Needed: Better IT Crystal Balls on Wall Street

One likely outcome of the financial market meltdown is an effort to improve the modeling tools that are designed to help firms spot risky business -- before it's too late.

Greed and a lack of regulatory oversight are getting most of the blame for the ongoing financial chaos on Wall Street. But there was also an IT component to the meltdown -- namely, a reliance on risk modeling and management tools that weren't up to the task of predicting just how risky some of the new forms of financial investments could be.

As a result, financial services firms are now expected to take an increased interest in developing or buying more effective risk management technologies -- partly for competitive reasons, and partly to help them comply with the new regulations that many people in the industry think are inevitable.

Financial risk models often have thousands of variables and require enormous amounts of computing power. As evidence of that need, market research firm IDC says that annual spending by financial firms on high-performance systems grew from $169 million in 2003 to $305 million last year.

As the complexity of the models increased, so did the confidence that federal regulators placed in them. For instance, The New York Times reported this month that at a little-noticed meeting in April 2004, the U.S. Securities and Exchange Commission -- acting at the behest of large investment banks -- voted to loosen its debt-limit rules and rely on the firms' computer models for assessing investment risks.

But as recent events have made clear, such confidence in the technology appears to have been misplaced.

Dennis Santiago , CEO and managing director of Institutional Risk Analytics, a consulting firm and software developer in Torrance, Calif., said the Wall Street crisis has exposed some fundamental shortcomings in current risk-modeling tools and data analysis techniques. "We have been pretty much using the same tools for a decade now," he said. "One of the things that is clearly beginning to show itself is that the techniques that worked in the last business cycle for managing risk don't work as well anymore."

For example, Santiago cited the way that companies model the risks associated with so-called structured finance transactions, such as loans made to a business based on its cash-flow history. Much of the statistical averaging applied in the risk models is done in "an almost blind fashion, on the assumption that it has worked all this time," he said. But, he added, that has become a faulty assumption, causing many institutions to lose confidence in such transactions.

What's needed now, according to Santiago, are increased investments in systems and analytical approaches that come closer to modeling the actual risks faced by financial services firms.

Risk Management Tenets

The current financial crisis highlights some "core tenets" for effective risk management, said Dave Hoag, director of clearing technology at CME Group Inc. , which operates the Chicago Mercantile Exchange, the Chicago Board of Trade and the New York Mercantile Exchange.

The most important one, Hoag said, is the need for fair and transparent visibility into the models, data and analytic techniques that are used to calculate the risks of transactions. For instance, he said that the methodologies and underlying numerical values that CME's tools use to generate portfolio risk scenarios are openly available to all of the participants in a planned financial transaction.

"One of the things we like to talk about is the transparency of what is going on in terms of who trades what and when," Hoag said. "Everybody understands what goes into the risk analysis." He added that he expects to see a heightened focus on making risk calculation processes more transparent -- as companies either elect or are driven to do so by new regulations.

Not everyone is convinced that big changes are needed in risk management systems and procedures. Glyn Holton , an independent financial risk management consultant based in Boston, acknowledged that there will likely be a greater emphasis on risk modeling, at least for a while. "There will be some more focus on strengthening risk management, some technology will be purchased, and probably monitoring will be increased," he said.

But Holton thinks the problems on Wall Street have far more to do with an absence of regulatory oversight. "Financial risk management makes a wonderful scapegoat," he said. "This is a cycle we go through whenever we have losses. We trot out the back-office risk management guys."

On the other hand, Suzanne Duncan, financial markets industry leader at the IBM Institute for Business Value, said there is broad recognition that new investments in risk modeling are needed.

In June, the IBM think tank and research operation and the Securities Industry and Financial Markets Association jointly surveyed about 500 IT professionals in the financial industry about their technology spending priorities. According to IBM, 67% of the respondents cited increased risk-transparency rules as the primary regulatory action that would affect IT. And a majority said that the crisis was a catalyst for increasing the amount of money their companies were investing in risk management projects.

Duncan said that could eventually drive even more demand for high-performance computers capable of crunching risk models.

Jaikumar Vijayan, Computerworld