“Risk management” has a nice ring to it. Not only does it suggest that a hedge fund team, for instance, has pretty much thought of all the things that could go wrong — it has also, bless its heart, managed those nasty surprises.

Arguably the science, or art, of risk management has never before been so integral a part of every financial organization. If you Google “risk management consultants,” you will find 38.8 million items — almost as many as pop up from Googling “Britney Spears.”

So how can it be that the financial industry has stubbed its toe so dramatically with all those sophisticated risk management systems and people in place?

Leslie Rahl, founder and president of Capital Market Risk Advisors and a board member of Fannie Mae, has an excellent perch from which to view the unfolding of this latest debacle. According to her Web site, her company is “the preeminent financial advisory firm specializing in risk management, hedge funds, financial forensics, and risk governance.”

Ms. Rahl graduated both from the Massachusetts Institute of Technology and its Sloan School of Management and was formerly head of Citibank’s derivatives group. She actually understands all those complex formulas that are supposed to identify risk. Numbers are to Ms. Rahl as Cheerios are to the rest of us: uncomplicated and easily consumed.

Her take? “Risk management is all about thinking about two or three standard deviations from the mean. No one ever expects events to fall beyond that. Once in a lifetime events that fall outside that parameter have exponential, not arithmetic, consequences. Risk management is built around models, and models are built around assumptions. The models will work if things behave the way you model them to — but they never actually do. These events are somewhat expected , but we keep forgetting. You can’t expect a computer model to anticipate changes. This is the big flaw — I keep reminding clients of this — that their assumptions are not the worst case.”

One of the comforting things that money managers will tell you is that they “test” the models all the time. The concept is that you plug in variables such as rising interest rates or a widening of credit spreads, and the model will indicate how these changes impact your portfolio. Unfortunately, the testing is done based on normal events and correlations, because the models can’t handle something not in the historical database.

This shortcoming, according to Andrew Davidson, head of a firm specializing in mortgage-backed securities, is compounded by the fact that financial modeling normally relies on implied data from the derivatives markets in addition to historical data. As no derivatives are based on mortgagebacked securities, historical input is even more important.

“By definition, most risk people are young quants,” Ms. Rahl said. Most, she said, do not carry their modeling back far enough to include similar events, such as the 1994 bankruptcy of Orange County, which she views as somewhat analogous to today’s situation.

“In 1994, the money funds broke the buck,” Ms. Rahl said, referring to the unthinkable: a money market fund that experiences such credit issues with its portfolio that it no longer trades at a dollar. A similar deterioration in shortterm instruments occurred over the past two months, as a few money market funds got into trouble. The credit problems in the early 1990s stemmed from holdings of “inverse floaters” and the “kitchen sinks” — the names given to the leftovers of collateralized mortgage obligations after they had been sliced and diced and the higher-grade parts of the securities had been bought by savvier investors. A professor of financial engineering at MIT, Andrew Lo, protests that many risk managers have data going back many decades. However, he argues that using historical data is complicated because “it’s very difficult to know what is relevant. For example, a currency trader might have data back to the 1970s, but at that time exchange rates were fixed, so the data would be useless. It’s like that expression, ‘You never step into the same river twice.'”

That may be, but the whole purpose of financial models is to forecast what may happen in the future using the past as a guide. The problem is that markets do change over time. A professor of economics at Princeton University, Yacine Ait-Sahalia, said: “A lot of the subprime stuff is quite new. What the models missed is that CMOs don’t behave in a correlation sense the same way as CDOs,” collateralized debt obligations.

CDOs, which consist of debt obligations from a wide range of corporations, do not exhibit much correlation. Generally, if one issuer has a problem, it is unlikely that another would also feel the impact, barring a significant recession. In the mortgage market, however, “it is not hard to realize that underlying assets could be correlated,” Mr. Yacine said. Extremely correlated, as it turns out. Mr. Lo points out that even if the authorities had better understood how interrelated the sectors of the mortgage market were, social priorities might still have argued for a laissez-faire response. “What would have been the alternative?” he asks. “There would have been fewer mortgage companies, and less aggressive lending practices. There is a social issue here that no one has picked up on. A significant fraction of the subprime loans were made to minorities.”

And yet, as mortgage defaults widen, social policy will take a backseat to financial realities. The aggressive lending practices that caused the subprime market to balloon involved, among other things, the use of secondary or “piggyback” mortgages to supply the 10% to 20% not covered by the primary mortgage, the portion that traditionally had been the down payment. He said a significant amount of fraudulent information also was included in the mortgage applications concerning the borrower’s income and ability to pay. In short: Garbage in, garbage out.

At the end of the day, we are reminded of the peril of investing in instruments so complicated that few could really understand them. “Even for me, who loves complex things, it’s very complicated,” Ms. Rahl said.

That’s all we had to know.

peek10021@aol.com