Wall Street & Technology is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Risk Management

09:57 AM
Connect Directly

Could Social Networks Help Risk Managers?

Why current risk models and practices don't work, and whether risk managers will be willing to share ideas in open forums.

Risk models don't work very well, do they? That's the thought we were left with as we departed a professional risk managers conference held at Goldman Sachs' headquarters last night.

David Rowe, executive vice president of risk management at Sungard, led the group through some of the reasons Wall Street firms' risk models failed them before and during the subprime crisis.

For one thing, he pointed out that the high ratings that the credit ratings agencies gave subprime-based securities gave everyone a false sense of confidence. "AAA was supposed to mean a one in 10,000 chance of an instrument failing," Rowe said. "We should have been asking ourselves, what empirical information did we have to know that that was true?" Coincidentally, this morning's Wall Street Journal reported that Eric Kolchinsky, a former analyst with Moody's Corp., has accused the agency of issuing inflated ratings and has taken his concerns to U.S. congressional investigators.

For another thing, risk modelers tend to look at the wrong things. "Looking at the middle of a distribution [such as a line chart of housing prices] does not give you an indicator of what's in the tail," Rowe said. Up until mid-2006, reasons for mortgage defaults were random events hitting individual families," Rowe said. "We should have been asking what unintended occurrence would have affected this pattern." He noted that the idea that home prices could drop was dismissed because most people thought that hadn't happened since the Great Depression, although in fact it did happen for a time in the early 90s. "If there's a contingency that you think would be a problem, you should begin to track those variables," Rowe told the risk managers.

Thirdly, risk managers tend to get too focused on individual markets and forget to look at the big picture, Rowe said. They also often rely on historical data for new, complex products, for which historical data becomes increasingly unreliable over time. Risk managers should always strive for an alternate means of valuation, he said.

Marcus Cree, North America solutions director for Sungard Trading and Risk Systems, suggested that risk managers should start using social networking tools such as LinkedIn discussion boards, Twitter and public forums, to share best practices and concerns. Sungard is trying to build an online community of risk managers at www.onlineriskforum.com. When a panel moderator asked about the risks of letting risk managers share information on such a forum, Cree said, "No one's looking for information about banks' risk exposure, but their risk methodology." A managing director at a large firm said, "If you express a concern, doesn't that tell me about your bank?" Cree answered that they could talk about general metrics, but that he understands the concern.

At the end of the conference, we turned to the person sitting next to us, who works at a large reinsurance company, and asked him if his models had helped his company navigate the subprime crisis. "Oh, no, not at all" he said, with a hearty laugh, and walked away.

Register for Wall Street & Technology Newsletters
Inside Abel Noser's Trading Floor
Inside Abel Noser's Trading Floor
Advanced Trading takes you on an exclusive tour of Abel Noser's New York trading floor, where the agency broker known for transaction cost analysis, is customizing algorithms for the buy side, while growing its fixed income trading and transitions business.