Skip to main content
LTCM 1998

LTCM 1998: Overview

Pomegra Learn

How Did the World's Most Sophisticated Hedge Fund Nearly Destroy the Global Financial System?

Long-Term Capital Management was built on the premise that superior quantitative analysis could identify and profit from small, persistent mispricings in financial markets. Its founders — including Nobel laureates Robert Merton and Myron Scholes, former Federal Reserve Vice Chairman David Mullins, and a team of PhDs from Salomon Brothers' legendary proprietary trading desk — were the most credentialed group ever assembled to manage a hedge fund. For four years, the premise worked magnificently: LTCM earned returns of 21, 43, 41, and 17 percent in its first four years, generating billions in profits for its investors and making its partners extraordinarily wealthy. In the summer of 1998, the fund lost $4.6 billion in four months — 90 percent of its equity — and came within days of defaulting on obligations that, through the dense web of derivatives counterparties, would have imposed catastrophic losses on virtually every major financial institution in the world. The Federal Reserve Bank of New York intervened not to save LTCM but to prevent a disorderly unwind of its positions from triggering a systemic cascade. Understanding what happened to LTCM — and why it mattered beyond a single fund's failure — requires understanding both the specific strategy that generated the vulnerability and the market mechanics that turned a manageable loss into a near-systemic crisis.

Convergence arbitrage: A trading strategy that identifies pairs of securities that are theoretically equivalent or should converge in price, then takes opposite positions in each (long the cheaper, short the more expensive), betting on the spread narrowing. Individually, the positions have small expected returns; leverage amplifies them into large returns relative to capital.

Key Takeaways​

  • LTCM was established in 1994 by John Meriwether (formerly of Salomon Brothers) and partners including Nobel laureates Merton and Scholes, with a strategy based on convergence arbitrage across global fixed income markets.
  • At peak, LTCM had approximately $125 billion in assets against $4.7 billion in equity — 25:1 leverage — with notional derivatives exposure exceeding $1 trillion.
  • LTCM's models were based on historical correlations that assumed market relationships would persist and that extreme moves in individual assets would not occur simultaneously across all its positions.
  • Russia's August 1998 default on domestic ruble debt was not LTCM's direct problem — the fund had limited Russian exposure — but it triggered a global flight to quality that caused correlations across all LTCM's positions to break down simultaneously.
  • Within weeks, LTCM had lost over $4 billion — 90 percent of its equity — because positions that were independently expected to converge all diverged at once in the panic.
  • The Federal Reserve Bank of New York orchestrated a $3.6 billion private sector rescue by 14 major banks, not to save LTCM's investors but to prevent the disorderly liquidation of $1 trillion in derivatives positions from devastating the counterparties.
  • LTCM's failure established concepts that would recur in 2008: model risk (historical correlations fail in crises), crowded trades (when everyone holds similar positions, exits become impossible), and too-interconnected-to-fail (derivatives exposure creates systemic risk regardless of regulatory status).

The Strategy in Detail​

LTCM's convergence arbitrage worked by identifying pairs of securities whose prices should theoretically be equivalent but were temporarily trading at different levels. Common examples:

On-the-run versus off-the-run Treasuries. The most recently issued US Treasury bond (on-the-run) trades at a slight premium to slightly older bonds (off-the-run) of similar maturity, reflecting its superior liquidity. LTCM would short the on-the-run bond and buy the off-the-run, betting the liquidity premium would narrow as the bonds aged together.

Interest rate swap spreads. The spread between interest rate swap rates and Treasury yields reflects credit risk and liquidity differences. When spreads were wide by historical standards, LTCM would take positions expecting reversion to historical norms.

Fixed income arbitrage across countries. Sovereign spreads between European government bonds and German bunds that were wider than historical norms, reflecting convergence expectations from European Monetary Union, offered arbitrage opportunities that LTCM exploited heavily.

Equity volatility. LTCM sold long-dated equity options, betting that implied volatility would decline toward its historical average. This position profited when markets were calm and volatility was high.

Each individual position had expected returns of basis points — tenths of percentage points. Combined leverage of 25:1 transformed these basis point returns into significant equity returns. The positions were also, by construction, long duration — they could take months or years to converge — so LTCM needed to survive short-term mark-to-market volatility without being forced to liquidate.

The Partnership and Its Culture​

John Meriwether had built Salomon Brothers' legendary Arbitrage Group in the 1980s, recruiting PhDs and mathematicians in a style that was unusual for Wall Street at the time. When Salomon's treasury bond bidding scandal forced Meriwether out in 1991, he eventually assembled most of his former team into LTCM, adding the academic prestige of Merton and Scholes.

The intellectual culture at LTCM was distinctive. Positions were established and sized by quantitative models; subjective judgment was subordinated to model output. Risk was measured by value-at-risk and expected shortfall metrics that were internally developed and among the most sophisticated in the industry.

But the culture also had a blind spot: a deep belief in the models' correctness that made the partners resistant to modifying positions when early signs of stress appeared. When LTCM's returns fell in 1997–98, the partners increased leverage to maintain target returns rather than reducing risk. The quantitative framework, which suggested the positions were still statistically attractive, was used to justify holding positions that market price action was signaling were problematic.

The fund's equity was approximately $7.5 billion in early 1998. Under pressure from investors worried about declining returns, Meriwether returned $2.7 billion to investors in December 1997 — reducing the equity cushion just as the period of greatest stress was approaching.

Russia's Default and the Correlation Collapse​

Russia defaulted on its domestic ruble-denominated government debt on August 17, 1998, and devalued the ruble simultaneously. The default was not entirely a surprise — Russia's fiscal position had been deteriorating for months, and spreads on Russian sovereign debt had been widening. But the specific decision to default on domestic ruble debt, rather than just allowing the ruble to depreciate, was unexpected: it was the first time a major country had chosen to default on debt denominated in its own currency, which it could theoretically always print.

The Russia default triggered a global flight to quality that was the proximate cause of LTCM's collapse. The mechanism was not LTCM's direct Russia exposure — the fund had reduced this to minimal levels — but the behavior of global markets in the aftermath.

In a genuine panic, the assumptions underlying convergence arbitrage strategies break down:

Correlation breakdown. LTCM's positions were designed to be low-correlation with each other — independent bets on different markets. In panic conditions, all risky assets fall simultaneously as investors sell everything to raise cash and buy safe assets. The on-the-run/off-the-run spread that should narrow actually widened as investors paid premiums for liquidity; the European convergence trades that should profit from EU membership widened as investors sold periphery sovereign bonds; the equity volatility positions suffered as implied volatility surged. Every position moved against LTCM simultaneously.

Liquidity vanishes. LTCM's models assumed that it could always trade at market prices. In panic conditions, market depth collapsed. Positions that might take hours to liquidate normally required days or weeks; in the meantime, prices continued moving against the positions.

Leverage amplification. At 25:1 leverage, a 4 percent adverse move in assets wipes out all equity. LTCM's assets moved far more than 4 percent against the fund between August and September 1998.

The Scale of Interconnection​

What made LTCM's situation a systemic concern rather than simply a large hedge fund failure was the scale and structure of its derivatives exposure.

LTCM had approximately $1.25 trillion in notional derivatives exposure spread across hundreds of thousands of individual contracts with dozens of major financial institutions. These were not one-sided bets; they were paired positions (long and short simultaneously). But the specific structure of the derivatives contracts meant that each major bank was a bilateral counterparty to LTCM's positions — and would face immediate losses if LTCM defaulted on its obligations.

The Federal Reserve Bank of New York estimated that a disorderly LTCM liquidation would impose losses of potentially hundreds of billions of dollars across the counterparties simultaneously — not sequentially. The losses would arrive as mark-to-market shock, not gradually. Several major financial institutions, already weakened by Asia-related losses, might not survive the additional shock.

The systemic concern was not LTCM's size in itself — at $4.7 billion in equity, it was not large by the standards of major financial institutions. The concern was the dense, bilateral derivatives network that meant LTCM's failure would transmit immediately to every major bank simultaneously.

The Federal Reserve's Orchestration​

On September 23, 1998, the Federal Reserve Bank of New York convened a meeting of the chief executives of 14 major financial institutions — including Goldman Sachs, JP Morgan, Merrill Lynch, Lehman Brothers, and others — to discuss LTCM's situation.

Fed officials were careful about the legal and institutional dimensions: the Fed had no regulatory authority over LTCM (a hedge fund, not a bank), and could not simply demand a rescue. Instead, Fed President William McDonough facilitated — his language was "arm's twisting" — a private sector rescue in which the 14 banks collectively invested $3.6 billion in LTCM in exchange for 90 percent ownership.

The recapitalized LTCM then conducted an orderly unwind of its positions over the following 18 months, with the bank consortium managing the process to minimize market impact. By early 2000, the positions were fully unwound and the consortium recovered approximately $300 million above their $3.6 billion investment.

The rescue was not a bailout in the public funds sense — no taxpayer money was committed. But the Federal Reserve's facilitation raised the same moral hazard questions that any government-organized rescue raises: had sophisticated investors been protected from the consequences of their own risk-taking, reducing the discipline that market losses provide?

Lessons Established by LTCM​

LTCM's near-failure established several concepts that would become central to financial risk management and regulation:

Model risk. The risk that the models used to price, hedge, and manage risk are wrong — particularly in extreme market conditions that differ from the historical data the models were calibrated on. LTCM's models were sophisticated and state-of-the-art in 1998; they failed because they assumed normal market conditions would persist. The concept of model risk, and the importance of stress testing beyond historical scenarios, became standard risk management doctrine after 1998.

Crowded trades. When multiple participants hold similar positions — as many banks and funds had established convergence trades similar to LTCM's — individual positions become correlated through market impact rather than underlying economic linkage. When the crowded trade reverses, all participants try to exit simultaneously, producing a liquidity crisis regardless of individual position size.

Systemic risk from interconnection. LTCM demonstrated that systemic risk can come from the density of financial interconnections rather than the size of any individual institution. A $4.7 billion fund — small relative to major banks — could threaten the system because of its trillion-dollar derivatives network. The 2008 crisis would dramatically amplify this lesson.

Leverage and liquidity interaction. High leverage combined with illiquid positions creates catastrophic risk: the leverage amplifies losses while the illiquidity prevents orderly exit. LTCM's 25:1 leverage was manageable when positions were liquid; it became unmanageable when markets seized.

Common Mistakes in Analyzing LTCM​

Blaming it on Russia alone. Russia's default was the trigger but not the cause. LTCM was vulnerable because of its leverage and position concentration; any sufficiently large shock to market correlations would have triggered a similar crisis. The Russian default simply happened to be the specific trigger.

Treating LTCM's models as simply wrong. LTCM's models were not wrong in ordinary market conditions; they described real relationships that had persisted for decades. They failed in crisis conditions because crisis conditions are qualitatively different from normal conditions — the statistical relationships that models are calibrated on do not apply when the entire financial system is simultaneously seeking liquidity.

Attributing the rescue to the Fed saving investors. The Fed's facilitation was explicitly aimed at preventing systemic damage from a disorderly liquidation — not at protecting LTCM's investors. The partners lost most of their wealth; the banks that did not participate in the rescue also lost through mark-to-market on their derivatives exposure. The rescue was a market stability intervention, not an investor protection.

Frequently Asked Questions​

Why didn't LTCM reduce its leverage earlier? As LTCM's returns declined in 1997, the partners' response was to return capital to investors (reducing equity) rather than reduce positions — effectively increasing leverage. The models suggested the positions were still attractive; the quantitative framework made leverage look manageable. The psychological and institutional barriers to accepting that the models might be wrong in a specific market environment were stronger than the early warning signals.

Were Merton and Scholes personally culpable? Both Merton and Scholes were partners who invested their own money and lost substantially. Neither engaged in fraud or misrepresentation. Their culpability was intellectual — they believed in the models more than the market signals warranted, and they did not adequately account for the possibility that the models could be systematically wrong in precisely the conditions that would generate large losses. Scholes subsequently argued that LTCM's failure was a "perfect storm" combination of events that the models could not have anticipated; critics argue that the failure was inherent in the strategy's leverage and correlation assumptions.

Did the LTCM rescue create moral hazard? Yes, to a degree. Financial institutions that had lent to LTCM and had derivatives exposure were partially protected from full loss by the managed unwind. The implicit signal that the Fed would facilitate rescues of systematically important entities — even non-banks — contributed to the subsequent decade's expansion of risk-taking by institutions that believed they would be protected from the consequences of failure. This moral hazard concern was explicitly raised in 2008 when the question of bank bailouts was debated.

Summary​

Long-Term Capital Management's 1998 near-collapse was a landmark event in financial history that established concepts — model risk, crowded trades, systemic interconnection through derivatives — that would define the subsequent decade of risk management and regulatory reform. LTCM's strategy of leveraged convergence arbitrage was intellectually sophisticated and empirically successful for four years; it failed catastrophically when Russia's August 1998 default triggered a global flight to quality that caused historical market correlations to break down simultaneously across all its positions. The fund lost 90 percent of its equity in four months. The Federal Reserve's orchestration of a $3.6 billion private sector rescue by 14 major banks prevented a disorderly liquidation that threatened the stability of the entire global financial system through the dense network of derivatives counterparties. LTCM established that systemic risk can originate from leverage and interconnection density rather than institutional size, and that models calibrated on historical market relationships provide no protection when crisis conditions produce qualitatively different market behavior.

Next​

→ The LTCM Strategy: Convergence Arbitrage and Leverage