Sunday, August 19, 2007

Sunday Morning Reflections

by Tanta on 8/19/2007 08:58:00 AM

I did a post yesterday that was ostensibly about loan modifications, but that was trying to make that an excuse to reflect on risk/behavior modeling. It didn't work, but life is like that in the risk business.

One of my points is that while all modeling of borrower behavior can be fraught with conceptual, mathematical, and data problems, modeling of what I call the "residual" borrower and the less polite call the "woodhead" borrower is more fraught than any other part of this. In any sufficiently large group of mortgage loans, you will get a "tail" of borrowers whose behavior you cannot predict or "solve for" (as the case may be) with intelligible results.

Some people simply will not behave the way theory says they will. As a risk manager in a financial instutition, I have always had some trouble dealing with this group. As a person, of course, I have always sought them out first in bars and parties. Theory predicts that I would be making you guys pay $100 an hour to listen to my insights about the world. In the real world, I'm bloggin'. Solve for that.

I suspect some of the friction that arises from time to time in our comment section involves the fact that I don't always like best the people I prefer to lend money to. I don't even always morally approve of them. But my job has always been to lend money safely and soundly at a reasonable profit, not to dole out rewards for good behavior or use loan commitment letters as a kind of Good Housekeeping Seal of Approval, a weird sort of Calvinist-banker thing that involves identification of The Elect. It's possible that uptight cheese-paring pompous authoritarian self-satisfied literal-minded parochial nosey-parkers are overrepresented in my approved borrower pool. That doesn't mean I'm willing to live in your neighborhood, listen to your homilies, or even drink your beer.

Some of the people I love best want to buy homes because if they own it, instead of renting it, they can paint it purple if they want to. Then they do, DIY, too, because they're frugal. Often the neighbors have a big meltdown over this, because it "brings down property values."

So many people want to be not just their own landlord, but everyone else's, too. I have a very vivid memory of the first time I saw the covenants and restrictions on a new PUD somebody actually wanted to buy a (bland, featureless, identical, not purple) home in. It prohibited hanging laundry in the backyard on a clothesline. I have been an apartment-dweller for a long time, and one of the two or three idle thoughts that would occasionally get me thinking about buying a home was wanting a place to hang my sheets out in the sun. I concluded that if we were going to make homeownership the equivalent of taking a shower with a raincoat on, I'd save myself the time and money.

At some level, the argument that a mortgage servicer should try reasonable alternatives to foreclosure because to flood the neighborhood with vacant REO punishes all the innocent bystanders makes some sense. At another level, it makes me uneasy. This is probably because I am a former lit major with an active imagination instead of a real high-class math wiz who understands only standard deviations. But I keep wondering whether some of these folks would be in such dire financial straits if they hung their laundry out to dry: the sun's free, but a gas dryer isn't. Lawns that look like lawns, not putting greens, cost less to maintain. Front doors that are an out-of-fashion color are in the sale bin at HomeDepot. Do you really want your neighbors to act prudently? Do you?

When I was in grad school I lived a few doors down from some folks who drove a '63 Plymouth Fury, gold-and-primer, with "Neighbors From Hell" painted on the side with Rustoleum. They'd get their shovels out and come help me dig my innocent white Volkswagen out of a snow drift when the need arose. It was a curious kind of hell they were from.

Yves over at naked capitalism has an interesting post up this morning on "cognitive bias" and risk assessment. This paper has some interesting comments on the subject for those of you who are more analytically rigorous than I'm in the mood to be this morning. I was, I admit, taken by the discussion of hindsight bias. It may be relevant to our preoccupations. I may either take it out on the patio to read over a cup of coffee, or get out a can of spray paint and go after my car. Some days it's a knife-edge.