Alan Krueger channels Burton Malkiel on hedge fund returns:
Hedge Funds Better at Managing Data Than Managing Money By ALAN B. KRUEGER - New York Times: December 9, 2004: HEDGE funds have grown at supersonic speed. In 1990, about $50 billion was invested in hedge funds; today, the amount is estimated at $1 trillion. Does superior performance explain the rapid growth? No, says Burton G. Malkiel, a professor of economics at Princeton University, and Atanu Saha, a managing principal at the Analysis Group, a consulting firm. The researchers recently completed a study that challenges the often-made claim that hedge funds, in general, produce lofty returns.
Hedge funds are a diverse set of investment funds that typically cater to wealthy clients and institutions. The funds pursue various strategies, like holding both long and short positions, and often employ substantial leverage. Their fees are usually much higher than those charged by mutual funds or other financial assets. Data on the performance of hedge funds comes from indexes like the CSFB/Tremont Index or the Van Hedge Fund Index. Those indexes are generated by companies that advise investors and operate funds. "Hedge funds in aggregate," Van Hedge Fund Advisors boasts on its Web site, "in most multiyear periods, have provided both superior returns and lower statistical risk than the S.& P. 500 or mutual funds."
The catch, according to Professor Malkiel, is that the information on performance is voluntarily provided to the organizations who track the funds. Because a good record helps attract investors, funds have a tendency to start reporting results only after they have achieved some success. Funds that are losers right out of the gate may never be represented in the database. Furthermore, when funds start reporting, they have the option of "backfilling" their data, or providing information on returns for previous months. If a fund was successful in preceding months, it has an incentive to backfill its data to increase its attractiveness to investors. This process creates a "backfill bias," because better results are overrepresented in the database. It is as if the Boston Red Sox waited until 2004 to report their World Series success, while the Yankees started in 1923; both franchises would look like smashing successes.
By analyzing statistics from TASS Research, which is owned by Tremont Capital and has perhaps the most comprehensive data on returns, Mr. Malkiel and Mr. Saha have shown that the backfill bias is substantial. The returns that were backfilled for a given year were 5.8 percentage points higher than the returns of other funds whose results were contemporaneously reported for that year. "I think there are a lot of people in the financial community who have a vested interest in showing only those pieces of data that help them sell products," said Professor Malkiel, who is also a director for the Vanguard Group.
Another problem he noted, called survivor bias, is a tendency for funds to stop reporting their monthly returns when they suffer losses and are on the verge of closing. Long-Term Capital Management, for example, did not report its losses to any of the database services from October 1997 to October 1998, a period when it lost 92 percent of its capital. (Long-Term Capital never reported to the TASS database.) Looking only at the past returns of hedge funds that are in existence today - that is, the surviving funds - it does appear as if hedge funds do produce generous returns. But this is tantamount to judging the success of a war by ignoring all the casualties. Mr. Malkiel and Mr. Saha have found that the funds that cease reporting their data, so-called dead funds, tend to have weak returns in the months before they cease reporting. The average annual return for dead funds was 7.4 percentage points less than that of surviving funds for the same years.
And hedge funds have a tendency to die - more than 10 percent stop reporting to the database each year. Although it is possible that some of these funds withdrew because they were so successful that they no longer desired further investors, the researchers found that smaller and underperforming funds were the most likely to cease reporting - not a profile of successful funds that were turning away business. Using data from 1996 to 2003, Mr. Malkiel and Mr. Saha found that correcting for backfill and survivor biases reduced the average annual return on hedge funds, after deducting fees, from 13.5 percent to, at most, 9.7 percent, which is almost three percentage points less than the return on the Standard & Poor's 500-stock index for that time period.
The lower return could be justified if hedge funds helped to diversify portfolios by providing an investment that did not move in lock step with other investments, and the researchers did find that hedge funds do not move closely with the stock market over time. Yet they also found that choosing a particular hedge fund entailed considerable risk because the funds exhibited enormous variability in performance in any given year. The best funds perform extraordinarily well, but the worst ones perform extremely poorly, with the spread between the best and worst greatly exceeding the spread between the best and worst equity or bond funds in a typical year.
"Clearly, there is a risk in investing in hedge funds that is far greater than the risk of investing in the other asset classes," the researchers said. Even the so-called fund of funds hedge funds, which try to diversify risks by investing in other hedge funds, display nearly as much variability in performance across funds in a given year as is exhibited across the entire universe of mutual funds. Moreover, from 1995 to 2003, the average fund of funds yielded only a 7 percent annual rate of return after deducting fees, well below that of the average mutual fund.
Picking a good fund is also dicey because there is little persistence in performance from one year to the next. The chance that a hedge fund that performed in the top half of the universe of funds in one year would do so again the following year is no better than 50-50, which raises the question of how the funds can command such high fees. Most hedge funds will be required to register and provide data to the Securities and Exchange Commission beginning in February 2006. While some people have argued that the S.E.C. already has too much to do, it would seem that collecting and disclosing information on performance is a small burden for the commission, and a great potential benefit to investors.
"As a free market person, I think markets work better when there is fuller and more accurate information," Mr. Malkiel said.