Debt service coverage ratio (DSCR) loans are residential investor property loans underwritten with reference to the cash flows generated by a property's rental income. Although they are exempt from qualified mortgage (QM) and ability-to-repay (ATR) rules, DSCR loans are generally considered part of the non-QM loan market because of their growing presence in non-QM residential mortgage-backed security (RMBS) transactions. DSCR collateral represented 50% (by balance) of the non-QM transactions rated by S&P Global Ratings in 2022--up from 22% for the 2019 vintage.

DSCR loans represent roughly half of the approximately 201,000 loans included in the non-QM securitizations that we rated between 2018 and February 2023. About 15% by count (25% by balance) of these DSCR loans had sub-unity DSCRs (i.e., a DSCR that is less than 1) or were "no-ratio" business-purpose investor loans. No-ratio loans (for which a DSCR is not provided), make up about 8% by count (12% by balance) of the DSCR loans. They generally have higher balances than investor property loans underwritten using a DSCR and loan-to-value (LTV) ratios that are two points lower on average. This report is based on the set of approximately 94,000 DSCR loans that excludes no-ratio loans. Of this set, about 9% by count (16% by balance) had a sub-unity DSCR value. As of March 2023, the 60-plus-day delinquency rate of the broader non-QM market was about 2%. DSCR loan delinquencies have been generally comparable, although sub-unity DSCR loans have performed slightly worse.

Why DSCR Loans?

The growing popularity of DSCR loans among originators and issuers, as well as the increase in sellers and originators (despite some consolidation and closures among originators and sellers in 2022), is partly due to the general interest in investor property ownership and management as a viable business. Much of the optimism stems from the relative strength in the housing market over the past decade, which has been supported by fundamental economic factors, especially supply-demand mismatches.

Indeed, the housing stock required to meet the needs of the growing U.S. population and the household formation among millennials has been inadequate in recent years. This has also been one of the main drivers of rising home prices since the Global Financial Crisis and one of the factors fueling the popularity of DSCR loans. Further, while the housing and mortgage markets appear to be undergoing a pricing correction, largely due to the rapid increase in interest rates that is eroding affordability, DSCR investor activity persists, although at a slower pace (see "2023 U.S. Residential Mortgage And Housing Outlook: Navigating A Softening Market," published Jan. 20, 2023).

Residential loans for investor properties have been around for many decades. However, the underwriting method used for DSCR loans offers a more direct--and thus quicker--assessment of the borrower and loan characteristics compared to the one used in the traditional agency program. Traditional investor property underwriting includes positive amounts related to the property's rent (subject to a haircut) as an addition to the borrower's personal income, and negative amounts as an "add back" to the liability when computing the debt-to-income (DTI) ratio. DSCR underwriting, on the other hand, is broadly based on the rental income needed to service the debt of the property. In this sense, the underwriting of a DSCR loan is similar to that of a commercial mortgage.

The Geographic Footprint

The volume of conventional mortgage lending, which is primarily to owner occupants, is roughly proportional to local U.S. populations, with larger regions seeing more conventional mortgage lending and smaller ones seeing less. On the other hand, investor property lending tends to be more geographically concentrated because certain regions are better suited to the DSCR loan operating model for rental units. We examined the geographic distribution of DSCR loans by cumulative original loan balance and observed that the regions with higher DSCR loan volumes include some of the country's larger metros, such as New York, Los Angeles, Miami, and San Francisco (see chart 1).

Chart 1

Where Are Rents And DSCRs Headed?

We expect single-family (SFR) and multifamily rent growth to decline this year, although the effect will likely be less pronounced in the SFR market. Rental property investments have enjoyed high risk-adjusted relative returns in recent years, even though the rent growth of SFR seems to have lower volatility when compared with that of multifamily units (see chart 2).

Chart 2

Although property values have appreciated and led to higher mortgage loan balances, LTV ratios have been relatively stable with an average of 65%-70% for DSCR loans that were securitized between 2018 and 2022. The recent rapid rise in interest rates seems to be the driving factor behind the current downward drift in DSCRs, even though higher loan balances are also a factor. It's possible that the lower DSCRs reflect calculations that are based on rental income data that haven't caught up to rental gains because of re-leasing. (For more on the single-family rental market and its relationship to interest rates, see "Real Estate Monitor: Rising Rates Driving Rental Housing Resiliency," published March 30, 2023.)

Sub-Unity DSCRs And The Impact Of Rising Interest Rates

DSCR is generally calculated either as the ratio of gross rental income to the sum of the mortgage payments, property taxes, and insurance on the property, or as the ratio of the gross rental income (less taxes and insurance) to the mortgage payment. However, for a substantial number of loans, the denominator exceeds the numerator, resulting in sub-unity DSCRs.

There could be several underlying reasons for this phenomenon, including rising interest rates, a lag in rental growth, choice in origination standards, and property location. For example, an investor might anticipate being able to increase rents after renovations, which would render the property more profitable. Alternatively, if interest rates are expected to decline, the investor might intend to refinance eventually and become cash flow positive. An investor could also anticipate property price appreciation, which could lead to a net gain through a property sale, despite a period of negative cash flow.

The recent trend in declining DSCR can largely be attributed to rising mortgage rates. Higher loan balances as a direct result of house price appreciation have only reinforced this trend. DSCR loans have tended to have coupons of 50 basis points (bps) to 100 bps above other non-QM loans, which themselves have historically had coupons of 150 bps to 200 bps over the conforming rate (see "Factors Affecting Non-QM Mortgage Interest Rate Spreads," published Feb. 20, 2020). Chart 3 shows the average 30-year fixed-rate mortgage, according to the Freddie Mac survey, as well as the inverse correlation between DSCR and interest rates.

Chart 3

To illustrate the relationship between interest rates and DSCRs, we conducted an analysis of a hypothetical mortgage with a starting monthly rent of $2,250 and a 70% loan-to-value (LTV) ratio on a $400,000 property with 2% in annual taxes and insurance. Table 1 shows that for every 50 bps increase in interest rates, an almost 5% increase in rent is needed to maintain a constant DSCR. It also shows that when mortgage rates were 5.5% for DSCR loans, monthly rent would need to be $2,250 to achieve a DSCR of unity. For a DSCR loan originated after the 300 bps increase in mortgage rates last year, a monthly rent of $2,810 (a roughly 25% increase above the $2,250 base) would be required to maintain a DSCR of unity.

Table 1

Interest Rates And The Diminishing Effect Of Interest-Only Loans

An interest-only (IO) feature can have a material impact on the DSCR calculated during the underwriting process. Qualification for a DSCR loan is usually based on the IO payment and not the fully amortized principal and interest payment, which, for a 30-year loan with a 10-year IO period, would result in a lower initial payment and consequently a higher DSCR. Because DSCR loans fall outside the scope of the ATR rules, underwriting guidelines do not require the use of the fully amortized payment, and we predominantly see originator guidelines use the IO payment in lieu of the fully amortized payment for DSCR loan underwriting.

In table 2, we repeat the exercise from table 1 but now examine the absolute difference in the DSCR obtained when using an IO payment rather than a 30-year amortization schedule to calculate the DSCR. When interest rates are low, say around 3%, a monthly IO payment is substantially lower than a fully amortized payment. In terms of DSCR, this means the advantage of an IO loan will increase materially as rental income grows. However, when interest rates are high, say above 7%, an IO payment is comparable to a fully amortized payment. This means that not only is there little advantage in terms of boosting DSCR, but also that the incremental gain in DSCR will not grow appreciably as rent increases in a high interest rate environment.

In practice, this advantage will be further muted because the IO rate is likely higher than the fully amortized rate, all else being equal. While the number of IO loans within the DSCR space increased with the jump in mortgage rates last year, it was not appreciably different from prior years when mortgage rate movements were smaller. By balance, IO loans represent about 20% (12% by count) of DSCR loans in securitized pools from 2018 to 2022.

Table 2

The Property Price Effect

There appears to be an inverse relationship between home value and rental yields (perhaps reflecting compensation for operational risk at different price points), which in turn affects DSCR. In the U.S., regions with lower home values tend to have higher rental yields, which we define as gross rent divided by property value (see chart 4). Western U.S. states tend to exhibit lower yields and lower DSCRs despite having higher home values. For example, when comparing metropolitan statistical areas (MSAs) by rental yield, those in California ranked the lowest, with an average yield of about 7% and a DSCR near unity (see chart 5).

Geographic variability in rental yield is generally consistent with that of DSCR, suggesting that LTV ratios are broadly uniform across the country. The variation in LTV ratios among DSCR loans securitized between 2018 and 2023 tends to lie within a relatively tight band. In 2022, the average rental yield was 9.4% (with a standard deviation of 3.2%) and the weighted average rental yield was 8.2% (weighted by property value).

Chart 4

Chart 5

DSCR Calculation Methods

DSCR is one of the factors we consider in our RMBS credit analysis when differentiating default likelihood of loans (see "Investor Property DSCR Loans: The Nonqualified Mortgage Exempt From Qualified Mortgage Rules," published Aug. 27, 2019). Unlike traditional conventional loans, for which underwriting is based on the borrower's DTI ratio, there can be substantial variability in underwriting when using DSCR. While the denominator in the DSCR calculation is typically standardized (except perhaps when IO payments are considered or when taxes and insurance are included in the numerator instead of the denominator), the numerator (i.e., gross rent) can vary according to underwriting guidelines.

The monthly gross rent is typically the lesser of the in-place lease payment obligations (i.e., the rent roll) and the market rent for the property provided on the appraiser's applicable property valuation report. A rent rolling property with submarket rents reported by the appraiser could lead to a lower DSCR. Alternatively, if the rent roll is associated with a 24-month lease in a region that has experienced strong rent growth since the lease was executed, DSCRs could be lower even though market rents are high. Some underwriting guidelines may consider the lesser of the market rent multiplied by a gross-up factor (e.g., 110%) and the rent roll. Alternatively, guidelines may apply a haircut to the market rent reported by the appraiser if the property is not rent rolling or is vacant.

To illustrate how DSCR depends on calculation method, we consider five scenarios. Each is based on a hypothetical loan with a monthly rent roll of $1,200, a reported monthly market rent of $1,000, monthly principal and interest payment of $800, and taxes and insurance of $200. Note that these comparisons do not account for variations due to an amortized or, if applicable, IO mortgage payment in the denominator.

• Scenario one: The DSCR numerator is the minimum of the market rent and the rent roll, resulting in a DSCR of 1.00.
• Scenario two: The DSCR numerator is the minimum of the market rent and the rent roll, but with a market rent grossed up by 10% (because the current rent exceeds the market rent), resulting in a DSCR of 1.10.
• Scenario three: The DSCR numerator is the minimum of the market rent grossed up by 10% and the rent roll less taxes and insurance, and the denominator is principal and interest, resulting in a DSCR of 1.125.
• Scenario four: The DSCR numerator is the market rent but with a 10% haircut, which is sometimes used when properties are vacant, resulting in a sub-unity DSCR of 0.90.
• Scenario five: While we have not observed this in practice, the DSCR numerator is the maximum of the market rent and the rent roll, resulting in a DSCR of 1.20.

Depending on the underwriting guidelines, an RMBS transaction may include or exclude loans with certain characteristics, such as no-ratio, sub-unity DSCR, or DSCR less than a certain minimum (below unity). Most of the sponsors we have assessed originate or acquire investor loans with sub-unity DSCRs (including no-ratio loans). Chart 6 shows the average DSCR and CLTV ratio for sponsors with DSCR loans in our rated transactions. Among the sponsors shown, 10 by count (12 by balance) have sub-unity DSCR or no-ratio loans that represent more than 25% of their DSCR collateral.

Chart 6

What's Next For DSCR Loans?

We believe tax reassessments by municipalities (due to strong home price appreciation) and rising mortgage rates could affect DSCR loan characteristics and performance this year. Demographic factors and supply constraints in the housing market will continue to support the rental market and drive demand in the sector, in turn spurring competition among originators. With positive but muted SFR rent growth forecast for this year, the rental income business model may prove to be a suitable volatility hedge even as home prices decline. DSCR lending growth will also depend on economic factors and housing market dynamics, and the extent to which originators continue to accept short-term rental activity as a viable business.