On the evening of 23 September 1998, the heads of Wall Street's biggest banks were summoned to a conference room at the Federal Reserve Bank of New York. Bitter rivals who normally met only across trading screens were asked, in effect, to pass the hat. The object of their reluctant charity was not a country or a bank but a single secretive hedge fund in Greenwich, Connecticut — a fund whose partners included two men who had collected the Nobel Prize in economics barely ten months earlier. The firm was Long-Term Capital Management, and its fall gave Roger Lowenstein's classic account its perfect title: When Genius Failed.
LTCM was founded in 1994 by John Meriwether, the revered former head of bond arbitrage at Salomon Brothers, who assembled what was instantly nicknamed a dream team: star traders, a former vice-chairman of the Federal Reserve in David Mullins, and the academics Myron Scholes and Robert Merton, whose work on option pricing — alongside the late Fischer Black — would win the 1997 Nobel prize while they sat in the fund's partnership. Investors needed $10 million minimum and had to lock their money away; they queued anyway. The fund returned roughly 40 per cent a year in 1995 and 1996, after fees.
The strategy was not wild speculation but its opposite. LTCM hunted tiny mispricings between nearly identical securities — a freshly issued US Treasury bond against a six-month-old one, Italian bonds against German ones as European interest rates converged — and bet that the gaps would close. Each trade earned pennies, so the pennies were multiplied by borrowing on a scale no fund had attempted: in early 1998 about $4.7 billion of capital supported well over $100 billion of positions, leverage of roughly 25 to one, with derivative contracts on top whose notional value ran past a trillion dollars. The models said the diversified bets were nearly independent of one another, and that a loss big enough to destroy the firm was, for practical purposes, impossible.
August 1998: the correlations go to one
On 17 August 1998, Russia devalued the rouble and defaulted on its domestic debt. The default itself was survivable; the panic was not. Investors everywhere fled anything risky or illiquid for the safest assets on earth, and every convergence trade on LTCM's books widened at once. Positions in Danish mortgages, Japanese bonds, US swap spreads and merger stocks — supposedly unrelated — all bled together, because in a panic the only trade that matters is the rush for the exit. The fund lost $1.9 billion in August alone, around 44 per cent of its capital, and could not unwind its vast positions without moving prices further against itself. By late September capital was down to a few hundred million dollars while the balance sheet had barely shrunk, driving effective leverage past 100 to one. Losses ultimately totalled about $4.6 billion.
Rescue nearly came from an unexpected quarter: a consortium of Warren Buffett's Berkshire Hathaway, Goldman Sachs and AIG offered to buy the portfolio outright, valuing the partners' stake at a humiliating $250 million, with a deadline measured in minutes. Legal structure and the ticking clock killed the deal. That left the New York Fed, whose president, William McDonough, feared that a fire-sale liquidation of a trillion-dollar book would tear through markets already raw from Asia and Russia. He brokered — but did not fund — a private solution: fourteen banks injected about $3.6 billion in return for 90 per cent of the fund. Bear Stearns, LTCM's clearing broker, famously refused to contribute; a decade later, when Bear itself was failing, Wall Street's memory proved long.
The dress rehearsal
No taxpayer money was spent, the banks were repaid as the positions were wound down by 2000, and Meriwether promptly launched a new fund — JWM Associates, which itself closed its flagship fund in the wake of the 2008 crisis, as if to underline the lesson. Yet the episode reads, in hindsight, like a dress rehearsal for 2008: extreme leverage hidden in plain sight, models that treated markets as better behaved than they are, every major bank exposed to the same counterparty, and a central bank concluding that some institutions are too entangled to be allowed a disorderly death. Critics argued that the rescue taught Wall Street a dangerous lesson — that someone would always convene the meeting. The Nobel laureates' equations were not wrong about normal days; they were silent about the days that matter.
Genius had not failed at mathematics. It had failed at imagining a world rude enough to ignore it.
Quiz nuggets
- LTCM was founded in 1994 by John Meriwether, former head of bond arbitrage at Salomon Brothers.
- Partners Myron Scholes and Robert Merton won the 1997 Nobel Prize in economics while at the fund.
- Russia's rouble devaluation and default of 17 August 1998 triggered losses totalling about $4.6 billion.
- On 23 September 1998 the New York Fed brokered a $3.6 billion rescue by a consortium of fourteen banks.
- Bear Stearns refused to join the 1998 rescue; in 2008 it collapsed and was sold to JPMorgan Chase.