A Failure of Imagination

Growth in a Time of Debt

Verified 8 Sep Carmen Reinhart & Kenneth Rogoff, Harvard · cited across European and US fiscal policy

What was claimed That when gross government debt exceeds 90% of GDP, growth falls sharply — the paper reported median growth turning slightly negative above that line. Presented as a threshold, it supplied the empirical case for rapid deficit reduction after 2008.

How load-bearing it was Cited by name by the Chairman of the US House Budget Committee, by the European Commissioner for Economic and Monetary Affairs, and across the UK Treasury's case for consolidation. It was, for a period, the single most influential empirical claim in Western macroeconomic policy.

What happened In 2013 Thomas Herndon, a graduate student at UMass Amherst, obtained the spreadsheet and found three problems. A coding error — the averaging formula omitted five countries. Selective exclusion of available country-years. And an unconventional weighting that gave one year of one country's data the same weight as nineteen years of another's. Corrected, the cliff disappeared: growth above 90% debt averaged around 2.2%, not the reported negative figure. The relationship became a gentle negative association, with the direction of causation unresolved — slow growth raises debt ratios at least as readily as the reverse.

The harm Not attributable with precision, and you should not pretend otherwise — austerity had political constituencies that did not depend on this paper. But the paper supplied the empirical authority, and the policy it authorised is associated with measurable outcomes: Greek GDP contracted by roughly a quarter, Greek and Spanish youth unemployment exceeded 50%, and studies have linked the period's cuts to excess mortality and to increases in suicide in several European countries. State it as authority supplied, not as deaths caused, and the claim is unanswerable.

Why it is the study's model record Because everything is documented and nothing is contested. The error is arithmetic. The citation chain is on the record. And the correction came from outside the institution, from a student who asked for the data — which is the same mechanism that overturned Study 329 in Pillar I after fourteen years, and forced out the Tamiflu clinical study reports. Three pillars, one countermeasure: somebody outside got the underlying data.

Who gained The authors, in influence — both remained eminent, and no professional penalty followed. Governments committed to consolidation, which acquired a citation for a position already held. Bondholders. And the discipline of economics, which gained a headline finding of the kind that rarely emerges from it. Note what did not happen: the paper was never retracted, because a spreadsheet error is not misconduct, and there is no mechanism for withdrawing a claim that has already moved a continent's policy.

The IMF's own multipliers

Verified 8 Sep International Monetary Fund · Blanchard & Leigh, 2013

What was assumed That the fiscal multiplier — the output lost per unit of spending cut — was around 0.5. On that assumption austerity was comparatively cheap, and the IMF's programme forecasts were built on it.

What happened Olivier Blanchard, the IMF's own chief economist, and Daniel Leigh published Growth Forecast Errors and Fiscal Multipliers as IMF Working Paper 13/001 on 3 January 2013 (verified 8 Sep). Examining the gap between forecast and actual growth across advanced economies, they estimated the multiplier for European countries in 2010–11 at around 1.6 — roughly three times the 0.5 the Fund had been assuming — with the relationship strongest early in the crisis. In plain terms, the institution formally established that its own forecasts had systematically understated the damage its programmes would do, after the programmes had been imposed.

Why it belongs beside Reinhart and Rogoff Because it is rarer and better than a critic's finding: an institution auditing itself and publishing the adverse result. That deserves recording as creditable conduct. It is also, unavoidably, an admission that hundreds of millions of people were governed for several years on a parameter the issuing body later found to be wrong by a factor of two or three — and no consequence attached to the finding.

Who gained From the original assumption: creditor institutions and governments needing programmes to appear survivable. From the correction: the IMF's own credibility, which is a real institutional asset and is why the paper could be published at all.

The Post Office Horizon prosecutions

Verified 8 Sep Post Office Ltd · Fujitsu · the Crown Prosecution Service · government as shareholder

The false fact That the Horizon accounting system was reliable, and that discrepancies in a branch's accounts therefore proved the sub-postmaster had taken the money. The Post Office asserted this in hundreds of prosecutions it brought itself, using a private prosecution power, over fifteen years.

What was actually true Horizon contained numerous defects, and Fujitsu staff could and did alter branch account data remotely, without the sub-postmaster's knowledge — a capability the Post Office repeatedly denied existed. In Bates and others v Post Office (2019) Fraser J found the system was not remotely robust, and that the Post Office's approach amounted to asserting that the earth is flat. He referred the matter to the Director of Public Prosecutions.

The harm More than 900 prosecutions, some 700 brought by the Post Office itself. Convictions for theft and false accounting. Imprisonment, including of a pregnant woman. Bankruptcies, home losses, marriages destroyed, communities turned against people who had run their post office for decades. At least four suicides, and dozens who died before exoneration. The Court of Appeal in 2021 quashed 39 convictions, finding the conduct an affront to the conscience of the court. Statutory exoneration followed in 2024, twenty-five years after the first prosecution.

Why it is the single most important record in the study Because it strips away every mitigating explanation the other pillars offer. There was no scientific uncertainty. No incomplete evidence. No forecast under conditions nobody could know. Sub-postmasters told the Post Office from the beginning; Computer Weekly published in 2009; Private Eye pursued it for years; the Post Office's own commissioned investigators, Second Sight, reported defects in 2013 and were dismissed and their work terminated. Every warning arrived, was understood, and was overridden — because accepting it meant accepting that the organisation had already imprisoned innocent people. The cost of admission rose with every year of denial, which is precisely why it did not come.

The generalisable finding That last sentence is the mechanism the whole study has been circling. It explains Semmelweis's rejection, the twenty-five years of antenatal X-ray after Alice Stewart, the eighteen years of DES after the negative trial, the withheld climate emails, and the infected blood cover-up. An institution's resistance to a correction scales with what admitting it would cost, not with the strength of the evidence against it. That is testable, and it is what your study should be trying to prove.

Who gained The Post Office, which recovered alleged shortfalls directly from sub-postmasters — it took money from the people it prosecuted, and the branch contracts made them personally liable. Fujitsu, whose contracts continued and were extended for two decades, and which continued to win UK government work throughout. Executives, on bonuses tied to performance the false accounting flattered. And the government as sole shareholder, which avoided the cost of admitting a flagship IT programme had failed. Compensation is now expected to exceed a billion pounds of public money — far more than honesty would have cost at any point. But the enabling beneficiary was not an organisation at all. It was a presumption: that computer evidence is reliable unless the contrary is shown. That doctrine put the burden on sub-postmasters to prove a negative about a system they were not permitted to inspect, and it did the work that no amount of Post Office malice could have done alone. Beneath it sat the institution's belief in its own character — a trusted British body could not be the kind of body that prosecutes the innocent in the hundreds, so the evidence that it was doing so was read as evidence about the accused.

Wirecard, and the regulator that attacked the reporters

Verified 8 Sep Wirecard AG · BaFin · EY · the Financial Times

What happened The Financial Times, principally Dan McCrum, reported from 2015 onward that Wirecard's accounts did not add up. In 2019 BaFin responded by banning short-selling of Wirecard shares for two months — the first time the regulator had ever extended that protection to a single listed company — and filing criminal complaints against two FT journalists on suspicion of market manipulation, alleging coordination with short sellers (verified 8 Sep). Investigators found the reporting fundamentally accurate, found no evidence of contact with short sellers or of insider information being passed, and prosecutors dropped the case in September 2020.

What was true €1.9bn of supposed trust account balances did not exist. Wirecard collapsed in 2020. Its chief operating officer absconded. The auditor had signed the accounts for years.

Why it is the study's sharpest instance of a warning refused Because the regulator did not merely fail to act. It used its statutory powers against the people who were correct, and did so in defence of a national champion. Set it beside the Post Office suing sub-postmasters and the Bendectin plaintiffs' bar in Pillar I, and the pattern is that institutional power is most dangerous when deployed to protect a position rather than to test one.

Who gained Wirecard's executives and shareholders, for five extra years. The auditor, in fees. Germany's financial-centre ambitions, briefly. BaFin's leadership was replaced afterwards, which is more consequence than most records here produce.

Personal pensions mis-selling

Verified 8 Sep UK government policy · life assurance industry · FIMBRA/LAUTRO

What was claimed That transferring out of an occupational pension into a personal one would leave savers better off. Government actively encouraged it, with incentives and advertising, as part of extending private provision.

What happened — corrected (verified 8 Sep) The Pension Review, established by the regulator in 1994, covered people who left or declined an occupational scheme for a personal pension between 29 April 1988 and 30 June 1994. Up to 1.2 million people had been awarded compensation by mid-2002, and the total cost of the clean-up reached around £13.5bn. Miners, nurses and teachers were among them, many giving up employer contributions and guaranteed benefits for products with front-loaded charges. My first draft said 2.4 million people and £11.8bn: the headcount was roughly double the awarded figure and the money was understated.

Why it belongs in a study of institutions Because this was not merely industry misconduct. Government policy created the incentive and the advertising campaign, and the regulatory architecture was self-regulatory and demonstrably inadequate. The state was upstream of the harm.

Who gained Life offices, on commission and charges. Sales forces, on commission structures that made a transfer worth more than good advice. And the Treasury, on a policy objective advanced at industry's expense rather than the Exchequer's. Name the objective, because it is the largest beneficiary here: the political conviction that individual ownership is superior to collective provision. That belief was sincerely held, widely shared and electorally popular, and it made leaving a good occupational scheme look like emancipation rather than loss. The commission explains which salesman knocked; it does not explain why the door was open.

Endowment mortgages

Verified 8 Sep UK lenders and life offices · regulators

What was claimed Pay interest only and invest the rest; the policy would clear the mortgage and leave a surplus. Illustrations used projected annual returns around 10–12%, presented to borrowers as expectations rather than as illustrations.

What happened — corrected (verified 8 Sep) By 2003 there were about 2.2 million households holding an endowment-linked mortgage with a projected shortfall on the loan secured against their home. My first draft said "around 5 million policies were sold" and I cannot support it. The redress picture also differed from the pensions scandal: the FSA declined a proactive industry-wide review, estimating the administrative cost at around £5bn and judging it disproportionate, and instead required firms to write to customers about projected shortfalls, leaving complaints to be pursued individually through the Financial Ombudsman Service. Around 80% of reviewed cases received redress, at average amounts roughly a fifth of those in pensions mis-selling.

The mechanism, familiar by now A projection generated under favourable assumptions, communicated without its range, to an audience with no means of evaluating it. That is the reasonable-worst-case failure from Pillar II with the sign reversed — here the best case was communicated as the expectation, because the incentive pointed that way. Same operation, opposite direction, same removal of the range.

Who gained Life offices, on premiums and charges. Lenders and brokers, on commission — an endowment paid the adviser far more than a repayment mortgage did, and that fact alone predicts the entire episode.

Payment protection insurance

UK banks · FSA

What was claimed Cover that would meet loan repayments if the borrower lost income. Sold alongside loans, mortgages and credit cards, frequently added without clear consent, and often to people the policy's own exclusions made ineligible to claim — the self-employed, and those with pre-existing conditions.

What happened The largest consumer redress episode in British history. Banks paid out £38.3bn in the period from January 2011 to the claims deadline of 29 August 2019 (verified 8 Sep); wider estimates including administration costs run considerably higher. Regulators had been warned about the product from the late 1990s, and the Citizens Advice super-complaint came in 2005; meaningful action took several more years.

The scale point worth making The PPI redress bill alone exceeds the entire cost of most disasters in the other three pillars combined. It attracted a fraction of the attention, because the harm was distributed in small amounts across millions of people rather than concentrated. That is a finding about what gets studied, not about what causes harm.

Who gained The banks, on margins reported at times above 80% — among the most profitable retail products ever sold in Britain. Sales staff on incentives. And the redress industry that grew up afterwards, which took a substantial share of the compensation.

Equitable Life

Verified 8 Sep Equitable Life Assurance Society · DTI · FSA · Treasury

What was claimed The world's oldest mutual life office, trading on 250 years of reputation, sold policies carrying guaranteed annuity rates and then, as rates fell, sought to avoid honouring them by cutting bonuses for those who exercised the guarantee.

What happened The directors realised in the early 1990s that the guarantees could not be met and resolved in 1993 to pay lower bonuses to the policyholders who held them. The House of Lords struck that down in Equitable Life Assurance Society v Hyman on 20 July 2000, holding the differential bonus policy unlawful; the society closed to new business that December (verified 8 Sep). Around a million policyholders lost substantial parts of their pensions. The Parliamentary Ombudsman's eight-year investigation reported in 2008 under the title Equitable Life: a decade of regulatory failure, recommending that public bodies apologise and that the government fund a compensation scheme. The scheme eventually paid a fraction of assessed losses; many policyholders died waiting.

Why it belongs The reassurance came from the reputation and from the regulator's continued authorisation, not from the product. A 250-year record is a claim about the past presented as a guarantee about the future, and it is the oldest form of the mis-sold projection.

Who gained Policyholders who left before the collapse, at the expense of those who stayed. Management, on the growth the guarantees attracted. And the Treasury, which limited its liability for a failure its own regulators had been found responsible for.

The Private Finance Initiative

Verified 8 Sep UK Treasury · successive governments · construction and finance consortia

What was claimed That private finance delivered better value than government borrowing, through risk transfer and private-sector efficiency. Around 700 projects were signed — hospitals, schools, roads.

What happened — with a correction to my own figures (verified 8 Sep) The National Audit Office reported on 18 January 2018 that there were more than 700 operational PFI and PF2 deals with a capital value of £60bn, annual charges of £10.3bn in 2016–17, and future charges running to the 2040s totalling £199bn. It found little evidence that the arrangements had delivered value for money, and no information establishing that PFI beat conventional procurement; it also put the cost of private financing at up to 40% above government borrowing. My first draft said "roughly £13bn of capital against over £200bn of obligations", which is wrong — I had the capital figure out by a factor of more than four and made the ratio look far more extreme than the NAO's own numbers support. Carillion's collapse weeks later showed the risk had not in fact transferred.

Why the accounting is the point PFI's durable attraction was that it kept the liability off the government balance sheet. It is the same operation as counting imported wood pellets as zero emissions in Pillar III: an accounting convention that made a real obligation invisible, chosen because it was invisible.

Who gained Consortia and their lenders, on index-linked returns over 25 to 30 years, with equity stakes traded repeatedly at large gains. Advisers, on fees. And every government of the period, which obtained buildings without recording debt. Two beliefs did the work the money could not. First, that private management is inherently more efficient than public — held across both front benches for thirty years, tested rarely and never as a condition of proceeding. Second, an accounting convention: that a liability not recorded is a liability not incurred. Neither was anybody's property, both were free to hold, and together they committed £199bn of public money against £60bn of assets.

AAA

Verified 8 Sep Standard & Poor's · Moody's · Fitch

What was claimed That mortgage-backed securities and collateralised debt obligations assembled from subprime loans carried the same default risk as sovereign debt. The AAA rating was not marketing — it was the legal and regulatory permission that let pension funds, insurers and banks hold the paper at all, and set the capital they had to reserve against it.

What happened Trillions of dollars of AAA-rated securities were downgraded, many to junk, within months. The Financial Crisis Inquiry Commission concluded that the credit rating agencies were essential cogs in the wheel of financial destruction (verified 8 Sep). Internal messages disclosed in litigation and to Congress included an analyst's remark that a deal could be structured by cows and we would rate it. S&P settled in February 2015 for $1.375bn with the Department of Justice, nineteen states and the District of Columbia, plus $125m to CalPERS; Moody's settled with the DoJ and twenty-one states for almost $864m in January 2017 (verified 8 Sep). Neither admitted a legal violation. Neither admitted wrongdoing and no individual was criminally convicted.

The structural fact The agencies are paid by the issuers of the securities they rate, and issuers can take a deal to a competitor. That is the identical arrangement as the private notified bodies that CE-marked the PIP breast implants in Pillar I, and the carbon-credit certifier in Pillar III. Three pillars, one design fault: the assessor is paid by the assessed, and the assessed can shop. It has never been fixed in any of the three markets.

Who gained The agencies, whose structured-finance revenue rose several-fold in the boom. Issuers, who obtained the rating that made the paper saleable. Banks, on origination and distribution fees. And regulators, who had outsourced a judgement they would otherwise have had to make themselves — which is why the arrangement was allowed to exist.

Long-Term Capital Management

Verified 8 Sep LTCM · Myron Scholes and Robert C. Merton, Nobel laureates 1997

What was claimed That option-pricing and convergence models permitted very large leverage at low risk, because the historical distribution of returns bounded the possible loss. Two of the principals received the Nobel Memorial Prize in Economic Sciences in 1997.

What happened In 1998 the Asian and Russian crises produced correlated moves the models treated as effectively impossible. The fund lost $4.6bn in under four months, and on 23 September 1998 the Federal Reserve Bank of New York convened fourteen institutions to put in $3.65bn, on the grounds that an unwinding threatened the system (verified 8 Sep).

The detail that connects it to Pillar I The Nobel preceded the collapse by one year. Egas Moniz received the Nobel for the lobotomy in 1949, at the height of a practice that no trial supported. A prize is an institution's assessment, made at a moment in time, and it does not update. Both prizes stand.

And the warning nobody acted on The lesson — that models calibrated on normal conditions fail precisely when they are needed — was available in 1998, published, and widely discussed. Ten years later a Goldman Sachs chief financial officer explained the crisis by saying the firm had seen 25-standard-deviation moves, several days in a row. Under the model's own assumptions such a run is impossible in the lifetime of the universe. The correct inference is that the model was wrong, and it was offered instead as an explanation of why nobody could have known.

Who gained The principals, on fees, until they did not. Counterparty banks, on business they extended on the strength of the names. And the intellectual prestige of quantitative finance, which was barely dented and expanded enormously over the following decade.

Enron and Arthur Andersen

Verified 8 Sep Enron · Arthur Andersen · Fortune · sell-side analysts

What was claimed That Enron was the most innovative company in America — named so by Fortune for six consecutive years, 1996 to 2001 — with audited accounts signed by one of the five great accountancy firms, and buy recommendations from most major banks weeks before bankruptcy.

What happened — with two corrections to my own text (verified 8 Sep) Losses and debt were concealed in off-balance-sheet vehicles. Andersen destroyed documents and was destroyed in turn; at its height it employed more than 85,000 people in 84 countries, and I previously wrote that 85,000 "lost their jobs", which conflates the firm's global headcount with the collapse's toll. At Enron itself roughly 25,000 employees lost their jobs, along with about $2bn in pension savings and $1.2bn in retirement funds — more than half of employees' 401(k) holdings had been in Enron stock. Second correction: Andersen's non-audit fees from Enron were about $27m a year against $25m in audit fees — consulting did exceed audit, but only slightly, and my phrasing implied a wider gap than the record shows. The Houston office was billing around $1m a week in total.

The generalisable finding Every independent check in the chain was paid by the party being checked — the auditor by the company, the analyst by a bank pursuing its underwriting business, the magazine by an advertising market that rewarded access. Sarbanes-Oxley followed in 2002 and separated audit from consulting; it did not change who pays the auditor, and that remains true today.

Who gained Executives, on stock sold while employees were locked in. Andersen, on consulting fees exceeding its audit fee. Banks, on underwriting and on structuring the vehicles that hid the debt. And an idea that everybody respectable held: mark-to-market as a truer accounting, and shareholder value as the sole measure of a firm. Enron was not a pariah being ignored — it was Fortune's most innovative company six years running, taught in business schools and held up by the business press as the model. The analysts, journalists and academics who praised it took no bribe. They were positions in an idea, and the idea was the thing being defended when the questions started.

The Exchange Rate Mechanism

Verified 8 Sep UK Treasury · Bank of England · both main parties

What was claimed That entering at DM2.95 and holding the parity would deliver low inflation and stability. The commitment was stated in the strongest terms available to a government, and the political consensus behind it was near-total.

What happened On 16 September 1992 the Bank raised base rates from 10% to 12% and announced 15% in a single day while buying its own currency, and then withdrew. Treasury papers released in 2005 put the cost of defending the currency at approximately £3.3bn (verified 8 Sep); the 15% announcement was withdrawn before it ever took effect. Households on variable mortgages absorbed the rate movements directly, during a period of widespread negative equity and repossession.

Why it is in a study about information Because the certainty was the policy. A currency peg's only mechanism is the market's belief that the commitment will hold, which means that expressing public doubt is itself the failure. That gives an institution a structural reason to state more confidence than it has — and it is the cleanest example in the study of a case where honest uncertainty was genuinely not an available option. Record it, because it is the strongest defence any institution in these registers can offer, and it applies to almost none of the others.

Who gained Speculators positioned against the parity, most famously to a reported profit of around $1bn. Longer term, the episode is credited with the framework that followed — inflation targeting and eventually an independent Bank — so the beneficiaries include the subsequent monetary settlement. Before that, the beneficiary was a doctrine: that an external exchange-rate anchor was the way to import credibility and break inflationary expectation. It was held with conviction across the Treasury, the Bank and both parties, and it is why the defence continued past the point of arithmetic — rates were raised twice in a single day in support of a parity the market had already priced as untenable. That is not what a calculation looks like. It is what a commitment looks like.

Brooksley Born and derivatives

Verified 8 Sep Commodity Futures Trading Commission · the Treasury · the Federal Reserve · Congress

The warning As CFTC chair, Born issued a Concept Release on 7 May 1998 inviting comment on whether existing regulation of the over-the-counter derivatives market — then about $29 trillion in notional value, with no reporting and no capital requirements — remained appropriate (verified 8 Sep).

What happened Within hours of publication, Treasury Secretary Robert Rubin, Federal Reserve Chairman Alan Greenspan and SEC Chairman Arthur Levitt issued an unusual joint statement declaring their grave concerns. They testified against her proposals and persuaded Congress to strip her agency of the relevant power; Congress imposed a moratorium on CFTC rulemaking. Born left at the end of her term in 1999 (verified 8 Sep). The Commodity Futures Modernization Act of 2000 expressly barred regulation of the market. Credit default swaps on mortgage securities grew inside that exemption and were central to 2008.

Why it is the pillar's Alice Stewart Correct, specific, in office, with the authority to act, and stopped by the most senior figures in the system — one of whom later told Congress he had found a flaw in the model of how the world worked. The information was not missing. The willingness to act on it was.

Who gained The banks took the money; the doctrine won the argument, and the doctrine was what made the money possible. What Greenspan, Rubin, Summers and Levitt defended was not a revenue stream but a worldview — that markets self-regulate, that counterparty discipline outperforms supervision, and that regulating an instrument nobody had yet been harmed by would destroy a market. Greenspan's later admission is as clean as the century offers: he was not bought, he was convinced. A purchased position can be exposed; a sincere one has to be refuted, and by the time the refutation arrived it cost several trillion dollars.

Harry Markopolos and Madoff

Verified 8 Sep US Securities and Exchange Commission

The warning Markopolos first approached the SEC in 1999, walked his findings into its Boston office in May 2000 and resubmitted with more documents in 2001, and was ignored each time. His 2005 memorandum — The World's Largest Hedge Fund is a Fraud, nineteen pages listing twenty-nine red flags — set out that the returns were arithmetically impossible, that the auditor was a one-man storefront operation, that there was no independent custodian, and that the near-absence of losing months was not credible (verified 8 Sep).

What happened Nothing, for nine years: on the 2005 memo the SEC opened a file, made a few calls and closed it. Its own Inspector General later documented that the agency had received credible and specific complaints and conducted inadequate examinations, and the Senate Banking Committee took evidence on that report in September 2009. The scale is usually quoted two ways and both are correct: about $65bn of fictitious account value on the fabricated statements, against roughly $17.5bn of actual principal invested. Charities, foundations and individual retirements were destroyed.

The finding This was not a failure of imagination in the British Academy's sense. It was a failure to act on arithmetic that had been handed to the regulator in writing, more than once, by someone showing his working. The Queen's question has a different answer here: somebody did notice, and told the people whose job it was.

Who gained Madoff, for a further nine years — and the SEC, which was spared the cost of investigating a former chairman of NASDAQ. His standing was itself treated as evidence: a man of that position was not the kind of man the hypothesis described, so the hypothesis was implausible before the arithmetic was examined. And Markopolos was a competitor, which let his motive be discounted instead of his mathematics being checked. An ad hominem did the work an analysis should have done, and it was very much cheaper.

Rajan at Jackson Hole

Verified 8 Sep IMF chief economist · assembled central bankers

The warning Has Financial Development Made the World Riskier?, delivered at Jackson Hole in August 2005 to the annual gathering of the world's central bankers (verified 8 Sep). The argument was specific: financial-sector pay structures combined with complex products gave bankers large cash incentives to take risks with other people's money and imposed no penalty on them for the losses.

What happened Larry Summers rose from the audience to call the paper's premise slightly Luddite and largely misguided, warning that regulation would cut the sector's productivity. Rajan later described the reception as being like an early Christian among lions.

Why the vocabulary matters The dismissal was not evidential. It was social — the argument was treated as unserious because of who was making it and where. That is the same move as the profession's response to Semmelweis, to Alice Stewart and to Barry Marshall in Pillar I. "That's alarmist" is a claim about the world, and it can be false, and it is made with the full confidence of the institution making it.

Who gained The room. What was defended at Jackson Hole was the intellectual settlement on which the assembled careers rested — that financial innovation had distributed risk rather than concentrated it, and that the Great Moderation was an achievement of policy rather than a run of weather. Nobody present was paid to believe that. They had built their professional lives on it, which is a stronger tie than payment and a much harder one to disclose.