The Economics of Financial Panics

In the summer of 1720, shares of the South Sea Company were trading at roughly ten times their January price. The company’s actual business — an exclusive contract to supply slaves to Spanish colonial ports, combined with a scheme to convert British national debt into company equity — could not plausibly generate returns anywhere near the implied valuation. This was known. Isaac Newton had invested early, made a 100 percent return, sold his shares, and then bought back in as prices continued rising, eventually losing £20,000 — approximately £3 million in today’s money — when the bubble collapsed in September. Newton’s observation on this episode — that he could calculate the motions of the heavenly bodies but not the madness of men — has become the most quoted response to financial folly in history. What Newton did not observe, because the historical record was not yet long enough for anyone to notice, was that the madness he observed in 1720 had a precise structural template, and that the same template would repeat itself in 1797, 1825, 1847, 1857, 1873, 1893, 1907, 1929, 1987, 2000, and 2008. The madness was not random. It was systematic.

Hyman Minsky, the economist who spent his career at Washington University in St. Louis being largely ignored by his colleagues, developed the most complete analytical account of this template. Minsky’s central claim, which he called the financial instability hypothesis, is that financial stability is self-undermining. Periods of economic stability encourage lenders to extend credit more generously, which allows borrowers to take on more debt, which inflates asset prices, which validates the generosity of the lending and encourages further extension of credit, in a feedback loop that inevitably produces a crisis because the asset prices being validated by credit expansion cannot be sustained once the credit expansion ends. This is not a story about irrational exuberance, though irrationality is present. It is a story about the structural dynamics of credit systems in which the extension of credit today changes the conditions under which future credit decisions will be made.

Minsky’s taxonomy of borrowers is the most useful analytical tool in the history of financial crisis literature. He distinguished three types: hedge borrowers, who can service both interest and principal from their income; speculative borrowers, who can service interest but must roll over their principal at maturity; and Ponzi borrowers, who cannot service either interest or principal from income and depend on rising asset prices to refinance their obligations. In a stable economy, most borrowers are hedge borrowers. As stability persists and lenders relax their standards, the proportion of speculative borrowers grows. As the boom matures and asset prices rise faster than any income-based valuation can justify, Ponzi borrowers proliferate. The system becomes structurally fragile at exactly the moment when everyone involved feels most confident, because confidence is what drove the relaxation of lending standards that allowed fragile borrowers to accumulate.

The pattern that Charles Kindleberger described in Manias, Panics, and Crashes, drawing heavily on Minsky, has five stages. The displacement is first: some genuine change in the economic environment — a new technology, a new trade route, a new financial instrument, a change in monetary policy — creates genuine investment opportunities and justifies genuine optimism. The South Sea scheme was a real attempt to refinance British government debt. Railway shares in 1845 represented real railways. American housing prices in 2000–2006 were responding to genuinely increased demand for homeownership backed by genuinely lower mortgage costs. The displacement is not a fabrication; the error comes later.

The credit expansion follows. Lenders see the opportunities the displacement has created and begin extending credit more generously to capture them. This is individually rational: each lender who extends credit profitably is rewarded; each lender who refuses credit while competitors extend it loses market share. The competitive dynamics of banking push toward credit expansion in boom periods regardless of the systemic consequences, because the benefits of lending accrue to individual institutions immediately while the risks accumulate on balance sheets that are harder to read. New financial instruments appear that allow credit to be extended to borrowers who would previously have been excluded — the country bank bills that financed the British Railway Mania, the radio stocks sold on margin in the 1920s, the collateralized debt obligations that packaged subprime mortgage pools in the 2000s. These instruments are typically described by their promoters as innovations that distribute risk more efficiently. In the structural analysis, they are mechanisms that obscure the accumulation of fragility.

Euphoria is the stage in which price behavior decouples from any fundamental valuation and begins to be driven purely by the expectation of further price rises. At this point the Minsky Ponzi borrower is the marginal buyer: someone who cannot justify the purchase on income or cash flow grounds but is buying in the expectation that prices will rise fast enough to allow a profitable sale before the financing burden becomes unmanageable. The euphoria stage is also the stage in which the most prominent skeptics are most publicly humiliated. John Kenneth Galbraith documented this carefully for the 1929 crash: economists and bankers who raised concerns about stock valuations in 1928 were dismissed as cranks and pessimists who had failed to appreciate the new economic reality. The ridicule of skeptics is a consistent feature of the euphoria stage, not because the skeptics are necessarily wrong about valuations but because rising prices provide continuous empirical refutation of their warnings until the very moment they are right.

Distress begins when something forces the Ponzi borrowers to realize that prices are not going to rise fast enough to service their obligations. The trigger varies — a change in interest rates, a credit event elsewhere, an earnings disappointment, a fraud revelation, a policy announcement — but the trigger is almost never the fundamental cause of the crisis. The crisis was built into the structure of debt and asset prices long before the trigger appeared; the trigger merely revealed it. What follows is a reversal of the feedback loops that had been operating during the expansion. Forced sellers appear in the market as borrowers who cannot service their debt are required to liquidate assets. Prices fall. The falling prices impair the collateral values supporting other borrowers’ debt, forcing further liquidation. Lenders who had been extending credit generously now refuse it entirely, even to creditworthy borrowers, because the panic has made it impossible to distinguish solvent from insolvent. The credit contraction reinforces the price decline, which reinforces the credit contraction, in a spiral that can continue until it has destroyed a substantial fraction of the wealth and productive capacity that the preceding boom had created.

Revulsion is the final stage, in which investors refuse to engage with the asset class that had been the object of the mania, regardless of price. Dutch tulip bulbs in 1637. South Sea Company shares in 1721. British railway shares in 1848. American technology stocks in 2002. The asset that had been bid to absurd prices is now regarded with contempt, and the contempt persists long after valuations have become genuinely attractive. This is not irrational — it is a rational response to the discovery that one’s judgment about this asset class has proved badly wrong and that the information available to assess it is systematically unreliable. But the revulsion stage can prolong recessions beyond what the financial damage alone would dictate, because the refusal to reinvest prevents the reallocation of capital to productive uses that recovery requires.

Walter Bagehot’s Lombard Street, published in 1873, provided the institutional solution to the acute phase of the panic. Bagehot observed that bank runs were self-fulfilling: if depositors believe a bank may fail, they withdraw deposits, which causes the bank to fail, which validates the belief. The individual rationality of each depositor guarantees the collective irrationality of the outcome. The only way to break this dynamic, Bagehot argued, is for the central bank to act as lender of last resort — to lend freely to any bank that can offer good collateral, at a penalty rate, without limit. The penalty rate prevents solvent institutions from borrowing unnecessarily; the unlimited supply prevents the panic from propagating; the good collateral requirement distinguishes illiquidity (a temporary cash shortage in an otherwise sound institution) from insolvency (an institution whose assets are genuinely worth less than its liabilities). Bagehot’s rule is simple: lend freely at high rates against good collateral. The difficulty is in distinguishing good collateral from bad in the middle of a panic, when asset prices are collapsing and balance sheets are opaque.

The 2008 financial crisis was in many ways a pure Minsky cycle: the displacement of low interest rates and financial innovation in mortgage markets, the credit expansion through securitization and CDO structures that allowed mortgage credit to be extended to borrowers who could not service it from income, the euphoria of rising house prices and the ridicule of those who doubted they could fall nationally, the distress triggered by rising delinquencies in subprime pools, and the revulsion that followed. The Federal Reserve applied Bagehot’s doctrine imperfectly but recognizably — lending freely to institutions that claimed collateral, at effectively zero rates, which violated Bagehot’s penalty rate requirement but arguably reflected the scale of the crisis. The financial system was stabilized; the broader economy was not, because the Bagehot doctrine addresses the acute panic phase but says nothing about the prolonged deleveraging and demand depression that follow the unwinding of a large credit cycle.

The question that the persistent recurrence of financial panics raises most insistently is why historical memory fails to prevent repetition. The answer is structural rather than psychological. Each generation of financial participants knows, abstractly, that credit cycles end in crises. The knowledge is not secret. Kindleberger’s book had been in print for decades before 2008. What each generation cannot know is when the current cycle will end, and the competitive dynamics of financial markets punish early exit as severely as late exit. A fund manager who reduced exposure to mortgage securities in 2004 because the valuations seemed unsustainable lost clients to competitors who remained invested through 2006 and 2007. The skeptic who is right eventually is often ruined before the market confirms the analysis. This is not irrationality. It is the structural consequence of the fact that financial markets reward performance relative to peers, and relative performance during a boom requires participation in the boom. Minsky’s insight is ultimately not that participants are foolish but that the incentive structure of financial markets systematically pushes rational participants toward collectively dangerous behavior. That structure has not changed, which is why the template has not changed, and why the next iteration of the cycle is not a question of whether but of when.