What’s this? A sudden wave of US housing optimism?
From HousingWire:
Serious delinquencies among US Alt-A residential mortgage-backed securities (RMBS) declined in April for the first time in four years, according to the latest data from Fitch Ratings.
Laurie Goodman, from Amherst Securities, is ready to quash it. In a 22-page note the analyst argues that US housing stats are massively distorted by mortgage modification programmes such as Hamp.
Here’s the basic problem:
The time from first delinquency to liquidation has been extending dramatically over the past 2 years. This is due primarily to a slow judiciary process and borrower-friendly government programs, which permit borrowers who have been delinquent for long periods to remain in their homes. We’ve also seen a huge rise in borrowers who are re-performing at the moment due to modifications (i.e., borrower was >60 days past due, and is now less than that), but the success rate on these modifications has been low. In fact, unless modification programs change substantially, it appears a near certainty that the vast majority of these modified loans will re-default.
By Amherst’s calculations once a loan becomes more than 60 days delinquent for the first time, borrowers have less than a seven per cent chance of catching up on it, on their own. About 20 per cent of them will get a mortgage modification and redefault at a rate of over 50 per cent per year.
As for the time to liquidation, it’s been stretched courtesy of mortgage modifications and a general overloading of foreclosures. For loans that liquidated in mid-2008, the time from the first missed payment to liquidation was about 15 months. Now it’s about 20.
Unsurprisingly then, the proliferation of mortgage modifications and the increased time lag between delinquency and liquidation can skew housing statistics (not to mention bank results).
Here’s what Amherst says:
The S&P/CS Index is great in that it is calculated from repeat sales; however, it does not distinguish distressed or REO sales from voluntary sales. As we all know, distressed home sales trade at a lower price than non-distressed home sales, which creates a fairly big impact on the indices as the contribution of distressed sales volume increases. Our data suggests that distressed home sales are generally suffering a 25% discount to non-distressed home sales. To the extent that the non-distressed sales component is a larger percentage of the total, home prices will appear to be firmer.
Exhibit 7 . . . shows two time series: the number of units in the private label securities market in REO and the number of units in serious delinquency or foreclosure. As you can see from that exhibit, the number of units in REO has been cut by approximately 50% over the past 18 months. Meanwhile the number of units in serious delinquency and foreclosure is up by nearly 3 times as much as the REO bucket has been cut. Homes are staying in the state of serious delinquency and foreclosure, and not making it to liquidation. As a result, distressed sales as a % of total sales is lower than it otherwise would have been. Exhibit 8 . . . is from the National Association of Realtors; it shows that total distressed sales, which were almost 50% of total home sales, are now ∼34%. If the loans moved through the delinquency foreclosure pipeline at the same speed they did 12-18 months ago the % of distressed sales would be much higher and the S&P/Case Shiller Price Index would likely be lower.
So basically the mix of distressed sales to non-distressed sales is impacting the index. The S&P/CS House Price index has improved but that’s likely due to an increase in non-distressed sales. If the mix changes again — say if distressed sales were to increase, the index could fall.
Amherst’s bottom-line:
We do not see housing as having bottomed. Extension of the liquidation timeline and the rise in modifications paint housing statistics a more rosy color than is really the case. Home price index stabilization is partially the result of a lower share of distressed sales, courtesy of the extension in the delinquency/foreclosure pipeline. Foreclosures have stabilized as a result of the slowing of the delinquency/foreclosure process. The drop in delinquencies is due to modification efforts. And, as we have shown, a large number of these will eventually default. And while transition rates are dropping, at least part of the drop is attributable to changes in loan composition.
It’s all very interesting but much of it we’ve heard before. In fact, some analysts have already argued that Hamp was very much about giving banks some time to build up their balance sheets while waiting for house prices to rise — extend and pretend and all that. Judging by the decline in real estate-owned, or shadow inventory, and bank NPLs peaking, it looks like it’s been a success in that sense.
But there are some loss considerations that perhaps haven’t been taken into account.
Yves Smith over at Naked Capitalism points out that mortgage servicers tend to advance interest payments and real estate taxes while a property is in default, and only recoup those costs when the property is sold. So a longer time to foreclosure could mean greater eventual losses for financials, if house prices and sales don’t meaningfully pick-up in the period.
Much more interestingly, Amherst also indicates that the stretched liquidation timeline can also mess up cash flows — and forecasts — for RMBS, the stuff still clogging plenty of bank balance sheets:
We also make the case that modifications and the extension of the liquidation timeline can distort cash flows in a RMBS securitization. That is, using current liquidation rates will give investors an expected loss number that is too low, as loans that have already transitioned and not yet liquidated are being ignored. Moreover, modifications not only alter cash flows on the loans, but also allow the servicer to recapture the principal and interest advances, which then lowers cash flows to the deal. These events do not affect all deals equally, as we have shown. Deals with shorter cash flows, and deals with few non-performing and re-performing loans are impacted less than other securities.
Full note, with much more technical detail, in the Long Room.
Related links:
Shadow bank losses – FT Alphaville
US homeownership minus negative equity = 61.6% - FT Alphaville
The slow death of Hamp, the summer of delinquencies – FT Alphaville
Hamp, what is it good for? – FT Alphaville