Only two trading days in the late 1990s saw the S&P 500 fall more than 4%. But those two days bookend the most dramatic global financial crises since the 1930s — a cascade that began with a currency devaluation in Bangkok, swept through Moscow, and nearly destroyed the Western financial system in a Greenwich, Connecticut office park.
The 1990s bull market is the greatest in American history. Between 1991 and 1999, the S&P 500 delivered positive returns every single year, compounding at an annualized rate that had no precedent. The internet was being born. The federal budget was moving toward surplus. The Cold War was over. And yet, in the middle of this extraordinary run, the global financial system came within days of collapse — not once, but twice.
The two -4% days in this episode are deceptive in their isolation. They make the 1990s panics look like brief interruptions — a bad Monday here, a rough August there. In reality, each crash day was the climax of weeks of deterioration, and the crises they represented threatened to unravel the entire framework of global finance. That only two days breached the -4% threshold is a testament to how quickly the Federal Reserve acted to contain the damage.
On July 2, 1997, Thailand devalued the baht. It seemed like a minor event — a small Southeast Asian economy adjusting its currency peg. But Thailand’s problems were not unique. Across the region, countries had borrowed heavily in U.S. dollars to finance domestic investments. When Thailand’s peg broke, investors suddenly questioned whether Indonesia, Malaysia, South Korea, and the Philippines could defend their own currencies. The answer, one by one, was no.
By October, the crisis had reached Hong Kong — a financial hub whose dollar peg was considered sacrosanct. On October 23, the Hong Kong stock market plunged 10.4%. The selling cascaded across time zones. European markets fell. And on Monday morning, October 27, it arrived in New York.
The S&P 500 fell 6.87% — from 941.64 to 876.99. For the first time in history, the NYSE’s circuit breakers — installed after Black Monday in 1987 — were triggered. Trading was halted twice: first at 2:36 PM when the Dow had fallen 350 points, and again at 3:30 PM when it had fallen 550 points. The second halt ended the trading day thirty minutes early.
Of 1,534 stocks in our universe, 1,364 declined — 88.9% of the market. The median stock fell 4.69%, almost exactly matching the S&P’s decline. The selling was broad and indiscriminate. Tech stocks were hit hardest: Autodesk fell 21.10%, AMD fell 19.79%, and a fourteen-month-old internet company called Amazon.com fell 15.04% — its stock price, split-adjusted, dropping to 21 cents.
The recovery was remarkably swift. The S&P bounced 5.1% the next day and reclaimed the 900 level within a week. By late November, the index was back above 940. The Asian crisis had produced a violent convulsion in American markets, but the fundamental engine of the U.S. economy — consumer spending, technology investment, low inflation — was undamaged. The 1997 crash was, in Wall Street parlance, a “buying opportunity.” The S&P 500 would gain 27% the following year.
If October 1997 was a fire that burned and was quickly extinguished, August 1998 was a fire in the walls — invisible, spreading, and far more dangerous than anyone realized until it was almost too late.
On August 17, 1998, Russia defaulted on its government bonds and devalued the ruble. In itself, this was not surprising — Russia’s economy had been deteriorating for months. But the default set off a chain reaction that exposed a terrifying vulnerability at the heart of global finance: Long-Term Capital Management.
LTCM was a hedge fund founded by former Salomon Brothers traders and two Nobel Prize–winning economists. It had $4.8 billion in capital and $125 billion in assets — a leverage ratio of 25 to 1. Its trading strategies relied on convergence trades: small discrepancies between related securities that were “guaranteed” to close over time. When Russia defaulted, global markets did the opposite of what LTCM’s models predicted. Spreads widened instead of narrowing. Positions that were supposed to be uncorrelated moved in lockstep. LTCM began hemorrhaging hundreds of millions of dollars per day.
August 31, 1998 — the S&P 500 fell 6.80%, from 1,027.14 to 957.28. The index had already been weakening all month — down 3.62% on August 4, down 3.84% on August 27. But August 31 was the day that fear of systemic collapse reached the stock market. Of 1,629 stocks tracked, 1,317 declined (80.8%), with a median return of -3.91%.
The top losers read like a who’s who of high-growth names: Amazon fell another 20.91% (to $0.70 split-adjusted), ASML fell 20.90%, Intuit fell 20.79%. The crisis was punishing exactly the stocks that had led the bull market.
The crisis climaxed on September 23, when the Federal Reserve Bank of New York orchestrated a $3.6 billion bailout of LTCM by a consortium of fourteen Wall Street banks. It was not a government bailout in the traditional sense — no taxpayer money was used — but it required the central bank’s direct intervention to prevent counterparty failures that could have cascaded through the entire financial system. LTCM’s $125 billion in positions were intertwined with every major bank on Wall Street.
The Fed also cut interest rates three times in quick succession — September 29, October 15, and November 17 — explicitly to prevent the crisis from spreading to the real economy. The combination of the LTCM rescue and the rate cuts worked. The S&P 500 bottomed at 957.28 on August 31, and by year-end it stood at 1,229.23 — a 28.4% gain from the trough. The 1998 crash became, like the 1997 crash before it, another buying opportunity in the greatest bull market ever.
| October 27, 1997 — Asian Contagion | |||
|---|---|---|---|
| Symbol | Company | Return | Note |
| ADSK | Autodesk | −21.10% | CAD software |
| AMD | Advanced Micro Devices | −19.79% | Semiconductor |
| UNH | UnitedHealth Group | −16.50% | Managed care |
| AMZN | Amazon.com | −15.04% | IPO’d May 1997 |
| ASML | ASML Holding | −14.21% | Chip equipment |
| TSM | Taiwan Semiconductor | −13.64% | Asia exposure |
| August 31, 1998 — Russia / LTCM | |||
|---|---|---|---|
| Symbol | Company | Return | Note |
| LNG | Cheniere Energy | −27.27% | Energy, Russia exposure |
| URI | United Rentals | −24.97% | Equipment rental |
| LEN | Lennar | −24.39% | Homebuilding |
| AMZN | Amazon.com | −20.91% | High-growth, high-beta |
| ASML | ASML Holding | −20.90% | Chip equipment |
| INTU | Intuit | −20.79% | Software |
Amazon appears on both crash-day loser lists. The stock that IPO’d at $1.50 (split-adjusted) in May 1997 fell to $0.21 in October 1997 and to $0.70 in August 1998. An investor who bought Amazon at its lowest price on either crash day would be sitting on one of the greatest investments in history. ASML, too, appears on both lists — it was trading below $12 in 1997 and was briefly in single digits in 1998. Today it trades above $700.
| Date | Return | Close | Prior Close | Decliners | Median | Peak→Trough | Catalyst |
|---|---|---|---|---|---|---|---|
| Oct 27, 1997 | −6.87% | 876.99 | 941.64 | 88.9% | −4.69% | −10.8% | Asian crisis / circuit breakers |
| Aug 31, 1998 | −6.80% | 957.28 | 1,027.14 | 80.8% | −3.91% | −19.3% | Russia default / LTCM |
The near-misses tell a broader story. August 1998 alone produced four days with declines exceeding 3%: August 4 (-3.62%), August 27 (-3.84%), and then September 30 (-3.05%) and October 1 (-3.01%) as the LTCM crisis peaked. The S&P 500’s 19.3% peak-to-trough decline in 1998 was far deeper than the 10.8% correction of 1997, reflecting the severity of the systemic threat.
The 1990s panics introduced a concept that would define every subsequent crisis: contagion. A currency devaluation in Thailand could crash the New York Stock Exchange. A bond default in Russia could destroy a hedge fund in Connecticut that could bring down every major bank on Wall Street. The world had become interconnected in ways that no model — not even Nobel Prize–winning models — had accounted for.
And yet, both crashes proved to be extraordinary buying opportunities. The S&P 500 gained 23% in 1997 and 27% in 1998, despite containing two of the ten worst trading days since Black Monday. The Fed’s willingness to intervene — cutting rates, orchestrating bailouts — established a pattern that would be repeated, on a much larger scale, in 2008 and 2020.
But the most important lesson of the 1990s panics was one the market would forget almost immediately: leverage kills. LTCM’s 25-to-1 leverage ratio was supposed to be safe because its positions were “uncorrelated.” In a crisis, everything correlates. That lesson, learned at a cost of $4.6 billion in 1998, would have to be relearned at a cost of trillions in 2008.