In the summer of 1998, a team of some of the world's brightest financial minds lost nearly four billion dollars in capital within a few months, putting the global financial system at real risk. The team included two Nobel laureates in economics for their work on option pricing theory, a former Federal Reserve vice chairman, and one of Wall Street's most respected bond investors. These people believed they had solved the risk mathematically. They were wrong.
Long-Term Capital Management (LTCM) was founded in 1994 by John Meriwether, the legendary bond investor at Salomon Brothers. He was quickly joined by Myron Scholes and Robert Merton, both of whom would win the Nobel Prize three years later for their Black-Scholes model of option pricing. With the addition of former Federal Reserve Vice Chairman David Mullins, LTCM gained unprecedented credibility in Wall Street history.
The fund's strategy was based on an approach called convergence trading. The idea was that when a temporary price difference formed between two very similar securities, that difference would eventually close. By buying the cheap and short-selling the expensive, it was possible to make a profit that was virtually risk-free, regardless of market direction – at least that's what the model predicted. The problem was that these price differences were usually very small.
The early years showed how powerful this strategy worked, with the fund delivering annual returns of 20%, 43%, 41%, and 17%. Investors lined up to join the fund, despite the minimum investment of ten million dollars and the three-year no-withdrawal requirement. Confidence was so high that at the end of 1997, the partners decided the fund had grown too large compared to the opportunities and returned $2.7 billion to investors, but kept the size of the positions the same. This decision increased the leverage ratio from approximately 25x to 28x, while capital decreased and risk remained the same.
On August 17, 1998, Russia announced it would not pay its debt denominated in its own currency and devalued the ruble. This event didn't directly affect LTCM's positions, but its secondary effects were severe. Global investors simultaneously fled risky assets and sought refuge in safe havens—a flight so intense that no model had predicted it.
LTCM's strategy was based on the assumption that assets that historically traded close together would eventually converge. But during this crisis, the opposite happened; the price spreads they were betting on widened to unprecedented levels, divergence instead of convergence. Positions that should have been market-neutral suddenly began moving in the same direction, and in the wrong direction. These highly leveraged positions put the fund in a difficult position. In a single month, almost half of the fund's value had evaporated.
Banks began demanding collateral, but LTCM didn't have enough. The positions were so large that trying to close them by selling would only drive prices down further, increasing losses. By mid-September, the fund's capital had fallen from approximately $4.7 billion to $600 million, a erosion of over 90 percent. LTCM's list of lenders and traders encompassed virtually all of Wall Street, transforming the fund's situation from a single company's problem into a systemic threat. The New York Federal Reserve Bank intervened, bringing together 14 major banks to organize a $3.6 billion bailout package without using public funds. The consortium seized 90% of the fund, effectively wiping out the shares of the original investors, including the Nobel laureate founders. The fund was formally liquidated in 2000.
The lesson LTCM left behind is not that mathematical models themselves are worthless, but that they operate within the limitations of historical data and become unreliable in events that transcend those limitations.
The fund's convergence strategy was based on the assumption that markets behave rationally and that relationships would return to normal over time. But in times of crisis, markets don't behave rationally; fear prevails, and assets that historically appeared independent suddenly begin to move in the same direction. The protection provided by diversification disappears precisely when it is most needed.
Leverage acted as a multiplier in this scenario. The amount of debt required to turn a small price difference into a significant profit meant that a move in the opposite direction would result in a large loss. And once concentrated positions started to incur losses, trying to close them by selling only pushed the market further against them, making the fund a prisoner of its own size.
Perhaps the most striking lesson is that intelligence and reputation don't protect against this trap. LTCM's founders were the brightest names in their fields, two of them Nobel laureates for their work. But just as Newton lived in the South Sea bubble, even the most superior analytical skills cannot account for the crowded behavior of the market and its irrational actions during times of crisis.
LTCM's clearest echo in the crypto world is seen in Three Arrows Capital, which faced serious financial difficulties in 2022. This similarity isn't just our observation; analysts at the time made the same comparison. Crypto asset manager Chris Zheng pointed out that both 3AC and LTCM were excessively leveraged funds, noting that despite the "geniuses" behind LTCM, including two Nobel Prize-winning economists, the fund had leveraged itself up to 25 times.
Founded in 2012 by Su Zhu and Kyle Davies, Three Arrows Capital had become one of the most respected and influential investment funds in the crypto world. As of March 2022, it was managing approximately $10 billion in assets. Like LTCM, 3AC had gained a certain aura of respectability and immunity in the market.
The fund's strategy also followed a familiar pattern: holding interconnected positions with high leverage. 3AC built a large, highly leveraged position in LUNA, the sister token of the algorithmic stablecoin TerraUSD, borrowing from counterparty funds. This position was worth approximately $560 million at its peak and was almost completely wiped out when the price crashed. But the real vulnerability was that 3AC was exposed to many interdependent positions simultaneously with such high leverage that it couldn't bear this loss alone. When the crypto market as a whole declined simultaneously, a correlation breakdown similar to what LTCM experienced occurred; the risk, thought to be spread across different assets, all moved in the same direction during the crisis.
When 3AC failed to meet margin calls, its positions were forcibly liquidated, contributing to a 48% decline in the total crypto market capitalization in May and June 2022. The fund was unable to pay its debt to Voyager Digital, which exceeded $670 million. As finance professor Nik Bhatia stated at the time, 3AC was seen as the "adult in the room" in the market, but ultimately, like LTCM, it became a prisoner of its own size and interconnectedness. The difference was that LTCM had a central bank behind it, and a bailout package was organized to protect the system. 3AC lacked such a mechanism; the fund was directly driven into bankruptcy, a process that affected many institutions, from Celsius Network to Voyager Digital.
This parallel tells us something important: even the most respected and experienced players in the market can fall into the same fragility when they hold sufficiently leveraged and concentrated positions. To put this into a self-applicable question, before trusting the past returns of a strategy or fund, you need to ask three things: How much leverage is used to generate that return? Are the positions in the portfolio truly independent of each other, or will they all shift in the same direction in a crisis? And if things go wrong, can those positions be quickly downsized without significantly impacting the market? The real failure of both LTCM and 3AC wasn't a prediction error; it was their failure to ask themselves these three questions in time.
