Your independent source for Harvard news since 1898 |

Risk without Reward

 
When an investment model melted down.

On October 14, 1997, Robert C. Merton (now McArthur University Professor at the Business School) and Myron Scholes were awarded the Nobel Prize in Economic Science. They had devised a method for pricing options, which are financial contracts that were designed to reduce risk, but that often produced disastrous losses for investors. On the day the prize was announced, Roger Lowenstein writes in When Genius Failed, “Merton, who was teaching a class at Harvard, got a three-minute ovation from his students. He humbly warned, however, ‘It’s wrong to believe that you can eliminate risk just because you can measure it.’”

That admonition was prophetic. A little less than a year later, Long-Term Capital Management (LTCM), the flashy Greenwich, Connecticut, firm that both Merton and Scholes had joined five years earlier—putting their options theories into practice—was hit with losses so severe the Federal Reserve Bank of New York was forced to call an unprecedented emergency meeting of Wall Street’s top powerbrokers, who ponied up $3.65 billion to prevent a worldwide financial catastrophe. What went wrong is the subject not only of a compelling and entertaining book by Lowenstein, a former reporter for the Wall Street Journal and author of an excellent biography of Warren Buffett, but also of the far less accessible Inventing Money by Nicholas Dunbar, who earned a master’s in earth and planetary sciences at Harvard and is now a financial editor in London. “Finance is often poetically just,” writes Lowenstein. “It punishes the reckless with special fervor.” That fact makes the story of LTCM particularly juicy. It goes like this: Young man from South Chicago works as caddy, gets graduate degree, joins Wall Street firm, is pushed out after a scandal, then starts his own firm, earns billions, gets caught in a market downdraft, nearly plunges world economy into chaos, loses practically all, rebuilds. And there’s even a moral.

John Meriwether joined Salomon Brothers in 1974, an ambitious young man with a business degree from the University of Chicago. He eventually became head of the investment bank’s arbitrage department, where he concentrated on making money off small anomalies between the prices of different assets, mainly bonds. Meriwether became a legend on Wall Street and later starred in Liar’s Poker, the bestseller by Michael Lewis that features him daring Salomon’s CEO, John Gutfreund, to play a single hand of poker for $10 million. Lowenstein sees the Lewis story as apocryphal. Meriwether was never reckless. He was, in fact, the “priest of the calculated gamble….cautious to a fault.” He knew he needed an edge to succeed in bond trading, and he found one:

Why not hire traders who were smarter? Traders who would treat markets as an intellectual discipline, as opposed to the folkloric, unscientific Neanderthals who traded from their bellies. Academia was teeming with nerdy mathematicians who had been publishing unintelligible dissertations on markets for years….That would be his edge.

So, in 1983, Meriwether hired Eric Rosenfeld, who was making $30,000 a year as an assistant professor at the Harvard Business School. He hired a Harvard colleague of Rosenfeld’s named William Krasker—plus Gregory Hawkins and Lawrence Hilibrand, with Ph.D.s from MIT, and others. The strategy worked. Meriwether’s arbitrage group became immensely profitable, and many of the academics became immensely rich. In 1989, for example, Hilibrand took home $23 million. But in 1991, things fell apart. A bond trader named Paul Mozer, who was working for Meriwether as head of the government-bond desk, confessed to his boss that he had submitted false bids to the U.S. Treasury in bond auctions. Neither Meriwether nor Gutfreund took Mozer’s revelations seriously enough, and a scandal ensued that cost all three their jobs.

But the scandal also presented Meriwether with an opportunity. The next year, with the help of Merrill Lynch & Company, he began raising more than $1 billion to start his own firm, enticing Hilibrand, Rosenfeld, and others to join him. He even bagged David Mullins, vice chairman of the Federal Reserve Board and a former professor at Harvard Business School, and sealed his coup by recruiting two of the biggest names in financial economics: Merton, who was teaching at Harvard, and Scholes, who had studied at the University of Chicago under Eugene Fama and Nobel laureate Merton Miller ’44, developers of the “efficient market hypothesis”—the idea that prices today are “correct” because they reflect all known information and that their movements in the future are essentially unknown, a “random walk.” But, like many random-walkers, Scholes believed, paradoxically, that he could beat the market.

And, of course, beating the market was the game that Meriwether and his cohorts were playing. LTCM concentrated on what are called “relative value” trades. Imagine, for example, that Ford and General Motors are essentially the same—in their profitability, riskiness, and prospects. Assume that both earned $2 per share last year. But also assume that today, Ford’s price in the stock market is $20 and GM’s is $30. Something is screwy here in the relationship between Ford and GM. So an investor might buy a share of Ford stock and sell short a share of GM. (In selling short, you borrow the share from another investor, selling it immediately for cash and promising to return the share later. Your hope is that the price of the share will fall, so you can buy it back more cheaply.) With this trade, you expect that the natural, logical state of affairs will soon come to pass—that the prices of Ford and GM will converge. You really don’t care if the market rises or falls—or whether Ford stands still and GM drops, or Ford rises while GM stands still. All that counts is the convergence—Ford’s price in comparison with GM’s.

The folks at LTCM focused their trades mainly on bonds, which have a purity that stocks lack. When the computer models developed by the academics signaled an anomaly between different kinds of bonds, the traders would swing into action, making bets that the relationship would eventually get back to normal—as it usually did. For instance, the computers might find that a plain-vanilla 10-year Treasury bond was yielding 6 percent at the same time that a certain sliced-and-diced mortgage security was yielding 9 percent. That “spread,” the model might say, is too large: the proper, historical relationship might be a difference of two percentage points. So the firm would speculate that the rates on the T-bond would rise and/or the rates on the mortgage security would fall. Such arbitrage plays were not very risky, but, since the spreads between the prices of the bonds were typically tiny, LTCM “would have to multiply its bet many, many times by borrowing” in order to make a lot of money. Such financial leverage boosts profits, but it also threatens to boost losses.

Moreover, there was a particular danger in this sort of investing. When a normal investor puts, say, $1 million into the stock of a company, only to find that it is not meeting its earnings projections, or a product has failed, or a key manager is leaving, the investor realizes the mistake and bails out. But LTCM’s principals had no mistakes to realize. Their computers recognized a situation of mispricing that would certainly be corrected. As Lowenstein writes in the very first sentence of his first chapter: “If there was one article of faith that John Meriwether discovered at Salomon Brothers, it was to ride your losses until they turned into gains.” In other words, if you have enough capital to stay in the game, to make the interest payments on your borrowings and even to increase your position, then, in the end, you will win the bet.

Nicholas Dunbar uses the gambling exploits of Giacomo Casanova in 1754 to explain the phenomenon. Casanova, playing faro (a game similar to roulette) in the casinos of Venice, used a technique called the “martingale.” Assume you are betting on red. Begin with a bet of $1. If black comes up and you lose, then double the wager to $2 the next time. If you lose again, double the bet again. Finally, on the fifth bet, with $16 at stake, suppose you win. You are paid $16 (and get your $16 bet back), and your total bets over the preceding four rounds come to $15 ($1 + $2 + $4 + $8 = $15). So you walk away a $1 winner.

The allure of the martingale as a betting system lies in its promise that you will win eventually, as long as you have enough capital to keep doubling your stake. Unfortunately, you might just go bankrupt before that happens. For example, in the unlikely event that the 26 black cards in the pack appeared in a row, Casanova would have needed to stake 67 million gold coins to win just one.

(I am well acquainted with the dangers of the martingale myself, having tried a version of it at roulette in Las Vegas. Betting on black, I lost eight times in a row. Even though I began with a bet of just $20, my get-ahead bet the next round would have had to be $5,120. I folded.)

As Dunbar puts it, the “small print of martingales” says that you are assured of a profit only if you have unlimited capital. LTCM had a lot of money, but its resources weren’t infinite. In April 1998, after four years of success, the firm’s own capital base totalled $4.9 billion. In fact, Meriwether decided that LCTM had too much capital and forced his institutional investors, against their vigorous protests, to take back $2.7 billion. The timing could not have been worse. Just a few months later, LCTM was stuck with some shockingly poor trades—hurt especially by the Russian debt default—and Meriwether had to scramble for cash to hold his positions. But firms like UBS, the huge Swiss bank, which had been begging to become an investor earlier, now turned him down. In August alone, LTCM lost 45 percent of its value and was bleeding to death at the rate of $500 million a week. One disastrous trade was a “risk arbitrage” bet that the telecommunications company Tellabs would successfully complete a takeover of its rival Ciena. Instead, the deal crumbled, and LTCM lost $200 million. By September 22, when William McDonough, president of the New York Fed, called together Wall Street’s barons (among them: David Komansky of Merrill, Jon Corzine of Goldman Sachs, Thomas Labrecque of Chase Manahattan, James Cayne of Bear Stearns, and Sandy Weill, who was about to merge his own empire with Citicorp), LTCM had debt totalling about $100 billion with just $773 million of its own equity left. In the end, a bailout was arranged—a fascinating story told well by Lowenstein and glossed over by Dunbar, whose main interest is the intricacies of derivatives trading—and a rolling disaster was averted. Without the extraordinary infusion, LTCM’s defaults could easily have caused other institutions to default, and on and on.

The firm’s principals saw their own investment in the fund plummet from $1.8 billion to $27 million. “Larry Hilibrand, the most cocksure of traders, who had previously been worth close to half a billion dollars, awoke to discover that he was broke,” writes Lowenstein. In fact, he was $24 million in debt. Rosenfeld, the former assistant professor at Harvard, “auctioned off caseloads of the wine collection he had once so lovingly assembled.” But the partners, while no longer superrich, weren’t reduced to penury. They kept their homes and lived to fight another day. In November 1999, JWM Partners (a firm that included Meriwether, Hilibrand, Rosenfeld, and others from LCTM) began circulating a document for “Relative Opportunity Value Fund II”—where leverage would be kept, it was promised, to just 15 to 1 and discipline would be tighter. Within a month, the fund had raised $250 million, and Meriwether “was off and running again,” showing that, indeed, there are third, as well as second, acts to American lives.

Despite Meriwether’s survival—and perhaps future prosperity—the Long-Term Capital Management story is a tragedy. Some very smart and talented people, including Merton and Scholes, gave their time and brainpower to an enterprise that lost billions of dollars and produced little more than heartache. There is an important lesson here, however, and it is not simply that the gods, as usual, punish Promethean hubris.

A few years ago, I came to the conclusion that there are two kinds of investors: outsmarters and partakers. Although many outsmarters acknowledge the efficiency of the market, they think they can beat it. Few can. One, certainly, is Warren Buffett, who became chief executive officer of Salomon after the Treasury bond scandal. He wrote in 1988: “Observing that the market was frequently efficient, [certain economists] went on to conclude incorrectly that the market was always efficient. The difference between the propositions is night and day.” Yes, but the temptation to try to beat the market has to be tempered with discipline that outsmarters rarely apply—for example, a regime for cutting losses and an aversion to leverage. In my financial writing, I tell investors not to borrow money to invest in the stock market—ever.

Partakers are merely along for the ride. They buy stocks, especially, to share in the success of individual businesses and in general economic growth. The record here is wonderful: Since 1926 a basket of stocks such as the Standard & Poor’s 500 Index, has returned, including both dividends and price appreciation, an annual average of 12 percent. Imagine a casino where the house advantage is one-eighth of all the cash that’s bet and where you are the house.

Despite LTCM’s fabulous record before its fall in 1998, the firm had done only slightly better than investors who had put their money into public index mutual funds that mimicked the S&P 500. And Long-Term Capital Management is not the only high-profile investment pool to fall on hard times. The celebrated George Soros has dismissed three of the top managers of his hedge funds after poor performance, and Julian Robertson actually closed his own hedge fund down following big losses. Markets for stocks and bonds—not to mention real estate and rare coins—will always have their ups and downs, but, in the end, partakers have the better strategy, doubling their money, on average, every six years, even without the help of John Meriwether and his band of geniuses.