The anatomy of a crisis: the Minsky and Kindleberger framework

Every financial crisis arrives dressed in the costume of its era. Bulbs, railway shares, Thai baht, mortgage-backed securities, deposits at a bank that could be emptied by an app. The costumes are what people remember, and they are the least useful part of the record, because the next crisis will wear a different one.

Underneath, the arithmetic repeats. Two writers described that arithmetic better than anyone else. Hyman Minsky, an American economist who spent most of his career outside the mainstream, explained why a financial system generates its own fragility from the inside, with no external shock required. Charles Kindleberger, an economic historian, borrowed Minsky's mechanism and fitted it to five centuries of episodes, producing a narrative sequence that reads uncomfortably like a script. This lesson sets out both, shows the balance-sheet arithmetic with numbers, and is honest about what the framework cannot do.

Minsky's three financing postures

Minsky's core idea is a classification of borrowers, not of assets. Take any indebted entity: a household, a property company, a bank, a state. Compare the cash flow the entity's assets actually produce against the cash it owes.

A hedge unit produces enough operating cash flow to cover interest and scheduled principal repayment out of income. It never needs to refinance, and it never needs to sell an asset to meet an obligation.

A speculative unit produces enough to cover interest but not principal. It is solvent on an income basis but must roll its debt when it matures. Its survival depends on funding markets staying open, not on its own earnings.

A Ponzi unit does not produce enough to cover even interest. It must borrow more, or sell assets, simply to service what it already owes. Minsky's use of the word is descriptive, not accusatory: a start-up burning capital or a developer holding land through construction can be a Ponzi unit for perfectly honest reasons.

Numbers make the boundaries sharp. Consider an office building bought for CHF 10 million, funded with CHF 8 million of debt and CHF 2 million of equity, generating net operating income of CHF 400,000 a year after costs.

At an interest rate of 2.5 percent, interest is CHF 200,000. Add contractual amortisation of CHF 100,000 and total debt service is CHF 300,000 against CHF 400,000 of income. This is a hedge unit, with CHF 100,000 to spare.

At 4 percent, interest is CHF 320,000. Income still covers it, but only CHF 80,000 is left against a CHF 100,000 amortisation requirement. The owner must negotiate, refinance or inject equity: a speculative unit.

At 5.5 percent, interest is CHF 440,000 against CHF 400,000 of income. The gap must be filled by borrowing or by selling. That is a Ponzi unit, and nothing about the building has changed.

Why calm produces fragility

The classification would be a static curiosity if the mix of the three postures were stable. Minsky's claim, the financial instability hypothesis, is that the mix drifts predictably in one direction during good times.

The mechanism is not psychological hand-waving. It runs through observed data. In a long expansion, realised default rates fall. Lenders price credit off realised losses, so risk premia compress and covenants loosen, because the covenant that was never breached looks unnecessary. Borrowers who took hedge-style leverage watch competitors take speculative leverage and earn higher returns on equity, and competitive pressure pushes them the same way. Rising asset prices raise the value of collateral, which mechanically expands how much can be borrowed against the same asset, which supports asset prices further.

Each step is individually rational and locally prudent. The aggregate result is that the share of the system in speculative and Ponzi postures rises for as long as conditions stay benign. Minsky's compressed statement of this is that stability is destabilising. That is the whole engine, and everything else in the framework is commentary on it.

The corollary matters for anyone reading risk indicators. Volatility, credit spreads and default rates are at their most reassuring exactly when fragility is at its highest, because they are backward-looking measures of a period during which nothing was tested.

Kindleberger's sequence

Kindleberger organised episodes into stages. The labels are his and Minsky's, refined across editions of Manias, Panics and Crashes and later carried on by Robert Aliber.

Displacement. An exogenous change alters profit opportunities in some sector: a new technology, a war's end, a deregulation, a durable fall in interest rates, the opening of a market. The change is usually real, which is why the story is credible.

Credit expansion. Funding grows to chase the opportunity, and it typically grows outside the regulated core, in whatever institution is least constrained at that moment.

Euphoria. Prices rise beyond what the displacement justifies. Participation broadens to people with no analytical stake in the asset. Valuation arguments shift from cash flow to comparison, and finally to the argument that supply is limited.

Distress. Insiders sell. Some prominent participant fails for reasons that look idiosyncratic. Prices stop rising, which is fatal for any unit whose plan required selling at a higher price.

Revulsion and panic. The rush for the exit. Credit is withdrawn from everyone in the affected class regardless of individual quality, because in a panic no lender has time to distinguish.

The final element in Kindleberger's account is the lender of last resort, and here he leans on Walter Bagehot's Lombard Street of 1873: in a panic, lend freely, at a penalty rate, against collateral that would be good in normal times. Kindleberger's addition was the observation that the lender of last resort must be ambiguous about whether it will act, or the expectation of rescue accelerates the next credit expansion.

The turn: why the down leg is faster

Manias take years. Panics take days. The asymmetry has a mechanical cause, not an emotional one.

On the way up, leverage is added deliberately, at a pace set by the borrower. On the way down, leverage is removed involuntarily, at a pace set by the lender. A margin call, a collateral haircut increase, a refusal to roll commercial paper, a deposit withdrawal: each forces a sale on a timetable the seller does not control.

Those sales lower prices, which lowers the value of the collateral held by every other borrower against the same asset class, which triggers the next round of calls. Irving Fisher described this loop in 1933 as debt deflation: the attempt to repay debt en masse lowers asset prices and raises the real burden of the remaining debt, so the more the system pays down, the more it owes in real terms.

Return to the building. At a capitalisation rate of 4 percent, income of CHF 400,000 supports a value of CHF 10 million. If required yields rise so that buyers demand 5 percent, the same income supports CHF 8 million. The property's value has fallen 20 percent, and the equity value has fallen from CHF 2 million to zero, a 100 percent loss on the equity. At a 5.5 percent capitalisation rate the value is roughly CHF 7.3 million and the equity value is negative by about CHF 0.7 million. The income never changed. Only the rate at which the market discounts it did.

What the framework cannot do

The framework is a diagnostic vocabulary, not a timing tool, and it should be held to that standard.

It has no clock. A system can sit in the euphoria stage for years, and a large amount of money has been lost by people who identified a mania correctly and early. It is also close to unfalsifiable in its narrative form, since almost any episode can be fitted to five flexible stages after the fact. Eugene Fama's objection to the word bubble is serious and worth holding onto: if the label can only be applied after the price falls, it carries no information beforehand.

The empirical repair is to test the credit-expansion step rather than the psychology. Moritz Schularick and Alan Taylor's work on more than a century of data across advanced economies found that rapid growth in bank credit relative to output is a statistically meaningful predictor of subsequent financial crises. That is the part of Minsky that survives contact with a regression, and it is why supervisors now watch the credit-to-GDP gap rather than watching sentiment. Carmen Reinhart and Kenneth Rogoff added the historical companion: the recurring belief that the current expansion is structurally different, which is where their title, This Time Is Different, comes from.

The Swiss version of the same mechanism

Switzerland is a useful place to see the credit channel operating slowly and visibly, because its household debt-to-income ratio is among the highest in the developed world and the reason is structural rather than speculative.

Mortgage interest is deductible from taxable income, and until recently the offsetting item was the imputed rental value, Eigenmietwert, an amount added to the taxable income of an owner-occupier representing the rent they implicitly pay themselves. The combination gave a durable incentive not to amortise: a mortgage carried indefinitely converts into an annual tax shield. Swiss voters accepted a change of system in a 2025 federal vote, with implementation to follow on a timetable set separately, but the stock of debt that the old rules encouraged is still on household balance sheets.

Against that, the supervisory architecture is unusually explicit. Bank self-regulation, recognised as a minimum standard by FINMA, requires a buyer to bring at least 10 percent of the lending value in funds that do not come from pillar-2 pension assets, and requires the mortgage to be amortised down to two-thirds of the lending value within a defined period. Affordability is tested not at the market rate but at an imputed rate several percentage points higher, which is a direct attempt to keep new borrowers in Minsky's hedge category even after rates rise. Switzerland was also the first country to activate the Basel III countercyclical capital buffer, in 2013 at 1 percent of risk-weighted domestic mortgage exposures, doubled the following year, released in March 2020 when the pandemic hit, and reactivated in 2022 at 2.5 percent.

Whether those tools are sufficient is an open question. What they demonstrate is that the Minsky framework has been absorbed into policy: the buffer is built up in calm periods precisely because calm periods are when fragility accumulates, and it is released in stress so that banks can absorb losses without contracting lending.

What to take away

• Minsky's classification is about cash flow versus obligations, not about asset quality: the same building is a hedge, speculative or Ponzi unit purely as a function of the interest rate and the amortisation schedule attached to it.
• The share of speculative and Ponzi financing rises during long calm periods because lenders price risk off realised losses, and realised losses are lowest exactly when nothing has been tested.
• Panics move faster than manias because deleveraging happens on the lender's timetable through margin calls and refused rollovers, while leverage was added on the borrower's timetable.
• A small change in the rate at which the market capitalises an income stream produces a very large change in equity value when the asset is levered: a 20 percent fall in asset value wiped out 100 percent of the equity in the worked example.
• The durable, testable residue of the framework is credit growth, which is why supervisors track credit-to-GDP gaps and countercyclical buffers rather than trying to measure sentiment.

Sources

• Hyman P. Minsky, Stabilizing an Unstable Economy (Yale University Press, 1986).
• Hyman P. Minsky, "The Financial Instability Hypothesis", Levy Economics Institute Working Paper No. 74 (1992).
• Charles P. Kindleberger and Robert Z. Aliber, Manias, Panics and Crashes: A History of Financial Crises (multiple editions since 1978).
• Walter Bagehot, Lombard Street: A Description of the Money Market (1873).
• Irving Fisher, "The Debt-Deflation Theory of Great Depressions", Econometrica (1933).
• Moritz Schularick and Alan M. Taylor, "Credit Booms Gone Bust: Monetary Policy, Leverage Cycles, and Financial Crises, 1870–2008", American Economic Review (2012).
• Carmen M. Reinhart and Kenneth S. Rogoff, This Time Is Different: Eight Centuries of Financial Folly (Princeton University Press, 2009).
• Swiss National Bank, Financial Stability Report (annual), and FINMA's recognition of the Swiss Bankers Association guidelines on mortgage financing as a minimum standard.