On Wednesday, September 23, 1998, in a conference room at the Federal Reserve Bank of New York at 33 Liberty Street, William McDonough (President of the New York Fed) presided over a meeting that began at 10:00 AM and concluded at approximately 16:00. The participants represented fourteen banks — Goldman Sachs, Merrill Lynch, J.P. Morgan, Morgan Stanley, Bear Stearns, Lehman Brothers, Salomon Smith Barney, Bankers Trust, Chase Manhattan, Crédit Suisse First Boston, Deutsche Bank, UBS, Société Générale, Paribas. The agenda was the rescue of Long-Term Capital Management, the Greenwich, Connecticut hedge fund whose positions had become so large and so distressed that liquidation would have produced systemic stress in major fixed-income, derivatives, and FX markets globally. By the meeting's conclusion, the participants had agreed to inject $3.625 billion into LTCM in exchange for 90 percent equity ownership. The fund's existing investors retained 10 percent.

This Desk has watched the LTCM episode and its consequences across the twenty-eight years since with the patience the historical record demands. The September 23, 1998 rescue was not a Federal Reserve action in the strict sense — no Fed funds were committed. It was a coordination action, with the Fed bringing together private-sector creditors who collectively had standing to organize the rescue. The structural significance was the Fed's willingness to use coordination authority to prevent systemic stress without committing public funds. The framework would become an institutional template that would be invoked repeatedly through the 2008 GFC and beyond.

Reading the September 1998 sequence in detail reveals what specific structural conditions produced the LTCM crisis, what the yen rally that month revealed about position concentration, and what 1998 taught about systemic linkages between hedge fund positions and central bank framework design.

What Specifically Configured the LTCM Crisis

LTCM had been founded in 1994 by John Meriwether (former Salomon Brothers partner) with a research staff that included Robert Merton and Myron Scholes (1997 Nobel Prize laureates) and David Mullins (former Federal Reserve Vice Chairman). The fund's initial capital was $1.25 billion. By end-1997, capital had grown to approximately $7 billion through fund returns and additional contributions.

LTCM's strategy was relative-value arbitrage. The fund identified pairs of related instruments where pricing diverged from theoretical relationships and took offsetting positions expecting convergence. The strategy was profitable when applied at moderate leverage; at LTCM's actual leverage levels (often 25-30x equity), small adverse moves on individual positions produced large absolute losses.

Specific LTCM positions by 1998:

  • Sovereign yield curve trades in major markets — bets on convergence between specific rates
  • Mortgage swap basis trades — bets on convergence between mortgage rates and swap rates
  • Equity volatility trades — bets on volatility levels being mean-reverting
  • Sovereign spread trades — bets on convergence between specific country spreads
  • Various FX-related positions including substantial yen-funded carry positioning

By summer 1998, the strategy had begun encountering stress. Russian sovereign default (August 17, 1998) produced sharp risk-off positioning across markets. Sovereign spreads widened (against LTCM convergence bets). Equity volatility rose (against LTCM volatility bets). Yen-funded carry positioning unwound (yen strengthened sharply).

Through August-September 1998:

  • Russian default August 17
  • Yen rallied from approximately 144 against the dollar to 117 within weeks — among the sharpest yen moves in modern record
  • Equity volatility VIX moved from approximately 20 to 45+
  • LTCM losses accumulated rapidly

By September 21, LTCM had lost approximately $4 billion in capital (down from $4.7 billion at start-September). Liquidation of remaining positions would have produced systemic stress given the scale of LTCM positions across multiple markets.

The September 23 Rescue Architecture

The Federal Reserve Bank of New York's coordination role required specific reconstruction.

Through September 21-22, the New York Fed under President McDonough had been monitoring LTCM stress through specific channels — banks with LTCM exposure had been reporting concerns. The combined exposure of major banks to LTCM (through derivative counterparty relationships, repo financing, and direct trading) was substantial. Failure of LTCM with disorderly liquidation would have produced losses across multiple major banks.

McDonough's framework: organize private-sector rescue through coordination, with Fed providing the convening authority and institutional credibility. Fed itself would not commit funds. The fourteen-bank consortium had collective interest in orderly resolution and standing to organize one.

September 23 meeting outline:

  • 10:00 AM: Participants convene at New York Fed
  • Morning session: McDonough outlines situation and Fed perspective. Banks present individual perspectives on LTCM exposure. Initial framework for rescue discussed.
  • Mid-day: Specific rescue terms negotiated. $3.625 billion injection, 90 percent ownership, governance arrangements with rescue committee.
  • Afternoon: Detailed terms finalized. Specific bank contribution amounts agreed (most banks contributed approximately $300 million; Bear Stearns notably did not participate).
  • Approximately 16:00: Agreement reached. Press release prepared.

The structural feature: Fed authority used to bring parties together, but the rescue itself was private-sector. The framework allowed crisis intervention without taxpayer commitment, while preserving Fed flexibility for future episodes.

The Yen Rally and Its Aftermath

The August-September 1998 yen rally requires specific note as it interacted with the LTCM crisis.

Pre-crisis context: yen had been weak against the dollar through 1997 and into 1998 — USD/JPY traded approximately 130-140 across spring-summer 1998. Yen-funded carry positioning was substantial — investors borrowed in yen at low Japanese rates, converted to dollars or other currencies, invested at higher yields elsewhere.

LTCM had substantial yen-short positioning (yen-funded carry positioning). When yen began rallying against unwinding flows in August, LTCM positions deteriorated rapidly.

Specific yen trajectory:

  • August 1998: USD/JPY approximately 145
  • Mid-September 1998: approximately 132
  • September 23 (rescue day): approximately 130
  • October 7-8, 1998: USD/JPY moved from 132 to 117 over 48 hours — among the sharpest 48-hour moves in major-currency history

The October 7-8 yen rally occurred two weeks after the LTCM rescue. The unwinding of LTCM-related and parallel hedge-fund positions produced explosive yen demand. The episode contributed to subsequent regulatory attention on hedge fund position transparency and systemic risk.

The yen subsequently stabilized around 110-120 through 1999-2000 before continuing post-bubble Japanese economic dynamics produced different trajectories.

What 1998 Specifically Taught

Three structural lessons emerged from the LTCM episode.

First, position concentration creates systemic risk independent of institution size. LTCM was a $4-5 billion equity hedge fund, small compared to major banks. But its positions across multiple markets at extreme leverage created systemic linkages. The failure of one institution holding such positions could produce stress across multiple market segments simultaneously.

Second, complex position interconnections defeat traditional risk management. LTCM's 25-30x leverage was supported by counterparty banks individually managing their LTCM exposure conservatively. But the combined position concentration produced systemic vulnerability that no single counterparty's risk management framework captured. The systemic risk emerged from interactions among individually-managed exposures.

Third, central bank coordination authority extends beyond traditional monetary policy. The Fed's use of convening authority to organize private-sector rescue established a template for crisis intervention that did not require fund commitment. The framework would be invoked repeatedly through 1999-2008 (Y2K coordination, post-9/11 framework, Bear Stearns 2008, AIG 2008) and continues to operate in 2026.

The 2008 GFC Inheritance

The LTCM framework directly informed 2008 GFC responses.

Bear Stearns rescue (March 2008). The structure mirrored LTCM in form — Fed coordination of private-sector resolution (in this case, JPM acquisition). The $30 billion Fed financing through the Maiden Lane facility extended the framework toward Fed fund commitment.

AIG rescue (September 2008). The $182 billion Fed/Treasury commitment was substantially larger than LTCM but operated through similar coordination logic. AIG's CDS positioning concentration created systemic risk requiring orderly resolution.

TARP framework (October 2008). The $700 billion Treasury authority operated alongside Fed framework. The combined response architecture drew partly on LTCM-derived institutional memory.

The 2008 framework was substantially larger and more interventionist than 1998, but the underlying logic — that systemic risk requires coordinated public-private response, with central banks playing convening role — was direct inheritance.

What 2026 Specifically Inherits

Three structural inheritances from the LTCM episode operate in 2026 systemic risk frameworks.

First, hedge fund and non-bank financial intermediation supervision. Post-1998 regulatory frameworks increased attention on hedge funds, family offices, and non-bank financial intermediaries. The 2026 framework includes Financial Stability Board, IOSCO, and national-level frameworks that explicitly trace to 1998 lessons.

Second, systemic risk framework as central bank function. The Fed's 2008 establishment of the Office of Financial Research, the Financial Stability Oversight Council, and various stress-testing frameworks build on the 1998 institutional template. Central banks in 2026 explicitly hold systemic-risk-monitoring responsibility.

Third, derivatives transparency frameworks. Post-2008 regulatory requirements (Dodd-Frank, EMIR) on derivatives reporting and central clearing reflect lessons from both 1998 and 2008 about position-concentration risk. The 2026 framework incorporates these requirements as ongoing infrastructure.

What 2026 does not inherit cleanly: the assumption that frameworks can prevent all systemic episodes. The 2020 Treasury market dysfunction, the 2022 LDI pension crisis in UK, periodic credit-market stress events demonstrate that systemic vulnerabilities continue evolving despite framework refinements. The 1998 lesson about complex interconnections defeating individual risk management remains operationally relevant.

What This Desk Tracks Through 2026

Three datapoints worth registering against the 1998 framework.

Hedge fund and non-bank financial intermediation activity. FSB monitoring data on non-bank exposure provides ongoing reading on whether 1998-style position concentration is reaccumulating elsewhere.

Yen carry trade positioning. The 2026 yen environment with 300 bp Fed-BoJ rate gap supports substantial carry positioning. The April 30 and May 7, 2026 BoJ interventions created some unwinding pressure. Whether positions continue accumulating or shift toward 1998-style unwind is the variable.

Systemic risk framework operation under 2026 stress conditions. Iran-conflict volatility, Q3 divergence stress test, and any specific crisis episodes test the post-1998-2008 framework architecture.

Honest Limits

This Desk reads the LTCM episode from publicly available Federal Reserve archives, congressional hearings, contemporary reporting in WSJ, FT, Reuters, Bloomberg, and substantial economic literature including Roger Lowenstein's "When Genius Failed." The 2026 references reflect current data through early May 2026. None of this constitutes investment guidance. Hedge fund and derivatives positioning carries substantial risk requiring qualified consultation.

Sources