What LTCM Taught Risk Management

Long-Term Capital Management was, on paper, the safest hedge fund imaginable. Its partners included two Nobel Prize winners in economics and some of the most experienced bond traders on Wall Street. Its trades were "relative value" bets — betting that two very similar bonds, temporarily mispriced relative to each other, would converge back together — the kind of trade that looks close to risk-free if you only look at price history. By 1998 the fund had grown its capital roughly thirty-fold through leverage, controlling over $1 trillion in notional positions on around $5 billion of equity. When Russia defaulted on its debt in August 1998, markets around the world moved together in a way LTCM's models said was nearly impossible, and the fund lost most of its capital in a matter of weeks, forcing a Federal Reserve-organized bailout by fourteen banks to avoid a broader meltdown.

What actually happened

LTCM's individual trades were, in isolation, reasonable bets that spreads would narrow. The problem was that the fund held dozens of such trades simultaneously across bond markets, equity volatility, and emerging markets, and treated them as diversified because their historical correlations were low. When the Russian default hit, investors everywhere rushed to the same trade at once — dump anything perceived as risky, buy the safest government bonds — and every one of LTCM's "diversified" positions moved against the fund at the same time. Spreads that had never widened this far in the sample data widened much further, because the sample data didn't include a global flight to quality. Leverage then did the rest: a loss that would have been a bad month for an unleveraged fund became an existential threat for a fund leveraged around thirty to one, because losses of that scale eat through equity almost immediately and force a fund to sell into a market that has already turned against it, deepening the loss further.

The lessons risk managers took away

Correlations estimated from calm markets are not a reliable guide to what happens in a crisis — assets that rarely move together can move together violently exactly when it matters most, because the same shock (a rush for cash and safety) hits everyone's book simultaneously. Leverage doesn't just scale returns, it scales the speed at which a loss becomes forced liquidation, and forced liquidation into an already-stressed market pushes prices further against you, creating a feedback loop. And size itself is a risk: LTCM's positions were so large relative to the market that unwinding them, even if the underlying thesis had eventually been proven right, was not something the fund could survive long enough to do.