and had substantially shaped the regulatory philosophy of the institutions responsible for financial stability.

The crisis did not merely cause large losses. It revealed that the risk models on which the financial system had been operating were wrong in ways that their designers had not adequately stress-tested against historical data outside the narrow window they had used for calibration. The instruments that had been rated as low-risk by credentialed analysts at credentialed institutions turned out to contain risks that the consensus framework had no adequate language for.

The Federal Reserve chairman of the period, Alan Greenspan, testified before Congress that he had found a flaw in his ideology — that his model of how the world worked had been incorrect. This was an unusually direct acknowledgment of the pattern this article has been tracing. It was not, however, the first time the pattern had appeared. It was, for anyone reading the record, entirely predictable that it would appear again.

What Deference Actually Costs

None of this argues for dismissing expertise. The person who refuses all medical consensus because physicians once prescribed mercury is making a worse decision than the person who consults current evidence carefully. The question is not whether to consult credentialed authority. The question is how to calibrate the weight you give it.

The historical record suggests a few readings worth keeping. Consensus formed within a community that controls its own credentialing and publication tends to be slower to self-correct than consensus subject to external accountability. Confident prediction in genuinely complex systems — macroeconomics, epidemiology, financial markets — has a documented track record that should inform how much certainty you accept from any single forecast. And the social cost of dissent within expert communities is real and measurable, which means that the absence of visible dissent is not strong evidence that no dissent exists.

History has been running this study for a very long time. The results are available. Reading them is not anti-intellectual. It is, in fact, the most intellectually serious thing you can do with them.