Debt Service Coverage Ratio (DSCR) and even non-DSCR ("no-ratio") loans have become increasingly abundant in nonqualified mortgage (non-QM) securitizations. These are investor property loans that are, in large part, 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/ability-to-repay (ATR) rules. However, "DSCR loans," as they are often called, are typically considered part of the non-QM loan sector because of their presence in non-QM transactions.

Non-QM RMBS can be viewed as comprising several subsets: prior credit event (PCE), alternative income documentation (e.g., bank statement), foreign national, DSCR/no-ratio, and debt to income (DTI) ratio over 43%/prime jumbo fall-out loans. DSCR loans bear a resemblance to the single-family rental (SFR) space, which gained popularity shortly after the financial crisis. Unlike SFR, however, the DSCR loan segment relies on smaller operating models. Moreover, it does not include certain provisions that characterize the SFR space (e.g., property substitution within the pool). In an apparent effort to create comparability within the non-QM/QM-exempt space (given the various sub-products within non-QM), we have been observing residential mortgage-backed securities (RMBS) issuances for which the entire pool is composed of DSCR/no-ratio loans.

RMBS Or Commercial Mortgage-Backed Securities (CMBS)?

The treatment of DSCR loans varies from that of traditional residential underwriting of investment properties. With DSCR loans the qualifying income and liability analysis focuses solely on the subject property cash flow instead of the borrower's DTI ratio, which would include numerous measures of income and obligations related to the borrower. As the name implies, DSCR loan underwriting incorporates a traditional commercial property underwriting measure: the DSCR. Nevertheless, the securitization structures and overall underwriting of these non-QM transactions skew heavily to the residential side. Moreover, the property types are predominately single-family or two-to-four unit.

Factors such as credit scores, loan-to-value (LTV) ratios, and property types are key considerations for underwriting DSCR loans. As such, the underwriting resembles that of traditional residential mortgage pools. The underwriting guidelines for some originators may include thresholds, such as credit minimums of 620 and maximum LTVs of 80%. The main feature that distinguishes the underwriting of traditional investment property loans from DSCR loans is the verification of qualifying income.

Traditionally, when underwriting a loan for an investment property in the residential space, lenders have relied on income sources such as W-2s, paystubs, and/or tax returns in order to determine a borrower's personal income. The lender would also consider a borrower's personal liabilities, such as those appearing on a credit report (e.g., mortgage, credit card, installment payments). The net rental income is used in the overall DTI ratio of the borrower, considering both personal and investment property income and expenses. For DSCR loans, however, the situation is different because the underwriter focuses solely on the income generated by property cash flow—determined via the DSCR, which is effectively monthly gross rent divided by the sum of the monthly amounts for principal, interest, property taxes, insurance, and applicable association dues related to the mortgaged property.

We have observed that gross rent is typically the lower of the annual market rent as listed on a market rent comparability form (typically Form 1007) and in-place rents as listed on the lease agreement. In the more extreme case, the lending decision for a property-focused no-ratio loan is based more on equity (i.e., the LTV ratio). We view these as akin to a DSCR loan for which the ratio is not calculated (i.e., zero DSCR). In both cases, FICO scores are commonly used in underwriting, as well as personal guarantees, which are obtained for closings in the name of an LLC.

Where Are They?

Traditional SFR strategies that grew out of the post-crisis housing price trough were focused in specific regions, such as Atlanta, Phoenix, and Texas. In the case of DSCR loans in non-QM RMBS we have rated, we also see clustering in expensive urban areas, such as New York, Los Angeles, and Miami, which leads to DSCR loans possessing higher average property values than was common several years ago in the SFR space. Chart 1 illustrates the regional clustering in New York, California, and Florida. The data are derived from a sample of the transactions we have rated, and include only DSCR/no-ratio loans. This state-specific focus is more pronounced for Florida but less so for California with DSCR loans compared to other non-QM segments--it is roughly the same in the case of New York. As with non-QM, DSCR loan pools have less geographical diversification than what one may see in prime and conforming RMBS.

Chart 1

How High Can They Go?

Of the S&P Global Ratings' rated non-QM portfolio, two shelves comprising six transactions (issued by Verus and Visio) consist of 100% investor loans (both DSCR and no-ratio). Our average 'AAA' loss coverage (LC) across these shelves is 35.5%, and the 'BBB' LC is 10.0%. The table below shows that DSCR loans/pools have higher loss projections across rating scenarios compared to loans in the general non-QM population, even though certain characteristics, such as the LTV ratio, might be stronger from a credit perspective. The elevated loss projections at different rating levels are primarily driven by the increase in projected foreclosure frequency (FF), itself derived from adjustments reflecting underwriting primarily to the DSCR as opposed to traditional factors on a residential investment property loan. While the losses in the table are higher for DSCR loans compared to the general non-QM population, projected loss severities are actually lower (given the typically lower LTV ratios). This is consistent with our view that the default likelihood (projected FF) of DSCR loans is higher, all else being equal.

Property type also plays a role given the greater popularity of the two- to four-unit designation (which is associated with rental income investment properties), for which we apply a 2.0x adjustment to our FF. On average, roughly 30% of the pools in the two DSCR loan-only shelves are two- to four-unit property types compared to around 10% for non-QM transactions. Original cumulative LTV ratios are, on average, near 65% for these shelves compared to roughly 70% for non-QM, generally.

When calculating the FF for DSCR loans we apply a DSCR adjustment factor that ranges from 3.15x to 6.0x. Our adjustment factor equals four divided by the applicable DSCR on the property, typically subject to a floor of 3.15 and a cap of six. The low end of the range (3.15x) was calibrated so that a DSCR loan with a high DSCR (e.g., greater than or equal to 1.27x) is treated similarly to a weak, traditionally underwritten investor property loan (i.e., underwritten to the borrower's income) with less than 12 months of income verification and poor DTI attributes (i.e., where the maximum adjustment factors for full-income documentation and DTI are assumed), all else being equal and given the limited performance history of DSCR loans through an incomplete economic cycle. We believe this is sufficient to analyze the expected credit risk for DSCR loans and includes the risks associated with occupancy type, documentation type, and DTI.

Historical residential credit modeling data has limited robustness as it relates to DSCR, which partly explains the calibration described above. Because DSCR loans are underwritten to the income from the mortgaged property rather than the borrower's personal income, we don't consider adjustments for personal income documentation type or DTI, as DTI relates more to personal rather than rental income. Furthermore, the risk associated with occupancy type is already captured by the DSCR adjustment factor because DSCR loans are, by definition, made on investor properties. We evaluate no-ratio loans as if they were DSCR loans with a DSCR of zero, which mathematically results in a 6.0x adjustment to the FF.

Where Do We Go From Here?

SFR investments have always existed in the U.S., but only within the past five or so years have we seen capital markets make substantial inroads into the space. After property prices reached a trough around 2012, the SFR market took off with large institutions acquiring rental properties at bargain prices. Shortly thereafter, companies with smaller operating models entered the SFR space, and now we are seeing the rental trend continue in the non-QM sector of non-agency RMBS.

Chart 2

The average price of DSCR loan properties is over $400,000, driven by the higher-priced metro areas in which we see the bulk of DSCR loans in our rated portfolio. This is higher than the national average and can create affordability challenges for the millennials that are now entering the housing market and forming families. However, this could translate into opportunities either for those renting to this demographic, or for millennials contemplating an investment in real estate.

With roughly 140 million housing units in the U.S., a homeownership rate below 65%, and increasing house prices leading to greater down-payment requirements, the single-family rental strategy can be a viable investment opportunity. Since the start of 2018, DSCR loan deals have generally increased in number and dollar volume. Moreover, DSCR loans now make up a substantial share of the RMBS in which they reside (see chart 2). These factors, coupled with the recent exponential growth of the non-QM market, suggest that the DSCR loan RMBS market will increase in popularity among issuers and investors.