I'm in DC to kick off this week, attending a PRMIA and FDIC Securitization Symposium that will surely fuel future columns -- but wanted to spend this one helping readers make real sense of last week's delinquency numbers out of LPS Applied Analytics.
There were some signs of life in the report, which saw mortgage delinquencies fall for the second month in a row--and most importantly, saw new rolls into delinquency drop sharply, as early-stage delinquencies also cured at a high rate. In fact, the rate of cures-to-current for loans less than 60 days delinquent hit an all-time high in March 2010, according to LPS data.
But, like usual, not all is as it seems here.
If you'll recall, last week, we took a trip through Wonderland as it related to housing prices. This week, we're going to jump down the proverbial rabbit hole again, but to look more deeply at foreclosure and delinquency statistics.
So, the real good news here? If anything, it's this: we're seeing seasonal effects re-emerge in the delinquency data.
Delinquencies almost always tend to decline month-over-month in February and March of each year, as households use their tax refunds to play catch-up and/or stay ahead of the collection calls (and as holiday foreclosure and REO eviction moratoriums expire, too). We saw this effect last year, actually, in most product types, too; the seasonal pattern in the DQ numbers had fallen apart in 2007 and 2008, however, as mortgage loans were going bad at what seemed to be a parabolic rate each and every month.
So, the fact that we've seen seasonal patterns re-emerge for two years now is actually good news, at least relative to what we had been seeing. But I fear that the seasonal tax effect might be enough to explain the surge in early-stage cures, as well as the drop in current-to-30 day delinquency rolls. Look for delinquencies to resume their upward trending later this year.
And despite the two-month downtrend garnering most of the press, it's important to note that overall delinquencies (excluding foreclosure) were up 15.7 percent year-over-year in March, while foreclosure inventories surged 32.9 percent. So I'm not sure that the WSJ deciding to report the LPS data as "encouraging" is truly the tack I would have taken; after all, comparing unadjusted monthly data and ignoring annual trends tends to be the epitome of economic folly.
Now, add this word to your vocabulary: recidivism
March 2010 will also be remembered as the month that the proverbial backlog of distressed loans began to work its way through the system -- LPS researchers clearly noted that the HAMP loan modification program has finally begun to lurch ahead, as servicers moved borrowers out of trial status and into permanent mods at an increased rate in March (much to the chagrin of servicer-chiding Treasury Secretary Timothy Geithner, I'm sure).
That said, I've always had a problem with the idea that a trial modification converting to performing ought to be called "permanent." That's a figment of slick political maneuvering, because it suggests to the public that a "permanent modification" is just what it says it is.
Nothing could be further from the truth.
Traditionally speaking, loan modifications have tended to kick the can of a borrower default a little further down the road -- 60% or so of loans successfully modified would re-default within 12 months, so a loan modification would help some borrowers, but not most borrowers. (And from an investor's perspective, it helped investors in the equity tranches of a given deal get a few more months of cashflow before being wiped out, but that's for another story.)
The investment term for this is recidivism, and we've seen it time and time again with government-led initiatives to modify loans, including the much-publicized FDIC takeover of IndyMac Bank and associated loan mod mandates stemming from the government's operation of the failed bank.
So, how are HAMP modifications performing? We don't yet know. For one thing, the Treasury isn't releasing recidivism data on permanent HAMP modifications. For another, it's too early to tell, even if the data was being released.
That said, I'd expect two things from HAMP this year: more "permanent" modifications, and a recidivism rate that mirrors what we've seen in other previous programs -- not because HAMP is a poor program, per se, but simply because it's a loan modification program nonetheless. Reduce interest rates, extend term, forbear principal. It's not rocket science.
I bring up recidivism because when we look at loans in delinquency falling, as we've seen, it's not entirely due to bad loans working their way through the pipeline and not entirely due to the tax effect I discussed earlier. The LPS data makes it amply clear that much of the drop in delinquencies is due to loans in the late-stage delinquency pipeline finding a cure -- and even if the modifications here outperform industry averages, history tells us that many of these loans will at some point re-enter the delinquency pipeline within the next 12 months.
How much? When? And what does that do to forecasting efforts? I'll save that discussion for the next rabbit hole. For now, I have some very smart people to go listen to here at the FDIC.
Paul Jackson is the publisher at HousingWire Magazine and HousingWire.com. Follow him on Twitter: @pjackson