On 2 July 1997, the Bank of Thailand stopped defending the baht. The IMF's news brief that day records the change: the exchange rate became a managed float, and the Bank Rate went from 10.5% to 12.5% to support it. By January 1998, the worst month, the baht had lost about half its value against the dollar, the Indonesian rupiah about four fifths, and the Korean won about half. The Asian financial crisis 1997 started with one currency and reached at least five.
The Asian financial crisis 1997 is the run on East Asian currencies, banks and stock markets that began when Thailand floated the baht on 2 July 1997 and spread to Indonesia, Malaysia, the Philippines and South Korea. Federal Reserve History counts $118 billion of international loans to Thailand, Indonesia and Korea. This is the ninth episode of our three centuries of markets series, and it follows Black Wednesday 1992: another peg, another central bank, and another market that sold faster than the bank could buy.
What caused the Asian financial crisis 1997?
The economies at the centre of the crisis did not look broken. Federal Reserve History notes that growth across the region had run above 5% a year, often close to 10%. The IMF's June 2000 review adds that government budgets were broadly balanced and inflation modest. This was not a fiscal crisis.
The weak points were in the private balance sheet. The IMF lists them plainly: weak bank loan portfolios, heavy reliance on short-term external debt, unhedged borrowing in foreign currency, risky investment in stock and property bubbles, implicit government guarantees, and thin supervision. Strong capital inflows funded all of it. And the dollar was rising, which dragged the pegged currencies up with it.
The arithmetic is simple. A Thai company borrows dollars because dollar rates are lower than baht rates. It converts them, lends or invests in baht, and earns the difference. As long as the peg holds, the trade looks like free income. If the peg breaks, the debt grows in baht terms at the same speed as the currency falls. If the baht loses half its value, the same dollar loan takes twice as many baht to repay, with no change in what the company owes in dollars.
Bank lending made the exposure concrete. A Reserve Bank of Australia research paper by Morris Goldstein and John Hawkins, published in May 1998, puts international bank lending to the crisis countries at $326 billion at the end of June 1997, with $97 billion of it from Japanese banks.
Why did Thailand float the baht on 2 July 1997?
Because it had run out of road to defend it. Pressure on the baht intensified in May 1997. The Bank of Thailand defended the peg in the forward market, selling dollars for future delivery rather than from its visible reserves. The RBA paper records that this forward intervention peaked at $24 billion. A reserve number that looks healthy on the balance sheet can be largely committed already.
When the forward book could not be extended further, the Bank floated the currency. The rate rise to 12.5% the same day was meant to make holding baht worth the risk. It did not stop the slide.
May 1997
Pressure on the baht intensifies
The Bank of Thailand defends the peg in the forward market. Forward intervention peaks at $24 billion.
2 July 1997
Thailand floats the baht
The exchange rate becomes a managed float. The Bank Rate rises from 10.5% to 12.5%.
20 August 1997
IMF support for Thailand
The IMF Board approves its arrangement. The total international package is $17.2 billion.
Late October 1997
Attack on the Hong Kong dollar
The peg holds. Short-term stock sell-offs follow across the globe.
5 November 1997
Indonesia stand-by
A three-year IMF arrangement. Sixteen insolvent banks are closed the same month.
4 December 1997
Korea stand-by, SDR 15.5 billion
About $21 billion from the IMF, 1,939% of Korea's quota. The won had lost about half its value by its low.
24 December 1997
New York creditor meeting
Banks meet at the Federal Reserve Bank of New York to roll over Korean short-term loans.
January 1998
The worst month
Depreciations across the region reach their deepest point. Indonesia guarantees all bank liabilities.
May 1998
Suharto falls
Indonesia's financial crisis becomes a political one.
2 September 1998
Ringgit fixed at RM3.80
Malaysia quotes its currency at a fixed rate after imposing capital controls.
How did the crisis spread from Thailand to Korea?
Contagion did not need a trade link. The IMF's review describes it as investors treating other countries as having the same weaknesses as Thailand: pegged currencies, short-term foreign debt, banks lending into property. Once one peg broke, every peg with that profile was tested.
The Philippines tightened its rules on offshore peso forwards in July. Indonesia limited foreign access to its swap market in August. By early October 1997, the IMF records, the rupiah had already lost about 30%. Taiwan let its currency go on 20 October. A few days later the Hong Kong dollar was attacked. Its peg held, but Federal Reserve History notes the attack set off short-term stock sell-offs around the world.
Korea came last. Its won lost about half its value, mainly in November and December 1997. On 4 December the IMF approved a three-year stand-by of SDR 15.5 billion, about $21 billion, equal to 1,939% of Korea's quota. The press release lists support of up to $10 billion from the World Bank, up to $4 billion from the Asian Development Bank, and more than $20 billion expected from 12 countries. On 24 December, Federal Reserve History records, the Federal Reserve Bank of New York hosted a meeting of banks to roll over Korea's short-term loans. The restructuring into medium-term loans was completed in April 1998.
How far Asian currencies fell against the dollar
Approximate fall from June 1997 to each currency's low
Source: Goldstein and Hawkins, Reserve Bank of Australia RDP 9805 (May 1998)
How much did the rescue packages cost?
Three countries went to the IMF in five months. The RBA paper gives the headline totals for each package: $17 billion for Thailand on 20 August 1997 (the IMF puts the total at $17.2 billion), $35 billion for Indonesia on 5 November, and $58 billion for Korea on 4 December. For comparison, the Mexican package of February 1995 was $52 billion.
Those totals bundle IMF money with World Bank, Asian Development Bank and bilateral lines. The IMF's own share was smaller. Its June 2000 review lists IMF commitments of $40.1 billion across the three countries, of which $34.8 billion had been disbursed by the end of May 2000.
$17.2bn
Thailand package
Approved 20 August 1997
$35bn
Indonesia package
Stand-by of 5 November 1997
$58bn
Korea package
Approved 4 December 1997
$40.1bn
IMF share committed
Three countries, as of 30 May 2000
$118bn
All international loans
IMF, World Bank, ADB and governments
The real cost was in output. The IMF records that GDP fell by about 14% in Indonesia and about 7% in Korea in 1998. In Indonesia the banking crisis ran into politics: 16 banks were closed in November 1997, a blanket guarantee of bank liabilities followed in January 1998, and President Suharto fell in May.
Why did Malaysia choose capital controls?
Malaysia did not take an IMF programme. It let the ringgit weaken under pressure, then, in September 1998, imposed capital controls. On 2 September 1998 Bank Negara Malaysia announced that the ringgit would be quoted at RM3.80 against the US dollar. The IMF's review notes that Malaysia replaced its holding-period rule with graduated exit levies in February 1999 and relaxed the controls further in September 1999.
The choice was contested at the time and remains debated. What matters for a trader is the mechanism. A government that cannot defend a price with reserves can still fix it by closing the exit. For anyone holding the position, the result is the same: the market price stops being the price you can get.
How long did the Asian financial crisis last?
About a year for Korea and Thailand, longer for Indonesia. The IMF records that financial markets in Korea and Thailand had stabilised by mid-1998 and that activity began to turn around from there. In September 1999 Thailand said it would not draw the remaining balance of its IMF arrangement. By the end of April 2000 Korea's reserves had reached $85 billion, and by the end of May it had repaid $13 billion of IMF financing.
Indonesia took longer. Its IMF programme was suspended in September 1999, and a new extended arrangement of about $5 billion was approved in February 2000. The crisis did not stay in Asia either. Federal Reserve History notes the spillovers to Russia and Brazil in 1998.
What does the Asian financial crisis 1997 mean for traders today?
The parallels are structural, not predictive. In 1971 the Nixon shock ended a peg from the centre. In 1992 the pound left a peg from the edge in one session. In 1997 the peg broke in Bangkok and the trade that depended on it broke in five capitals. The pattern is the same each time: a price held by a promise, a carry that pays while the promise holds, and a reversal that arrives in one move rather than in steps.
Many pegs and managed currencies still exist. The lesson is not about which ones will hold. It is about what a position looks like when its risk depends on someone else's decision.
What your journal would have shown
Take an illustrative demo account through 1997. Illustrative demo data, not a record of any real trader.
The demo trader runs a carry position: short dollars, long baht deposits, sized to risk 1R on a stop 2% beyond the peg. From January to June the position collects the interest differential. Twenty-six weekly entries in the journal, each between 0.08R and 0.15R, all green. The win rate reads 100%. The equity curve rises in a straight line. Then comes 2 July. The baht gaps through the stop on the float, and in the demo the exit fills at -7R. Half a year of gains, about 3R in total, and a single day that costs more than twice that.
A read-only journal would not have stopped the trade. It is not built to. What it would have shown is the shape before the gap. Twenty-six wins and no loss is not an edge until you see the size of the loss the strategy is exposed to. In Socius Trades, trading expectancy is computed from your full history, in R, so one -7R trade weighs what it really weighs against 26 small wins. The day itself shows up as one bar on the drawdown chart. You can see how much of your record depends on a price someone else promised to hold. Plans start with a Free tier and the paid plans are listed on the Socius Trades pricing page.
We look. We never touch.
Trading involves risk of loss. Socius Trades is an analytics tool, not investment advice.
