Insurance & 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.

News

09:12 AM
Connect Directly
Facebook
Google+
Twitter
RSS
E-Mail
50%
50%

Risk Modeling and Wishful Thinking

The fact is that risk modeling, like so many other control methods, can be used to mask as well as reveal reality. As useful, indeed as indispensable, as risk modeling is to project future probability it is nevertheless subject to one of the cardinal maxims of information technology: garbage in, garbage out.

If one wanted to identify a single factor that has had the greatest influence in the current financial crisis it would be poor evaluation of risk. Many people, especially those outside the financial industry who manage risk in their own businesses, are now asking how Wall Street's professionals could manage theirs so catastrophically, especially considering the potential consequences.The debacle also raises the question of the role of technology. One can easily understand "irrational exuberance" or plain greed skewing risk calculation, but financial services today rely on computers to perform vastly complex calculations that can plot the probability of outcomes and precisely evaluate risks. Or so we thought.

The fact is that risk modeling, like so many other control methods, can be used to mask as well as reveal reality. As useful, indeed as indispensable, as risk modeling is to project future probability it is nevertheless subject to one of the cardinal maxims of information technology: garbage in, garbage out.

No financial experience is necessary to realize that "subprime" mortgages were an especially risky instrument to commoditize, or that risk multiplies as the left side of leverage ratios increases. In late 2007, notes Robert Samuelson, Lehman Brothers' leverage ratio was about 30:1; Fannie Mae and Freddie Mac ran ratios as high as 60:1. Samuelson comments:

It wasn't that Wall Street's leaders deceived customers or lenders into taking risks that were known to be hazardous. Instead, they concluded that risks were low or nonexistent. They fooled themselves, because the short-term rewards blinded them to the long-term dangers. Inevitably, these surfaced. Mortgages went bad. The powerful logic of high leverage went into reverse.

The self-deception Samuelson describes is not surprising in itself, but it becomes troubling when one reflects that self-deception was incorporated into pseudo-scientific arguments for the viability of risky investments. Sellers and buyers could defend their actions because they could prove that the investments made sense through risk modeling.

Thus they present a cautionary tale about the abuse of risk modeling. In the first place, if the assumptions of risk models are wrong, their output will be too. Also, as David West, senior vice president of Valen Technologies observes, the number and quality of sources of input will affect the reliability of output. "A consortium approach to data acquisition and subsequent modeling is necessary to avoid building models that have very limited scope," he says. Furthermore, models have limited shelf life. As the multiple factors influencing risk change, the risk of a given investment may also change, and sometimes dramatically.

On top of all these challenges and limitations, "models do not give black and white predictions - everything is described in varying shades of gray," West adds. "Yet banks and insurers want clear-cut decisioning."

Of course decisions themselves must draw a clear line. However, that line can be drawn under the influence of prudence or wishful thinking. Financial services companies exercising due diligence will refine their risk modeling competence and increase their use of sources to make their models more complete. But the distinguishing characteristic of the best financial services professionals will always be good old common sense.

What else explains the fact that some insurers have been decimated by subprime exposure and others have escaped unscathed? Some months ago, the CEO of one of the nation's largest life insurance companies told me how his company had avoided subprime exposure. He said:

"We never thought it made sense to lend money to people who couldn't put any money down and couldn't afford to pay for the mortgage, no matter what all the models were saying."
The fact is that risk modeling, like so many other control methods, can be used to mask as well as reveal reality. As useful, indeed as indispensable, as risk modeling is to project future probability it is nevertheless subject to one of the cardinal maxims of information technology: garbage in, garbage out.

Anthony O'Donnell has covered technology in the insurance industry since 2000, when he joined the editorial staff of Insurance & Technology. As an editor and reporter for I&T and the InformationWeek Financial Services of TechWeb he has written on all areas of information ... View Full Bio

Register for Insurance & Technology Newsletters
Slideshows
Video