Recent Lessons for the QRM

On October 1st, the loan limits at which the FHA, Fannie Mae, and Freddie Mac can finance and refinance loans were lowered in 669 counties across the US.[1] The result was dramatic and provides a valuable insight into the impact of another potential change to the mortgage finance system: the qualified residential (QRM) mortgage rule.

The QRM rule is a set of requirements with which mortgages that are to be securitized and sold into the secondary market must comply.  If a loan within a securitization does not meet one of the rules, then the securitizer must retain 5% of the ownership in that securitization.  Any loss on loans in the non-QRM securitization would first be absorbed by the securitizer’s 5% ownership stake before impacting private investors who purchase the other 95% of the securitization.  In this way, the securitizer has “skin in the game”.  The group of regulators who wrote the proposed rule, including the Federal Reserve, the Treasury, and the FDIC, believe that the 5% retention will align incentives of securitizes with those of investors, a critical component of the mortgage finance system that was missing during the recent housing boom.  Over the last two decades, private securitizations have played an increasingly important role in directing funds from private investors to home buyers.  In the wake of the housing bust, the FHA and the GSEs gained much of the market share from private label securitizers.  However, if the government’s role in mortgage finance declines in the decade to come, as many policy makers hope, private securitizations will once again grow in importance.  Thus, providing the transparency, recourse, and aligned incentives that protect investors from undue and unseen risks is important.

An overly stringent set of rules could hurt the economy, though.  Excessive downpayment requirements or high mortgage rates will inhibit homebuyers from utilizing the flow of funds from the private sector.  Sluggish demand for housing in turn hurts the economy, which historically has averaged roughly 19% of the US’s annual GDP, but has recently fallen to just 15%.  Thus, the final QRM rule must balance the desire to improve safety to investors with the need for economic stability and growth.

Since an overly rigid QRM rule could weigh on the economy, it is important to understand how large of an impact it would have.  The recent change to the loan limits at which the FHA and the GSEs can finance mortgages provides a valuable insight.  Those borrowers impacted by the change now face at least some of the impacts that a non-QRM qualified borrower would face.  For instance, the downpayment requirement for a person who qualified for an FHA loan prior to the change in limits jumped from 3.5% to an average effective downpayment of 17.7% or more[2] with a private loan or GSE backed mortgage.  Likewise, the rates faced by borrowers in high cost areas have increased.  A survey conducted by NAR Research in the 669 counties affected by the change in the limits found that the average increase in mortgage rate for an affected borrower was 79 basis points. This trade off, larger downpayment or higher mortgage rates, will be the same choice faced by borrowers if the proposed QRM rule becomes law.

According to the same NAR survey results, 16% of borrowers in the affected areas who fell into the impacted price ranges chose to give up their home search as a result of the higher rates or increased downpayments.  To estimate the impact of QRM, this figure must be adjusted by the share of the market covered by the FHA, which is exempt from the rule, as well as the share of non-exempt loans that will qualify as a QRM.  Roughly 18% to 22% of loans financed by the GSEs between 1997 and 2003 would have qualified as a QRM according to the FHFA[3] and the FHA is exempt from the rule, which will drive more borrowers that direction.  If the FHA maintains a historically high 25% share of the mortgage market, these assumptions would imply that 9.6% of sales would fall out under the QRM.[4] NAR Research has estimated the impact on rates from the proposed QRM rule to be in a range of 85 to 180 additional basis points.  The 79 basis point impact and effective increase in downpayment from 3.5% to 20% from the loan limit change would roughly match the lower end of the estimated impact of the QRM rule.  On a national level, a reduction in home sales of 9.6% would translate to roughly 480,000 fewer home sales.  This assessment is in line with estimates from other groups like Moody’s Analytics[5] of 423,000 fewer home sales for a 100 basis point increase.  Such a decline would have a depressing impact on the housing market and economy, stifling demand which could place downward pressure on prices and possibly touch off another round of delinquencies and foreclosures.  What’s more, the 9.6% lower sales rate would not be a one-time event, but would drag on sales in every year that the QRM is in place.

The impact of the QRM rule could be far worse.  NAR’s estimate of an additional 85 to 180 basis points is at the middle range of current estimates for the QRM’s impact.  The American Securitization Forum, the trade association representing both securitizers of asset backed securities as well as the investors who purchase them, estimated in their comment letter to the regulators on the QRM rule that the impact could be as high as 200 additional basis points.[6] Subsequently, Moody’s Analytics estimated that more stringent language from the regulators on some of the nuances of future securization structures suggests that the impact might be even higher, in the range of 100 to 400 additional basis points, and would create a disincentive for banks to originate 30-year fixed rate mortgages.[7]

Finally, the FHFA’s historical estimates of the share of borrowers impacted by the QRM may understate the case.  According to statistics from NAR’s Profile of Home Buyers and Sellers, the share of repeat buyers who relied on savings for their downpayment rose from 40% to 59% between 2005 and 2011. Simultaneously, the share of repeat buyers who used proceeds from the sale of their primary residence fell from 66% to 41% over this same time period while the share of repeat buyers who used loans from their 401k/pension or gifts increased.  This shift toward dependence on savings swung beyond the pre-boom share of 51%, while proceeds from the sale of a prior home are also historically low.  The loss of equity from recent price declines has impacted repeat buyers’ ability to save and slow price growth over the next half decade will likely hamstring housing equity as a conduit for trade up buying.  Furthermore, the use of gifts and inheritance as well as the tapping of retirement savings like 401(k)s, IRA, and stocks has increased and in the latter case was much higher in 2011 than in 1997.  The heightened downpayment requirement under QRM could accelerate the diversion of funds from retirement to the downpayment.  The high downpayment requirement of the QRM would hurt repeat buyers and not just first-time buyers.

In the QRM rule, regulators must strike a balance between protecting investors, the banking system, and the flow of funds to homebuyer, while not stunting the economy’s ability to grow.  In the years to come, the QRM rule if enacted in its current form is likely to have a substantial negative impact on the housing market and the economy.  The recent change to the loan limits at which the FHA and GSE can finance mortgages provided a litmus test for the magnitude of the QRM’s impact, which suggests that it could undermine a housing recovery and create headwinds to economic growth.

[1] The limits were subsequently raised for the FHA, but not for the GSEs

[2]Based on the average county level decline in the FHA loan limits for the 669 affected areas and FHFA data.

[3] FHFA’s Mortgage Market Note of 3/31

[4] 100%-25% = 75% is the non-FHA share of the market; 75%*(1-20%) = 60% is the non-FHA, non-QRM qualified portion of the market, where the 20% is the average of the 18%-22% pre-boom/bust QRM share estimated by the FHFA; thus 60%*16% = 9.6% is the portion of the total market that will drop out under QRM


[6] page 49


Ken Fears, Director, Regional Economics and Housing Finance

Ken Fears is the Manager of Regional Economics and Housing Finance Policy. He focuses on regional and local market trends found in the Local Market Reports and the Market Watch Reports . He also writes on developments in the mortgage industry and foreclosures.

More Posts

  1. Thomas Lawler

    It would seem that the logical solution would be to eliminate the concept of “qualified residential mortgage” completely — it never made sense to begin with. If an issuer has to retain 5% of the risk of a pool of “low risk” mortgages (NOT defined by some arbitrary QRM standard), then the issuer retain 5% of a “low” amount. If an issuer has to retain 5% of the risk of a pool of “high risk” mortgages, …

    This whole QRM issue emerged because various folks argued that issuers should have “skin in the game,” but various lobbyist “pushed” for exemptions. Regulators, seeing that the “skin in the game” issue was a desired feature, but faced with the “silly” insertion of the concept of exempted “qualified” residential mortgages, were faced with a dilemma: how do you define a “QRM” that is SO devoid of ANY risk, that an exemption to the desired risk retention rule would be “ok.” It actually is no surprise that regulators defined a QRM so restrictively, but it has led to a disastrously “marketing” result; loans so devoid of risk that risk retention, desired in the vast bulk of transactions, was defined to include only a modest share of mortgages!

    The obvious solution is to completely get rid of the concept of “qualified residential mortgages,” or QRMs. As noted before, issuers with a pool of low risk mortgage retain 5% of very little risk. That is not a BFD (big financial deal).

    Potential issuers argue that any risk retention under the “old” private label market “is tough,”
    as it’s harder to “arb” the old credit ratings and hard to “do deals” if you actually have to hold ANY capital against the securities you issue.

    But, as we all learned last decade, the old private label market model was “built to fail” and should be completely scrapped, and a new model needs to be built. It can work, but it can’t rely on “arbitraging” faulty credit rating models and can’t have the “conflicting” incentives inherent in the old model that have HAMPered the ability to deal effectively with the millions of delinquent loans created last decade.

    If the NAR is arguing that the whole concept of the QRM be thrown out, I am on board.

  2. T.J. Doyle, Marketing & Communications Manager

    Good morning Tom. Pasted below is a reply from the author of this post:

    Thank you for the question. Let me first clarify your point. You have argued that the QRM, which acts as an exemption from the 5% risk retention rule, should be scrapped, so that everyone, no matter how good their credit or how large a downpayment they make, would have to pay the higher mortgage rates under a retention scheme.

    No, NAR is not arguing that the QRM should be scrapped. Rather, the NAR is arguing that the LTV and DTI requirements of the QRM be eliminated, thus expanding the exemption. Removing the exemption to the risk retention (e.g. eliminating the QRM) would actually expose the entire market (rather than just 80% of the non-FHA portion) to the higher cost of retaining risk. In addition, it would expand the monopoly of the handful of securitizers with portfolios large enough to handle 5% risk retention. As you point out, the borrowers who qualify under the current QRM are the least risky, so it makes little sense to impose those costs on them. In addition, eliminating the DTI and LTV widens the exemption to a large pool of borrowers who already meet the heightened credit requirements and stability conditions (e.g. no interest only, negative amortization loans etc.) of the QRM rule.

    It’s important to remember that the base requirements of the QRM will rule out risky, exotic mortgages, high rate re-sets, and will raise the credit requirements to a high bar. Both DTI and LTV have an impact on mortgage delinquencies, but less than other factors like credit quality. Furthermore, the QM statute should help to clarify a borrower’s ability to repay a loan up front as well as make the details of the origination more transparent for investors, making the investor less dependent on the ratings agencies.

    No system is perfect, but we can’t ignore the mistakes of the past. Yes, the regulators are incentivized to focus their attention on making the QRM as tight as possible, but they must be careful not to create a substantial drag on the economy or new risks in doing so.

  3. Tom Lawler

    A risk retention rule does not necessarily create a “monopoly,” and should not, if structured correctly, result in much of an increase in mortgage rates.

    Consider a very simple example: an issuer acquires/originates $100 of “low-risk” mortgage with a servicing fee of 0.25%, and plans to issue a straight pass through security (carving up cash flows creates conflicting incentives, as evidenced by the horrific problems in dealing with troubled loans in exisitng private label securities) Suppose the issuer believes that investors will purchase this security at a yield of, say, 4.25%. If the issuer must retain 5%, what would the issuer have to charge borrowers?

    Well, if investors get 4.25% on the $95 sold to them, and servicers get 0.25% on the $100, if the issuer/originator “charged” borrowers, say, 4.625%, then here’s the “yield”: breakout:

    Borrower cost: 4.625% on $100
    Servicer: 0.25% on $100
    Investor: 4.25% on $95
    Issuer: 6.75% on $5

    (100*.25% + 95*4.25% + 5*6.75% = 4.625%)

    So, on a very low risk pool of mortgages, the issuer would, by charging borrowers 0.125% more, would get a “yield” on its 5% retained of 6.75%, more than 150 bp more than private investors in the security would get for the same risk!

    By the way: it is a fallacy to say that “exotic” mortgages are the only mortgage that are “high risk.” E.g., FHA-insured loans endorsed in fiscal years 2007 and 2008, the vast bulk of which are 30-year fixed rate mortgages, have a serious delinquency rate of 20%+, and the recent actuarial study projects that ultimately the FHA will pay claims on 22.4% of 30-year fixed rate mortgages endorsed in 2006, 25.5% on 30-year FRMs endorsed in 2007, and 20.8% of 30-year FRMs endorsed in 2008.

    These were low down payment/high DTI loans — but 30-year fixed rate mortgages!

  4. Thomas Lawler

    The risk retention rule, of course, would not result in materially higher mortgage rates, OR an increasing “monopoly” (oligopoly?) of originators/issuers.

    E.g., If an issuer/originator were to originate $100 of “low risk” mortgages and had to retain just 5% of the loans/risk, and if that issuer were able to sell a security backing that low risk mortgage pool to investors at, say, 4.25% (with 25 bp going to servicing), then if that issuer were to charge 0.125% more to borrowers for the small amount of risk the issuer was taking, then that issuer would retain a low risk pool of mortgages that would yield the issuer 6.75%, vs. the 4.25% that private investors would get on their 95%.

    Of course, it is a fallacy to say that only “exotic” mortgages have performed poorly and been high risk. Serious delinquency rates on FHA-insured mortgages originated in 2007 and 2008 are over 20%, and a recent actuarial study suggests that the FHA will have to pay claims on over 20% of 30-year fixed rate mortgages it insured/endorsed in 2007 and 2008. These were low down payment and for the most part high DTI loans, but they were 30-year fixed-rate mortgages.

  5. T.J. Doyle, Marketing & Communications Manager

    Good afternoon Tom. Pasted below is a reply from the author:

    Yes, oligopoly is the proper descriptor, but I think the readers of this site are likely more familiar with the concept of a monopoly and monopolistic pricing, hence the word choice.

    As for whether an oligopoly would occur, the expansion of the spread between rates in the primary and secondary market since 2007, a period of consolidation among mortgage originators, suggests that the presence of fewer market participants would result in oligopolistic pricing pressure. Why would there be fewer securitizers under QRM? The increase in portfolio requirements to retain the 5% piece of a security would limit the opportunities for smaller securitizers.

    In the current environment of low growth and returns, your example of a “low risk” securitization might work. However, as the economy improves, returns will improve and the opportunity cost to the securitizer of holding the 5% in retention at 6.75% will grow. From discussions with market participants, I’ve gleaned that the return on equity expected by securitizers prior to the recession were between 10% and 15%. I’ve also been told that 10% is the rate at which REITs will jump in. That is important because of the need for a large number of market participants to reduce the oligopolistic pricing issue. So, the securitizer would likely demand a greater ROE on the security than the 6.75% that you suggest. There is also a liquidity issue. Even for the most riskless borrowers, without some standard that distinguishes superior quality, securitizations trade at a discount relative to those with the perception of less risk (e.g. GNMA vs. FNMA). In addition, would banks choose to forgo other fixed investments on their portfolios in lieu of the retain piece? If not and even if they did, they might require a liquidity premium since they cannot sell these for cash in a pinch. Finally, there still remains the question of why charge the least risky borrower more if, as you suggest, they are no BFD (big financial deal)?

    As for the exotics, those structures (e.g. 2-28s, negative ammortizations, etc.) are high risk and the rule recognizes this and accounts for it. They are not the only high risk mortgages as you point out. However, the high delinquency rates on FHA issuances in 2007 and 2008 reflect fraud. The seller funded downpayment assistance program (SFDPA) was at its peak during this period and represented roughly 23% and 19% of loans in those two years, respectively. Under this program, the seller could finance the downpayment for the borrower through a third party. As early as 2005 the GAO produced a study on this issue. The GAO found that the SFPDAs resulted in fraud as the price of the home was bid up by the amount of the seller funded downpayment, inflating the price of the home above its value. This program was ended in 2009, but its legacy continues to weigh on the FHA’s books. In the 2010 FHA actuarial, it was pointed out that SFPDA loans have a default rate 2-3 times that of non-SFDPA, FHA loans. So, yes the FHA did have high delinquency rates on 30-year loans in 2007 and 2008, but that result was skewed by a large share of loans that involved fraud which has since been rectified even though the problem lingers on the FHA’s books. The GSEs did not allow this type of seller assistance. This is one area where the QRM/QM might be improved…an explicit prohibition of indirect seller assistance through non-profits that boost the sales price above value.

  6. Thomas Lawler

    Don’t mean to be a stickler, but I think you characterization of third party down payment assitance resulting in a higher price for the property as “fraud” is a little strong. It is true, of course, that if a borrower gets down payment assitance and as a result the borrower puts nothing at all down, the borrower may be willing to pay more for the property. But “fraud” is a strong word. Not surprisingly, claims rate for other down payment assisted purchases have also been higher than “no-gift/assistance” FHA loans, where there was no “fraud,” but simply that the borrower had no equity at the start.

    It would seem as if what you REALLY want is no risk retention, as opposed to some exemption, even for loans with little down and high DTIs — which have been characterized and documented as “high risk” relative to “typical” loans based on data from the last several decades. That, by the way, is a more reasonable argument that exempting certain loans from the risk retention rule: as I’ve said before, if you have to retain 5% of a pool of low risk mortgages you retain very low risk (5% of “low”), while if retain 5% of a pool of high risk mortgages, you retain more risk. (5% of “high”)

    Given that there really is no merit to the argument for an exemption to the risk retention rule, let’s focus on the more logical argument you might want to make: the risk retention rule will not be effective, and should be scrapped.

    If the argument that having originators have “skin in the game” is not effectively dealt with by a risk retention rule — which a number of analysts have suggested — what are the alternatives? Stronger regulations/more effective enforcements of existing laws? Higher capital requirements for orignators combined with no-exceptions enforcements of reps and warrants? I don’t know the right answer, but you might want to consider an alternative strategy/argument, as the “QRM” argument from a math perspective just doesn’t make any sense

  7. T.J. Doyle, Marketing & Communications Manager

    Good morning Tom. A reply from the author:

    This is really a side point, but I don’t want to leave the impression that all assistance programs are troubled. Claims rates for loans that received no-gift/assistance were lower than those who received gifts from family members or loans from governments, employers, and non-profits. However, the claims rates for assistance from non-profits were significantly higher than those from governments and employers. The GAO analysis posited that the reason for this observation was that the prices of the homes with seller-funded downpayments were bid up over their value, which was not the case for other types of assisted loans. In addition, the share of FHA originations that were from employer and government assistance programs was less than 2% in 2008 vs. nearly 19% for non-profit.

    NAR is in favor of a QRM with no LTV or DTI. While those factors are important in determining risk, they are not the only or most important factors. The QRM/QM handles these other factors and improves transparency. We can’t eliminate all risks from the system, but we can reduce it significantly through more traditional underwriting and improved transparency, while not over-burdening the economy.

    Scrapping the risk retention is an alternative path, but I agree, there must be another mechanism that effectively aligns the incentives of buyers, lenders, securitizers and investors in place to remediate the problems of the last decade.

  8. Thomas Lawler

    As I noted before, the QRM exemption concept makes no sense though the risk retention “scrappage” may have merit, if there are other ways to ensure that originators have a stake (skin in the game) on the loans they originate.

    But on the “fraud” issue: HUD reported last year that loans with “high” seller contributions (NOT third party) had materially higher claims rates that loans with “low” seller contributions, which makes senses in that sellers would not contribute a large $ amount to closing costs, etc. if the seller were able to sell the home to someone else at the same price without paying concessions — meaining, of course, that high seller concessions inflate the selling price.That was HUD’s conclusion. Based on that data, HUD rightfully said it wanted to scale back the allowable seller concessions from a silly 6% to a still too high 3%. The “correct” number, of course, is zero.
    And on “fraud”: do sellers that make large concessions/paying closing costs in order to sell the home at a higher price than would otherwise be the case engage in “fraud?” Or are they just playing by “rules” that don’t make prudent lending sense? If I understand your logic, a seller who pays closing costs/makes other concession payments in order to sell at a price higher than would be the case if the seller did not pay closing costs/other concessions is engaged in “fraud.” Oh, my

  9. Sandra K

    I don’t get it. Again, Congress is trying to ensure that Banks make sound loans that are not so apt to blow up, by forcing them to hold 5% of the loan amt back from securitization as their own risk in that loan. But you’re saying that the consumer now has to put that much more $$ down to entice the Banks to make the loans so that in fact they are playing with somebody else’s money!? If they are fully funded of their 5%, then they will be fat & happy as they have their profit assured without long-term worries. So again the banks get fatter & more profitable on the backs of the consumers and our troubled real estate economy. As for me, I’d like to see an alternative to the VERY PROFITABLE banks that won’t ‘Vampirize’ consumers. Ideas anyone? Securitizing residential loans will continue the confusion that MERS created, with nobody knowing who owns their loan, nor ensuring that when they make their payments, the correct investor gets their payment. The Good Ole Days were not so bad after all; interest rate & price stability; you could find & contact the holder of your note, and EVERYBODY had a stake in the Borrower making their payments.

  10. Ron Napier

    I think what lenders may be doing, is looking for more Equity, in case they have to take the property back. More equity would help them in absorbing the 5%. In other words, if it was a 100% loan like before, and they take it back, they will probably have to resell it for less, and will eat up their original 5% for sure. More equity reduces that risk, and they may even recapture their 5%. So it’s just a Game. Lenders are sandbagging the consumer. Any type of “Lender Securitization” should be tied into the Appraised value, by comment in the law.