LTCM 1998: When 25-to-1 Leverage Meets a Liquidity Crisis

Key Takeaways

• LTCM ran roughly 25-to-1 balance-sheet leverage on $4.7 billion of capital, meaning a 4% adverse move could wipe out all equity.
• In August 1998 alone LTCM lost 44% of its capital; a single day saw a $553 million loss against an internal VaR estimate of $35 million.
• Investors mistake low historical volatility for low risk, a strategy with a 2% standard deviation leveraged 25 times carries the risk of a 50% vol strategy.
• LTCM's supposedly uncorrelated trades all moved against the fund simultaneously when the common factor, liquidity stress, emerged after Russia's default.

What It Is

LTCM ran highly leveraged convergence trades in fixed-income markets. The fund identified pairs of securities whose prices had diverged from historical norms, bought the cheaper one, shorted the more expensive one, and waited for the spread to close.

The trades looked low-risk in normal markets. Spreads were small, so the fund used enormous leverage, reportedly reaching around 25:1 on the balance sheet and much higher including off-balance-sheet derivative exposures. At the start of 1998, LTCM managed about $4.7 billion of capital supporting roughly $125 billion of balance-sheet positions and trillions in notional derivatives.

On August 17, 1998, Russia devalued the ruble and defaulted on its domestic debt. Global credit and liquidity spreads widened sharply. LTCM lost 44% of its capital in August alone and kept bleeding into September. On September 23, fourteen banks and brokerages led by the New York Fed committed $3.6 billion to recapitalize the fund and wind its positions down in an orderly way.

The Intuition

LTCM is the textbook case of how low-volatility strategies explode when correlations flip. In normal conditions, LTCM's trades were nearly independent. Italian bond spreads, US mortgage spreads, Japanese swap spreads, and equity volatility trades seemed uncorrelated. In a panic, all of them moved against the fund together because every trade shared a common factor: selling cheap securities and buying expensive ones is what stressed investors do.

The deeper lesson is that leverage magnifies tail risk non-linearly. A 25:1 leveraged book can tolerate small adverse moves. A single 4% move wipes out equity. When the move arrives, forced selling at bad prices prevents the trade from ever recovering.

How It Works

Four features made the collapse possible:

• Convergence trades. Buy a less liquid but similar instrument, short the more liquid one, capture the liquidity premium. Works until liquidity premiums blow out, at which point both legs move against you.
• Extreme leverage. LTCM used balance-sheet leverage around 25:1 plus massive derivative notional. A 4% adverse move on the underlying assets could wipe out all equity.
• Risk model calibration. Internal VaR models estimated a maximum daily loss of around $35 million. In August 1998 the fund lost $553 million on a single day, a multiple-sigma event under their assumptions. The models used recent-period volatility that badly understated tail risk.
• Concentration among a few counterparties. Major banks had sizable exposures to LTCM. A disorderly unwind threatened to transmit losses through the derivatives market to counterparties that were systemically important. This is what prompted the Fed to coordinate a rescue.

Worked Example

Consider LTCM's on-the-run versus off-the-run Treasury trade, a typical position. The newest 30-year Treasury traded at a small premium to a slightly older 30-year Treasury because it was more liquid. LTCM bought the off-the-run bond, shorted the on-the-run bond, and captured the roughly 10 to 20 basis-point spread as it converged over months.

In August 1998, frightened investors rushed into the most liquid securities. The on-the-run premium widened from 10 to 15 basis points to 30 to 40 basis points. LTCM was losing money in exactly the scenario its models had assumed was vanishingly unlikely. Because the position was leveraged, a spread move that looked small translated into massive dollar losses. Other LTCM positions widened similarly. The fund could not close positions without crystallizing losses that would breach margin with every counterparty.

Common Mistakes

• Mistaking low volatility for low risk. A strategy with a 2% standard deviation leveraged 25 times has the risk of a 50% vol strategy unleveraged. Historical volatility from calm periods is a poor guide to stress-period risk.
• Ignoring correlation regime change. Supposedly uncorrelated trades can all move together when the underlying factor is liquidity or counterparty stress. Diversification that works in calm markets often disappears in panics.
• Trusting Nobel-endorsed models blindly. Merton and Scholes's option-pricing work is correct in its domain. Applied to leveraged arbitrage, the same models underestimated tail behavior. Prestige is not a risk control.
• Underestimating systemic interconnection. LTCM itself was privately held, but its counterparty network included every major dealer. The Fed intervened not to rescue investors but to prevent a disorderly cross-market unwind. Opaque linkages between funds and banks remain a systemic issue.
• Treating the rescue as a success template. The private rescue prevented a messy liquidation but arguably encouraged the risk-taking that preceded 2008. Moral hazard from visible bailouts changes the behavior of the next generation of leveraged funds.

Frequently Asked Questions

Q: What was LTCM in simple terms? LTCM was a hedge fund that borrowed heavily to bet on small price gaps between similar securities closing over time. The strategy worked in calm markets but depended on being able to hold positions through stress. When Russia defaulted in 1998, the gaps widened and LTCM could not hold on, losing most of its $4.7 billion in capital in weeks.

Q: How does LTCM affect investment decisions today? It demonstrated that leverage turns a small adverse price move into an existential event. It also showed that diversification based on historical correlations fails in stress because all trades can share a hidden common factor, like needing to sell liquid assets when everyone else needs liquidity too.

Q: What is a real-world example from the LTCM collapse? LTCM's on-the-run versus off-the-run Treasury trade saw spreads widen from 10–15 basis points to 30–40 basis points as frightened investors rushed into liquid securities. Because the position was leveraged 25-to-1, a basis-point move that looked small translated into hundreds of millions in daily losses.

Q: How can investors apply LTCM lessons to avoid similar blowups? Cap leverage so that the worst plausible historical drawdown cannot wipe out equity. Test correlations under a stress scenario, not just normal conditions. Treat models built on recent-period volatility as optimistic baselines, not risk ceilings.

Q: How is LTCM different from a typical hedge fund failure? Most hedge fund failures involve bad fundamental bets. LTCM held technically correct convergence trades that would have paid off if it could hold them. The failure was purely structural: too much leverage against too little time horizon, in a world where every counterparty faced the same stress simultaneously.

Sources

1. Federal Reserve History. Near Failure of Long-Term Capital Management. https://www.federalreservehistory.org/essays/ltcm-near-failure
2. Lowenstein, R. (2000). When Genius Failed: The Rise and Fall of Long-Term Capital Management. Random House. https://www.penguinrandomhouse.com/books/164406/when-genius-failed-by-roger-lowenstein/
3. PRMIA. Case Study: Long-Term Capital Management. https://prmia.org/common/Uploaded%20files/ORM%20Designation/PRMIA_LTCM_062321.pdf
4. Richmond Fed Economic History. LTCM retrospective. https://www.richmondfed.org/-/media/richmondfedorg/publications/research/econ_focus/2009/summer/pdf/economic_history.pdf

Disclaimer

This article is educational content only and is not financial advice. Nothing here is a recommendation to buy, sell, or hold any security. Consult a licensed advisor before making investment decisions.