From Harvard To Hedge Funds

Text from Bloomberg Markets, April 2004, pp 48-54.

From Harvard To Hedge Funds

Six years after Long-Term Capital Management collapsed, college professors are back running hedge funds. Some find the marketplace more invigorating than the classroom.

By Michael Peltz

On Friday, Jan. 9, computer science professor John Moody was stuck inside his yellow Victorian home in the West Hills section of Portland, Oregon. He was trapped there by the biggest snow and ice storm to hit the U.S. Pacific Northwest in a decade. Moody, who runs a hedge fund when he's not doing academic research, was also having one of his best trading weeks in a year. ``We were up about 4 percent,'' says Moody, who has a Ph.D. in theoretical physics from Princeton University. ``We managed to ride the currency wave further, made money in metals and even got back into cattle.''

Moody is among a group of leading academics who are putting their theories into practice in the world of hedge funds - - lightly regulated investment pools that tap university endowments, pension funds and wealthy individuals for capital. Some of those professors, like Moody, have science backgrounds. Others, such as Andrew Lo, a professor at the MIT Sloan School of Management, and John Campbell, a professor in the economics department at Harvard University, are trained in finance. Still others, like Richard Clarida, a Columbia University economics professor and former assistant secretary of the U.S. Treasury under President George W. Bush, bring a macroeconomic perspective to their work with hedge funds.
Some of the most successful hedge fund managers have come from academia, including former Columbia computer science associate professor David Shaw and Wharton School finance professor Sanford Grossman. The dean of academics who have become hedge fund managers is James Simons, 65, former chairman of the mathematics department at the State University of New York at Stony Brook and a winner of the prestigious Oswald Veblen prize in geometry, which is given by the American Mathematical Society every five years for outstanding research. In 1982, Simons founded hedge fund company Renaissance Technologies Corp. in East Setauket on New York's Long Island. Since 1988, his flagship $5.2 billion Medallion Fund, which focuses on futures trading, is up almost 35 percent a year after fees.

One reason professors are flocking to the hedge fund world is the chance to get rich. A finance professor at a top university typically makes $150,000 to $225,000 a year in base salary.``Their salary could go up substantially at a hedge fund,'' says Georges Holzberger, a partner at executive search firm Sextant Search Partners LLC. ``The head of quantitative research at a major fund can make well into the seven figures.''

Holzberger says the potential rewards are even higher for an academic who starts a fund. Managers typically charge an annual fee of 1.5 percent of assets under management and get 20 percent of the profits. ``I'm sure everybody is getting involved in it partly for the money and partly to see if you can really put your money where your mouth is,'' says Richard Thaler, a professor of behavioral economics and decision-making at the University of Chicago's Graduate School of Business. Thaler, 58, is a principal at Fuller & Thaler Asset Management Inc. The firm manages $2.6 billion, including $600 million by using what's known as a long/short hedge fund strategy. In such a fund, a manager hedges its long positions -- equities bought in the hope their price will rise -- by selling short. That means borrowing securities, often from an investment bank, in the hope of buying them back later at a lower price.

Daniel LeVine, senior managing director of quantitative research at Citadel Investment Group LLC in Chicago, says he expects hedge fund managers looking for ways to maintain their edge to go after more academics in the future. ``As markets become more liquid, more electronic and more continuously traded, a higher degree of analytics is necessary in order to compete,'' says LeVine, who has a Ph.D. in applied mathematics from Brown University. At Citadel, which manages more than $9 billion in hedge funds, a majority of the 50 people in the quantitative analysis group have Ph.D.'s, including a dozen who have taught.

Academia is a good training ground for hedge funds, says Thomas Schneeweis, 56, a finance professor at the University of Massachusetts, Amherst. ``An academic background offers some insight into why and when certain investment strategies make money,'' he says. Professors also need to be good at presenting ideas, a useful skill when working with traders to implement a strategy or when explaining that strategy to investors. Even so, says Schneeweis, academic professionals with little market experience have no particular edge in the investing world. ``Just because an academic is involved doesn't mean you can change the rules of nature,'' he says. ``All investment strategies make money in certain markets and lose money in other ones.''

Professors who try to maintain ties to academia face added challenges, because top universities expect professors to publish research regularly, says Frank Meyer, 60, who recently retired as chairman of Chicago-based Glenwood Capital Investments LLC, which manages about $5 billion in funds that invest in hedge funds. ``It's hard to serve both masters,'' says Meyer. ``At a hedge fund, you want to keep your ideas proprietary, because that's how you maintain your edge.''

The collapse of Long-Term Capital Management LP in 1998 underscored the risks and rewards of hedge funds for academics. Former Salomon Brothers Inc. Vice Chairman John Meriwether set up LTCM in 1993, reassembling most of his old Salomon arbitrage trading group, which included several Ph.D.'s. He also recruited Myron Scholes, a finance professor who at the time was cohead of the fixed-income derivatives group at Salomon, and Robert Merton, a finance professor at Harvard Business School. Merton had worked with Scholes and Fischer Black at MIT in the early 1970s to develop the Black-Scholes model for pricing options. ``LTCM had arguably the best academic finance department in the country,'' MIT's Lo says. From 1994 to 1996, Meriwether and his team delivered annual returns after fees of 20 percent, 43 percent and 41 percent. The firm used its computer models and extensive databases to identify pairs of financial assets whose values had temporarily moved apart -- and that LTCM expected to come together over time regardless of whether the overall market went up or down.

LTCM would buy the underpriced asset, sell short the overpriced one and then wait for the prices to converge. LTCM's convergence strategy required extensive leverage, because the price differences between the two assets tended to be small. For some trades, the firm borrowed as much as 40 to 50 times its capital, says Donald MacKenzie, a University of Edinburgh sociology professor who has published two papers on LTCM. In 1997, Scholes and Merton were awarded the Nobel prize in economics for their work on options theory. Their acclaim offered little comfort to LTCM investors in August 1998, when Russia partially defaulted on $40 billion of its ruble-denominated debt. Even though Russian bonds accounted for only a small percentage of the LTCM portfolio, the Russian default caused investors in markets around the globe to seek safer investments. That, MacKenzie says, proved to be LTCM's undoing. ``Even the trades that ought to have gone in LTCM's favor went against them, because others who had been imitating them were trying to unload similar portfolios,'' he says.

LTCM lost 44 percent of its capital in August 1998. Meriwether and his band of academics had to be bailed out by a consortium of 14 U.S. and European banks put together by then Federal Reserve Bank of New York President William McDonough. In September 1998, the banks injected $3.6 billion of new capital into LTCM in return for 90 percent of its equity and oversaw the orderly unwinding of its portfolio.

The collapse of LTCM had a sobering effect on academics who were thinking of going into money management. ``You couldn't start a company at that time and just say, `We're smart academics. Trust us. Give us your money,''' says Harvard's John Campbell, cofounder of money management firm Arrowstreet Capital LP. Campbell, 45, was born in London and grew up in Oxford, where his father taught American history. Campbell's paternal grandfather taught theology at Cambridge University.``Being a senior academic is a very nice lifestyle,'' says Campbell, stretching out his 6-foot-2-inch frame in a chair in his wood-paneled office at Harvard. ``The freedom and job security one has is unbeatable.''

In 1999, Peter Rathjens, a former doctoral student of Campbell's who had become chief investment officer at PanAgora Asset Management Inc., and Bruce Clarke, PanAgora's president, decided to form their own money management firm. They asked Campbell to join them. Harvard restricts its professors' outside business activities to no more than one day a week, or 20 percent of their time. Clarke and Rathjens agreed with Campbell that he should remain a full-time professor. ``John's a great academic, so why change that?'' asks Rathjens. ``Moreover, he would have better access to the broader flow of ideas about capital markets if he stayed in academia.''

Campbell, Clarke and Rathjens founded Arrowstreet in July 1999 to develop mathematical and statistical models for investing in global equity markets. Clarke is president and runs the now 26- person shop. Rathjens, chief investment officer, oversees the portfolios at Arrowstreet, which has about $4.4 billion in assets. Campbell, a managing partner, heads up research.
Campbell has led the development of Arrowstreet's quantitative models, which try to identify and exploit behavioral and informational errors by investors. The models look for instances when investors blunder by overreacting to news, following the herd or failing to account for new information. To find them, they analyze three factors: price momentum, earnings indicators such as changes in analyst forecasts and value measures like price-earnings ratios.

Arrowstreet looks at about 4,600 stocks around the world, which it groups into some 200 baskets by country and industry type. Late last year, the firm started buying Norwegian energy company Norsk Hydro ASA, partly because it met one of the firm's chief value criteria: Its dividend yield, about 2.4 percent, was larger than the dividend yields of other natural resource companies. The stock also had good momentum; that is, its price was going up. ``We feel that indicates a herd is forming behind the stock, and it's likely to outperform for some period,'' Rathjens says. By last December, when Arrowstreet was building its position, Norsk Hydro's share price had risen about 9 percent from its September 2003 low. As of Feb. 24, the shares were up another 26 percent.

Two years ago, Campbell and his partners launched a long/short hedge fund strategy. ``For every dollar of stocks invested long, there's one dollar of stocks invested short,'' says Campbell, who generally works two mornings a week at Arrowstreet, which is just a five- minute walk from his Harvard office. Arrowstreet's long/short strategy uses the same return forecasting and risk models as the firm's long-only strategies. Rathjens says one of the firm's best shorts was Newell Rubbermaid Inc. In April 2003, the firm's models picked up that the herd was starting to turn against the U.S. consumer products maker. Arrowstreet's traders began shorting Newell at about $30 a share and closed out their position in October in the low $20s. Rathjens declines to disclose how much profit the firm made.

The initial performance of Arrowstreet's hedge fund has been mixed. Campbell says the firm hopes to deliver, before fees, a return of 10 percentage points above the interest rate clients could get investing in a short-term instrument such as 90-day Treasury bills. The fund surpassed its objective in 2002 and fell short last year; Rathjens won't reveal specific performance numbers. Like Harvard, MIT follows the one-day-a-week rule. Compliance is left largely to the honor system; faculty members are expected to inform their department heads of all outside professional activities and are required to fill out an annual form detailing those activities. Professors don't have to disclose how much money they made from them. Richard Schmalensee, dean of the MIT Sloan School, says Andrew Lo has been a model citizen when it comes to accounting for his time. Lo, 43, teaches finance and investments at Sloan. He's also director of the MIT Laboratory for Financial Engineering, a research center that uses mathematical, statistical and computational models to study financial markets.

Lo founded AlphaSimplex Group LLC, a Cambridge, Massachusetts-based investment management firm, in March 1999. Today, AlphaSimplex employs 15 people -- mostly Ph.D.'s like Lo -- and uses two different long/short hedge fund strategies: one that invests only in U.S. equities and another that invests globally in stocks, bonds and currencies. ``For me, getting into the practical side of things was a natural progression,'' Lo says, sitting in his fourth-floor office at MIT overlooking the Charles River. ``My research has always been about investments.''

In the summer of 1999, Lo took a sabbatical from MIT, signed a lease for office space in nearby Kendall Square and began looking for financial backing. In October, he met Donald Sussman, chairman and founder of Paloma Partners LP, which manages about $2 billion in hedge funds in Greenwich, Connecticut. Lo says they hit it off immediately, and on Nov. 1, 1999, he signed a deal to launch a fund for Paloma that would invest long and short in U.S. equities by using AlphaSimplex's quantitative models.

Lo says it took him two years to rewrite the software that he would need at AlphaSimplex for research, development and trading. ``Even though the software was pretty similar to what I was using at MIT, I didn't want to have any intellectual property conflicts,'' says Lo, who was born in Hong Kong and immigrated with his family to New York at the age of five. ``What I do at MIT with MIT funding belongs to MIT.''

Lo's title at AlphaSimplex is chief scientific officer. About half of the firm's 6,000-square-foot Cambridge office is devoted to computers, which Lo and his team use to scour markets for price anomalies. They look at fundamental factors such as dividend yields, as well as behavioral signals such as reactions to earnings surprises. Unlike the strategy at LTCM, which was designed to find long-term arbitrage opportunities, AlphaSimplex's models, Lo says, try to capitalize on short- and medium-term imbalances in the supply and demand of securities.
AlphaSimplex started trading its long/short fund for Paloma in April 2001. Although neither Lo nor Paloma will comment on performance, the fund has done well enough to attract interest in a second one from large investors such as Global Asset Management, a subsidiary of UBS AG that manages $12.8 billion in hedge funds. In December 2003, AlphaSimplex launched a global hedge fund that plans to invest in stocks, bonds and currencies in 10 countries. The $250 million fund closed to new investors on Feb. 1.

Some career academics who take the plunge into hedge funds have long dabbled in the markets. John Moody says he began trading futures contracts -- agreements to buy or sell a commodity or financial instrument at a set price and date in the future -- in the mid-1980s as a postdoctoral student at the Institute for Theoretical Physics at the University of California, Santa Barbara. He continued to trade for his own account at Yale University, where he taught computer science from 1987 to 1992. One of his best trades was a bet on palladium futures in 1989 following news that two Utah scientists had successfully used palladium rods to create cold nuclear fusion. He says he rode the price up amid the resulting speculative furor and then shorted the futures a day before it became clear that the results of the experiment couldn't be replicated -- making a profit of almost half a year’s salary.

Moody, who has published more than 65 research papers on computer science, physics and finance, left Yale in 1992 to teach at the Oregon Graduate Institute in Portland, where he had grown up. This past autumn, he decided to take a break from full-time academia to focus on JE Moody & Co., a hedge fund he started almost three years ago. He left OGI and joined the Algorithms Group at the International Computer Science Institute, a nonprofit research center affiliated with the University of California, Berkeley.

One of Moody's specialties is machine learning, which involves the development of algorithms, or series of instructions, that enable a computer to learn to solve a problem through trial and error. Programmers at International Business Machines Corp. used machine learning to help teach its Deep Blue chess program, which beat world chess champion Garry Kasparov in 1997. ``The system utilized machine learning because brute-force search methods weren't sufficient,'' says Moody. ``In machine learning, the computer doesn't try to memorize everything; rather, it tries to distill qualities of decisions that give good results.''

Moody says that just as machine learning algorithms can teach a computer to play chess, they can enable computer programs to learn to make trading decisions based on direct experience. ``A good trader must learn to recognize and act on opportunities, to trim positions when markets are uncertain or volatile and, most of all, to avoid the risk of ruin,'' he says.
Moody calls his computerized trading programs RoboTraders, and he teaches them by means of simulations. ``Without risking capital, a RoboTrader can acquire 10 to 20 years of simulated trading experience across a range of markets and market conditions,'' says Moody, who has hired three researchers and a human trader to execute the trades. ``To graduate and go live, we require that the strategies learned by a RoboTrader are easily interpretable by us humans and that they make sense.''

Last October, Moody bought futures on the Australian dollar, which at the time had risen about 10 percent from its September lows. A RoboTrader recommended buying Australian futures as the best way to play the weakening U.S. dollar, because short-term interest rates were higher in Australia than in the U.S. and the Reserve Bank of Australia was expected to raise them further, which it did in November and December. In mid-January, the RoboTrader said it was time to sell, as the dollar had started to show short-term strength against all major currencies. On Jan. 15, Moody closed out his futures position, thereby making 10 percent on the trade.
Moody says his computerized trading system is right about 60 percent of the time and that the winning trades tend to make much more money than the losing ones lose. In 2003, his fund outperformed the benchmark Barclay CTA Index of 359 commodity trading advisers, which was up 8.6 percent. He says the fund made money in energy, currencies, platinum, silver and copper.

Professors who make the transition from academia to a hedge fund often confront a more hectic lifestyle. Columbia's Richard Clarida, 46, says one of the biggest changes for him was the noise. ``It can get pretty loud,'' says Clarida, chief economic strategist at New York-based hedge fund manager Clinton Group Inc., which invests primarily in fixed-income securities. Although he has a private office, Clarida says, he spends 99 percent of his time at a desk in the middle of Clinton's 50-person trading floor, sitting next to the firm's founder and president, George Hall.
On the morning of Jan. 13, the Clinton trading floor was less noisy than usual because many of the traders were listening to Alan Greenspan. They wanted to hear whether the Federal Reserve chairman, who was speaking in Berlin, would express any worries over the near-record U.S. current account deficit. Greenspan didn't, sparking a rally in the dollar. ``Had he expressed serious concern about the ability of the U.S. to finance its current account deficit, the dollar might well have broken the other way,'' Clarida says.

Clarida's academic background has equipped him to interpret the Fed's actions. His research has focused on global macroeconomics, and he has published several papers that modeled Fed monetary policy under Greenspan. He also served as assistant secretary of the Treasury for economic policy in the Bush administration from February 2002 to May 2003. As the Treasury's chief economist, Clarida says, he had several private briefings with Greenspan.
Clarida has multiple roles at Clinton. He's there to consult with traders and analysts when news breaks or the government releases economic figures. He's developing new trading strategies using some of his academic research on the link between exchange rates, interest rates and yield curves. He has even gotten involved on the marketing side of the business, meeting with clients when they come to New York and giving them his take on the economy.
Since last autumn, Clarida has been saying that the Fed will be in no hurry to raise interest rates as long as U.S. economic growth continues to be driven by strong productivity growth without any significant improvement in employment. ``Once the Fed does start to hike rates, there will be a lot of tightening in the pipeline,'' he says. ``That makes this a challenging environment on the fixed-income side.''

It's especially challenging for Clinton Group, whose records have been under review by the U.S. Securities and Exchange Commission and the Commodity Futures Trading Commission following the resignation of a Clinton trader last October. The trader, Anthony Barkan, said in an e-mail that he quit because of a disagreement with senior management about how some of the firm's bonds were being priced.

Clinton hired auditing firm PricewaterhouseCoopers LLP to investigate the matter. Hall, Clinton's president, said in a letter e-mailed to investors on Nov. 26 that the results of the internal review had been furnished to government authorities. In a separate note e-mailed the same day, Clinton's directors said the investigation had shown that the firm's methods for valuing its assets ``were consistent with industry practice'' and ``materially accurate. ''Clarida declines to comment other than to say that his job as economic strategist does not entail the pricing or valuation of any securities.

Clinton struggled last year as its flagship Trinity Fund Ltd., which invests in mortgage bonds, was down 22 percent. Clinton began 2004 with $2 billion in hedge funds, down more than 60 percent from a peak of $5.5 billion in August 2003, as many investors pulled out late last year.
Clarida stands by his decision to join Clinton. ``Hedge funds provide important liquidity to the markets and enable them to function better,'' he says. ``Hedge funds are here to stay for some very sound and fundamental reasons.'' Still, his experience highlights the fact that academics who stray off campus don't always make a score. Even with glittering academic credentials, sophisticated quantitative models and computers that can learn, professors can still get trumped by the market.