The Asian Crisis, 1997Start here

Currency pegs, borrowing in a currency you do not earn, and the fastest reversal of capital flows on record. The original template for every emerging-market crisis since.

What happened

• Early 1990s — several fast-growing Asian economies maintain exchange rates fixed or tightly managed against the US dollar. Growth is rapid and capital flows in heavily.
• The peg creates an incentive. Domestic borrowers take dollar loans at lower dollar interest rates, because the fixed rate appears to remove the currency risk. Banks and corporates accumulate large foreign-currency liabilities against local-currency assets and revenues.
• Early 1997 — current-account deficits widen, export growth slows, and property and equity markets weaken. Pressure builds on the pegs.
• 2 July 1997 — Thailand abandons its peg after exhausting reserves defending it. The currency falls sharply.
• July–December 1997 — pressure spreads across the region. Currencies fall by large multiples of anything that had been contemplated; equity markets fall alongside them.
• 1997–98 — deep recessions, banking failures, IMF programmes with conditions that remain debated to this day, and political change in several countries.
• 1998 — the shock propagates further, contributing to Russia's default and to the collapse of LTCM.

The mechanism: a currency mismatch on a balance sheet

• The peg did not remove currency risk; it concentrated it. A fixed rate makes the risk invisible day to day and enormous at the moment it breaks. Borrowers behaved as though a policy commitment were a hedge.
• Devaluation multiplies a liability that is fixed in foreign terms. A company earning local currency with dollar debt sees its debt double in local terms when the currency halves — with no change in its business at all.
• That converts a currency event into a solvency event, which becomes a banking event, which becomes a credit contraction, which deepens the recession that started it.
• Defending a peg is expensive and finite. It requires selling reserves and raising interest rates — and high rates damage the domestic economy the peg was meant to support. The covered interest parity calculator shows what the rate differential implies for the forward, which is the market's price on the peg surviving.
• The short-term maturity profile was the accelerant. Much of the foreign borrowing was short-dated and had to be rolled. When lenders declined to roll, the outflow was immediate — a liquidity mismatch at national scale.

Contagion: why unrelated countries fell together

• Common lenders. The same international banks and funds were exposed to several countries. Losses in one forced deleveraging in all — correlation created by the holder rather than by the assets.
• Reassessment. One peg breaking caused investors to re-examine every similar arrangement, and several looked similar.
• Trade and competitive links. A large devaluation by one exporter pressures its competitors, which is a real economic channel rather than a sentiment one.
• The general lesson: correlations that appear low in calm periods converge under stress, because the mechanism linking assets is the behaviour of their holders — the second pattern in the failure taxonomy, and the reason to stress-test with the correlation tool.

What it teaches

• Borrowing in a currency you do not earn is a leveraged position on that currency, whether or not anyone calls it one. This is true for countries, companies and individuals with foreign-currency mortgages alike.
• A fixed exchange rate is a policy, and policies end. Treating one as a permanent feature of the world is the same error as treating any assumed bound as contractual.
• Cheap foreign funding is compensation for a risk, exactly as in any other yield. The lower rate was the price of the currency risk being transferred to the borrower.
• Growth does not immunise. These were genuinely fast-growing economies with high savings rates. The vulnerability was in the structure of the financing, not in the quality of the growth.
• Crises propagate through balance sheets, not through borders. The map of who is exposed matters more than the map of countries.

What changed afterwards

• Reserve accumulation. Many affected countries built very large foreign-currency reserves in the following decades — expensive self-insurance, bought deliberately.
• More flexible exchange rates, which distribute the adjustment continuously rather than in one break.
• Deeper local-currency bond markets, directly addressing the mismatch: borrowing in the currency you earn removes the mechanism entirely.
• The pattern still recurs wherever foreign-currency borrowing is cheap and the exchange rate looks stable, which it does periodically and always for a while.