The second mortgage market hasn't shut down, but it's shrinking fast. Which means that, on the other end of the see-saw, mortgage insurers are back and dictating the credit terms for primary financing on about 80 percent of a house's value.
This is a sea change for housing finance that, so far, has gone mostly unnoticed. It also explains why, in May, both Fannie Mae
[stock FNM][/stock] and Freddie Mac
[stock FRE][/stock] could abruptly reverse the maximum financing "haircuts" that they had earlier adopted for transactions in declining housing markets last December.
It's this simple: when a mortgage exceeds 80 percent of a home's value, the GSEs are required by charter to have credit support on the portion of the loan above 80 percent in the form of mortgage insurance, resource to the seller or seller participation greater than 10 percent. The most common form of credit support on higher LTV loans securitized or held in portfolio by the GSEs is -- what else? -- mortgage insurance.
Likewise, in the traditional lending markets of yore, lenders also commonly required MI on primary non-GSE conventional loans with LTVs over 80 percent.
As most of us now know, the traditional conventional mortgage lending paradigm of 20 percent down or MI (and full documentation, too) was eroded by growing acceptance starting in the mid-1990s of so-called purchase money seconds -- or "piggy-backs," as they've come to be called. Borrowers -- or their brokers -- could finance up to 100% (or more) of the traditional down payment with secondary financing in the form of a closed-end loan or a HELOC. Brokers and lenders aggressively marketed the so-called 80-20 (80 percent first, 20% second lien for 100 percent financing) and 80-10-10 (80 percent first, 10 percent second lien, 10 percent down) and other loan combinations as more affordable than MI.
The result? Piggy-backs became extremely popular.
Although GSE pool data doesn't tell us the extent of secondary financing in their loan universe, we do know from analysis of loan level data off private securitizations that piggy-back lending was a feature of much non-agency lending. For example, in first half 2006, data provided by UBS analysts in January of this year found that the percentage of piggy-backing recorded for prime ARMs was 31 percent, Alt-A ARMs 54 percent, subprime ARMs 35 percent, Option ARMs 30 percent, Prime Fixed 24 percent, and Alt-A fixed 26 percent. That's a lot. Only fixed rate subprime deals disclosed a relatively low 8 percent rate of piggy-back financing.
As lenders reached for more business, and borrowers reached for more house, leading lenders to reach for more business, subsequent vintages tended to have yet ever-higher percentages of piggy-backs. As piggy-backing expanded, MI volumes shrank. By one estimate, piggy-back lending ate over 40 percent of MI providers' previous market share before the loan market imploded.
Now lenders aren't eager to make seconds, for obvious reasons. The current rate of price depreciation is great enough in many housing markets to wipe out a second in a year or two; worse yet, there is no securitization market for them, and they're turning into deadly sludge on bank and thrift balance sheets to boot.
Detached seconds (a term used when the second is held by a different lender relative to the first) are also a problem for borrowers attempting to refinance -- if the second isn't refinanced, the lender will very likely decline to retain second lien position.
There is little quantitative evidence that secondary financing of home purchases and refinancings is down. But anecdotally we know it is. One second lien lender we know says 100 percent CLTV is still possible, but the underwriting is very tough and the documentation is scrutinized.
Freddie Mac's new loan and pool level disclosures do indicate a steep decline in piggy-backing (and leverage) in their universe (alas, Fannie does not disclose as richly). Based on the weighted average of original LTV and CLTV data across all Freddie Mac pass-throughs (ARM, Fixed, etc.) by origination month, borrowers' leverage peaked in 3rd quarter 2007 at 76.9 percent LTV and 80.2 percent CLTV, according to data provided by eMBS, Inc.
Tighter credit conditions have shaved both measures sharply since that time, to 70.1 percent LTV and 72.9 percent CLTV as of May 2008 originations.
Making our point -- that piggy-backs are fading -- the spread between CLTV/LTV was cut in half, from a peak of 3.8 percent in late spring 2007 to 1.8 percent as of May originations!
There are more hard numbers available to support MI's recent surge. MICA, the trade association representing the private mortgage insurance industry, began reporting rising volume monthly after February 2007. For example, mortgage insurers wrote 190 percent more business this year, through April, than in the comparable period of 2006, when subprime/Alt-A were in their heyday.
To put that sort of gain into proper context, consider that even GSE production is only up 160 percent -- and they are doing an estimated 80 percent of all new mortgage lending. By inference, MI providers have made huge
gains in market share.
The return of mortgage insurers doesn't mean a return to traditional lending standards, but it does make insurers the final arbiters of credit requirements for over 80 percent LTV.
This is important, so I'll restate: the responsibility for tightening credit standards for home financings levered more than 4x has now been taken out of the hands of the GSEs, who are subject to terrific political scrutiny and pressure. That responsibility now sits squarely with private mortgage insurers, whose various state regulators are charged with ensuring the MI companies can pay claims -- and not with putting a safety net under housing markets and household wealth.
And mortgage insurers have been steadily lowering their risk, cutting programs, raising rates and adopting "declining" (aka restricted, distress, and soft) market criteria that obviate the need for the GSEs to persist with their own -- and make it possible for the GSEs to avoid any political heat for so-called red lining (from consumer groups) or increasing risk to taxpayers (from GSE bears/hawks). Nor do the GSEs have to look like they are "pro-cyclically" restraining credit in the current crisis, when the Bush administration/Treasury/OFHEO have made it clear that "good" housing GSEs should behave in a "counter-cyclical" fashion, tightening credit and building reserves in boom times, while private lenders rake in the profits, and salvaging the market when private capital withdraws.
So, Freddie on May 2 issued a Seller Servicer Bulletin announcing that maximum LTV requirements need not be reduced for purchases or no-cash-out refis in declining markets on single-unit primary residences when the mortgage recieves an "Accept Risk Class" from its automated underwriting system, Loan Prospector.
On May 15, Fannie made a headline splash with its announcement of a Single National Down Payment Policy, rescinding its declining markets maximum financing rules, effective June 1. The announcement included an assertion that the new D.U. Version 7.0 -- which went into effect on June 1 as well -- "will limit risk layering and assess each loan more precisely." (DU 7.0 still can't handle jumbo loans, BTW; they must still be underwritten manually. Freddie Mac's LP platform was jumbo-ready when the programs first went into effect.)
To market insiders, this stuff about DU and risk laying really seems like posturing. Risk layering is about piggy-backs and stated-income, no-income documentation, especially when combined with loans MI won't insure -- things like 2/28 ARMs, short term IO, Option ARMs, and so forth. In other words, with MI now running the tables on low-equity loans, risk layering isn't happening anyway.
Editor's note: Linda Lowell is a mortgage market veteran, and principal of Offstreet Research LLC.
Disclosure: The author was long FRE, and held no other positions of interest, when this story was originally published. HW reporters and writers follow a strict disclosure policy, the first in the mortgage trade.