Inman

Using hard data, psychology to predict mortgage default

Paying your mortgage is all in your head.

Well, really, it’s all in your wallet or checkbook, but what’s going around in your brain is whether you intend to pay that mortgage, and savvy loan servicers are trying to figure out what’s up there swirling about the cranium.

Take Ocwen Financial Corp., for example. It is now the largest servicer in the country of subprime mortgages, which are, of course, risky loans, and it behooves Ocwen to try to figure out a way to keep the borrowers — many of whom are distressed — paying their mortgages.

Mortgage servicers collect loan payments on behalf of investors or lenders that own the loans, and when those loans appear to be troubled, mortgage servicers will either remedy or foreclose.

Ocwen has for a long time employed a team of social psychologists who write the scripts used in the call centers.

As writer Ruth Simon wrote about Ocwen in the Wall Street Journal last year: "Psychologists are parsing the words borrowers use for clues to their emotional and intellectual states. That information could help Ocwen determine whether someone will respond better to a brief recitation of facts or a more detailed discussion."

The Ocwen "psych squad" made me think of the WillCap program unveiled back in 2010 by Santa Ana, Calif.-based CoreLogic.

WillCap is referred to as a "decisioning" system that predicts a distressed borrower’s desire and ability to make mortgage payments. It’s less social psychological and more behavioral psychological than what Ocwen does, but the hoped-for result is the same: predicting continued mortgage payments.

Or, as Michael Bradley, CoreLogic’s vice president of modeling and analytics, tells me, "WillCap built a variety of behavioral models to capture a borrower’s willingness and capacity to pay."

This decisioning system uses hard data to predict a particular behavior — i.e., whether the client will maintain a current mortgage or bail.

"The capacity really captures over a rolling six-month period the amount of income the borrowers devote to debt payments, and the willingness captures their interest in either staying out of trouble or quickly getting out of trouble should they find themselves in it," Bradley said.

In a sense, it all sounds a bit touchy-feely, but there’s a sharp edge to it all.

"If you understand why a borrower is defaulting, then (you) can design an appropriate treatment to help them respond in a way you would like them to respond," Bradley said.

When a borrower is at risk of going 90 days or more past due, the servicer will devise a loan treatment plan that would involve payment reduction, principal reduction or some combination of the two.

"You get a loan modification or we go to a short sale because there is nothing else to be done. This borrower is so distressed that I could come up with all kinds of loan modification terms, but the redefault rates would be astronomical," Bradley said.

"For some borrowers, there is no appetite to work with the servicer, so then you have to take them down the foreclosure route."

WillCap analyzes the different cash flows in a probabilistic sense as to what will happen with the different kinds of modifications, as well as other kinds of treatments, then ranks them as to which would probably be more successful. Is this a borrower that can successfully use a loan modification, and if so, what are the terms of the mod.

If no mod can be done, can the client go through a short sale? If so, what would be the value of that short sale? Finally, if you take the homeowner through the foreclosure process, what will be the expected outcome?

Although WillCap was launched two years ago, it’s taken a while to be picked up by the large banks, for a couple of reasons: long sales cycles and the banks’ need to internally test the model.

"Our model that predicts loans will go 90 days or more past due significantly outperforms the current at-risk models at the money-center banks," Bradley said.

Fate and fortune predictably have changed the program.

"The nice thing about what’s happening now and the reason why WillCap has gotten considerably more traction over the past quarter is because mandated programs are winding down," Bradley said.

Going forward, major loan servicers will have more discretion as to how mortgages are treated.

"Even though banks knew what they were doing, may not have been optimal — the banks had to do it because it was mandated," Bradley said.

Which leads to an interesting point, Bradley said. "We are talking to hedge funds and people involved in buying and selling whole loan portfolios because to the extent the financially regulated institutions are following disposition strategies that were suboptimal but couldn’t do anything about it.

"Another entity not nearly so regulated could buy those distressed assets from the banks at below what the assets might be worth. That creates an arbitrage opportunity for the acquirer."

To my way of thinking, this could be a useful tool for the mortgage industry. Someone comes in and wants a $400,000 loan to buy a single-family residence in Boulder, Colo. The bank runs WillCap and sees that if the loan is made there is a distinct possibility it will go 90 days delinquent at some point in the future. The bank doesn’t make the loan.

"WillCap could be used for new mortgages," Bradley said. "That’s just not our current focus, which is helping people manage distressed assets either in their portfolio or buying and selling distressed-asset portfolios."

Well, OK. But I did have a good idea.