When Genius Failed: The Rise and Fall of Long-Term Capital Management

by Roger Lowenstein

In short, Merton assumed a perfect, risk-free arbitrage. This assumption may approximate real markets when they are calm—but only then.

Merton’s theories were seductive not because they were mostly wrong but because they were so nearly, or so nearly often, right.

Merton was prey to this very trap. His “continuous-time finance” seemed to wrap the financial universe in a tidy ball. On paper it solved, or pointed to the solutions for, virtually every problem in finance: how to value junk bonds, how much to pay for deposit insurance, you name it. His theories seemed to behold the elegance and order he had always craved. “Not all that is beautiful in science need also be practical,” he wrote with satisfaction. “But here we have both.”6

The beauty of cards is that the universe is known; there are fifty-two cards in a deck, and only fifty-two. Life insurance is a bit different: since new people are always being added to the universe, actuaries rely on samplings. They aren’t perfect, but they work, because the sample of people is very large and mortality rates change only very slowly. But in markets, we are never sure that the sample is complete. The universe of all trades looked one way throughout the 1920s and another way after the Great Depression. The pattern changed again during the inflationary 1970s, yet again in the effervescent 1990s. After which of these periods was the picture “normal,” and how do we know that the next new period won’t change the story again?

Eugene Fama, Scholes’s thesis adviser, wondered what his old student was up to as well. Fama had all the admiration in the world for the option formula as a model. But trusting people’s money to such models was different. In the early 1960s, Fama had written his thesis on the price movements of the thirty Dow Jones Industrial Average stocks and discovered a remarkable pattern: for every stock, there were many more days of extreme price movements than would occur in a normal distribution. Fama’s stocks were like a world in which most people were average height but every twentieth person was either a giant or a dwarf. There were too many extreme readings, or “outliers,” that random distribution could not explain. As Fama put it: If the population of price changes is strictly normal, on the average for any stock . . . an observation more than five standard deviations from the mean should be observed about once every 7,000 years. In fact such observations seem to occur about once every three to four years.9 Graphically, Fama’s stocks depicted a bell curve with the ends swollen, representing too many incidents at the extremes (known as the “tails”).

By the time Long-Term was formed, it was well documented that virtually all financial assets behaved like the stocks that Fama had studied.

A key condition of random events is that each new flip is independent of the previous one. The coin doesn’t remember that it landed on tails three times in a row; the odds on the fourth flip are still fifty-fifty. But markets have memories. Sometimes a trend will continue just because traders expect (or fear) that it will. Investors may slavishly follow the trend for no other reason than that they think enough others will do likewise. Such momentum trading has nothing to do with logically appraising securities; it doesn’t fit the ideal of rational investors in efficient markets. But it’s human. After three bad “flips” in the market, the fourth flip may no longer be completely random. Some traders may have taken losses and be forced to sell; other investors, looking over their shoulders, may panic and decide to beat them to it—as happened with Treasury bonds during Long-Term’s inaugural spring.

This human spirit was totally absent from Merton’s markets. Though he acknowledged that his models represented an ideal state, his writing was suffused with the notion of perfect prices and rational investors. Even his prose was wooden and dry, as if he were trying to boil off life’s emotional content and reduce it to a controlled abstraction.

He had the disease of perfect belief, which makes compromise impossible.

The traders at Long-Term knew the models were imperfect, of course. “We know this doesn’t work by rote,” said Robert Stavis, the former member of the Arbitrage Group at Salomon. “But this is the best model we have. You look at the old-timers who went with their gut. You had this model, you had these numbers, and in the end you thought they were a lot more powerful than a guy’s gut.”

In a document dated October 1994, the same month as Meriwether’s letter, Morgan admitted that markets did not appear to be random or independent in the way of coin flips and that volatilities “are themselves quite volatile.” Nonetheless, the firm, and its imitators all over Wall Street, continued to use Value-at-Risk. Morgan couldn’t find any “persuasive alternatives,”18 the bank explained—as if that would make up for its shortcomings.

A central tenet of the partners’ philosophy was that markets were steadily getting more efficient, more liquid, more “continuous”—more as Merton had envisioned them.

Thanks to Italy and the IO mortgage trade in particular, Long-Term in 1995 earned an astounding 59 percent before fees and 43 percent after. A little over half of the gains came from Europe. During its first two years, in which Long-Term had earned a remarkable $1.6 billion, Italy contributed an estimated $600 million in profits. All told, it was the fastest, most impressive start by any fund ever.

To William F. Sharpe, a Nobel Prize–winning economist and an adviser to one of Long-Term’s investors, the returns seemed surreally smooth. “We distinctly asked, ‘What’s the risk?’ ” Sharpe recalled. “Myron [Scholes] said, ‘Well, our goal is to get the risk level [the volatility] of the S&P 500.’ He said, ‘We’re having trouble getting it that big.’

Including the money from new investors, the firm’s equity capital had, in less than two years, virtually tripled, to a total of $3.6 billion. Long-Term’s assets had also grown, to the extraordinary sum of $102 billion. Thus, at the end of 1995, it was leveraged 28 to 1.

As Long-Term’s strategies grew more diverse, the fund felt comfortable taking bigger positions. It focused on its portfolio as a whole, and thus it was willing to pile bigger trades onto each of its “risk” dollars back in the vault. The theory, of course, was that the likelihood of many leveraged trades collapsing simultaneously was slim, just as an insurer does not expect that all of its clients will file claims simultaneously. Indeed, Long-Term thought of itself in exactly those terms: as an insurer of financial risk.

The question the partners could have asked themselves was, how hard had they stretched to make that 59 percent, and what sort of claims might they face after a storm?

The partners strode to the plate and rapped hit after hit, day after day. More confident than ever, they lengthened their swing, which is to say they leveraged further. By the spring of 1996, Long-Term had an astounding $140 billion in assets, thirty times its underlying capital. Though still unknown to 99 percent of Americans, Long-Term was two and a half times as big as Fidelity Magellan, the largest mutual fund, and four times the size of the next largest hedge fund.¹

As the business relationship grew closer, Long-Term began to push Merrill Lynch for more financing. Leahy, who managed the relationship for Long-Term, would say, “If you want us to trade, you got to give us more balance sheet.” Gradually, Merrill opened the spigot. By 1996, it was providing $6.5 billion of repo financing (roughly equal to Merrill’s total equity) and proportionately more on the derivative side.

“Everybody was enamored with their intellect,” Dunleavy, the salesman, recalled. “It was like Kennedy’s inner circle—Camelot! They had the best and the brightest.”

By 1996, Long-Term had grown to well over a hundred employees and the partners were building immense fortunes. All along, they had been deferring the 25 percent of the outside investors’ profits that they took as fees, thus leaving the money in the fund, where, untouched by the tax man, it could compound all the faster. As their third year drew to a close, the partners collectively had a stake in the fund of $1.4 billion, nine times their initial $150 million investment. It was an incredible fortune to have made in so little time—and all from bond spreads!

J.M., Hilibrand, Rosenfeld, and the rest left every nickel on the table. When they needed spending money, they simply redeemed a bit of the fund, which they treated like a personal checking account.

The pursuit of money may have been central to their lives, but as is often the case, it went far beyond any conceivable lifestyle needs. The money was a scorecard, a proof of their superlative trading skills. For Merton and Scholes, it added a worldly validation to their academic successes.

The arbitrageurs were blind to the value of contributions that didn’t drop straight to the bottom line; only financial points counted.

The fund earned 57 percent in 1996 (41 percent after the partners’ fees) thanks to leveraged spread trades on Japanese convertibles, junk bonds, interest rate swaps, and—again—Italian bonds. Also, when French bonds began to trade above German bonds (implying, curiously, that France bore less inflation risk than Germany), the partners cleverly, and successfully, bet on Germany to make a relative comeback. Their total profits in 1996 were an astounding $2.1 billion.15 To put this number into perspective, this small band of traders, analysts, and researchers, unknown to the general public and employed in the most arcane and esoteric of businesses, earned more that year than McDonald’s did selling hamburgers all over the world, more than Merrill Lynch, Disney, Xerox, American Express, Sears, Nike, Lucent, or Gillette—among the best-run companies and best-known brands in American business.

Despite its scientific pretensions, economics still remains more of an art than a science.

Lured by the scent of the fantastic profits being earned in Greenwich, other banks were reaching for the same nickels as Long-Term. Inevitably, they whittled away the very spreads that had attracted them; thus do free markets punish success.

Although pricing a bond can largely be reduced to math, valuing a stock is far more subjective. Wall Street (and academe) had devised many a formula to forecast the market, but none, no matter how esoteric or rigorous, had worked. Over the short run, stocks are subject to the whim of often emotional traders. Over the long run, they vary with business performance, which is subject to great uncertainty and is notoriously hard to forecast. It requires judgment—not merely math—of the sort that no computer has ever mastered.

Given such uncertainties, most players limited paired-share trades to moderate size. But with its coffers burgeoning, Long-Term was developing a sense of proportion all its own; like a man who pays for dinner with hundred-dollar bills and never asks for change, it had lost the habit of moderation.

But the size of Haghani’s position was stunning. Long-Term bet $2.3 billion—half of it long on Shell, the other half short on Royal Dutch—without, of course, any assurance that the spread would contract. In practical terms, a position that large was totally illiquid.

Still, there is no explaining the size of the Royal Dutch/Shell trade, other than to accept that Haghani was beginning to believe in his own invincibility.

The partners debated hiring a risk-arbitrage specialist but didn’t. Meanwhile, Hilibrand bought deal stock after deal stock. And he confidently bought them in very big size, despite the growing discomfort of a half dozen of his partners. Scholes and Merton argued that merger arbitrage—particularly on such a scale—was excessively risky for the obvious reason that Long-Term was playing in a field in which it had absolutely no expertise. Meriwether and his traders knew the bond world inside out. In merger arbitrage, J.M. had no edge; indeed, it was Long-Term’s rivals who had the advantage.

J.M.’s personal loyalties weighed on portfolio considerations, a serious flaw in a risk manager.

“Our business looks very attractive next to their business, where they trade to make a couple of basis points,” Tisch noted. “The only trouble is, if you’re wrong on a government bond the spread may change by a half a point. If you’re wrong on risk arb, you can lose half your position.” In short, the reason that deal spreads were so much wider than bond spreads was that you could lose a lot more money on merger arbitrage. Tisch left feeling that Long-Term didn’t know what it was doing. It was both inexperienced and highly leveraged—a potentially explosive combination.

Apart from CBS, Long-Term’s biggest position was MCI Communications, which in 1996 agreed to be acquired by British Telecommunications. Both CBS and MCI had to clear regulatory thickets, and each deal dragged on for longer than expected. In the case of CBS, Long-Term continued to snap up shares even when CBS stock had inched to within 62 cents of the deal price. Long-Term’s leverage on the trade was 20 to 1—without its having any specialized knowledge of the merger. “You’re picking up nickels in front of bulldozers,” a friendly money manager warned Rosenfeld, alluding to the risk that one deal or another would collapse. Eventually the CBS merger was completed, which only emboldened Long-Term to do more such trades. Its portfolio would grow to thirty deals.

What mattered to J.M. and company was that they could make a huge investment in CBS with no money down and without having to make all of the usual disclosures. And despite Reg T, it was perfectly legal. The Fed, after all, merely restricted loans toward the purchase of stocks. Long-Term wasn’t purchasing anything; it was making side bets on the direction of stocks—which amounted to the same thing.

By 1995, when Meriwether’s traders were happily ensconced at Long-Term, the group had a total derivative book worth $650 billion. Within two years, the total doubled, to an astounding $1.25 trillion. Given the opaque nature of Long-Term’s (and everyone else’s) disclosures, it was impossible to pinpoint the fund’s derivative risks according to specific trades. And since many of its contracts were hedges that tended to cancel each other out, it was impossible to calculate Long-Term’s true economic exposure. One could say only that it appeared to be growing very quickly—as were exposures up and down Wall Street. Almost imperceptibly, the Street had bought into a massive faith game, in which each bank had become knitted to its neighbor through a web of contractual obligations requiring little or no down payment.

In the words of Nicholas Brady, the former Treasury secretary, “Every time there’s been a fire, these guys [derivative traders] have been around it.”6

A bit of liquidity greases the wheels of markets; what Greenspan overlooked is that with too much liquidity, the market is apt to skid off the tracks.11 Too much trading encourages speculation, and no market, no matter how liquid, can accommodate all potential sellers when the day of reckoning comes.

Save for the Fed, the only ones who could restrain derivative lending were the banks. But Wall Street never polices itself in good times.

No borrower had to account for its total exposure; no lender asked.

Swiss banks were cloistered bureaucracies in which only Swiss nationals could hope for employment and promotion.

But there is nothing like success to blind one to the possibility of failure. At the very time when the outlets for their narrow skills were closing, the partners tossed off their innate caution, which had served them so well, like an old suit.

But even that was not enough. To leverage even further, LTCM, the partners’ management company, borrowed a total of $100 million from a trio of banks—Chase, Fleet Bank, and Crédit Lyonnais of France—which money the partners plowed right back into the fund.16 Their hunger to turn millions into billions knew no bounds, nor did it recognize any risks. For men who prided themselves on being disciples of reason, their drive to live on the edge seemed inexplicable, unless they believed that becoming the richest would certify them as also being the smartest.

The UBS warrant raised $1 billion of equity for Long-Term at the worst possible moment—when the fund was struggling to find places to invest the money it already had and when it was fighting off more and more competitors in arbitrage. In the first half of 1997, it earned only 13 percent before fees—still impressive, though well below its prior average. Its leverage (again, not counting derivatives) fell from 30 to 1 to 20 to 1, evidence that opportunities had grown scarce. A group of skeptical partners, including Scholes, were growing increasingly uncomfortable with the fund’s portfolio.

Mahathir had some of the right culprits but the wrong crime. Foreigners could hardly be blamed for deserting a listing ship; their sin was having provided the too-easy short-term credit that had launched the ship at flank speed. Comforted by Treasury Secretary Robert Rubin’s bailout of Mexico, Western banks and investors had shoveled money at Asia, fueling speculative investments and underwriting the region’s corrupt behavior. Rubin had paved the way by urging Asian governments to lift controls on capital and thus to allow the money to pour in—despite the area’s appalling lack of corporate disclosure and regulatory oversight. “The simple fact is that very sophisticated banks loaned to Indonesian companies, without any real knowledge of their financial condition,” noted James Wolfensohn, president of the World Bank.21 Though Western investors had repeatedly made the same mistake—today romanticizing the developing world, tomorrow coming to rue their naïveté—the bankers never learned. In 1996, $93 billion of foreign capital had flooded Indonesia, Malaysia, the Philippines, South Korea, and Thailand even as their economies were slowing. Now that same money was draining out, unnerving markets around the world. From Long-Term’s point of view, both Asia’s uphill and downhill slopes were instructive. Easy credit, which hedge funds also enjoyed, had led to ill-advised and excessive investments. Then, when the tables had turned, Greenspan’s much-vaunted liquidity was nowhere to be f...

Long-Term’s plan was to return, at the end of 1997, all profits on money invested during 1994, its first year, and to return all money (principal and profits) invested after that date. It excluded the partners and employees and partially excluded some big strategic investors such as the Bank of Taiwan. Jimmy Cayne, chief executive of Bear Stearns, Long-Term’s clearing broker, got an exception, too. Others demanded to stay in, but Long-Term turned them down. They were naturally angry that Meriwether and his boys, supposedly charged with watching out for their investors’ interests, were giving themselves preferential treatment. “This is outrageous!” one investor thundered to Merrill Lynch. “We were there when they needed us.”

In retrospect, the outside investors’ pleas to stay in the fund would seem ironic, and the forced redemption of their money would come to seem a godsend. It was the partners who would suffer the most. But without the benefit of hindsight, the partners’ obsessive pursuit of wealth carried more than a whiff of greed. Now that they had the scale to operate worldwide, they had no interest in managing money for others and largely froze them out. They added an unusually self-serving touch, tacking on fees, for the first time, to the bonus money invested by some of their own employees! The partners justified the forced redemption as a sensible response to diminished prospects. Myron Scholes said it was ironic that Long-Term was giving money back at the same time that newcomers were piling into the field.23 The implication was that Long-Term was prudently scaling back in the face of shrinking spreads. Yet the fund was not going to shrink by one iota. Returning capital only reduced the equity supporting its assets, which weren’t, in fact, shrinking at all. Indeed, at a time when opportunities were drying up and the portfolio was bloated with 7,600 positions, the fund was defiantly and ill-advisedly leveraging itself further, like an Icarus chasing the sun.24 Moreover, by forcing outsiders to sell, the partners were, once again, increasing their personal leverage in the fund. The personal debt, on top of the leverage in the management company and the debt in the fund itself, made fo...

“Sandy spent nine billion dollars to get a piece of paper from Warren Buffett saying what a great investor he was. He was running around showing it to people like a kid in a candy store.”

Their responses were 180 degrees apart. By merging Salomon with a more diversified partner, Buffett had diluted the Salomon shareholders’ interests in arbitrage and in the rest of Salomon’s business. Now they would own pieces of a much larger Travelers. Long-Term’s partners had made the exact opposite decision: to a great extent, they had bought out their partners, thus redoubling their bets on arbitrage. The irony was that Buffett had converted a chronic loser into $9 billion while Long-Term’s partners had converted a consistent winner into a giant—and still unrealized—bet on the future.