Skip to main content
LTCM 1998

The Collapse: LTCM's Four Months to Ruin

Pomegra Learn

How Did $4.7 Billion Become $400 Million in Four Months?

Few financial collapses have been as precisely documented as LTCM's implosion between May and September 1998. Roger Lowenstein's detailed account, the Congressional testimony, and the post-mortem analyses have produced an unusually granular record of how one of history's most sophisticated investment funds destroyed 90 percent of its equity in 120 days. The narrative arc has the quality of a Greek tragedy: each decision that seemed rational in the moment — returning capital to reduce investor pressure, maintaining positions that models said would eventually converge, seeking rescuers who ultimately declined — brought the fund closer to the catastrophe that the models said was essentially impossible. The collapse was not a single event; it was a cascade of losses that fed on themselves, each day's loss triggering margin calls that forced liquidations that drove positions further against the fund, which triggered more margin calls. Understanding the collapse mechanics — not just the headline losses — illuminates why systemic risk cannot be managed at the individual institution level.

Mark-to-market (MTM) loss: The loss recognized when financial instruments are valued at current market prices rather than original cost or theoretical value. For LTCM, daily MTM losses required posting additional collateral to counterparties or facing liquidation — creating a direct connection between paper losses and immediate cash demands.

Key Takeaways​

  • LTCM lost approximately $461 million in May 1998, a loss attributed partly to early Russian stress and market volatility — it should have been a warning but was treated as temporary by the partners.
  • Following Russia's August 17 default, LTCM lost $1.9 billion in a single month — 44 percent of its remaining equity — as all positions moved adversely simultaneously.
  • By mid-September 1998, LTCM's equity had fallen from $4.7 billion to approximately $600 million against $125 billion in assets — a leverage ratio of approximately 200:1.
  • Goldman Sachs, in mid-August, made a consortium bid for LTCM — Warren Buffett's Berkshire Hathaway, Goldman, and AIG offered to buy the fund for $250 million and recapitalize it with $3.75 billion. LTCM rejected the offer; the bid later lapsed.
  • On September 23, the Federal Reserve Bank of New York orchestrated the $3.6 billion consortium rescue by 14 banks — the only option remaining to prevent immediate default.
  • Without the rescue, the New York Fed estimated the disorderly liquidation could have imposed losses of potentially hundreds of billions on counterparties and could have forced several major financial institutions to the brink of failure.

Early Warning Signs: May 1998​

LTCM's problems began in May 1998 — three months before Russia's default — when the fund lost approximately $461 million. The losses were attributed to widening spreads in several markets where LTCM held convergence positions.

The May losses coincided with increased stress in emerging markets as the Russia situation began deteriorating and Asian crisis contagion continued. Global risk appetite was declining; investors were beginning to reduce exposure to riskier assets.

LTCM's partners interpreted the May losses as a temporary fluctuation — exactly the kind of adverse short-term move that their long-duration capital was designed to survive. The models still showed the positions as attractively priced; the expected value of the trades was positive. The appropriate response, the partners believed, was to maintain positions and wait.

What LTCM did not do was reduce leverage. Total assets remained approximately $125 billion against $4.7 billion in equity. The May losses reduced equity by approximately $461 million — roughly 10 percent — but the partners did not proportionally reduce the position book.

August: The Catastrophe​

Russia's August 17 announcement triggered losses that were qualitatively different from May's. Between August 17 and the end of the month, LTCM lost approximately $1.9 billion.

The losses came from multiple positions simultaneously:

On-the-run/off-the-run spread. The flight to quality dramatically increased demand for the most liquid Treasury instruments. On-the-run bonds rallied as investors sought liquidity; off-the-run bonds fell as investors sold less liquid instruments. The spread that LTCM expected to narrow instead widened dramatically — the opposite direction from LTCM's position.

Swap spreads. The yield difference between interest rate swaps and Treasury bonds widened sharply as investors paid premium prices for Treasury safety. LTCM's short Treasury / long swap position lost substantial value.

European convergence trades. As investors fled to German bunds for safety, yields on Italian, Spanish, and other European sovereigns rose relative to Germany. The convergence trades reversed sharply.

Equity volatility. Implied equity volatility surged as markets became fearful. LTCM had sold long-dated equity options; these positions were deeply underwater as the options' mark-to-market value increased.

The correlation that the models said was near zero. All four trade categories moved against LTCM simultaneously. The portfolio diversification that the risk models had quantified was non-existent in the crisis. The model-predicted loss with 99 percent confidence was approximately $45 million per day; the actual losses were approximately $100–200 million per day.

The Death Spiral: Margin Calls and Forced Selling​

The August losses triggered a mechanical process that fed on itself.

When mark-to-market losses reduced LTCM's equity, counterparties that had provided financing (through repo arrangements and prime brokerage) had contractual rights to demand additional collateral. LTCM had to either post additional cash or liquidate positions.

To post cash, LTCM had to sell positions to raise liquidity. But selling positions in a market where the same positions were already falling — and where other similarly positioned funds were also selling — drove prices further against the fund. Each forced liquidation made the remaining positions more deeply underwater, which triggered more margin calls, which required more liquidation.

This feedback loop — losses → margin calls → forced selling → more losses → more margin calls — is the dynamic that makes balance sheet crises self-reinforcing. LTCM was experiencing the same mechanism that had devastated Asian banks in 1997, though through derivatives and hedge fund mechanics rather than banking system channels.

By the end of August, LTCM's equity had fallen from $4.7 billion to approximately $2.3 billion. The leverage ratio had risen from 25:1 to approximately 55:1 as equity contracted faster than assets.

The Failed Goldman Rescue Attempt​

In mid-August, Goldman Sachs — which was both a major LTCM counterparty and a competitor that had been trying to recruit LTCM's talent — approached LTCM about a potential acquisition.

The consortium Goldman assembled included Warren Buffett's Berkshire Hathaway and AIG, representing the most financially credible potential buyers that could be assembled quickly. The offer: $250 million for LTCM (effectively buying the partners out), followed by $3.75 billion in recapitalization from the consortium.

The offer would have ended the crisis for the financial system — the consortium had the resources to manage the positions in an orderly way without threatening counterparties. But it would have essentially wiped out the LTCM partners, who would receive $250 million for a firm they had built and that they believed was still worth far more.

The LTCM partners rejected the offer. Meriwether and his team believed — understandably given their quantitative framework — that the positions would eventually converge to their theoretical values and that LTCM's equity was worth far more than $250 million at appropriate valuations.

The Goldman bid lapsed. Several weeks later, with LTCM in far worse condition, there would be no acquirer willing to take the risk on a standalone basis. The Fed-orchestrated consortium was the only remaining option — but at far worse terms for LTCM's partners.

September: Final Approach to Zero​

Between September 1 and September 22, LTCM lost another $2.1 billion. By September 22, equity had fallen to approximately $400–600 million against over $100 billion in assets. The leverage ratio had reached approximately 200:1.

At this level of leverage, a 0.5 percent adverse move in assets would eliminate all remaining equity. Daily price moves in the affected markets were multiples of this; the fund was functionally insolvent on a mark-to-market basis even if the long-run theoretical values of the positions remained positive.

The specific loss breakdown for 1998 cumulative:

  • Equity volatility: approximately -$1.3 billion
  • Russian and emerging markets: approximately -$0.4 billion
  • Merger arbitrage: approximately -$0.5 billion
  • Fixed income arbitrage: approximately -$1.6 billion
  • Other strategies: approximately -$0.8 billion

The diversification across strategies — intended to limit correlated losses — failed entirely. Every category contributed substantial losses.

The Fed's Intervention​

On September 23, 1998, William McDonough — President of the Federal Reserve Bank of New York — convened the meeting of 14 major financial institution CEOs that produced the rescue.

The Fed's position was carefully defined: it was not lending its own money to LTCM, and it was not guaranteeing LTCM's obligations. It was facilitating — insisting, perhaps — on private sector coordination that the firms would not have achieved independently.

McDonough's argument to the banks was straightforward: if LTCM defaulted and its positions were liquidated disorderedly, every bank in the room would face immediate losses from their derivatives exposure. The rescue was in each bank's individual interest. The question was whether coordinating to rescue together was better than each bank trying to exit first.

This was precisely the prisoner's dilemma that the Korea voluntary rollover had also addressed. The Fed's convening authority broke the dilemma by making the coordination focal: every bank at the table could observe every other bank's commitment.

Fourteen banks — including Goldman, JP Morgan, Merrill Lynch, Deutsche Bank, and ten others — agreed to contribute $3.6 billion in exchange for 90 percent ownership of LTCM. The Lehman Brothers and Bear Stearns abstentions (which declined to participate) were noted.

The Orderly Unwind​

With the $3.6 billion consortium recapitalization in place, LTCM's positions were unwound over the following 18 months under the management of the consortium's oversight committee.

The unwind was deliberately slow. Liquidating LTCM's trillion-dollar notional derivatives book rapidly would have required massive market transactions that would have moved prices against the positions, crystallizing losses for the consortium investors. Instead, the positions were reduced gradually as market conditions allowed — taking advantage of natural maturities, offsetting new trades, and selling when markets provided reasonable prices.

By early 2000, the positions were fully unwound and the fund was dissolved. The consortium recovered approximately $300 million above its $3.6 billion investment — a modest profit that reflected both the eventual convergence of some positions and the careful management of the unwind.

LTCM's original investors and partners lost the majority of their investment. The partners — who had retained personal wealth invested in the fund — lost hundreds of millions collectively. The investors received a small portion of their original investment back.

Common Mistakes in Analyzing the Collapse​

Treating the losses as caused by bad positions. LTCM's positions were not bad in a fundamental sense; many of the convergence trades that LTCM held did eventually converge to their theoretical values after the crisis. The problem was leverage and capital insufficiency — LTCM could not survive the time required for convergence. The positions were right; the capital structure was wrong.

Ignoring the alternative consequences. The rescue has been criticized as protecting sophisticated investors from their own mistakes. The relevant comparison is not "rescue versus no losses" but "rescue versus disorderly liquidation." A disorderly LTCM liquidation would have imposed potentially hundreds of billions in immediate losses across the 14+ major banks with derivatives exposure, potentially triggering defaults at several institutions. The rescue prevented a systemic crisis, not just LTCM's failure.

Overstating the Fed's role. The Fed did not commit money or provide guarantees; it provided convening authority and credibility for the coordination exercise. The actual rescue was entirely private sector. Characterizing the Fed's role as a "bailout" misrepresents the mechanics.

Frequently Asked Questions​

Why did LTCM reject the Goldman/Buffett offer? The partners believed their positions would eventually converge to theoretical value, making the fund worth substantially more than $250 million. Given their quantitative framework, this was a defensible analysis. The psychological difficulty of accepting that the models might be wrong — and that the fund was worth only $250 million — was also real. The decision proved catastrophically wrong; three weeks later, LTCM had no choice but the Fed-orchestrated rescue at worse terms.

Did any LTCM employees become wealthy from the rescue? The rescue did not make partners wealthy; it preserved a fraction of their pre-collapse wealth. Partners who had invested personal wealth in LTCM lost the majority. Some received a small amount from the eventual liquidation, but the collapse represented severe personal financial losses for the senior partners. Meriwether himself lost over $100 million.

Were there legal consequences for LTCM's management? No criminal charges were filed. The SEC reviewed LTCM's disclosures and risk management but did not find actionable violations. The losses were the result of a strategy that failed catastrophically in extreme conditions, not of fraud or deliberate misrepresentation.

Summary​

LTCM's collapse from $4.7 billion in equity to approximately $400 million in four months was a mechanical consequence of extreme leverage meeting extreme market stress. Russia's August 17 default triggered a global flight to quality that simultaneously moved all of LTCM's convergence positions in the adverse direction — the exact scenario that the fund's diversification was intended to prevent. The self-reinforcing dynamics of margin calls and forced selling amplified losses beyond what any model had considered plausible. The Goldman/Buffett consortium offer — which would have resolved the crisis at a fraction of the eventual rescue cost — was rejected by LTCM's partners who believed their models' assessment of long-run value. By September 23, with equity at $400 million against $100 billion in assets and trillion-dollar derivatives exposure, the Fed-orchestrated private sector consortium was the only option remaining. The $3.6 billion rescue prevented a disorderly liquidation that the New York Fed estimated could have imposed catastrophic losses on virtually every major financial institution simultaneously.

Next​

→ Model Risk and the Failure of Historical Correlations