The non-qualified mortgage (non-QM) sector is the fastest-growing segment of the U.S. private-label residential mortgage-backed securities (RMBS) market, with issuance expected to roughly double in 2019, to $25 billion. S&P Global Ratings has assigned credit ratings to the majority of the non-QM securitizations issued over the past several years by almost 20 different issuers. As such, we have analyzed loan and pool data for a large portion of this space.
State Of The Mortgage Market
The rise of non-QM
The concept of the qualified mortgage (QM) and the related ability-to-repay (ATR) rule were introduced by the Consumer Financial Protection Bureau (CFPB) in early 2014 as mechanisms to protect borrowers from risky loan products and practices. One of the consequences of the new rules was that many would-be buyers found themselves unable to obtain home financing because their borrower profiles did not meet conventional guidelines. This gave rise to a market for individuals whose underwriting characteristics fell outside the QM specifications. By 2015 a new niche residential mortgage-backed securities (RMBS) segment emerged, and it has since been dubbed "non-QM." Because of its novelty and rapid growth (see chart 1), there has been great interest in non-QM by both market participants and the media.
Chart 1
Non-agency RMBS making a comeback
The non-agency RMBS market makes up less than 10% of overall annual RMBS issuance volume, down substantially from pre-crisis days when the market share of new issuance peaked at roughly an even split with the agencies. The non-agency market is growing, however, having increased to $70 billion in 2017 and $86 billion in 2018 from $34 billion in 2016. We are forecasting roughly $100 billion in non-agency RMBS issuance by year-end 2019 and, of this total, we expect non-QM to make up 25% (see chart 1). If this is realized, non-QM issuance will have surpassed that of other non-agency RMBS sectors, as well as government sponsored enterprise (GSE) issued credit risk transfer (CRT) deals, as shown in chart 2.
Chart 2
As the non-QM RMBS sector grows, we expect continued comparisons to mortgage products that were popular more than 10 years ago. (In our Dec. 4, 2017, paper, "How Do Non-QM Loans Stack Up Against Pre-Recession Mortgage Products," we provide a collateral-level breakdown examining the extent to which non-QM resembles Alt-A, subprime, and prime jumbo over two different pre-crisis periods.)
The Increase In New Issuers And Originators
As the non-QM sector becomes more popular, more issuers are entering the space. Only a couple of years ago, there were about half a dozen non-QM RMBS issuers. As of second-quarter 2019, the number had increased to almost 20. The number of non-QM mortgage originators selling into securitizations has increased substantially over the past few years, with a growth profile that appears to mirror that of issuers (see chart 3).
Chart 3
The increasing number of originators catering to this market segment is a reflection of the diverse collateral make-up and the broad range of credit quality (from prime to non-prime) within the non-QM space. Indeed, non-QM loan characteristics differ in various ways from those of typical agency-eligible collateral. This collateral spectrum is part of the reason that our 'AAA' attachment points have such a wide range in the case of non-QM RMBS. For example, a high-quality loan with an interest-only feature or one that is missing a document related to income (small in the total scope of income qualification) can trigger a non-QM designation. As a result, 'AAA' credit attachment points can vary widely, but still have the non-QM designation. The average 'AAA' loss projection for the non-QM securitizations we rate is approximately 25% (the range is roughly 10%-35%).
Three General Issuer Model Categories
While the rate of growth in the count of distinct issuers jumped roughly a year ago (see chart 3, above), we have recently seen it trailing off somewhat. Currently, the pace of arrival of new entrants is more or less steady. In addition to a growing non-QM RMBS issuer base, we have observed an evolution of the business models for individual issuers. The models typically fall into one of three different categories:
- Single originator to an issuer for which a vertical integration component exists.
- Issuers that aggregate loans from multiple originators (in some cases close to 100 in securitizations) utilizing a seller approval process and internal credit underwriting guidelines.
- Issuers that leverage a securitization platform and focus on viable securitization strategies by acquiring varyious whole loan pools with less emphasis on issuer-specific guidelines.
The significant growth in the number of individual loan sellers to ultimate issuers could be attributable to the increase in lenders originating the product. It could also derive from increased competition to originate and/or the increase in aggregation platforms sourcing collateral and driving up the seller count. Another factor could be connected to the increase in mortgage rates late last year. At that time, originators across the U.S. may have become inclined to fold non-QM into their lending platforms to fill production pipelines as conventional rate and term refinancing dried up with the disappearance of the interest rate incentive.
Our view is that non-QM lending and origination was initially approached with caution because of the additional layers in the loan manufacturing process. Examples include ATR demonstration and the complexities associated with funding loans that varied from the more traditional GSE/Ginnie Mae styles that have prevailed since 2009.
The growth in both securitization volume and number of issuers indicates that some of these concerns have subsided. S&P Global Ratings factors the greater potential for assignee liability given a non-QM status into the loss modeling it performs when assigning ratings on non-QM RMBS.
Five Subsectors
Our general categorization of non-QM refers to the collateral type. We recognize that strong credit attributes may coincide with certain products, such as interest-only loans (which, in these cases, are often used for wealth management), and may therefore fall more into a prime jumbo designation. For the most part, however, non-QM lending has evolved to comprise five distinct subsectors (see table 1).
Table 1
| The Five Main Non-QM Subsectors | |||
|---|---|---|---|
| Non-QM sub-sector | Description | Credit view | General view |
| Prior credit event (PCE) | Includes loans to borrowers for which they experienced a housing-related credit event (e.g., foreclosure/short sale) or bankruptcy and, therefore, cannot qualify for traditional financing (e.g., Fannie/Freddie/FHA loan). S&P Global Ratings has typically used a three-year discharge and two-year discharge time frame for housing and bankruptcy related events, respectively. | Notwithstanding that such loans should have a lower FICO score given the credit event, we may assume such loans have a greater default likelihood being that such behavior was already exhibited. As a result, we may increase default likelihoods by assuming the loan is similar to a 30-day delinquent loan. | Given the continual seasoning and passage of time since the housing downturn, as well as benign macroeconomic conditions in recent years, the volume of PCEs is expected to decline. In particular, the peak period of foreclosures occurred more than five years ago. |
| Alternative income documentation | Includes alternate sources of income verification (e.g., bank statements), which vary from the traditional income underwriting forms/documents, such as W-2s, paystubs, and tax returns. The variation is due to the use of nontraditional sources of documentation. There may be several reasons why traditional income verification is not used, such as type of employment (e.g., self-employed). | Given the income documentation for this subsector could be considered to be more subjective, less accurate, and entail less continuity elements than traditional types, we adjust the default likelihood in our credit modeling based on the duration for which the income documentation was obtained. The adjustment typically ranges from 1.75x to 2.25x. | Alternative income documentation is probably the largest portion of non-QM, with the average transaction composition being roughly 50%. Given that many people in the U.S. may not have more traditional types of employment/documentation, we expect this subsector to continue to make up a large share of non-QM. |
| Foreign national | Includes, under S&P Global Ratings' description, borrowers that are not U.S. citizens or that are not permanent resident aliens. Conventional loan guidelines may have limitations on residency status. | Because it may be more difficult to service and interact with borrowers that do not consistently reside in the U.S., our credit modeling applies an adjustment factor of 1.5x for these loans. | While foreign investment in the U.S. was more popular in prior years than currently, there may be a decline in this subsector. |
| Debt service coverage ratio (DSCR) | Includes loans made on investor properties for which the income portion of the underwriting uses the cash flow expectations of the mortgaged property only, as opposed to the personal income and liabilities of the borrower. Such loans are typically QM-exempt, but are included under the general label given their inclusion in non-QM securitizations. Some of the business purpose loans may not have a DSCR; these are often referred to as "no-ratio loans." | Given the limited performance history of DSCR loans, we treat the best of these (DSCR higher than 1.27x) similar to a weak, traditionally underwritten investor property loan. Our FF adjustment factor equals 4.00 divided by the applicable DSCR on the property, typically subject to a floor of 3.15 and a cap of 6.00. | We expect this subsector of the market to grow because of factors such as potential GSE reform, demographic shifts, and household formation. |
| Other (includes main factor being debt to income (DTI) over 43%, prime jumbo fall-out, etc.) | Includes loans for which the main fallout reason, in terms of non-QM designation, is not one of the above, but rather DTI greater than 43%, or points and fees or interest-only (IO) features. | No specific default adjustment is applied based only on the designation, but factors that caused the non-QM designation would be embedded in the credit modeling (e.g., higher-DTI and IO loans have a greater likelihood of default). | This subsector could expand depending on GSE reform related to the QM patch expiry. |
A prevalent theme in the non-QM space is change. In certain cases, a loan is now designated non-QM simply because it appears commonly in non-QM RMBS pools--not because it falls outside QM guidelines. For example, we have observed movement into specific unique shelves for debt service coverage ratio (DSCR) loans. These are investor property loans that are typically underwritten with reference to the cash flows generated by the property rental income. Because these mortgages are secured by investor properties, they are usually exempt from QM/ATR rules. Nevertheless, they are considered part of the non-QM loan sector because of their presence in non-QM transactions (see "Investor Property DSCR Loans: The Nonqualified Mortgage Exempt From Qualified Mortgage Rules," Aug. 27, 2019).
Another change we have observed concerns the growth and variation in alternative documentation, which now accounts for roughly half of securitization pools, on average. Alternative documentation has evolved from loans predominantly underwritten based on income calculation involving personal/business bank statements to, in certain cases, using profit and loss statements--with or without a few months of bank statements to corroborate such income (see "Key Factors For Assessing U.S. Non-Qualified Mortgage Bank Statement Loans," April 11, 2019).
Chart 4 shows how non-QM loan type concentrations have evolved since 2017. For each of the 2017, 2018, and 2019 vintages, the chart contrasts the loan type concentrations as they were at securitization with the existing population today. Table 2 shows the FICO/combined loan-to-value (CLTV) characteristics at securitization versus today.
Chart 4
Table 2
| WA FICO And CLTV | ||||||
|---|---|---|---|---|---|---|
| WA FICO | WA CLTV | |||||
| 2017 | 693 | 71.9 | ||||
| 2017 (existing) | 688 | 70.8 | ||||
| 2018 | 716 | 66.8 | ||||
| 2018 (existing) | 717 | 66.5 | ||||
| 2019 | 712 | 69.4 | ||||
| 2019 (existing) | 712 | 69.4 | ||||
| WA--Weighted average. CLTV--combined loan-to-value ratio. | ||||||
One noticeable trend is the decline in PCEs, which is consistent with the passage of time since the housing crisis. Correspondingly, our default likelihood adjustment factor for the pools (PCE factor) has also been declining as a result of the decrease in the PCE loan count. Non-QM loans can possess multiple non-QM characteristics; our designation in Chart 4 classifies the loan as a member of each subsector when multiple characteristics are present. For example, a bank statement loan with 45% debt to income (DTI) would be classified as both alternative income documentation (bank statement) and DTI over 43.
The Credit Quality Variance In Non-QM Pools
As discussed above, the 'AAA' attachment points in non-QM RMBS can vary widely due to the range in credit quality. Evidence of this lies in dispersion of average FICO/CLTV combinations. Certain shelves among transactions rated by S&P Global Ratings have specialized in particular FICO/CLTV profiles, which can result in different attachment points for the RMBS (see charts 5 and 6). The FICO/CLTV distribution reinforces the notion that non-QM can have credit characteristics ranging from prime to non-prime.
Chart 5
Collateral Performance
The U.S. economy is currently enjoying the longest expansion in its history. Given that QM/ATR is in only its sixth year, the non-QM sector has only known benign economic conditions, complemented by strong home price gains on a national level. Correspondingly, performance of non-QM RMBS as an asset class has been relatively good, with some variation across issuers (see chart 6).
Chart 6
Chart 6 also shows that pool factors have fallen relatively quickly for most issuances, which can, in turn, affect delinquency percentages, as they are based on outstanding balance. While delinquencies are higher than for a 2.0 RMBS prime jumbo cohort, the ratio of 'AAA' loss projections to 30+ day delinquencies is roughly 5:1 for both (25:5 on average for non-QM and 5:1 on average for jumbo).
Given the novelty of the non-QM sector and some of its subproducts, we isolated certain non-QM subsectors (which are neither mutually exclusive nor exhaustive), including full income documentation, in order to determine whether there were performance differences among them. Table 3 shows that while there was some variance, there was no obvious pattern. The decreasing trend in delinquency rate with the more recent vintages is expected because the newer loans have had less time to underperform.
Table 3
| Delinquent Population As % Of Existing Population(i) | ||||||||
|---|---|---|---|---|---|---|---|---|
| Characteristic | 2017 | 2018 | 2019 | |||||
| Alternative documentation | 7.82 | 3.06 | 2.47 | |||||
| Full documentation | 10.32 | 3.72 | 2.30 | |||||
| Other documentation(ii) | 3.88 | 2.81 | 3.30 | |||||
| DTI>43 | 6.22 | 3.90 | 2.69 | |||||
| Prior credit event | 8.49 | 4.97 | 1.86 | |||||
| Foreign national | 8.31 | 4.09 | 3.75 | |||||
| (i)Delinquent population includes 30+ days delinquent/foreclosure/real estate owned (including bankruptcy). (ii)"Other" includes debt service coverage ratio, no ratio, and asset depletion loans. DTI--Debt to income. | ||||||||
Even though there doesn't appear to be a prevailing trend when looking at the data in the aggregate, there are sometimes patterns in individual transactions. This could be attributable to other credit characteristics (such as FICO/CLTV), which are not broken out in the subsectors listed in the table.
The Prepayment Story
Non-QM pool factors are currently low for many deals considering how recently they were issued. This is because, rather than defaulting (which appears to be rare), the loans are prepaying quickly in the aggregate. The average non-QM prepayment speed was roughly at a 35% conditional prepayment rate (CPR) last year. In contrast, prime jumbo speeds have been in the neighborhood of 5%-15%. While the non-QM pools have slowed somewhat for the more recent vintages (see chart 7), the aggregate speeds are still near 30% CPR. We have determined that the main drivers of prepayment in non-QM RMBS are income documentation type, occupancy type, duration of income verification, interest rate, and FICO score (see "Non-Qualified Mortgage Prepayments: Is It Life In The Fast Lane, Or Will They Start To Take It Easy?," Sept. 21, 2018). In particular, full income documentation loans tend to have faster prepayment speeds, while investor property loans with prepayment penalties tend to be slower.
Chart 7
Over the past year, the non-QM market has doubled in size, with a host of new issuers and originators competing in the space. This has reduced some of the spread in available borrower note rate, which could dampen prepayments going forward. Speeds can vary greatly over time within a vintage. This is likely because the borrower is typically unable to obtain the best market rate (depending on the non-QM subsector) at origination. As a result, changes in either credit/loan attributes or prevailing non-QM market rates can incentivize prepayments more than changes in prevailing conforming/jumbo rates. The chart above shows that the first-half 2017 and 2019 vintage prepayments jump early on. Contributing factors could include interest rate movements, credit curing, and increased refinancing options as more non-QM lenders have surfaced.
The Geographic Footprint
Unlike CRT, non-QM RMBS tend to be geographically clustered--similar in some cases to prime jumbo. However, prime jumbo may contain representation in a greater number of states than non-QM. The bulk of non-QM originations are either in the New York metro area or in L.A. and the Bay Area in California (see chart 8). One of the reasons for this could be that these higher-priced regions may provide fewer financing alternatives for jumbo balance loans relative to GSE or FHA channels. The average non-QM loan balance over the past several years is roughly $425,000, which is substantially higher than the average U.S. mortgage balance of approximately $250,000 in the CRT space.
When conducting our credit analysis for RMBS, we consider geography via what we call our "over/under analysis" (see "Will the Froth in U.S. Housing Bubble Over Again? We Think Not," published Feb. 22, 2019). Our analysis of newly issued U.S. RMBS involves a determination of the degree to which regional home prices could decline under different economic scenarios. We assess whether a market is over- or under-valued by calculating the ratio of affordability to its long-term average, based on a regional price-to-income (PTI) ratio.
Further Near-Term Sector Changes May Arise
The non-QM sector has evolved over the past five years and there could be more changes to come. Non-QM securitization, which makes up only a small fraction of annual residential mortgage originations, has doubled in growth annually since its inception over five years ago. Assuming the economy remains strong and housing fundamentals persist, we expect this pace to continue.
The regulatory factor will be a key determinant in the future of the non-QM space, both in terms of the QM patch expiration for the GSEs and any corresponding change to the QM designation. Loans with DTI over 43% account for more than a quarter of all GSE acquisitions. Because the GSEs securitize more than half of all mortgages, the removal of the non-QM patch could result in a substantial spillover into the private-label RMBS space, even if some loans migrate to FHA, qualify with lower DTI, or go unfilled (see "The Credit Effects Of The Temporary QM Patch Expiration On The U.S. Mortgage Market," Sept. 4, 2019).
This report does not constitute a rating action.
| Analytical Contacts: | Jeremy Schneider, New York (1) 212-438-5230; jeremy.schneider@spglobal.com |
| Sujoy Saha, New York (1) 212-438-3902; sujoy.saha@spglobal.com | |
| Noury Fekini, New York + 1 (212) 438 0446; noury.fekini@spglobal.com | |
| Research Contact: | Tom Schopflocher, New York (1) 212-438-6722; tom.schopflocher@spglobal.com |
No content (including ratings, credit-related analyses and data, valuations, model, software or other application or output therefrom) or any part thereof (Content) may be modified, reverse engineered, reproduced or distributed in any form by any means, or stored in a database or retrieval system, without the prior written permission of Standard & Poor’s Financial Services LLC or its affiliates (collectively, S&P). The Content shall not be used for any unlawful or unauthorized purposes. S&P and any third-party providers, as well as their directors, officers, shareholders, employees or agents (collectively S&P Parties) do not guarantee the accuracy, completeness, timeliness or availability of the Content. S&P Parties are not responsible for any errors or omissions (negligent or otherwise), regardless of the cause, for the results obtained from the use of the Content, or for the security or maintenance of any data input by the user. The Content is provided on an “as is” basis. S&P PARTIES DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, ANY WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE OR USE, FREEDOM FROM BUGS, SOFTWARE ERRORS OR DEFECTS, THAT THE CONTENT’S FUNCTIONING WILL BE UNINTERRUPTED OR THAT THE CONTENT WILL OPERATE WITH ANY SOFTWARE OR HARDWARE CONFIGURATION. In no event shall S&P Parties be liable to any party for any direct, indirect, incidental, exemplary, compensatory, punitive, special or consequential damages, costs, expenses, legal fees, or losses (including, without limitation, lost income or lost profits and opportunity costs or losses caused by negligence) in connection with any use of the Content even if advised of the possibility of such damages.
Credit-related and other analyses, including ratings, and statements in the Content are statements of opinion as of the date they are expressed and not statements of fact. S&P’s opinions, analyses and rating acknowledgment decisions (described below) are not recommendations to purchase, hold, or sell any securities or to make any investment decisions, and do not address the suitability of any security. S&P assumes no obligation to update the Content following publication in any form or format. The Content should not be relied on and is not a substitute for the skill, judgment and experience of the user, its management, employees, advisors and/or clients when making investment and other business decisions. S&P does not act as a fiduciary or an investment advisor except where registered as such. While S&P has obtained information from sources it believes to be reliable, S&P does not perform an audit and undertakes no duty of due diligence or independent verification of any information it receives. Rating-related publications may be published for a variety of reasons that are not necessarily dependent on action by rating committees, including, but not limited to, the publication of a periodic update on a credit rating and related analyses.
To the extent that regulatory authorities allow a rating agency to acknowledge in one jurisdiction a rating issued in another jurisdiction for certain regulatory purposes, S&P reserves the right to assign, withdraw or suspend such acknowledgment at any time and in its sole discretion. S&P Parties disclaim any duty whatsoever arising out of the assignment, withdrawal or suspension of an acknowledgment as well as any liability for any damage alleged to have been suffered on account thereof.
S&P keeps certain activities of its business units separate from each other in order to preserve the independence and objectivity of their respective activities. As a result, certain business units of S&P may have information that is not available to other S&P business units. S&P has established policies and procedures to maintain the confidentiality of certain non-public information received in connection with each analytical process.
S&P may receive compensation for its ratings and certain analyses, normally from issuers or underwriters of securities or from obligors. S&P reserves the right to disseminate its opinions and analyses. S&P's public ratings and analyses are made available on its Web sites, www.standardandpoors.com (free of charge), and www.ratingsdirect.com and www.globalcreditportal.com (subscription), and may be distributed through other means, including via S&P publications and third-party redistributors. Additional information about our ratings fees is available at www.standardandpoors.com/usratingsfees.
Any Passwords/user IDs issued by S&P to users are single user-dedicated and may ONLY be used by the individual to whom they have been assigned. No sharing of passwords/user IDs and no simultaneous access via the same password/user ID is permitted. To reprint, translate, or use the data or information other than as provided herein, contact S&P Global Ratings, Client Services, 55 Water Street, New York, NY 10041; (1) 212-438-7280 or by e-mail to: research_request@spglobal.com.