Long-Term Capital Management (LTCM)
Company Overview
Long-Term Capital Management was a Greenwich, Connecticut-based hedge fund founded in 1994 by John Meriwether — formerly of Salomon Brothers — and staffed with an extraordinary collection of financial talent: two Nobel laureates in economics (Myron Scholes and Robert Merton, co-developers of the Black-Scholes options pricing model), former Federal Reserve officials, and the most sophisticated quantitative finance practitioners of the era. In 1998, after four years of exceptional returns averaging over 30% annually, LTCM collapsed in what nearly became a systemic financial crisis — requiring a coordinated rescue organized by the Federal Reserve Bank of New York.
For Charlie Munger, LTCM's collapse was not a surprise but a confirmation of theoretical predictions he had derived from his models of human psychology and institutional behavior. He cited it repeatedly — in Wesco annual letters, in speeches, and in DJCO meetings — as the most dramatic demonstration of the Lollapalooza Effect in financial market history: a case where multiple cognitive biases and institutional incentive failures operated simultaneously and with devastating mutual reinforcement. The lesson, in Munger's framework, was not about LTCM specifically but about the structural vulnerability of any system that combines high leverage with overconfident models of future behavior.
Investment Story
The strategy was intellectually rigorous, academically respected, and had produced extraordinary returns in its first four years. The problem was not the strategy's logic but its assumptions: that historical correlations between instruments would hold under stress conditions, that liquidity would be available to maintain positions during temporary adverse price moves, and that the firm's models of market behavior were sufficiently robust to justify the leverage ratios it employed.
As spreads widened dramatically — moving in the wrong direction from LTCM's perspective — the firm's massive leveraged positions generated catastrophic losses. LTCM had accumulated positions with notional values exceeding $1.25 trillion against an equity base of approximately $4.7 billion — leverage of roughly 25:1 on balance-sheet assets and far higher on notional exposure. As losses mounted, the firm's attempts to reduce positions pushed prices further against it, creating a feedback loop that threatened not just LTCM's solvency but the financial system's stability.
Excessive self-regard tendency: LTCM's founders — Nobel laureates and former Salomon bond traders — had overwhelming confidence in their models and their ability to manage risk. This confidence was rational given their individual track records but became dangerous when extended to models of market behavior under unprecedented stress conditions.
Overconfidence in the reliability of quantitative models: LTCM's models were calibrated on historical data that did not include conditions of simultaneous global flight to quality. Extrapolating from normal market conditions to stress conditions using parameters derived from normal conditions is not a technical error but a conceptual one — the assumption that models calibrated in one regime will perform in a fundamentally different regime.
Excessive leverage as incentive-caused bias: LTCM's partners were compensated on returns, and leverage magnifies returns (as well as losses). The incentive to employ maximum leverage was built into the compensation structure.
Social proof and authority: LTCM's investors — sophisticated institutions — were comfortable with the fund's risks partly because of the authority of its founders. If Myron Scholes and Robert Merton — the architects of modern options theory — were comfortable with the leverage ratios, surely they must be defensible.
Commitment and consistency: Having publicly committed to a strategy and having defended it to investors as rigorous and well-controlled, reversing course became psychologically and institutionally costly.
Munger's Own Words
"Similarly, the hedge fund known as 'Long Term Capital Management' recently collapsed, through overconfidence in its highly leveraged methods, despite I.Q.s of its principals that must have averaged 160. Smart, hard-working people aren't exempted from professional disasters from overconfidence. Often, they just go around in the more difficult voyages they choose, relying on their self-appraisals that they have superior talents and methods."
"Yeah, what was interesting about that one is how talented the people were. And yet, they got in so much trouble. I think it also demonstrates that — I'd say, the general system of finance in America involving derivatives is irresponsible. There's way too much risk in all these trillions of notational value sloshing around the world. There's no clearing system, as there is in a commodities market. And I don't think it's the last convulsion we're going to see in the derivatives game."
Investment Lessons
Quantitative models fail most catastrophically in the regimes they were not designed for. LTCM's models performed as intended in normal market conditions — the conditions under which they were calibrated. When a genuinely unprecedented event (simultaneous global flight to quality) occurred, the models provided no useful guidance because they were extrapolating from a regime that no longer applied. The lesson is not that quantitative models are useless but that model users must understand the regime assumptions embedded in their parameters and recognize when those assumptions no longer hold.
Leverage is the variable that converts survivable errors into catastrophic ones. Any investment strategy will produce periods of adverse price movement. With modest leverage, adverse price movements are uncomfortable but survivable — positions can be maintained until prices revert. With extreme leverage, adverse price movements trigger margin calls, forced selling, and feedback loops that can render a fundamentally sound strategy terminally insolvent before the prices revert. LTCM's strategy may have been sound in a world without the leverage it employed; with 25:1 balance-sheet leverage and much higher notional leverage, adverse price movements were not survivable.
Extraordinary intellectual credentials do not guarantee rational behavior. The presence of Nobel laureates and former Federal Reserve officials at LTCM provided social proof that the fund's risk management was defensible. But credentials measure mastery of existing knowledge, not immunity to cognitive bias. Myron Scholes and Robert Merton understood options pricing better than anyone alive; they were not immune to overconfidence in models they had built or to the incentive-caused bias created by the compensation structure they operated within. Intelligence is not rationality.
The Berkshire bid that wasn't. A footnote Munger's audiences found fascinating: during the rescue weekend, Berkshire made a firm bid for LTCM's portfolio — one hundred billion-plus of balance-sheet assets and over a trillion of derivative contracts — while Buffett was in an Alaskan canyon chasing a satellite phone signal and Munger was unreachable in Hawaii. The bid lapsed; the Fed brokered its consortium instead. The episode is a quiet demonstration of the liquidity advantage: Berkshire was perhaps the only private entity on earth that could write that bid on forty-five minutes' notice, because decades of float discipline had built a balance sheet that was itself a kind of insurance company for financial panics.
You only have to get rich once. Buffett's summary at the same 1999 meeting — that very bright people risked something that had enormous utility to gain something with no utility to them at all — is the behavioral twin of Munger's structural analysis. LTCM's partners were already rich beyond any consumable purpose. Leverage offered them nothing but a larger number; it threatened everything that made the number meaningful. That they took the trade anyway is the cleanest proof in Munger's case files that this particular human tendency does not yield to intelligence, experience, or self-interest correctly understood.
Mentioned In
- Speech to the Foundation Financial Officers Group, 1998 (the "I.Q.s that must have averaged 160" verdict)
- Berkshire Hathaway Annual Meeting, 1999 (the derivatives-system indictment and the Berkshire bid story)
- Poor Charlie's Almanack, Talk Nine: Investment Practices of Leading Charitable Foundations (LTCM alongside the GM fourth-door case)
Source: Charlie Munger Knowledge Base — Wesco Financial annual letters 1998–1999, USC speeches