Long-Term Capital Management
Featuring John Meriwether
Two Nobel laureates, a roster of the most sophisticated quant traders alive, and a track record that printed money for years. Long-Term Capital Management, founded in 1994 by John Meriwether, built models that spotted tiny pricing gaps between related instruments and bet they would converge, then layered on leverage that at times exceeded $25 borrowed for every dollar of equity. The models assumed correlations would stay roughly stable, even under stress. In the summer of 1998, Russia defaulted, investors fled to safety all at once, and the relationships the firm had counted on broke down simultaneously.
For founders and operators, this is a study in how confidence in a framework becomes most dangerous exactly when everyone shares it. It sharpens the decision of how much to trust your own operating model, and how to stress-test the assumptions that would only fail on a genuinely bad but entirely plausible day.
Frequently asked questions
What was Long-Term Capital Management and what happened to it?
Long-Term Capital Management was a hedge fund founded in 1994 by John Meriwether that included two Nobel laureates and elite quant traders. It built models to spot tiny pricing gaps between related instruments and bet they would converge, then layered on enormous leverage. It nearly collapsed in 1998 when the Russian default broke the relationships its models relied on.
How much leverage did LTCM use?
LTCM at times borrowed more than $25 for every dollar of equity to amplify its convergence bets. Its models assumed that correlations between instruments would stay roughly stable, even under stress. That assumption combined with extreme leverage left no cushion when markets moved against the firm all at once.
Why did LTCM's models fail in 1998?
LTCM's models assumed correlations would stay roughly stable even under stress, but in the summer of 1998 Russia defaulted and investors fled to safety all at once. The relationships the firm had counted on broke down simultaneously, so its diversified-looking bets all moved the wrong way together. Confidence in the model was most dangerous exactly when everyone shared it.
What can founders learn from Long-Term Capital Management?
The lesson is that confidence in a framework becomes most dangerous when everyone shares it, so you should question how much to trust your own operating model and stress-test the assumptions that would only fail on a bad but entirely plausible day. Stable-looking correlations can break all at once under stress. CaseBook turns this into a move you apply to your own company, with an AI coach that reads your answer.