Playing with Fire: The CFPB's Proposed Repeal of the "GSE Patch"

CFPB recently put out an advance notice of proposed rulemaking to amend the Qualified Mortgage (QM) Rule by letting the "GSE Patch" expire.  What the Bureau is proposing is potentially very dangerous.  While I haven’t liked some of the Bureau's other proposed rules (including under the Cordray Directorship), none of them were an all-out ideological gamble with the economy. This one, in contrast, is really playing with fire.  

The QM Rule is a safe harbor to the Dodd-Frank Act's Ability-to-Repay requirement for mortgages.  (See some of my previous blogs about QM here, here, here, and here.)  Since 2011 it's been illegal to make a mortgage without considering the borrower's ability to repay.  The QM rule provides a safe harbor for compliance with the ability-to-repay requirement.  QM is a bit complicated rule—there are a few different ways to qualify and two different levels of safe harbor—but currently loans that are eligible for GSE purchase or guaranty do not have to meet three of the six basic QM requirements: a 43% debt-to-income ratio (DTI) cap, verified income/assets, and underwritten based on the maximum interest rate possible in the first five years.  This exception for GSE-eligible loans is known as the "GSE Patch."  

 

The GSE Patch is currently set to sunset on January 10, 2021 (before a new administration will be sworn in).  The GSE Patch was not originally intended to be permanent because the assumption at the time it was adopted in 2013 was that Fannie Mae and Freddie Mac would not still be in conservatorship and that GSE reform would necessitate a revision of QM. That assumption has failed, but the Bureau is still proposing to allow the GSE Patch to lapse.

 

It's not clear at this point if the Bureau will simply let the GSE Patch lapse without other amendments to the QM Rule. If it only lets the GSE Patch lapse, we're headed for real trouble in the housing market. If the Bureau lets the GSE Patch lapse, but amends the QM Rule (such as eliminating the DTI requirement or raising the DTI cap), the impact will be more muted.  

 

To understand just how risky it is to play with the GSE Patch, it's important to recognize that around 20% of the loans purchased by the GSE have a DTI ratio of greater than 43%.  That 20% figure will rise, however, if interest rates rise because monthly payments (the “D” in DTI) will go up accordingly. 20% of GSE loans translates to around 10% of the US mortgage market in recent years. Without the GSE Patch, it’s not clear if those loans would get made, and the CFPB admits as much in the rulemaking notice.  

 

There is a separate FHA/VA Patch to QM, which is not expiring, so some of the high DTI loans might migrate to FHA, which allows up to 57% DTI, but not all of the loans are within FHA conforming loan limits.  Moreover, those that do will then have to pay for FHA insurance, which reduces what borrowers can actually buy.  Moreover, most FHA/VA loans get financed through Ginnie Mae, which is already stretched to manage seller/servicer risk.  The Bureau suggests that portfolio lenders and private-label securitization market might pick up the slack, but all of that’s very wishful.  

 

Banks have been getting out of the mortgage business, and private-label securitization of residential mortgages is moribund for lots of reasons--it only financed only $29 billion of mortgages last year.  We're talking about perhaps $150 billion of mortgage loans that will need a new financing channel if the GSE Patch expires, and if they don't those loans don't get made.  Perhaps borrowers instead seek more modest loans that are within the 43% DTI limit, but that will reduce housing demand. 

 

Let's suppose that a quarter of the loans currently exempted under the GSE Patch don’t get made as the result of the expiration of the GSE Patch without any other amendment to QM.  That's a 2% or so drop in housing demand.  That's going to have a serious downward effect on home prices, and that's the sort of thing that ripples through the economy.  

 

What’s more, the expiration of the GSE Patch disproportionately harms minorities—black, Hispanic, and Asian borrowers make up a disproportionately high percentage of borrowers with >43% DTI. (These borrowers are already at a disadvantage with any DTI cap because a disproportionate share of them have 1099 income, which lenders tend to discount.)

 

So why is the Bureau doing this?  The Bureau does need to address the expiration of the GSE Patch in some way.  But the apparent motivation for letting it expire is to achieve “a more transparent, level playing field,” which I read to mean "shrinking the GSEs' market share."  That's WAY outside of the CFPB's lane.  The CFPB's job is not to regulate market share in the secondary mortgage market, and its proposed move starts to force Congress's hand on GSE reform.

 

Whatever one thinks about a GSE-dominated market, the optimal structure of the secondary housing finance market should have zero role in the CFPB's decision-making process about a consumer protection rule.  Not only is the Bureau going outside its lane, but it's neglecting its statutory duties: the Bureau is charged with ensuring consumer access to markets for financial products and services and with preventing discrimination in consumer finance markets. Yet the proposed rule cuts off access, particularly for minority consumers.

 

One can only speculate why the Bureau would chose to pursue a truly risky gamble with the US housing market, but I'll note that some of the key members of the Bureau's senior political staff previously worked for Jeb Hensarling when he chaired the House Financial Services Committee and kept up a Sisyphusian attempt to eliminate the GSEs and privatize the mortgage market, and that some conservative commentators have called for the elimination of the GSE Patch as a countercyclical measure to discourage excessive leverage. 

 

Neither the QM Rule nor GSE Patch is perfect. There are improvements that can be made to both, but they are generally incremental and modest. If one is worried about excessive housing leverage, for example, incorporating an LTV component into QM or revising the GSE Patch to have a DTI limitation would seem a better approach to throwing out the GSE Patch entirely. (I'm not endorsing either of those approaches, just illustrating more moderate moves.) This seems like a pretty drastic way to attempt to undertake countercyclical regulation, and the Bureau isn't justifying the proposal on that ground.  

 

Alternatively, one might think of eliminating the the 43% DTI cap, which would make the GSE Patch of relatively little importance, so its lapse would not matter. The effect of eliminating the 43% DTI cap, however, is to make another QM provision, a 150 basis point spread for the mortgage rate over the averaged offered prime rate, which is required to qualify for the strongest level of the QM safe harbor, have much more bite.  (Notice that this would underscore how QM operates as a de facto usury regulation.)   

 

Irrespective, any sensible approach would keep the GSE Patch largely in place until the GSE conservatorship is ended on the basis that creating instability in the housing market and the economy at large is not consumer protection. Whatever the motivation, letting the GSE Patch lapse is a really dangerous and irresponsible ideological gamble with the housing market and, by extension, with the economy as a whole.

 

I never thought that I would see an appropriate situation for the Financial Stability Oversight Council to exercise a veto over a CFPB rulemaking, but this might well be such an instance... other than that the Bureau doesn't have to actually undertake a rulemaking to cause a problem. Simply by failing to act and letting the GSE Patch expire, the Bureau could create real economic instability. My read is that regulatory inaction is outside of FSOC's veto ambit.   

 

Figuring out how to define ability-to-repay is tricky. But the starting point for the Bureau should be "first, do no harm." The regulatory equivalent of knocking out load-bearing columns and structural supports is not the way to do that.