Ginnie I, II, or...III: Combining the Programs and the Advantages to Originators; Credit Unions Weigh in on QM
Among the nation's 30 largest metro areas included in a recent Zillow study, Pittsburgh (39%), Tampa (33%), New York (30%), Cleveland (29%) and Miami (29%) had the highest percentage of free-and-clear homeowners. It appears that areas with lower home values generally have higher outright homeownership rates, as smaller loan amounts are easier to pay back more quickly, but the age and credit rating of primary borrowers also influence free-and-clear homeownership rates. Zillow found that 65- to 74-year-olds are most likely to be free-and-clear (21%), followed by 74- to 84-year-olds (18%). Interestingly, when examining free-and-clear ownership rates as a percentage of homeowners in various age groups, Zillow found 35% of 20- to 24-year-old homeowners are free of debt.
And one last unsurprising housing stat for Realtors: price and proximity to work are key concerns for first-time homebuyers, while trade-up buyers tend to be most focused on the design of the home and the neighborhood, according to "Characteristics of Home Buyers," an analysis of the recently released 2011 American Housing Survey (AHS) by the National Association of Home Builders.
Does whether or not a loan meets CFPB-mandated Qualified Mortgage (QM) criteria matter? You bet it does, and in the future, if a borrower needs a loan that fits into the QM box, they may not have much luck at the local credit union. Mortgage News Daily reports that "nearly half of members responding to a National Association of Federal Credit Unions survey say they are already planning to drop loans from their product line that do not meet Qualified Mortgage (QM) guidelines. Members were asked about a range of impacts anticipated under new Dodd-Frank Act rules being implemented by the Consumer Financial Protection Bureau in a survey conducted for the May Economic and CU Monitor. Respondents said they were getting a head start on complying with the new Ability-to-Repay/QM mortgage rule. A large majority, 92.9 percent, said they have seen regulatory burdens rise under the rule and 88.1 percent reported increasing compliance costs." Read all about it.
There are renewed concerns/hopes regarding the possibility of reintroducing the Boxer-Menendez bill with a the provision to extend the HARP eligibility cut-off date to May 31, 2010 from the current May 31, 2009. Although the original version of the Boxer Menendez bill included extending the eligibility date to May 31, 2010, significant pushback from mortgage market participants had resulted in the bill being amended to remove the eligibility date change. It now appears, however, that the Boxer-Menendez team is considering re-introduction of the language around the eligibility date extension with possibly some restrictions around re-HARPing. Their thought process appears to be that adding a provision to limit re-HARPing addresses some of the common concerns expressed by the industry on extending the HARP eligibility date to May 31 2010. Before the champagne corks start popping among the originator ranks, experts and analysts believe that the bill with an extended HARP eligibility cut-off date has a low probability of passing the House. It is important to note that the FHFA has the power to change the eligibility date without legislative action. With the nomination of Rep. Mel Watt to be the FHFA Director, there is elevated concern in the market that the HARP cut-off date may be extended even without legislative action.
I know that I am dating myself, but for most of my career in the secondary markets FHA & VA loans were put into Ginnie Mae I securities (aka Ginnie 1's, GNMA I, and GN1), and those loans with odd coupons or buydowns were put into Ginnie Mae II's (GNMA II, GN2). Historically, GN1 issuance generally outpaced GN2 until 2010, but during the height of the financial crisis, this dynamic reversed. Beginning in mid-2009, GN1 volumes began to decline significantly, while growth in GN2 accelerated. As proposed capital constraints under Basel III made holding excess servicing less favorable, the note rate flexibility provided by the GN2 program was likely a key reason driving this shift in issuance toward GN2s. In fact, since mid-2011, monthly GN1 production has averaged less than $5 billion compared with more than $22 billion per month for GN2 ($25 billion per month since January 2012). Why even offer a Ginnie I program?
Low GN1 production volume (about 15% compared to GN2) is likely one of the key reasons for GNMA's exploration of winding down the program. GNMA is only in the preliminary stages of seeking industry input on the feasibility and potential paths forward in phasing out GN1, and perhaps combining it with GN2 into a single MBS issuance platform. There are a lot of entities with a vested interest in this, including lenders, investors, and broker dealers, and of course they all want their voice heard. But one can't miss the correlations between this discussion and the talk of a single platform for Freddie & Fannie loans, often advocated by the MBA.
There was a meeting last week between GNMA and SIFMA. GNMA must determine what is best for the market and then figure out how to operationalize its plan. Ultimately, GNMA will need to obtain regulatory approval, which will lengthen the implementation timeline. The idea is to create an MBS program that is responsive to the market and provides the best liquidity possible, and any modernization would, ideally, build on the best attributes of the two existing programs by probably using the GN2 program as the template.
Moving to a single program based on GN2 has several advantages. One key benefit of the GN2 program is that only a single large pool per coupon is produced each month (this excludes smaller "custom" pools that are also originated monthly), so prepayments for GN2 pools generally have very little idiosyncratic risk. This also limits adverse selection for TBA delivery, though this comes at the expense of worse liquidity and execution for specified pools.
But, as analysts from Barclays and Nomura point out, the GN1 and GN2 programs differ in several key respects, and ironing those out will take some clever work. For example, the loans in a given GN1 pool must all have exactly the same underlying mortgage rate, which must be precisely 50 basis points higher than the net coupon of the pool (i.e., 4% loans go into a 3.50% security). But for GN2s, the note rate on each underlying loan need not be the same; in fact, it can be 25 basis points and up to 75 basis points above the pool's coupon rate.
There are multi-issuer pools for each coupon every month. In addition, GN2 allows for the issuance of "custom" pools, typically single-issuer pools that do not meet GN1 guidelines (see paragraph above). Each program has a different payment date and, thus, different payment delays. (Remember that just because the borrower sends in their payment on the 1st doesn't mean the ultimate investor receives it on the 2nd - there is a delay that benefits the securitizer in earning a little interest income.) GN1 investors receive payments on the 15th of the following month and GN2 investors them on the 20th, resulting in payment delays of 45 days and 50 days, respectively. Receiving principal and interest five days early is worth only approximately one tick (.03125), but it will have to be resolved. Most believe that the GN2 program offers several advantages over GN1 for issuers and investors, including increased flexibility in loan origination, less idiosyncratic prepayment risk, less adverse selection for the TBA deliverable and generally better float and liquidity.
Does this mean anything to borrowers or loan originators? Yes, but indirectly. Existing price differences, often difficult to explain, might eventually go away, especially between incremental mortgage rates on rate sheets. So the current pricing differences between GN1 and GN2 arising from technical and fundamental factors would go away if an optional conversion mechanism to exchange existing GN1s into the new combined platform security was worked out. Although Ginnie II issuance has been close to 85% of all GNMA MBS issuance, Ginnie II pass-through are still priced off Ginnie I pass-through, which are in turn priced off FNMA MBS pass-through! Stop the madness! This mechanism introduces an additional layer of unnecessary bid-offer spread in the GNMA market from investors' perspective. If the two GNMA MBS programs are combined, GN IIs should be directly priced off FNMA MBS.
Overall, if this takes place, there appears to be more flexibility for originators. Because the GN2 program allows more flexibility on loan terms, originators can offer borrowers a wider selection of note rates and target loans for delivery to specific coupons for the best execution - just like for Fannie & Freddie loans now. For example, issuers may deliver a 3.75% mortgage into a GN2 3.5% or 3% MBS. Such "coupon slotting" gives issuers the ability to selectively deliver mortgages into various coupons to garner the best pricing. And on the servicing side, there will be less excess servicing. Remember that by definition, GN1 has 19 basis points of excess servicing because the weighted average coupon of the underlying mortgages must be exactly 50 basis points above the GN1 coupon, as mentioned above. (After taking out the 6 basis point guarantee fee and mandatory 25 basis points servicing fee, the remaining excess servicing is 19 basis points.) GN2 issuers, however, can actively manage the amount of excess servicing they book each month through note rate flexibility. This may be attractive to issuers seeking to minimize capital effects of excess servicing.
The market thinks that combining the two GNMA MBS programs is a step in the right direction for improving liquidity/float situation in the GNMA market. The matter is not inconsequential given the amount of existing securities (about $1.2 trillion) and securities being created every month - just like combining Fannie and Freddie securities but on a smaller scale. A major portion of recent issuance of GNMA MBS has gone into CMO deals or has been purchased by the Fed, domestic banks and overseas investors, but it is generally thought that the liquidity/float advantages associated with having a single GNMA MBS program easily overwhelm the disadvantages of eliminating the GN I program from new issue perspective.
So what are the options? Stay tuned for Monday's edition!