Fortune's Formula (33 page)

G
AMBLING RAN IN
John Meriwether’s family. As a boy, he learned blackjack from his grandmother and was permitted to place bets at the racetrack and on sports. Always looking for an edge, John would check the weather forecast for wind velocity at Wrigley Field and use that to decide how to bet on Cubs games.

Born in Chicago in 1947, Meriwether was a bright, mathematically inclined kid educated by priests. He attended Northwestern University on a scholarship for golf caddies. Meriwether taught a year of high school math, then got a business degree at the University of Chicago. His first job out of business school was trading government bonds at Salomon Brothers in New York. Pre-Milken, bonds were pretty boring, and government bonds the most boring of all. Meriwether found much to keep his interest. New York City came close to defaulting on its bonds. The bond market panicked, and all government bonds took a hit. Meriwether reasoned that New York’s financial woes were irrelevant to the credit of municipalities elsewhere. He therefore bought government bonds at bargain prices, expecting to see them rebound. When they did, Meriwether suddenly looked like a genius.

In 1977 Meriwether started Salomon’s Arbitrage Group. This was bond arbitrage, and it became the firm’s biggest profit maker. A shy man, Meriwether gained a share of fame in Michael Lewis’s 1989 memoir,
Liar’s Poker
. It was Meriwether who bluffed his way out of a high-stakes game of liar’s poker with Salomon chairman John Gutfreund. In fact Meriwether’s tastes ran more to horse racing. He boarded racehorses on his 68-acre estate in North Salem, New York, and at Belmont racetrack. To hedge the bets he made every working day, Meriwether kept a set of rosary beads in his briefcase.

Meriwether left Salomon Brothers during a scandal-driven shake-up in which Meriwether was, it appears, innocent of wrong-doing. He decided to start a hedge fund.

It was a good time to do that. Princeton-Newport’s long run had convinced many wealthy investors of the possibility of beating the market while containing risk scientifically. Scores of new hedge funds were started in the early 1990s. Of all these new funds, Meriwether’s Long-Term Capital Management was to become the best known.

Ed Thorp first heard of Meriwether’s fund through a mutual friend. The friend knew some of the people who were writing software for the new fund. “It’s gonna be a great investment,” Thorp was told, “and for ten million dollars you can get into it.”

Like most of the new group of fund managers, Meriwether promised better-than-market returns through science and software. Meriwether did not himself possess a first-rate mathematical mind. Instead, he recruited the top academic talent. No finance professor was more respected than Robert C. Merton. Merton had consulted for Salomon Brothers, so Meriwether already knew him. He agreed to come on board. Meriwether’s other great coup was recruiting Myron Scholes. As journalist Roger Lowenstein said, that was like putting Michael Jordan and Muhammad Ali on the same team.

Thorp decided not to put any of his money in the fund. He was concerned that Merton and Scholes, brilliant as they were, had little experience investing other people’s money. It didn’t help that Merton was second only to Samuelson as a critic of the Kelly criterion. Thorp also had heard that Meriwether was a “martingale man.” “The general chatter was that he was a high roller, and it wasn’t clear that the size of his bets were justified,” Thorp recalled. “The story was that if he got in the hole, if things went against him, he’d bet more. If things still went against him, he’d bet more.”

L
ONG
-T
ERM
C
APITAL
M
ANAGEMENT
(LTCM) was the first fund to raise a billion dollars. It did this by projecting a 30 percent annual return net of fees—better than even Princeton-Newport had done. LTCM’s partners charged 25 percent of profits (rather than the usual 20) plus 1 percent of invested assets per year. The 25 percent fee was a deal-breaker for the trustees of the Rockefeller Foundation, who decided they did not have that kind of money to burn. Other wealthy and in some cases glamorous investors didn’t seem to mind. Harvard University, which had had money in Princeton-Newport, put some in LTCM. LTCM investors ranged from Merrill Lynch to the Kuwaiti state pension fund, the Bank of China (the People’s Republic of China) to Hollywood agent Mike Ovitz.

LTCM hit the ground running in March 1994. By the end of the calendar year, its investors had racked up a 20 percent return after fees.

In 1995 the return was 43 percent after fees. The next year, it was 41 percent.

These were good years for the stock market, too. The S&P 500 sprinted 34 percent in 1995 and another 20 percent in 1996. LTCM was 9 and 21 points ahead of these already rich returns.

Zillionaires were begging Meriwether to take their money. It didn’t do them much good. The fund was closed to new investment. Some people were so desperate to own a piece of Wall Street’s hottest property that gray-market LTCM shares sold for about 10 percent above asset value.

In 1997 LTCM made a 17 percent return after fees. That is superb by any reasonable standards, but 1997 was not an especially reasonable year. The S&P 500 shot up 31 percent.

By October 1997, the fund’s capital had mushroomed from $1.2 billion to $7.1 billion. After the lukewarm 1997 showing, Meriwether decided to return the money of some of his investors in the hope of boosting future performance.
Fortune
magazine reported that “many went kicking and screaming, and at least one protested so angrily that LTCM allowed him to stay onboard.” By the end of December, the fund’s capital was down to $4.7 billion.

LTCM’s trading strategies were secret. It is startling how much money Meriwether was able to raise while disclosing almost nothing about what he intended to do with it. One thing was disclosed. LTCM used a lot of leverage. That was how they were able to obtain better-than-market returns from a nearly efficient market.

Adherents of the efficient market hypothesis generally allow that small mispricings can arise and persist because they’re too small for anyone to bother with. The transaction costs would eat up any profit. LTCM’s strategy was to use leverage to multiply these small profit opportunities to the point where they were big enough to matter.

Paul Samuelson said he had doubts about LTCM when he first heard of it. It appeared that the fund was placing a lot of faith in the random-walk model. The leverage left little room for any misfit of theory and reality. Myron Scholes, however, threw himself into his new role as hedge fund pitchman. In presentations to potential investors, Scholes said they were vacuuming up nickels no one else could see. As he said this he would snatch an imaginary nickel out of the air.

The core of LTCM’s business was convergence trades, the long-and-short hedged trades that many other hedge funds ran. LTCM favored government bonds, the area where Meriwether had such success at Salomon Brothers. One type of trade was known as “on the run, off the run.” A brand-new thirty-year U.S. Treasury bond is said to be “on the run.” It costs $10,000 and is good for a full thirty years of semiannual interest payments and repayment of the original $10,000 at the end of that time. An older bond, with some of the interest payments already made, is “off the run.” The market price of an older bond depends on a lot of things, most important the current interest rate. Meriwether had found that off-the-run bonds were usually a bargain compared to new bonds. As with cars, people pay an irrational premium for the shiny new models. Once you drive a car off the lot—once a bond becomes last year’s model—it takes a hit in price.

Meriwether’s people bought older bonds and sold short brand-new bonds. Then they waited for the prices of the two bonds to converge. In time, the new bonds would become “old” and move closer in price to the old bonds. When this happened, they could realize a minuscule profit. It took leverage to inflate this to the kind of sky-high returns investors were expecting.

In 1996 one of LTCM’s investors spoke by phone with several of the partners. The investor asked exactly how much return they were making on the dollar. The answer was 67 basis points. The return was 0.67 percent.

The LTCM investor also learned that the fund was using leverage of about thirty times. For every dollar of investor money, the fund borrowed $29 more. This meant that the fund achieved thirty times the profit. After paying off the lender, it had thirty times the 0.67 percent profit on the original dollar, or 20 percent.

Despite their value in selling the fund, Merton and Scholes played modest roles in the day-to-day decision-making. The fund’s investors surely understood that neither great scholar sat at a desk barking trades into a phone. It is less clear whether investors believed that the two famous economists had created the fund’s detailed financial models (they had not).

One LTCM road-show presentation was held at the insurance company Conseco in Indianapolis. Andrew Chow, a Conseco derivatives trader, interrupted Scholes. “There aren’t that many opportunities,” Chow objected. “You can’t make that kind of money in Treasury markets.”

Scholes snapped: “You’re the reason—because of fools like you we can.”

Traveling with Scholes were some Merrill Lynch people who were experts in raising investment funds. They advanced the expert opinion that Scholes should apologize. Another LTCM partner, Greg Hawkins, doubled up with laughter. Conseco did not invest in LTCM.

T
HIRTY TIMES LEVERAGE
sounds like a lot. On many of the fund’s positions, the leverage was higher than that—effectively infinite. As much as possible, LTCM tried to operate like an infomercial real estate guru who walks into a city without a dollar in his pocket, buys real estate on credit, and makes a positive cash flow—all with “no money down.” When you don’t have to put up any money, you make a return on investment of
infinity
. Or
negative infinity
, when things go wrong.

When a trader buys on credit, the securities are themselves collateral. The bank or other lender has a right to repossess the securities and sell them in the event of a serious loss. Because securities can drop in value quickly, this right might not be enough to protect the lender’s interests. The trader is therefore normally required to put up a down payment called a “haircut.” It works much like the down payment on a house. When you put up 20 percent of the purchase price of a house, the bank can be reasonably confident that it will be able to sell the property for at least the amount of the 80 percent mortgage. The bank will not end up with a loss.

Haircuts are also required when selling short. A short-seller can, theoretically, lose an unlimited amount of money. Collateral is required to protect against that, too. Since all long-short hedged trades involve selling short, collateral requirements are an integral part of the game, even when leverage is not used.

The size of the haircut depends on security law, the type of securities being bought or sold, and the trader’s credit and negotiating skills. Investment banks routinely borrow 99 percent of the cost when purchasing treasuries. This is one hundred times leverage, and it is not necessarily considered reckless.

It was a point of pride with LTCM’s people that they paid
zero
haircuts on many of their deals. This is testimony to the fund management’s ability to romance creditors.

Zero haircuts do not change the facts of life. LTCM was simply in the position of a gambler who goes to a casino where the pit boss extends him unlimited credit.

You might take the position that with unlimited credit it’s irrelevant how much money you’ve got in your pocket or bank account. The more you bet, the more you win. Therefore any wager, no matter how high, is justified.

This argument might hold water in the case of a casino with literally unlimited credit and bet size. You wouldn’t even need an edge in a casino like that. Martingale would work.

In the real world, “unlimited credit” is a figure of speech. What the pit boss means is roughly: “I know this guy, and he’s okay. Don’t bother running a credit check. Let him start gambling right away. Of course, check with me if he wants a lot of money or is losing heavily.”

The pit boss has no intention of lending more money than the casino can readily collect, should it come to that. So it was with LTCM’s banks. A home buyer puts up a down payment only once. A hedge fund’s collateral requirements are constantly adjusted. Each day, the value of the account is recomputed at current prices (“marked to market”) and collateral figured from that. When the value of the account rises, the trader is allowed to withdraw collateral from the margin account. When the value of the account falls, the bank demands that more collateral be put in the account. Should the trader be unable to do this, the bank may sell some of the account to raise collateral.

LTCM had a sophisticated system for handling collateral requirements. When a particular trade showed a profit, less collateral was required. This money could be withdrawn and wired to meet the collateral requirements of a losing trade.

One term for gambler’s ruin among traders is
blowing up
. To blow up an account is to lose everything in high-risk trades with borrowed money. A stellar career can end in a few miserable days or hours. Blown-up traders are Wall Street’s undead. They have failed at the most important judgment a trader can make, namely how much money to commit to a risky trade.

LTCM’s people were well aware that multiplying profits through leverage also multiplies risk of ruin. They told investors that they had risk under control through their financial engineering. LTCM used a sophisticated form of the industry standard risk reporting system, VaR or “Value at Risk.”

After the Black Monday crash of 1987, investment bank J. P. Morgan became concerned with getting a handle on risk. Derivatives, interest rate swaps, and repurchase agreements had changed the financial landscape so much that it was no longer a simple thing for a bank executive (much less a client) to understand what risks the people in the firm were taking. Morgan’s management wanted an executive summary. It would be a number or numbers (just not
too
many numbers) that executives could look at every morning. Looking at the numbers would reassure the execs that the bank was not assuming too much risk.

Two of Morgan’s analysts, Til Guldimann and Jacques Longerstaey, devised Value at Risk. The concept is as simple as it can be. Compute how much a portfolio stands to lose within a given time frame, and with what probability. A VaR report might say that there is a 1-in-20 chance that a portfolio will lose $1.64 million or more in the next day of trading.

Want more numbers? VaR’s got as many numbers as you want. Make a spreadsheet. The cells of the spreadsheet are the possible losses, for different time periods or various thresholds of likelihood. Throw in color charts, print it out on the good paper, and hand it to the client.

Morgan’s management liked the idea. Practically everyone else did, too. Other banks began hiring “risk managers” to prepare daily VaR reports. The Basel Committee on Banking Supervision—head-quartered in the city of the Bernoullis—endorsed VaR as a means of determining capital requirements for banks.

VaR migrated downstream to private investment managers. By calculating VaR, a money manager shows the client that she is serious about managing risk. She’s got it all down in numbers, and numbers are good, right? When the investor scans the figures and raises no fuss, he has implicitly signed off on those risks. Should something terrible happen later on, the money manager can always pull out the VaR report, point to cell D18, the 5 percent risk of a 37 percent loss. As a ritual between portfolio manager and client, calculating VaR is not such a bad idea in a litigious society where many well-off people don’t know much math.

In October 1994, LTCM sent its investors a document comparing projected returns to risks. One reported factoid: In order to make a 25 percent annual return, the fund would have to assume a 1 percent chance of losing 20 percent or more of the fund’s value in a year. A 20-percent-or-more loss was the worst case considered.

The chapter on Value at Risk in the popular finance textbook
Paul Wilmott Introduces Quantitative Finance
begins with a cartoon of the author shrugging. “I’ve got a bad feeling about this…” he says.

Wilmott isn’t alone. There are at least two problems with VaR. One is that it plays into the mystique of numbers. The consumer of VaR reports is led to believe that the numbers are reliable because smart people have gone to a lot of trouble to work them out. The numbers are only as good as the assumptions underlying them. When the assumptions are bad, VaR is a case of garbage in, garbage out.

The other problem exists even when the assumptions and numbers are right. VaR does not tell you everything you ought to know about risk. It sidesteps the two questions that are central to John Kelly’s analysis: What level of risk will lead to the highest long-run return? What is the chance of losing
everything
? (A VaR report
could
address the second question. In practice it rarely does. Who wants to freak out the client with scare talk?)

Every Tuesday at LTCM’s Greenwich, Connecticut, headquarters, the fund management held a meeting on risk. These meetings centered on printouts from a top secret program called the “Risk Aggregator.” Most of the fund’s employees never saw these reports, and apparently none of the investors did.

The Risk Aggregator was capable of diverse what-if calculations. “We spent time thinking about what happens if there’s a magnitude ten earthquake in Tokyo, what happens if there’s a 35 percent one-day crash in the U.S. stock market,” said LTCM’s David Modest. “We certainly spent hours and hours thinking about it.” According to Modest, the worst-case outcome the model ever projected was a loss of $2.5 billion, or about half the fund’s capital. In the end, people shrugged and went back to their trading.