Hedge fund gave warning signs, experts say

By Mark Gimein

updated 5:58 p.m. ET Dec. 12, 2008

Bernard Madoff, one of Wall Street's best-known brokers and money managers, was arrested  on Thursday after allegedly confessing to his sons in a near nervous-breakdown-type meeting that instead of the $17 billion they thought he had in his funds, there was pretty much zip. The whole thing was—as the criminal complaint quotes Madoff himself saying—"basically, a giant Ponzi scheme" in which investors who wanted their money back got paid with earlier investors' money.

There's no shortage of cases in which shady hedge-fund managers have disappeared with hundreds of millions of dollars, but the size of what seems to have happened with Madoff is well beyond anything in Wall Street memory. The amount of money involved—at least $17 billion, and maybe as much as $50 billion (a number that Madoff himself put on the money lost)—is bigger than the losses that took down Bear Stearns, bigger than the $7 billion in hidden losses in Societe Generale's Jerome Kerviel scandal, and these were big banks with thousands of clients. Madoff seems to have had one or two dozen clients, in accounts he closely watched over himself.

The question that everybody with big chunks of money parked with exclusive fund managers on Wall Street will be asking today is whether there are other possible Bernard Madoffs out there: high-profile managers who've been lying about their returns for years. The answer to this is going to be "yes," which leads to the second question of, "Is there any way to spot them?" Or, in other words: Is there a way to know whether a money manager's returns are too good to be true?

It's a question everyone wonders about but that few investors, even the big ones, ask directly, because there seems to be no way to answer it except by looking a fund manager in the eye and hoping you trust him. Ordinary intuition tells us that just by looking at a string of numbers in a fund manager's reports, there's no way to know if those numbers might be made out of whole cloth. But in fact, it turns out that there may be a good way to guess—and it could well have helped the investors who'd given their money to Madoff see what was going on earlier.

The key here is not looking just at how well Madoff seemed to perform. It's how consistently he seemed to be doing it. Stories in both the New York Times and the Wall Street Journal both noted the pattern. According to the stories, he seemed to make a return of 10 percent or 11 percent a year, year in and year out. And it wasn't just an annual return kind of thing. Almost every month, the WSJ story says, Madoff made somewhere between 0 percent and 2 percent. Hardly any losses, no really outsize gains.

Professional investors are taught to value steady returns—it implies that managers aren't taking excessive risks. The latest research on hedge funds, however, reveals that overly smooth returns may not indicate that the risks are small, but that they are hidden.

The key concept here, developed by MIT professor and noted hedge-fund theorist Andrew Lo, is "serial correlation." Simply put, serial correlation is the degree to which each month's returns in a fund mirror the results of the month before. A fund that returns the exact same amount every month is perfectly serially correlated. Madoff's returns were strikingly consistent month after month, year in and year out. That kind of performance—a nice, smooth line going up no matter what the market does—is a really good sign that you should look more closely.

The extraordinary thing that Lo does in the third chapter of his book "Hedge Funds," published earlier this year, is to demonstrate mathematically that an excessive degree of serial correlation is a powerful indicator that the holdings of a fund aren't being reported realistically. What Lo shows from the pattern of historical returns in hedge-fund databases is that when funds' returns grow too consistent, it is a sign that the investments are either very hard to value accurately and the returns are just guesses, or, worse, that they've been manipulated in a way that smoothes them artificially. What Lo creates is a mathematical model for judging what "looks too good to be true." Lo's work turns a lot of the conventional thinking about what's safe on its head. It shows that the evenness that investors have traditionally been taught indicates safety and reliability can actually be the best sign risk is being hidden or that the data are unreliable.

MORE FROM PERSONAL FINANCE

Personal finance Section Front

U.S. debt could push interest rates higher Answer Desk: Adding up bank bailout losses Americans getting better at managing credit Therapy tries to achieve life, account balance FDIC: Minorities struggle to access banks Vanguard: 60 percent of 401(k) have recovered Answer Desk: Repairing consumer credit  New $100B safety net for jobless in works Congress scrambles to extend estate tax Auto loan delinquency rate edges higher Personal finance Section Front   Add Personal finance headlines to your news reader: