Institute for Financial Transparency

Shining a light on the opaque corners of finance

Math Wizards Working On Spells to ‘Cure’

By Scott Patterson

The financial engineers are at it again.

Critics may complain that these math wizards started the trouble in the first place by designing securities that couldn’t withstand the market’s turbulence. But they also may have the expertise to help fix the problem.

“Airplanes fail, too,” says Peter Cotton, founder of Julius Finance, a structured-finance firm in New York. “That doesn’t mean you don’t fix them.”

Mr. Cotton is one of many such engineers trying to solve a seemingly intractable problem before the government: how to design a system for buying up assets shunted into a massive “bad bank.” The government doesn’t want to pay too much and banks don’t want to sell for too little.

How big are these spreads? Last week, a triple-A student-loan auction-rate security was offered for 95 cents on the dollar by its owner. The highest bid: 50 cents.

For years, Mr. Cotton worked on Morgan Stanley‘s structured-finance desk, often designing collateralized-debt obligations, large pools of loans that are one of the prime culprits in the banking collapse. A longtime critic of credit models, he left Morgan in early 2007 to start his own company.

Mr. Cotton says the models most banks and ratings firms used to price CDOs were poorly designed. “They are superficial,” he says, and “often spit out prices that don’t capture the underlying value of the assets.”

Using those same failed models now, says Mr. Cotton, most banks are “essentially just making up numbers.”

At Julius, he has been designing new systems that dig deeper into the underlying loans of CDOs. The models use a variety of data points crucial to valuing these assets, such as the relationships between underlying slices of debt with different maturities in the assets.

For instance, a CDO containing many slices of corporate debt, or derivatives tied to that debt, can be priced by looking at where a large number of baskets containing these assets are trading and implying the behavior of the slices from these prices. The resulting CDO price closely matches similar assets changing hands on the open market. The model most banks use today is calibrated to a small subset of available data points.

Other financial engineers are working on new methods to price troubled assets. Richard Field, managing director of structured-finance firm TYI, has designed a system that provides real-time loan-performance data investors can consult to more accurately price the securities.

If investors can peer into the underlying loan-performance of their assets on a daily basis, they’ll have a much better idea of their present value, Mr. Field argues. Now, investors usually have to rely on stale loan-performance data that’s often updated on a monthly basis.

Getting a better idea of the value of these securities will also go a long way toward understanding the value of the institutions holding them. “Without real time transparency on which to base independent valuations, the market can’t determine if the banks are adequately capitalized,” says Mr. Field.

More broadly, financial engineers are struggling to reorient themselves to a world that shuns mathematical gizmos. It’s a worthy goal. First, they need to focus on cleaning up the problems their gizmos left behind.