The Asian Financial Crisis: Which Pegs Broke and Which Held

Key Takeaways

• The break was violent and over in a day. The baht closed at 24.52 per dollar on 1 July 1997 and at 30.18 on 2 July, an 18.8 percent loss of dollar value in one session.
• The two worst falls were not simultaneous. The won hit its weakest close on 23 December 1997 and the baht three weeks later on 12 January 1998, so an investor who thought Thailand had found a floor still had Korea ahead.
• Hong Kong held its link and paid for it elsewhere. The Hong Kong dollar moved 0.2 percent, but on 23 October 1997, the day rates were raised to defend it, the Hang Seng fell 10.41 percent.
• The devaluation that should have helped exporters ruined borrowers first, because the debts were in dollars and the revenues were not. That asymmetry is why these countries tightened policy into a collapse instead of easing.
• US equities ignored it for months. The Securities and Exchange Commission records broad US indexes setting new closing highs on 7 October 1997, three months after the baht broke.
• When Asia did reach Wall Street it arrived through Moscow. Russia defaulted in August 1998, Long-Term Capital Management lost 44 percent that month, and the Federal Reserve cut three times in seven weeks, then reversed all three during 1999.
• Recovery was not one number. Korea regained its pre-crisis output in 1999 and Indonesia in 2003, from a shock weeks apart.

What Happened When Thailand Devalued the Baht on 2 July 1997?

Federal Reserve History puts it in one sentence: on 2 July 1997 Thailand devalued its currency relative to the US dollar, following months of speculative pressure that had substantially depleted the country's official foreign exchange reserves. The daily record shows why that is treated as the start. In the H.10 series the baht closed at 25.15 on 30 June and 24.52 on 1 July. On 2 July it closed at 30.18. A currency that had spent years inside a narrow band lost 18.8 percent of its dollar value between one session and the next.

What made that a crisis rather than a policy adjustment is what the peg had been doing while it held. Thai banks and companies had borrowed against it: foreign money came in short and in dollars, and went out long and in baht, into property and domestic credit. While the rate held, that trade paid the gap between a low foreign interest rate and a higher domestic one, and the currency risk was invisible because it had never been realised. On 2 July it was realised at once, and every dollar of external debt on a Thai balance sheet became 23 percent dearer to repay in local earnings, before the currency had finished falling.

Federal Reserve History describes what followed as a twin balance-of-payments and banking crisis, with currency, equity and property markets weakening together. Those were not two problems that happened at once. They were one problem seen from two sides, because the devaluation that fixed the external imbalance destroyed the domestic borrowers who had funded themselves abroad.

Chronology of the episode

Dated events, each established by one of the sources listed at the end of this page. Exchange rates are closing values from the Federal Reserve H.10 daily series.

That sequence is not a wave travelling outward from Bangkok. It jumped, stalled for weeks at a time, skipped the largest market in the region without breaking it, and reached the United States more than a year later by way of a Russian default that had nothing to do with Asian currencies.

How Did a Thai Currency Problem Reach Seoul, Kuala Lumpur and Hong Kong?

Thailand is a small economy and its devaluation alone was not an event of global consequence. What made it one was that the creditors who had funded Thailand had funded its neighbours on the same reasoning, and when they revised it they revised it everywhere at once.

Federal Reserve History is explicit about the channel. The Thai episode showed how banking sector problems could trigger a pullback by foreign investors, setting off a spiral of depreciation, recession and further banking weakness, with creditors then withdrawing from other countries in the region seen as having similar vulnerabilities. Note the wording: seen as having similar vulnerabilities. The mechanism was not trade, and not a mechanical linkage between Thai and Korean banks. It was a change in how a category of borrower was assessed, and once lenders began sorting the region by the features Thailand had displayed, being sorted into the wrong pile was enough.

The same account adds a quieter second channel. Japan's own deteriorating position played a role, because Japanese banks had been an important source of credit to the region and were pulling back. A regional funding crisis needs a regional funder, and this one was in trouble at home for unrelated reasons.

The third element was the defence itself. Federal Reserve History records that foreign exchange intervention often proved counterproductive, with some countries depleting the bulk of their official reserves and then suffering even larger subsequent depreciations. Spending reserves to hold a rate does not merely fail; it can make the eventual move worse, because the buffer that would have cushioned the adjustment is gone by the time it arrives, and because the market can watch it going.

Korea is where all three met. The won had been drifting rather than collapsing: 992 per dollar on 14 November 1997, 1,188 on 1 December. Then 1,342 on 8 December, 1,565 on 10 December, 1,710 on 12 December and 1,960 on 23 December. That is not a valuation adjustment. That is a funding market closing on banks whose short-term foreign obligations had to be refinanced continuously.

What broke the spiral was not a price. Federal Reserve History records that following a meeting on 24 December 1997 hosted by the Federal Reserve Bank of New York, US banks with the largest exposures to South Korean banks voluntarily committed to roll over their short-term loans and restructure them into medium-term debt, with similar outreach in other G-10 countries and the restructuring completed in April 1998. In the H.10 series the won closed at 1,960 on 23 December, 1,835 on 24 December, 1,506 on 26 December and 1,450 on 29 December, a 35 percent recovery in three sessions. Nothing about Korean fundamentals changed that week. A small number of creditors agreed not to leave at the same time.

Which Currencies Broke, and by How Much?

The phrase "Asian currencies collapsed" hides most of the information. Here is the same window measured the same way for five currencies, from the Federal Reserve H.10 daily series, referenced to 30 June 1997, the last business day before the devaluation.

Local currency units per US dollar, closing values. Fall against the dollar is the loss in the local currency's dollar value from its 30 June 1997 close to its weakest close between 1 July 1997 and 31 December 1998.

Four things come out of that table that the phrase "Asian currencies collapsed" would never give you.

The two deepest falls were three weeks apart. The won bottomed on 23 December 1997 and the baht on 12 January 1998. What is remembered as one regional event was, at the time, a sequence in which each apparent floor was followed by a different country's worse one.

Singapore fell too, by a fifth. It is absent from the Federal Reserve's list of countries whose currencies were forced down and from the list of the three that needed the international loan package, yet it lost 20.4 percent of its dollar value anyway. Being adjacent to a regional funding crisis has a price even when none of the specific vulnerabilities apply, which is the practical case for sizing an emerging markets index position by region rather than by country.

The recoveries were only partial, and what was lost stayed lost. At the end of 1999 the baht was at 37.75, still 33 percent below its pre-crisis dollar value, and the won at 1,136, still 22 percent below. A currency that breaks a peg does not generally return to it, because the old rate was not a level the market had chosen.

The ringgit did not recover, it was stopped. The H.10 series shows it at 4.195 on 31 August 1998, 3.895 on 1 September, 3.81 on 2 September, and then 3.8000 on 319 of the next 333 business days to the end of 1999, with the fourteen exceptions never further than three tenths of one percent from that rate. Bank Negara Malaysia records the policy behind that flat line: selective exchange controls on 1 September 1998 and the rate fixed at 3.80 the next day, not replaced by a managed float until 21 July 2005. Malaysia is the one member of the group that answered a capital-flow reversal by restricting capital flows instead of raising the price of money. Whether that was the better trade neither table settles: Malaysia regained its pre-crisis output in 2000, a year after Korea and a year before Thailand.

Which Warning Signs Were Visible Before 2 July 1997, and Which Only After?

Federal Reserve History gives both halves of this in consecutive sentences, and quoting only one of them is how most retrospectives of this crisis go wrong. The events, it says, generally caught market participants and policymakers by surprise. And: while some vulnerabilities were well recognised beforehand, especially in Thailand, these economies were also viewed as having many strengths, and the most affected were among the world's most successful of the preceding decade, with growth rates exceeding 5 percent and often approaching 10 percent. Something was visible. It was not the thing that mattered most.

Pre-crisis evidence, classified by whether it was usable in advance.

The most instructive contemporaneous evidence is what the American market did with all this. The Securities and Exchange Commission's reconstruction of October 1997 notes that US share prices kept rising through much of the year despite mounting losses on the smaller Southeast Asian markets, that Indonesian and Malaysian equities had reached year-to-date losses of almost 23 percent and 35 percent by late August 1997, and that this held even through July and August as the Thai currency crisis accelerated. Broad US indexes then set new closing highs on 7 October 1997. None of it registered.

Then it registered at once. The Commission traces the turn to the week of 20 October, when the instability reached the much larger Hong Kong market and share prices there fell 14.43 percent over three sessions on doubts about the peg. Even that produced only a slow reaction elsewhere. What changed the picture was the defence: when Hong Kong short-term rates were raised sharply on 23 October, the Hang Seng fell 10.41 percent that day, the Nikkei 3.03 percent, the FTSE 3.06 percent and the Dow 2.33 percent. Four days later the Dow fell 554.26 points, 7.18 percent, to 7,161.15, the tenth largest percentage decline in the index since 1915, and the cross-market trading halt procedures adopted in 1988 were used for the first time in their existence.

Hindsight check. Short-dated dollar borrowing behind a fixed rate is so legible as a vulnerability once written down that the 1997 warning signs look obvious in retrospect. Test it against a date. On 30 June 1997 the observable facts were Thai equities in severe decline, Thai reserves under pressure, and a decade of high regional growth. From that set, what position would you have taken in Korean assets? Nothing in the pre-crisis information set singled Korea out, and the market did not single it out either until November. Cognitive biases in trading explains why the retrospective version feels so much more determined than the real one.

Why Did Devaluation Bankrupt Borrowers Instead of Helping Exporters?

A weaker currency is supposed to be an adjustment mechanism: exports get cheaper abroad, imports dearer at home, and the economy grows into its problem. In 1997 that ran into a balance sheet.

Federal Reserve History states the mechanism directly. Heavy foreign borrowing, often at short maturities, exposed corporations and banks to significant exchange rate and funding risks that longstanding currency pegs had masked, and when the pegs proved unsustainable, firms saw sharp increases in the local currency value of their external debts, leading many into distress and even insolvency.

Follow one hypothetical Thai company through 2 July 1997 and the trap is obvious. It earns baht. It borrowed dollars, because dollar rates were lower and the peg made the currency risk look theoretical. Its interest coverage was comfortable on 1 July. On 2 July, with no change in its sales, costs or management, the baht value of its debt rose 23 percent, and by 12 January 1998 it had more than doubled. The devaluation that made its exports cheaper made its liabilities unpayable, and the second effect landed immediately while the first took quarters to reach the order book.

That is the feature separating this crisis from every other episode in this library, and it produces a policy trap with no clean exit.

• Cutting rates to support borrowers weakens the currency further, which increases the local value of the very debts you are trying to help them service.
• Raising rates to defend the currency protects the external debt burden but destroys domestic demand, collateral values and the banks holding the domestic loans.
• Spending reserves to hold the rate can leave you with neither the reserves nor the rate, which Federal Reserve History records as the actual outcome for several countries.

Federal Reserve History describes the mix that was chosen: countries hiked interest rates to stabilise currencies and tightened fiscal policy to speed external adjustment and cover the cost of bank clean-ups, loosening both only later as markets settled. Read against the tables above, that is a set of economies tightening into contractions of 5 to 13 percent of output. It is the exact inverse of the American response in 2008 and 2020, and not because the policymakers were less wise. The debts were denominated in a currency none of them could issue.

The lesson is narrower than "currency risk matters" and more useful for it. Currency exposure is not principally a translation effect on the value of an overseas holding. It is a solvency variable for anyone whose assets and liabilities sit in different units, and it is invisible in every ratio you can compute while the exchange rate is fixed. Currency hedging sensitivity covers measuring that exposure inside a company, and international ETFs, currency risk and hedging the fund-level version.

What Did Hong Kong Do in 1998 That Nobody Else Did?

Every other currency in the table moved. Hong Kong's moved 0.2 percent across eighteen months, while Federal Reserve History records several large but unsuccessful attacks on the peg, the first triggering short-term stock market sell-offs around the world. Why that link survived is more instructive than cataloguing the ones that did not, because the difference was not the absence of an attack.

The first defence, in October 1997, worked and the bill arrived the same day: rates up sharply on 23 October, the Hang Seng down 10.41 percent that session and a further 13.70 percent on 28 October to close at 9,059.89. The peg held. The equity market absorbed the entire adjustment.

The second defence, in August 1998, has no parallel anywhere else in this library. The Hong Kong Monetary Authority's own retrospective describes a "double play" in which the currency, the stock market and the futures market were attacked together, so that pressure on the peg would force interest rates up, which would push shares down, which would pay off short positions in the index. Defending only the currency would have meant paying the attackers on the other two legs.

So the government bought the index. The HKMA records that over ten trading days from 14 August 1998 it mobilised HK$118 billion, about 18 percent of the Exchange Fund's total assets, to buy 33 constituent stocks of the Hang Seng Index, plus further money in index futures unwound by the end of September. The operation ended on 28 August with the index at 7,830, about 18 percent above where the intervention began, on turnover that day of over HK$79 billion at a point when, by the HKMA's own description, it was almost the only buyer in the market.

The Hong Kong result supports two conclusions that pull against each other. The link survived because the cost was paid elsewhere, deliberately. A currency board can always defend the rate; the question is what it will sacrifice, and Hong Kong sacrificed domestic interest rates in 1997 and a fifth of its reserve assets in 1998. The countries whose pegs broke were not less committed in principle. They ran out of what they were spending before the pressure stopped. And it is not a repeatable template. Hong Kong entered the episode with an Exchange Fund large enough that 18 percent of it could turn the Hang Seng, and with a linked exchange rate run as a currency board rather than as a managed rate a central bank had to talk up. Thailand, Indonesia and Korea had neither. Whether a peg holds is decided by the size of the buffer relative to the pressure, and that is knowable in advance far more often than the timing of an attack.

How Did an Asian Currency Crisis End Up on the Federal Reserve Agenda?

For an American investor the most surprising fact about this crisis is how long it took to matter, and then how it arrived.

The Federal Reserve's published record shows no target change between 25 March 1997, when the funds target was raised to 5.50 percent, and 29 September 1998. Through the Thai devaluation, the October 1997 circuit breaker halt, the Korean funding collapse and the April 1998 loan restructuring, it did not move once. Federal Reserve History notes that the direct trade impact on the United States proved manageable and was partly offset by lower inflation pressure from cheaper Asian imports and weaker commodity prices, and lower bond yields from a flight into dollar assets. For most of a year the crisis was, from a US policy perspective, mildly helpful.

What changed had a Russian address. Federal Reserve History records that several emerging economies in Latin America and Eastern Europe, Brazil and Russia among them, faced significant balance-of-payments pressures in 1998, reflecting spillovers from Asia together with vulnerabilities of their own. In August 1998, by the Federal Reserve's account of what followed, Russia suddenly devalued its currency and stopped payments on its debt, which sent investors looking for safer and more liquid assets.

The Federal Reserve's account of Long-Term Capital Management picks up the thread: in 1998 the crisis that had started in Southeast Asia the previous year intensified, and when Russia stopped paying, investors moved into safer and more liquid assets. LTCM had positioned for spreads to converge and in almost every case they diverged. The fund lost 44 percent of its value in August alone, carrying roughly 30 dollars of debt for every dollar of capital after returning capital to investors at the end of 1997 without reducing its positions. On 23 September 1998 fourteen banks and brokerage firms put 3.625 billion dollars into it at the New York Fed in exchange for 90 percent of the fund, to prevent a fire sale rather than to rescue the partners, whose claim was cut to 10 percent.

Only then did the rate move: three consecutive 25 basis point cuts, to 5.25 percent on 29 September 1998, 5.00 percent on 15 October and 4.75 percent on 17 November. The whole easing was 75 basis points, and three increases during 1999 returned the target to 5.50 percent, exactly where it had been before any of this began.

Federal funds target changes from the last pre-crisis move to the completion of the reversal, from the Federal Reserve Board's open market operations archive.

The HKMA's account describes creditor banks withdrawing credit lines from leveraged funds that autumn, forcing a massive unwinding of positions regardless of loss and triggering a wave of buying back short yen. The H.10 series shows the yen going from 135.75 per dollar on 1 October 1998 to 117.00 on 9 October, a 16 percent appreciation in six sessions, in a currency whose own economy was not improving. That is a positioning unwind, not a change of view about Japan.

Thailand to Moscow to a Greenwich hedge fund to the Japanese yen has no direct economic link at any step. It runs through balance sheets: the same leveraged investors held all of it, and their creditors did not distinguish between the positions when they pulled the lines. How correlations change across regimes covers why stress-period correlations differ so sharply from calm-period ones, and the dollar, rates and cross-asset transmission traces the modern version of the same plumbing.

How Long Did Each Economy Take to Get Its Output Back?

Most case studies on this site answer "how long did it take to recover" with an index and a date. This one cannot, and the reason is worth more than the answer would be.

Annual real GDP growth, and the year in which real GDP in constant local currency first exceeded its pre-crisis peak. From World Bank national accounts data.

Four things in that table are hard to reconcile with the single-crisis account.

The currency ranking and the damage ranking are different. The baht and the won fell almost identically, 55.2 percent and 54.6 percent. Korea contracted 4.9 percent and was back above its output peak within two years. Thailand contracted 7.6 percent in 1998 on top of 2.8 percent in 1997 and did not recover until 2001. Anyone using the exchange rate as a proxy for economic damage was reading the wrong variable.

Thailand was already contracting before the peg broke. Its real output peaked in 1996 and it shrank 2.8 percent during 1997. The devaluation is dated as the start of the crisis because it is the visible event, but Thai output had turned down first. The peg was the last thing to give way, not the first.

Indonesia is a different event that shares a name. A 13.1 percent contraction is roughly three times the Korean one, and output took six years to return. Federal Reserve History describes Indonesia as gradually falling into a multifaceted financial and political crisis, which the exchange rate alone does not capture. Grouping it with Korea under one label loses the most important fact about it.

Korea rebounded harder than it fell, growing 11.6 percent in 1999 after the 4.9 percent contraction. A V-shaped recovery is possible after a funding crisis, and this is the clearest instance in the library, which is also why Korea is a poor guide to what the other four experienced.

None of those figures are stock market recovery dates, and the omission is deliberate: no source verified this session supplied continuous index history for the Thai, Korean, Malaysian or Indonesian markets. Several ran far longer than the output recoveries above, but "far longer" is a direction and not a number, and it is left as one. The drawdown and recovery calculator covers the arithmetic of what a given fall needs to get back; the date it happened is a claim that needs a source.

Common Myths About the Asian Financial Crisis

"It was a crisis of government profligacy." Federal Reserve History describes the affected economies as run with business-friendly policies and cautious fiscal and monetary management, producing high savings and investment and growth often approaching 10 percent. The borrowing that broke them was private and external, not public and domestic. A strong government balance sheet is not evidence that a country is safe from a funding reversal.

"The whole region collapsed." The Hong Kong dollar moved 0.2 percent. Singapore lost a fifth of its currency value and never needed a programme. The Philippines contracted 0.5 percent and was back above its output peak in 1999. Treating "Asia" as the unit of analysis produces a story in which everything failed, and loses the interesting question, which is what separated the outcomes.

"It started in July 1997." The devaluation started in July 1997. The Bangkok SET index was already down more than 44 percent for the year by 19 June, and Thai real output had peaked in 1996. What happened on 2 July was the last defence being abandoned, which is the most visible moment in a sequence and rarely the first one.

"Devaluation fixed it." Devaluation turned a currency problem into a corporate insolvency problem, because the debts were in dollars. The competitiveness benefit was real and arrived quarters later, by which time many of the borrowers who would have exploited it were in distress. A cheaper currency can help an economy and destroy the firms inside it, on different time horizons.

"Everyone saw it coming." Federal Reserve History states that the events generally caught market participants and policymakers by surprise, and Korea went from 992 won per dollar in mid-November to 1,960 in late December. The retrospective sense of inevitability is manufactured entirely by knowing which countries turned out to be on the list.

"It has no bearing on developed markets." Its best-known casualty inside a developed market was a Connecticut hedge fund, and the intervention that ended the acute phase happened at the Federal Reserve Bank of New York, by way of a Russian default that was itself partly an Asian spillover. Crises transmit through the balance sheets holding the positions, not through the map.

What a Reader Can Actually Carry Forward

This is the episode least likely to repeat in its original form, because a pegged exchange rate over an open capital account with heavy short-dated dollar borrowing is a configuration far fewer economies now run. That makes it a better source of mechanisms and a worse template than any other case here.

What generalises

• A fixed price is a promise that stops being honoured at the worst moment. A peg is valuable to a borrower because it lets them ignore currency risk, and it does so right up until the risk is realised in full, at once, at the point of maximum distress.
• A mismatch between the unit of your assets and the unit of your liabilities is a solvency variable. It appears in no leverage or coverage ratio while the relationship holds, so a company can look conservatively financed and be one policy decision from insolvency. It is the specific thing to look for in country and sovereign political risk analysis, and it is measurable before the event.
• Defending a price with a finite buffer can make the eventual move worse. Federal Reserve History records countries depleting the bulk of their reserves and then suffering larger depreciations than if they had not.
• Runs are coordination problems and can be stopped by coordination. The won recovered 35 percent in three sessions after a meeting at which creditors committed to roll over rather than withdraw.
• Contagion travels through who owns the exposure, not through geography. Thailand to Korea to Russia to a US hedge fund to the yen is not an economic chain. It is a list of positions held by the same leveraged investors and financed by the same creditors. Stress testing and scenario analysis is how you check whether your own holdings are connected that way.

What does not generalise

• The policy response. These countries raised rates into contractions of 5 to 13 percent of output because they could not issue the currency their debts were in. An economy borrowing in its own currency faces different choices entirely, which is what makes the 2008 crisis and the 2020 crash misleading comparisons for this one.
• The speed of the recoveries. Korea grew 11.6 percent the year after its contraction and Indonesia needed six years, from the same shock in the same months.
• Hong Kong buying its own index. That worked because of a currency board, an unusually large reserve fund and negligible foreign currency sovereign debt. It describes one jurisdiction's options, not a general tool.
• "Emerging markets are the fragile ones." Fighting this crisis with a country label repeats the 1997 error, which was sorting the region into safe and unsafe piles by resemblance. The variable was the funding structure, and funding structures move.

The one question worth asking now

Rather than "am I exposed to an emerging market crisis," ask it as the Thai company would have had to in June 1997: in what currency do my assets earn, in what currency are my obligations denominated, and what happens to the second if the first moves 50 percent? For a household holding an unhedged international fund the answer is usually mild, because there is no liability on the other side. For anyone with a fixed obligation in one unit and a variable income in another it is the whole question, and unlike a forecast it does not require knowing when anything will happen. International investing covers the portfolio-level version and risk management the sizing.

References

Every figure on this page was verified against the following sources, each retrieved on 26 August 2026:

• Federal Reserve History: Asian Financial Crisis: the 2 July 1997 devaluation and the reserve depletion before it, the July 1997 to December 1998 dating, the pre-crisis growth rates, the account of credit growth and pegs masking exchange rate risk, the surprise finding, the contagion mechanism and Japanese bank retrenchment, the counterproductive-intervention finding, the 118 billion dollar loan total, the 24 December 1997 New York Fed meeting and April 1998 restructuring, the unsuccessful attacks on the Hong Kong peg, the policy mix, and the 1998 pressures on Brazil and Russia.
• Federal Reserve Board: Foreign Exchange Rates, H.10, Country Data: every exchange rate quoted here, including the baht, won, ringgit, Singapore dollar, Hong Kong dollar and yen closes and the percentages computed from them. The historical country files are the right destination rather than the current weekly Federal Reserve Board: Foreign Exchange Rates, H.10 release, which no longer carries Thailand, South Korea, Malaysia, Singapore or Hong Kong at all.
• Federal Reserve Board: Open Market Operations Archive: the 25 March 1997 increase to 5.50 percent, the three 1998 cuts and the three 1999 increases in the table above.
• US Securities and Exchange Commission: Trading Analysis of October 27 and 28, 1997: the Bangkok SET decline by 19 June 1997, the Indonesian and Malaysian year-to-date declines, the 7 October US closing highs, the October Hong Kong falls with the accompanying Nikkei, FTSE and Dow moves, and the 27 and 28 October figures with the first use of the circuit breakers adopted in 1988.
• Federal Reserve History: Near Failure of Long-Term Capital Management: the August 1998 Russian devaluation and halt in debt payments, the 44 percent August loss, the roughly 30 to 1 leverage at end-1997, and the 23 September 1998 investment of 3.625 billion dollars by fourteen firms.
• Hong Kong Monetary Authority: Difficult Decisions in the Disposal of Shares After Stock Market Operation: the double play, the ten trading days from 14 August 1998, the HK$118 billion and its 18 percent share of Exchange Fund assets, the 33 constituents bought, the 28 August close at 7,830 on HK$79 billion of turnover, and the short yen unwind.
• Bank Negara Malaysia: Significant Milestones in the Malaysian Foreign Exchange Market: the selective exchange controls of 1 September 1998, the ringgit fixed at 3.80 the next day, and the managed float from 21 July 2005.
• World Bank: GDP Growth, Annual Percent: the 1996 to 1999 growth rates for Indonesia, Thailand, Malaysia, Korea and the Philippines.
• World Bank: GDP, Constant Local Currency Unit: the pre-crisis output peak year for each of the five economies, and the year each one first exceeded that peak. This is the series behind the 1999 to 2003 recovery dates, and it is a different indicator from the growth series above.

Figures deliberately not stated. No depreciation figure for the Indonesian rupiah, which the H.10 series does not carry. No per-country split of the 118 billion dollar total, because the institutional pages holding that breakdown could not be retrieved. No reserve levels, short-term debt to reserve ratios, non-performing loan ratios, institution failure counts or unemployment rates, and no Asian equity index drawdowns or recovery dates beyond the Hang Seng values the Commission report establishes. Where a direction is documented and a magnitude is not, the direction is described and the number left out rather than estimated.

Method note: exchange rate levels, percentage falls and the dates of weakest closes were computed by Swoopr Investment from the complete daily series of the Federal Reserve H.10 release, retrieved on 26 August 2026. A fall against the dollar is the loss in the local currency's dollar value, from the reciprocal of the quoted rate, close to close from 30 June 1997 to the weakest close between 1 July 1997 and 31 December 1998. Output recovery years were computed from World Bank GDP in constant local currency as the first year exceeding the highest pre-1998 figure. Growth rates are the World Bank published series and are subject to revision.

This is educational content about a historical episode. It is not investment advice, it is not a forecast, and nothing here should be read as a claim about how any future currency or market decline will behave.

Frequently Asked Questions