September 1998EconomicHistoric

Long-Term Capital Management (LTCM) Collapse 1998

Year

1998

Region

United States (global spillover)

Markets

Macro، Banking، Prop Risk، Commodities، Stocks

Crisis

Liquidity Crisis، Leverage Unwind، Contagion، Volatility Spike

Severity

Historic

Reading

4 min

Sources

Core lesson

Leverage is a magnifier, not an edge. The same borrowing that made LTCM's tiny spreads profitable also guaranteed that a normal-looking move would be fatal.

◵ Practice the pattern · Practice framework

⏱ 30-second summary

A hedge fund run by Nobel laureates and star traders blew up in weeks when its highly leveraged "sure thing" bets all moved against it at once — forcing a Fed-organized rescue.

Timeline

1994
LTCM launches with star power
John Meriwether, ex-Salomon Brothers, founds the fund with Nobel laureates Scholes and Merton and a roster of elite traders. Early returns are spectacular, and investors and banks line up to lend.
Early 1998
Leverage reaches extreme levels
On roughly $4-5 billion of capital, LTCM holds over $100 billion in positions plus around $1 trillion in derivatives — a balance-sheet leverage near 25-to-1 and far higher counting derivatives.
August 17, 1998
Russia defaults
Russia defaults on its debt and devalues the ruble. Investors worldwide flee to safety at once, and credit spreads widen everywhere instead of converging — exactly the opposite of LTCM's bets.
Late August 1998
Losses snowball
The fund loses a large share of its capital in days as nearly every position moves against it together. Shrinking equity sends the leverage ratio spiraling toward and past 100-to-1.
September 23, 1998
Fed-organized rescue
The Federal Reserve Bank of New York convenes fourteen banks and brokerages to inject about $3.6 billion and wind LTCM down in an orderly way. The Fed lends none of its own funds.
2000
Orderly wind-down completed
With markets calmer, the rescued positions are slowly unwound and the fund is dissolved. The episode becomes the canonical case study in leverage, liquidity, and systemic risk.

What happened?

Long-Term Capital Management was founded in 1994 by a former Salomon Brothers bond chief, with Nobel-winning economists Myron Scholes and Robert Merton on the team. Its strategy was "convergence" trading: spotting two very similar bonds priced slightly differently and betting the gap would close. The edge per trade was tiny, so to make big returns the fund borrowed enormously — by early 1998 it controlled positions worth over $100 billion on roughly $4-5 billion of capital, plus around $1 trillion in off-balance-sheet derivatives.

The models said these spreads were stable and the positions were diversified. But in August 1998 Russia defaulted on its debt and devalued the ruble. Frightened investors everywhere rushed into the safest, most liquid assets at the same time, so spreads that were "supposed" to converge instead blew wider across many markets at once. LTCM's supposedly independent bets turned out to be the same bet wearing different costumes — and they all lost together.

Losses fed on themselves. As capital shrank, the fund's leverage ratio exploded past 100-to-1, and it could not sell its huge, illiquid positions without crashing the very prices it was marked against. Banks feared that an LTCM fire sale would drag down counterparties across Wall Street and freeze global markets. On September 23, 1998, the Federal Reserve Bank of New York organized fourteen banks and brokerages to inject about $3.6 billion to wind the fund down in an orderly way. The Fed lent none of its own money — but the episode became the textbook case of how one over-leveraged fund can threaten the whole system.

Why markets reacted

The fear was contagion, not just one fund's losses. LTCM's positions were so large and so widely mirrored by the banks that traded with it that a forced fire sale could have crashed bond and derivatives prices everywhere at once, inflicting losses on dozens of counterparties.

Because so many banks held similar relative-value bets, LTCM unwinding was effectively the whole market trying to exit the same crowded trades through one narrow door. Liquidity vanished, prices gapped, and a single fund's failure threatened to become a system-wide freeze — which is why regulators stepped in to coordinate an orderly rescue.

What traders usually get wrong

Sizing on the model instead of the disaster. They sized positions to the spreads the model expected, not to the loss a once-in-a-decade shock could deliver.

Treating low historical volatility as low risk. Calm recent history made the bets look safe right up until they were not.

Assuming exit liquidity would always be there. Positions too big to unload quietly are only profitable on paper until you have to sell.

Overfitting to the past. Models tuned tightly to historical relationships broke when the world produced an out-of-sample regime.

The risk lesson for traders

Leverage is a magnifier, not an edge. The same borrowing that made LTCM's tiny spreads profitable also guaranteed that a normal-looking move would be fatal.
Diversification is an illusion if your positions share a hidden common driver. In a panic, correlations jump toward one and "unrelated" bets become one big bet.
Liquidity is part of your risk, not a given. A position you cannot exit without moving the price is not really worth its marked value when you need to sell.
Brilliant models built on calm-market history can fail out-of-sample. Markets can do things your backtest never saw.

Practise this lesson in Map.Trade

Practice framework

Practice the pattern, not the exact event. Rehearse how a leveraged book behaves when correlations spike and exit liquidity disappears — and size so that being early or wrong cannot wipe you out before you are proven right.

Replay Lab →Practice scenarios →Testing Workflow →Risk review →