Skip to content
Menu
Finabase Learn

Articles · Market history

Long-Term Capital Management (1998): how leverage sank a genius fund

How LTCM made convergence trades at about 25 times leverage, lost 44% in August 1998 after Russia’s default and was recapitalized by 14 firms.

On this page
  1. How the fund made money
  2. The arithmetic of leverage
  3. What went wrong in 1998
  4. The rescue
  5. Lessons for investors
  6. LTCM and its neighbors
  7. How Finabase helps
  8. Frequently asked questions
  9. Summary
  10. Check your understanding

Long-Term Capital Management was a hedge fund founded in 1994 by John Meriwether, who had run bond arbitrage at Salomon Brothers, with partners who included two economists, Myron Scholes and Robert Merton, who shared the 1997 Nobel Prize in economics. For its first years it was one of the most successful funds in the world. In 1998, after Russia defaulted on its debt, it lost most of its capital in about five weeks, and on 23 September the Federal Reserve Bank of New York brought fourteen banks and brokers together to put $3.6 billion into the fund rather than let it fail.

LTCM is the textbook case of how leverage and crowded trades turn small price moves into a crisis. Its strategies were not reckless one by one; most were bets that small price gaps between related securities would close. What sank it was the size of its positions relative to its capital, the fact that many others held similar trades, and a market that, for a few weeks, behaved unlike anything in its models.

How the fund made money

LTCM's core trades were relative-value bets. A classic example: two US Treasury bonds with almost the same maturity, one newly issued ("on the run") and one issued a few months earlier ("off the run"). The newer bond is easier to trade, so investors pay a little more for it. That gap usually narrows as the newer bond ages. LTCM would buy the cheaper older bond, sell the newer one short, and collect the narrowing. It ran similar trades in interest rate swaps, mortgage securities, European government bonds before the euro, pending mergers and stock-index options.

Each trade earned a small amount, often a fraction of a percentage point, with what the models said was little risk because the two sides moved together. To make those small spreads worth while, the fund used leverage: borrowing against the bonds it held, and using derivatives that required little capital up front.

The results were strong. Federal Reserve History reports returns of 43% in 1995, 41% in 1996 and 17% in 1997. At the end of 1997 the partners returned about $2.7 billion to outside investors, judging that the opportunities no longer justified so much capital. That made the fund more leveraged, not less: the positions stayed roughly the same size against a smaller base.

The arithmetic of leverage

Equity left after a fall in asset values, by leverage (worked example) 0% 25% 50% 75% 100% -10% -8% -6% -4% -2% 0% Assets 1× equity Assets 5× equity Assets 25× equity At 25×, a 4% fall erases the equity Change in the value of the assets held
Worked example: how much equity is left after a fall in the assets it supports.

Leverage multiplies both directions. With $1 of capital supporting $25 of assets, a 1% gain on the assets is a 25% gain on capital, and a 4% loss on the assets wipes out the capital entirely. For a fund whose trades were supposed to move only slightly, that seemed acceptable. It assumed, though, that the two sides of each trade would keep moving together and that the fund could keep borrowing.

A worked example: a spread that widens before it narrows

Suppose LTCM buys $1 billion of an older bond and sells short $1 billion of the newer one when the yield gap is 0.10 percentage points, expecting it to shrink to 0.05. On bonds of about ten years' duration, a 0.05-point change in the gap is worth roughly 0.5% of the position, about $5 million.

Now suppose a panic makes investors pay up for the easy-to-trade bond, and the gap widens to 0.30 points instead. That 0.20-point move against the fund costs about 2% of the position, $20 million. With the trade held on, say, $40 million of capital (25 times leverage), half the capital behind it is gone before the "safe" trade has a chance to work. If lenders then ask for more collateral, the fund may have to close the trade at the worst moment, locking in the loss.

What went wrong in 1998

The trouble began before Russia. Spreads had been widening for months in 1998, and Salomon Brothers, which ran similar trades, closed its US arbitrage desk that summer, selling positions that resembled LTCM's. Then on 17 August 1998 Russia devalued the ruble and defaulted on its domestic debt. A crisis that had begun in Asia a year earlier (The Asian financial crisis of 1997: pegs, hot money and contagion) suddenly spread to every market where investors were taking risk.

Investors around the world sold risky and hard-to-trade assets and bought the safest, most liquid ones. That was exactly the opposite of LTCM's book: it owned the cheaper, less liquid side of each pair and had sold the expensive, liquid side. Every gap widened at once. Trades that the models treated as unrelated turned out to be the same bet on calm markets and plentiful liquidity.

According to Federal Reserve History, the fund lost 44% of its value in August alone. The President's Working Group put its capital at $2.3 billion at the end of August, against more than $125 billion of balance-sheet assets and well over $1 trillion of derivatives notional. By late September, as reported in the Wall Street Journal, its capital was down to about $600 million.

Cboe Volatility Index (VIX), monthly highs and closes, 1996–1999 20 40 Oct 1998 high 49.5 Monthly high Monthly close 1997 1998 1999
VIX, monthly bars, 1996–1999. Data: Finabase chart API, frozen 2026-11-18.

US stocks fell too: the S&P 500 dropped about 19% from its July 1998 high to its October low, and the VIX, a measure of expected stock-market volatility, reached about 49.5 in October in our data.

Why the losses fed on themselves

  • Everyone held the same trades. Other banks and funds had copied LTCM's positions. When they cut risk, they sold the same securities, widening the gaps further.
  • The market knew. Once it was clear that LTCM was in trouble, other traders could expect it to be forced to sell, which made its positions harder to exit.
  • Liquidity vanished. The securities LTCM owned were, by design, the less liquid ones. In a panic there were few buyers for them at any price close to the models' values.

The rescue

The Federal Reserve Bank of New York was concerned that a disorderly default would force LTCM's many lenders and trading partners to sell collateral at the same time, pushing prices down further for everyone. On 22–23 September it brought together the fund's largest counterparties. They agreed to invest $3.6 billion in exchange for about 90% of the fund and control over winding it down. Bear Stearns, the fund's clearing broker, declined to take part.

No public money was lent to LTCM. The Fed did cut interest rates three times in the autumn of 1998, partly to calm markets. The consortium ran down the positions over the following months, and most of the trades eventually converged, too late for the original investors. The fund was wound up by early 2000.

Lessons for investors

LTCM and its neighbors

Episode Year What the leverage was in What broke it
Silver Thursday 1980: the Hunt brothers and a silver corner that broke 1980 Silver futures Higher margins, position limits
LTCM 1998 Bond and swap spread trades Russia's default, a flight to safety
The 2008 financial crisis: housing, Lehman, TARP and the 57% fall 2008 Mortgage securities at banks Falling house prices, a run on funding
The 2010 Flash Crash: what happened on May 6 and what changed 2010 Not leverage: thin liquidity A large sell order in a fragile market

How Finabase helps

The VIX chart with the maximum range puts 1998 beside 2008, 2020 and later spikes. On the Portfolio page you can see how concentrated your own holdings are; the Stress testing a portfolio: what would a past crisis cost? lesson shows how to ask what a 1998-style shock would do to them.

Frequently asked questions

What was Long-Term Capital Management?

LTCM was a hedge fund founded in 1994 by John Meriwether, whose partners included Nobel laureates Myron Scholes and Robert Merton. It made leveraged bets that small price gaps between related bonds and derivatives would close.

Why did LTCM fail?

After Russia's default in August 1998, investors fled to safe and liquid assets, so the price gaps LTCM bet on widened instead of narrowing. With about 25 times leverage, the losses consumed most of its capital within weeks, and its lenders and trading partners faced the risk of a disorderly default.

Was LTCM bailed out by the government?

Not with public money. The Federal Reserve Bank of New York organized a meeting at which 14 banks and brokers agreed, on 23 September 1998, to invest $3.6 billion of their own money for about 90% of the fund.

How much leverage did LTCM use?

The President's Working Group reported equity of about $4.8 billion at the start of 1998 and balance-sheet assets of more than $125 billion at the end of August, more than 25 to 1, plus over $1 trillion of derivatives notional. It put leverage at 28 to 1 at the end of 1997; Federal Reserve History puts the fund's debt at about $30 for every $1 of capital then.

Summary

LTCM was built on careful, mostly sensible trades that each earned little, made worth while by borrowing about 25 times its capital. When Russia's default in August 1998 set off a flight to safety, every trade moved against it at once, it lost 44% in a month, and fourteen firms put in $3.6 billion to wind it down in an orderly way. Its lessons are about leverage and crowding: a position that must survive the path can be right in the end and still fail.

Check your understanding

4 questions on this lesson. Each answer comes with the reason, and nothing is scored.

  1. A fund holds $25 of assets for every $1 of capital. What fall in the value of the assets wipes out the capital?

    1. 25%
    2. 4%
    3. 1%
    Show the answer

    B. 4% $25 × 4% = $1, all of the capital. High leverage turns small price moves into total losses.

  2. LTCM’s convergence trades bought the cheaper, less liquid security and sold the dearer, more liquid one. Why did a flight to safety hurt every trade at once?

    1. Because in a panic investors pay up for liquid, safe assets, so the gaps widened in all the pairs together
    2. Because the trades were all in Russian bonds
    3. Because interest rates were frozen
    Show the answer

    A. Because in a panic investors pay up for liquid, safe assets, so the gaps widened in all the pairs together The pairs looked unrelated but shared one bet: that liquidity premiums would shrink. A rush to safety widened them all.

  3. Many of LTCM’s trades eventually converged after the rescue. Why did that not save the fund?

    1. Because the trades were closed at a profit
    2. Because with high leverage it had to survive the interim losses and could not wait for the gaps to close
    3. Because regulators banned the trades
    Show the answer

    B. Because with high leverage it had to survive the interim losses and could not wait for the gaps to close Leverage removes time. A position can be right in the end and still fail if losses along the way exceed the capital.

  4. Why did the New York Fed organize a private recapitalization rather than let LTCM default?

    1. To protect LTCM’s partners from losses
    2. Because a disorderly default could force many lenders to sell collateral at once, pushing prices down for everyone
    3. Because the Fed owned part of the fund
    Show the answer

    B. Because a disorderly default could force many lenders to sell collateral at once, pushing prices down for everyone The worry was fire sales across markets. The 14 firms used their own money; the partners lost most of their stakes.

Sources and further reading

  1. Federal Reserve History — Near Failure of Long-Term Capital Management
  2. President’s Working Group on Financial Markets — Hedge Funds, Leverage, and the Lessons of Long-Term Capital Management, April 1999 (capital, assets, leverage, $2.7 billion returned)
  3. Lowenstein — When Genius Failed: The Rise and Fall of Long-Term Capital Management, Random House, 2000
  4. MacKenzie — An Engine, Not a Camera, MIT Press, 2006 (chapter 8, LTCM)
  5. Wall Street Journal — reporting of 24–25 September 1998 on the LTCM recapitalization
  6. Yardeni Research — S&P 500 bull and bear market tables
  7. Finabase chart API — ^VIX monthly bars, frozen 2026-11-18