Long-Term Capital Management"The Enemy Is Us"
How a fund run by two Nobel laureates and the brightest minds on Wall Street earned billions through mathematical arbitrage — and then lost nearly everything in five weeks. Based on Sebastian Mallaby's More Money Than God, Chapter 10.
Origins
From Salomon Brothers to Greenwich
John Meriwether built a legendary bond trading group at Salomon Brothers in the 1980s. Unlike traditional traders who relied on gut instinct, he hired PhD economists and finance professors — including Robert Merton and Myron Scholes, who would later win the Nobel Prize. By the end of the decade, this small academic unit generated 90% of Salomon's profits.
In 1991, a Treasury bond scandal forced Meriwether to resign. In February 1994, he launched Long-Term Capital Management (LTCM) in Greenwich, Connecticut — "Salomon without the bullshit." The idea: strip away the overhead of a large bank and concentrate purely on the quantitative trading strategies that had made his team so profitable.
The Core Belief
The Strategy
Convergence Arbitrage at Scale
LTCM's core approach was convergence arbitrage — finding two nearly identical securities priced differently, buying the cheap one, shorting the expensive one, and waiting for the gap to close. The profits per trade were tiny, so LTCM used massive leverage to amplify them.
New bonds trade at a liquidity premium over slightly older ones. Both will pay identical cash flows — so the gap must close. Buy the cheap old bond, short the expensive new one.
Tax rules suppressed foreign demand for Italian bonds, making them cheap. Partner with an Italian bank to bypass the tax, earn a 1% spread. Contributed $600M in LTCM's first two years.
Two listed companies (Dutch and British) representing claims on the same profits — but trading at different prices. Buy the cheap British shares, short the expensive Dutch ones.
Sell S&P 500 options at 19% implied volatility — far above the historical 10–13% range. Collect premiums from nervous investors. Morgan Stanley dubbed LTCM 'the central bank of volatility.'
The Leverage Logic
Annual Returns: LTCM vs S&P 500
LTCM's returns were uncorrelated with the market — until 1998, when they collapsed catastrophically.
- LTCM
- S&P 500
The Collapse
Three Acts of Destruction
The IMF bailout of Indonesia faltered, the Suharto regime collapsed, Japan entered recession, and Russia tripled interest rates. Investors fled to U.S. Treasuries, widening the spread between safe and risky assets. LTCM lost 6% in May and 10% in June — far beyond its VaR estimates. The partners trimmed positions and dismissed it as a freak event.
Russia defaulted on its debt, triggering a global 'flight to quality' of unprecedented scale. Every single LTCM trade was based on the premise that irrational spreads would converge. Instead, spreads widened violently across every market simultaneously. The on-the-run/off-the-run Treasury spread — which normally moved less than 1 basis point per day — widened by 8 basis points in a single day. LTCM lost $1.9 billion (44% of its capital) in August alone. Their models said this should happen less than once in the lifetime of the universe.
Once the market knew LTCM was in trouble, every bank and hedge fund began trading against it. Goldman Sachs, conducting due diligence on a potential rescue, allegedly downloaded LTCM's entire position book — and Goldman's trading desk began selling the same positions, front-running LTCM's inevitable liquidation. Meriwether approached Warren Buffett, Saudi Prince Alwaleed, Michael Dell, and Merrill Lynch — all refused. By mid-September, LTCM was losing $500 million per day. The New York Fed brokered a $3.6 billion bailout by 14 major banks.
Treasury Spread Widening (basis points)
On-the-run vs off-the-run spread — LTCM's core trade — exploded in Aug–Sep 1998.
LTCM Capital Base ($M)
From $4.7B at start of 1998 to under $600M by September — a 87% destruction of capital.
The Central Lesson
The Enemy Is Us
The chapter's title comes from the Walt Kelly comic strip character Pogo: "We have met the enemy and he is us." LTCM's very success had created its own destruction.
By being so widely imitated, its trades became monstrously crowded. When the crisis hit, every arbitrageur was a forced seller simultaneously — driving prices further from fair value, making LTCM's losses worse. The Slinky effect had reversed: instead of prices snapping back, they were being pushed further apart.
Eric Rosenfeld, one of the partners, later identified this as "the fund's central error": the failure to anticipate trader-driven correlation. LTCM's models assumed its trades were uncorrelated because they were economically different. But they were all held by the same type of fund — so in a crisis, they all lost money together.
VaR models based on historical data cannot anticipate unprecedented events. LTCM's models said August 1998 losses should occur less than once in the lifetime of the universe.
Diversification failed — not because trades were economically similar, but because they were all held by the same type of fund. In a crisis, all positions lost money simultaneously.
LTCM's success attracted imitators. By 1998, every bank that might buy LTCM's positions had already bought them. When LTCM needed to sell, there were no buyers.
The more LTCM tried to raise capital, the more it signaled distress, the more predators attacked its positions, the more capital it lost. A self-reinforcing death spiral.
LTCM Leverage Ratio Over Time
As capital eroded in 1998, the effective leverage ratio exploded — making every adverse price move more devastating.
- Leverage (×)
Pre-Class Assignment
Draft Answers — Click to Expand
These are draft answers grounded strictly in the case material. Question (c) requires your personal opinion — please revise before submitting.
Key Numbers to Remember