Long-Term Capital Management was the most modern of collapses: rigorous models, cheap funding, and a world that politely behaved—until it didn’t. The fund did not implode because its trades were clownish; they were sensible, even elegant. It failed because they were financed as if normal times were a law of nature. When normality broke, everything that looked diversified turned out to be the same bet—and leverage made it existential.
A trade that worked—until it didn’t
LTCM’s core idea was convergence. When two securities were economically similar but priced a little differently, the fund would buy the cheap one and short the rich one, harvesting the spread when markets reverted. It was market-making without a storefront, executed at institutional scale across government bonds, interest-rate swaps, and credit. The edge was statistical and the fuel was repo financing: borrow cash against securities, lever up a low-volatility position, and let time do the work.
The problem was structural, not intellectual. Spreads are tiny and mean reversion is slow, so the only way to earn hedge-fund returns is to borrow a lot and count on stable correlations. That works—until a macro shock widens those tiny spreads, counterparties demand more collateral, and the slowness of reversion collides with the speed of margin calls. In 1998, a sovereign default and a global flight to safety turned “uncorrelated” trades into one trade: long illiquidity, short Treasuries. Models that treated 10-sigma weeks as legends met markets that delivered them.
When liquidity is just leverage in nicer clothes
Funding fragility, not theoretical error, did most of the damage. Repurchase agreements are designed for smooth times. Haircuts rise and financing dries up precisely when you need both. Mark-to-market losses force sales, sales widen spreads, and spreads increase losses. The cycle is mechanical and rapid. What looks like a diversified book becomes a single liquidity position when everyone elsewhere is selling the same things for the same reasons. Even robust legal netting can’t offset the raw speed of a margin spiral.
The rescue that wasn’t a bailout—and the template it offered
The New York Fed didn’t write a check; it convened one. A consortium of major dealers injected capital into the fund in exchange for control, creating time for an orderly unwind. It was controversial—private losses socialized among private counterparties under public supervision—but it avoided a fire sale. The lesson was less about heroics than choreography: when trades are crowded and financing is brittle, coordination is a public good.
Why the lesson endures
The cast changes; the script doesn’t. Crowding in safe-haven basis trades, leveraged volatility harvesting, and liability-driven strategies all rhyme with convergence funded by short-term borrowing. The 2022 gilt turmoil showed how quickly collateral calls can turn a tranquil, “hedged” portfolio into forced liquidation. Risk is a triangle—leverage, liquidity, correlation—and you only discover you’ve been standing on one leg when it gives way.
What to do differently? Less faith in past correlations, more humility about liquidation timelines, and funding that matches the true half-life of the trade. If you need time to be right, you must be able to buy time.
Our take
LTCM’s real legacy isn’t a morality play about hubris; it’s a field manual. In tranquil markets, the cheapest risk to ignore is liquidity risk. The next panic will have new acronyms, but it will punish the same assumption: that the market will give you time to exit. It never does.
