Strange accounting

By Megan McArdle

I want to revisit the notion, now popular in my comments and emails, that mortgage cramdowns are good because they force banks to value the houses securing their mortgages at true market value.

This is very odd because, first of all, that's not really how mortgage accounting works--banks don't mark houses to market, and a very good thing, because their balance sheets would have been puffing up like blowfish for most of the decade.  They very sensibly do not do this because in a rising market, they are unlikely to get their hands on a house in foreclosure to sell at a profit.  So what they look at instead is the potential cash flow from the mortgage.


In a falling market, of course, they are much more likely to end up with the house, and the loss.  But what matters is still not the absolute size of the change in the house's value; it's how that affects your future cash flow.   We don't want to just write down every mortgage bond by how much the price of the underlying houses has fallen; we want to plug that change into our model, which will increase the losses we have to allow on each default.

But here's the thing:   even if we get a more accurate picture of how much each default is going to cost us, we still don't know how many defaults we're going to have; that depends on a lot of other things, like whether the homeowner could ever afford the payments, whether they lose their job, and how willing they are to torpedo their credit rating in order to get out from under an overpriced mortgage. 

Nonetheless, it wouldn't be a bad idea to get a better handle on the value of the houses securing these mortgages, forcing the bankers to increase the allowance for bad debts.  But the other problem with this notion is that rewriting the bankruptcy code does not achieve that laudable goal.  It writes down the value of only some houses--the ones where the owner is willing to declare bankruptcy fairly immediately.  And it writes them down completely--new interest rate, new principal.  We're not forcing banks to recognize the potential losses on these mortgages; we're forcing them to take the actual loss.  Meanwhile, all the loans where the borrowers haven't declared bankruptcy remain on the books at their original value, representing a larger potential loss than our models currently allow.

Let's remember, too, that the possibility of a one-time-only cramdown of your underwater mortgage makes bankruptcy more attractive.  That means we'll get more of them.  Customers who might have surrendered the house in foreclosure will take them, but so, probably, will marginal borrowers who are simply struggling, who might have sent the wife back to work or gotten a second job.  Suddenly, bankruptcy isn't just a way to get your monthly debt service down to a manageable level--it's a way to substantially improve the net asset position of the household, by tens or hundreds of thousands of dollars.  People talk about feeling like they're idiots for not just mailing in the keys and walking away, but when you do that you end up with nowhere to live and a gigantic crater in your credit report.  With cramdowns, you take a big ding on your credit report, but you get to keep a house, and any future appreciation, in exchange.

To sum up, the cramdown doesn't solve the problem of unknown risks lurking on the balance sheets.  Meanwhile, it turns potential losses into real losses, and probably increases them, because Chapter 13 is now definitely a much more attractive option than foreclosure, and possibly a more more attractive option than continuing to make your payments.  It forces us to recognize losses in the banking sector, yes--the losses you just created by government fiat.   

This article available online at:

http://www.theatlantic.com/business/archive/2009/01/strange-accounting/185/