Sunday, April 01, 2007

A Walk Down the Subprime Memory Lane

by Tanta on 4/01/2007 09:01:00 AM

CR may have posted on this study by researchers at the OCC and Federal Reserve Bank of St. Louis early last year just after it was published; at the time I was distracted by my post-surgical morphine pump (and you think I’m a Luddite?) and quite honestly was paying no attention to such matters. It’s not just that I’m too lazy to check the archives. Some questions have arisen in the comments recently about the history of subprime lending, as well as the extent to which depository institutions are implicated in it. So this may be worth a read even if we’ve been there before.

It’s a good paper; it’s also worth thinking about how off the mark some of its predictions about the direction of the subprime market turned out to be, just a year or two later. The authors were using data sets up to about mid-2004 and were writing in 2005, at what now appears the top of the real estate market in a lot of places. Even the list of subprime originators is sorely out of date: note the statement that “Household Financial Services, one of the original finance companies, has remained independent and survived the period of rapid consolidation. In fact, in 2003 it was the fourth largest originator and number two servicer of loans in the subprime industry.” Of course HSBC, a depository, was busy negotiating the purchase of HFS at the moment that sentence was written.

So the discussion here of the recent history of subprime originating is a bit dated in spots. What disappoints me is that the authors do not address one of the key facts about the situation in the 90s that I happen to remember: how the prime-lender depositories got their feet wet in subprime to start with.

Remember that back in the early-to-mid 90s, FICOs and AUS were still in development or not widely used, and nearly every loan was originated as “full doc” or what was called “alt doc,” which simply meant that it used borrower-provided documents (pay stubs, bank statements) instead of the probably more accurate but more expensive and time-consuming employer-provided or bank-provided documents (the Verification of Employment and Verification of Deposit forms a lender mailed to the relevant third party and received back, in the mail, with original signatures and in-depth information). Credit history was analyzed by looking at a full credit report (the old “RMCR” or “long-form” credit report ordered specially for residential mortgage applicants). Appraisals were all submitted on paper, with ink signatures and original photographs. Processing and then underwriting loans was time-consuming, and the application fee charged to a borrower was generally only enough to cover the actual cost to the lender of obtaining the credit report and appraisal. So the effort made by the lender to order verifications and evaluate the loan was sunk cost if the loan request was denied.

Several contexts converged here: after the S&L fiasco, prime depository lenders were finding themselves under some more attentive than usual regulatory supervision on safety and soundness issues, which, combined with RE busts, tended to make the declination rate for loan applications go up. At the same time, Fair Lending laws were getting enforced, which tended to make lenders less casual about having a high declination rate that they couldn’t justify on strictly credit-related grounds. For a lot of banks, specifically, the mortgage lending department (as opposed to other sources of loan assets like commercial, consumer, etc.) was a cyclical overhead-sink when rates went up (as they did in 1994-1995) and the refi boom of the early 90s came to an end.

Therefore, the issue of how one might recover some of those costs expended on loan applications one had ended up denying converged with the by-now commonplace idea that the prime-denied borrower class is an “underserved borrower segment” to create a willingness to dabble in subprime. If you, like me, had ever been handed a coffee mug with “There is no problem; there is only opportunity” emblazoned on it in some corporate pep rally in those days, you are familiar with the approach.

But the prime depositories of those days weren’t necessarily interested in placing these loans in their own investment portfolios, and they weren’t eligible for sale to the GSEs, which were the major buyers of loans the depositories didn’t hold. So banks and thrifts started developing whole-loan servicing-released loan sale programs, where the “fallout” from the prime pipeline could be recaptured by making a subprime loan and then selling it outright to a subprime conduit. Everybody gets a loan; all costs are eventually recouped; opportunity thrives. The subprime conduits could securitize the stuff and lay off the risk. It was great.

Except. There’s always an except. One unforeseen difficulty was that it became possible for certain participants who had always lived in the prime world to compare the profit margins on good old Fannie Mae fixed rates (maybe 50 bps if you were good at it) to those subprime deals (easily 150-200 bps if you were fair-to-middlin’ fastidious). Increased subprime lending could even improve those prime margins: as more and more of your weakest loans fell out of the bottom tier of your GSE loans and into the top tier of your subprime loans, you got paid better by the GSEs in the form of improved guarantee fees (since your average credit quality was so much better) and by your subprime investors (since your average credit quality was so much better). There was, in short, a moral hazard in play: in the shift from non-price to price rationing, more borrowers got mortgages, but it wasn’t always clear that they got the cheapest mortgage they should have gotten.

Those of us who were there at the time that the story about how “subprime is a way of serving the poor” got written do, then, tend to be somewhat more skeptical of this claim than others. The fact that many participants did not start out with the intention of preying on borrowers doesn’t change the fact that predation became widespread, or that a form of lending that had once been reserved for people with a lot of equity became associated just a few years later almost exclusively with people who had no money at all.

We can certainly debate the extent to which price-rationed lending provides true benefit to the historically credit-constrained. I’m game. I just think that market participants with short institutional memories are a menace. To all of us.