How to tackle foreclosures and unemployment at the same time

Back in mid-December, after I wrote a post about some numbers from the Office of the Comptroller of the Currency (OCC) that seemed to show loan modifications don't work so well, a reader of this blog suggested that the data release might be politically motivated—that the Bush administration, no fan of Sheila Bair's crusade to rewrite mortgage terms, was trying to show how unproven loan modifications are for the long-term. True that. At least the "unproven" part.

But that doesn't necessarily mean loan modifications don't work when done conscientiously. In fact, in this week's magazine I have a story about a subprime servicer that's finding modifications work just fine—for homeowners, investors in mortgage-related securities, and the company's own bottom line.

In reporting that piece I dug back into the data. It's a pretty quick dig. The Hope Now Alliance has told me in the past that it wants to get into details like what servicers mean when they say they've "modified a loan," but so far I haven't seen anything come out of that effort. And, as it turns out, details like definitions are kind of important. For instance, in the past, modifying a loan has often meant spreading missed payments out over the remaining life of a mortgage, thereby raising a homeowner's monthly payment. That's probably not what we thought we were talking about.

As I was going back through the OCC's numbers, and talking to other people about them, I realized another definition we've really got to firm up: re-default.

When the OCC first reported its statistics, it focused on the fact that more than half of the people who receive modifications are back to being late on their mortgages within six months. Technically, that's true. The thing is, that's looking at what percentage of people are 30 or more days late. Even in normal times, plenty of people are a little late mailing in their checks—especially among subprime borrowers, who make up a decent chunk of the overall numbers. If you look at the percentage of people who are 60 or more days late—an indication of a more-serious problem—the figure drops back down to 37%. I'm not sure that's the conspiracy our blog reader posited, but it does underscore the importance of being meticulous about how we talk about the numbers.

Rod Dubitsky, the head of asset-backed securities research at Credit Suisse who probably knows more about loan modifications than anybody else (you decide on your own what this says about our policymakers), thinks it's ridiculous that no one has sorted all this out yet. He's been pushing for a coordinated national effort around modifications since at least early November. He also wants banks that have rolled out their own programs—like JP Morgan Chase, Bank of America and IndyMac (under the direction of Bair's FDIC)—to release standardized data about what, specifically, they're doing. He wants disclosure around the type of modification (e.g., interest-rate reduction, extended amortization), the documentation that was used to determine how much a homeowner can pay, other loss mitigation options that were considered (for example, HUD's Hope for Homeowners program) and the math that shows the decision that was made is the best one, assumptions about future home prices, and a lot of other things.

Keep in mind that Dubitsky's constituency is made up of investors in asset-related securities. Helping people keep their homes is great, but there's also a purely capitalistic motive to collecting clean data and analyzing it and figuring out what works and what doesn't.

And that's how we get around to chipping away at unemployment. In talking with Dubitsky, he made a remark about the government needing a "data swat team." He pointed out that there are quite a few unemployed Wall Streeters who, guess what, know something about mortgage securities and sifting through data and building computer models. I liked what he was getting at: evidence-based public policy. Fascinating thought, no?

Barbara!