When Genius Failed: The Rise and Fall of Long Term Capital Management. Roger Lowenstein

to sell, they were pushing spreads wider, creating just the gaps that Meriwether was hoping to exploit. “The unusually high volatility in the bond markets … has generally been associated with a widening of spreads,” he chirped in a—for him—unusually revealing letter to investors. “This widening has created further opportunities to add to LTCP’s convergence and relative-value trading positions.”³ After two flat months, Long-Term rose 7 percent in May, beginning a stretch of heady profits. It would hardly have occurred to Meriwether that Long-Term would ever switch places with some of those panicked, overleveraged hedge funds. But the bond debacle of 1994, which unfolded during Long-Term’s very first months, merited Long-Term’s close attention.

      Commentators began to see a new connectedness in international bond markets. The Wall Street Journal observed that “implosions in seemingly unrelated markets were reverberating in the U.S. Treasury bond market.”⁴ Such disparate developments as a slide in European bonds, news of trading losses at Bankers Trust, the collapse of Askin Capital Management, a hedge fund that had specialized in mortgage trades, and the assassination of Mexico’s leading presidential contender all accentuated the slide in U.S. Treasurys that had begun with Greenspan’s modest adjustment.

      Suddenly markets were more closely linked—a development with pivotal significance for Long-Term. It meant that a trend in one market could spread to the next. An isolated slump could become a generalized rout. With derivatives, which could be custom-tailored to any market of one’s fancy, it was a snap for a speculator in New York to take a flier on Japan or for one in Amsterdam to gamble on Brazil—raising the prospect that trouble on one front would leach into the next. For traders tethered to electronic screens, the distinction between markets—say, between mortgages in America and government loans in France—almost ceased to exist. They were all points on a continuum of risk, stitched together by derivatives. With traders scrambling to pay back debts, Neal Soss, an economist at Credit Suisse First Boston, explained to the Journal, “You don’t sell what you should. You sell what you can.” By leveraging one security, investors had potentially given up control of all of their others. This verity is well worth remembering: the securities might be unrelated, but the same investors owned them, implicitly linking them in times of stress. And when armies of financial soldiers were involved in the same securities, borders shrank. The very concept of safety through diversification—the basis of Long-Term’s own security—would merit rethinking.

      Steinhardt blamed his losses on a sudden evaporation of “liquidity,” a term that would be on Long-Term’s lips in years to come.⁵ But “liquidity” is a straw man. Whenever markets plunge, investors are stunned to find that there are not enough buyers to go around. As Keynes observed, there cannot be “liquidity” for the community as a whole.⁶ The mistake is in thinking that markets have a duty to stay liquid or that buyers will always be present to accommodate sellers. The real culprit in 1994 was leverage. If you aren’t in debt, you can’t go broke and can’t be made to sell, in which case “liquidity” is irrelevant. But a leveraged firm may be forced to sell, lest fast-accumulating losses put it out of business. Leverage always gives rise to this same brutal dynamic, and its dangers cannot be stressed too often.

      Long-Term was doubly fortunate: spreads widened before it invested much of its capital, and once opportunities did arise, Long-Term was one of a very few firms in a position to exploit the general distress. And its trades were good trades. They weren’t risk-free; they weren’t so good that the fund could leverage indiscriminately. But by and large, they were intelligent and opportunistic. Long-Term started to make money on them almost immediately.

      One of its first trades involved the same thirty-year Treasury bond. Treasurys (of all durations) are, of course, issued by the U.S. government to finance the federal budget. Some $170 billion of them trade each day, and they are considered the least risky investments in the world. But a funny thing happens to thirty-year Treasurys six months or so after they are issued: investors stuff them into safes and drawers for long-term keeping. With fewer left in circulation, the bonds become harder to trade. Meanwhile, the Treasury issues a new thirty-year bond, which now has its day in the sun. On Wall Street, the older bond, which has about 29½ years left to mature, is known as off the run; the shiny new model is on the run. Being less liquid, the off-the-run bond is considered less desirable. It begins to trade at a slight discount (that is, you can purchase it for a little less, or at what amounts to a slightly higher interest yield). As arbitrageurs would say, a spread opens.

      In 1994, Long-Term noticed that this spread was unusually wide. The February 1993 issue was trading at a yield of 7.36 percent. The bond issued six months later, in August, was yielding only 7.24 percent, or 12 basis points, less. Every Tuesday, Long-Term’s partners held a risk-management meeting, and at one of the early meetings, several proposed that they bet on this 12-point gap to narrow. It wasn’t enough to say, “One bond is cheaper, one bond is dearer.” The professors needed to know why a spread existed, which might shed light on the paramount issue of whether it was likely to persist or even to widen. In this case, the spread seemed almost silly. After all, the U.S. government is no less likely to pay off a bond that matures in 29½ years than it is one that expires in thirty. But some institutions were so timid, so bureaucratic, that they refused to own anything but the most liquid paper. Long-Term believed that many opportunities arose from market distortions created by the sometimes arbitrary demands of institutions.⁷ The latter were willing to pay a premium for on-the-run paper, and Long-Term’s partners, who had often done this trade at Salomon, happily collected it. They called it a “snap trade,” because the two bonds usually snapped together after only a few months. In effect, Long-Term would be collecting a fee for its willingness to own a less liquid bond.

      “A lot of our trades were liquidity-providing,” Rosenfeld noted. “We were buying the stuff that everyone wanted to sell.” It apparently did not occur to Rosenfeld that since Long-Term tended to buy the less liquid security in every market, its assets were not entirely independent of one another, the way one dice roll is independent of the next. Indeed, its assets would be susceptible to falling in unison if a time came when, literally, “everyone” wanted to sell.

      Twelve basis points is a tiny spread; ordinarily, it wouldn’t be worth the trouble. The price difference was only $15.80 for each pair of $1,000 bonds. Even if the spread narrowed two thirds of the way, say in a few months’ time, Long-Term would earn only $10, or 1 percent, on those $1,000 bonds. But what if, using leverage, that tiny spread could be multiplied? What if, indeed! With such a strategy in mind, Long-Term bought $1 billion of the cheaper, off-the-run bonds. It also sold $1 billion of the more expensive, on-the-run Treasurys. This was a staggering sum. Right off the bat, the partners were risking all of Long-Term’s capital! To be sure, they weren’t likely to lose very much of it. Since they were buying one bond and selling another, they were betting only that the bonds would converge, and, as noted, bond spreads vary much less than bonds themselves do. The price of your home could crash, but if it does, the price of your neighbor’s house will likely crash as well. Of course, there was some risk that the spread could widen, at least for a brief period. If two bonds traded at a 12-point spread, who could say that the spread wouldn’t go to 14 points—or, in a time of extreme stress, to 20 points?

      Long-Term, with trademark precision, calculated that owning one bond and shorting another was one twenty-fifth as risky as owning either bond outright.⁸ Thus, it reckoned that it could prudently leverage this long/short arbitrage twenty-five times. This multiplied its potential for profit but—as we have seen—also its potential for loss. In any case, borrow it did. It paid for the cheaper, off-the-run bonds with money that it borrowed from a Wall Street bank, or from several banks. And the other bonds, the ones it sold short, it obtained through a loan, as well.

      Actually, the transaction was more involved, though it was among the simplest in Long-Term’s repertoire. No sooner did Long-Term buy the off-the-run