In Depth Analysis: CalculatedRisk Newsletter on Real Estate (Ad Free) Read it here.

Thursday, August 07, 2008

2007 Vintage: Nowhere to Go?

by Tanta on 8/07/2008 08:51:00 AM

The Wall Street Journal continues our run of bad news about the 2007 mortgage vintage:

An analysis prepared for The Wall Street Journal by the Federal Deposit Insurance Corp. shows that 0.91% of prime mortgages from 2007 were seriously delinquent after 12 months, meaning they were in foreclosure or at least 90 days past due. The equivalent figure for 2006 prime mortgages was just 0.33% after 12 months. The data reflect delinquencies as of April 30. . . .

Data on other classes of mortgages suggest the same trend. Freddie Mac reported Wednesday that 1.38% of the 2007-vintage loans it purchased were seriously delinquent after 18 months compared with 0.38% of 2006 loans at the same point in their life. Freddie Mac generally purchases loans made to creditworthy borrowers.

Last month, J.P. Morgan Chase & Co. said it expects losses on prime mortgages that weren't securitized and remain on its books to triple from current levels. The increase in bad loans is driven mostly by jumbo mortgages originated in the second half of 2007, a company spokesman said. . . .

Economists and industry officials say several factors may account for the dismal performance of the class of 2007. Home prices were falling sharply in much of the country by 2007, meaning many borrowers who took out loans in that year for nearly the full price of the home now owe more than the home is worth. These borrowers are particularly vulnerable to a weakening economy, and have difficulty selling or refinancing if they lose their job.

Questionable business practices may have played a role, too. Some of the 2007 loans "were knowingly originated as really bad loans," says Chris Mayer, a professor of real estate at Columbia University's business school. Mortgage originators who profited handsomely from the housing boom "realized the game was completely over" and pushed mortgages out the door, says Mr. Mayer.

As credit began to tighten last year, some mortgage brokers and borrowers tried to circumvent tougher restrictions by inflating borrowers' credit scores and appraisal values, says Jay Brinkmann, vice president of research and economics for the Mortgage Bankers Association.
No doubt all of these factors are in play, even though I'm not yet convinced that they were that much more in evidence in 2007 than in 2006.

It seems to me that one thing that would help us understand this marked difference in performance between these two vintages is an analysis of the "cure" rate and method of cure of the 2006 vintage delinquent loans, as well as some analysis of the actual loan life (number of months to payment in full) of the higher-risk 2006 loans. I think we need this before we conclude that either the 2007 vintage contained worse loans than 2006 or that house price depreciation itself is an unproblematic "cause" of the elevated early delinquencies in 2007.

What I am saying is that when you compare two vintages like this, you want to know whether the loans in the earlier vintage experienced a lower serious delinquency rate because fewer of those loans were "bad," or because more of those loans had an "exit" short of foreclosure when they went bad. Another way to say this is that we are not simply asking about what origination practices or loan characteristics were at the time of origination of these loans; we are looking at what mortgage market (and RE market) conditions are at the time of first delinquency of these loans.

I am personally not ready to believe, without more data, that inflated FICOs, inflated appraisals, fraudulent income claims, etc. were more prevalent in the 2007 vintage than in 2006. I think it's possible that the marked difference in the early serious delinquency rate is more a function of the choices that a delinquent borrower had in mid-2007 compared to mid-2008. Assuming for the sake of argument that these two vintages were of either comparable quality at origination--or that the 2006 vintage was even worse at origination than 2007--you can still get a higher serious delinquency rate at 18 months for the 2007 vintage just because at 18 months out, 2006 borrowers could still refinance, get a HELOC, or sell their homes when they were still current or only mildly delinquent. No doubt some of those 2006 borrowers refinanced--in 2007, meaning that they just got "revintaged." But the 2007 vintage is hitting its 18-month history right now, when they cannot "escape" into the 2008 vintage or sell or get a HELOC to make first-lien mortgage payments with.

That's just a way of saying that credit tightening will in the nature of things--along with home price drops--increase the serious delinquency rate of a book of mortgages compared to earlier books even if the original credit quality is similar across books. The classic metaphor is "musical chairs."

One thing that was in the Freddie Mac investor slides we didn't look at yesterday was some data on "roll rates" from 2007-2008. I sure wish we had comparable charts from 2006 for comparison purposes. The "roll rate" is the percentage of loans that were in a given status last month and a given status this month. For instance, the "30 to 60" roll rate tells you what percentage of loans that were 30 days delinquent last month became 60 days delinquent this month. You need to bear in mind that a couple of things could have happened to the loans that didn't "roll to 60": they could have become current (the borrower caught up on the missing payment), or stayed at 30 days (the borrower made the next month's payment but never caught up on the missing payment). When you begin to get into the roll rate of serious delinquences, especially 90 to FC, you can also have loans that didn't roll to the next status because of a workout (modification, forbearance, repayment plan).

These roll rates are based on Freddie's total portfolio, not just the 2007 vintage. What they show is that roll rates from 30 to 60 and 60 to 90 increased from January 2007 to June of 2008 for any loan in Freddie's total portfolio in that delinquency category. The 90 to FC roll rate also increased, but seems to have hit a plateau in 2008. I suspect that is because of Freddie's major efforts in the workout department.

But very few if any 30-day or 60-day loans get workouts. Loans that "cure" from a 30-day or 60-day delinquency are almost exclusively a matter of the borrower making up missed payments from his or her own funds, whatever the source of those funds. One possible explanation of the rising roll rates here is that those funds in at least some cases were coming from HELOCs or credit cards until those got maxed out or frozen. Again, that doesn't necessarily mean that the loans that rolled to serious delinquency were "worse" at origination than the loans that cured; it may simply mean that the most recently-originated loans had fewer opportunities to avoid serious delinquency.

Roll rate analysis like this has a major drawback: it doesn't tell you about prepayments. Roll rates are calculated on how many loans you still have on your books today that were in a certain status last month. It is possible to have a rising roll rate but a more stable delinquency rate: the loans you still have on the books get worse (roll to a more serious delinquency at a higher rate), but if at the same time a lot of loans that were mildly delinquent last month paid off this month, your total percentage of seriously delinquent loans can be unchanged or rise at a much slower rate than your roll rate.

We do know that 2006 vintage loans prepaid at a faster rate than 2007 vintage loans. One way of looking at the matter is that you simply have to expect delinquency levels to be higher for 2007 than for 2006 simply due to loan life: the fewer high-risk-at-origination loans in the vintage that refinance (or sell the home) in the first 18 months, the higher the serious delinquency rate will be just because these loans got old enough to go bad.

We have to think about that because we have to understand that the process of credit tightening inevitably forces delinquency rates up. This is the thing that a lot of our politicians just don't get: you cannot "return to sane lending standards" and still prevent the "insane" loans from earlier vintages from ending up in foreclosure. You have to consider the possibility that at least some of the nasty performance of the 2007 vintage is a function of lenders having originated fewer high-risk loans in 2007 than in 2006, not more. It's just that the bad loans they didn't originate in 2007 were things like HELOCs that 2007 borrowers might have used to stave off serious first-lien delinquencies in the first 18 months of their loan lives. Obviously any first-lien loan that basically requires the availability of high-CLTV HELOCs in order to perform for a year and half is not a "good" loan. I'm just not sure that more of that kind of loan was originated in 2007 than 2006. I think it's possible that more of them are getting "flushed out" earlier because of credit tightening in 2008 is putting a stop to their ability to limp along as earlier vintages did.