Although the U.S. housing market has been subdued over the past two years, we expect mortgage origination and securitization activity to increase year over year in 2024 as the economy begins to normalize. Despite affordability constraints, the housing market has proven resilient, with the national home purchase-only seasonally adjusted Federal Housing Finance Agency home price index up 6.2% between December 2022 and October 2023. This relatively strong figure is partly attributable to lack of supply, which has been exacerbated by weak existing home sales. During 2023, the upward price pressure from supply constraints dominated the downward pressure from elevated mortgage rates. In some cases, price fluctuations increased due to population movement in certain regions of the country: a factor that was evident prior to the COVID-19 pandemic and was amplified during it. While we expect mortgage rates to decrease in 2024, it is unlikely that demand from first-time buyers will be adequately met in either the new or existing home markets. Therefore, we expect limited net downward pressure, if any, on home prices over the near term.

Part of our focus in the new year concerns the consumer's decision to buy or rent. In third-quarter 2023, the U.S. rental vacancy rate was 6.6%, below the 50-year average of 7.4% and well below the peak of 11.1% in 2009. While this figure is up from the levels in late 2021 and early 2022, it is still low by historical standards, with vacancy rates comparable to those in the mid-1980s. There is a complicated interplay between the rental and purchase markets, which potential buyers must navigate. The recent vacancy rate is so low, in part, because of the limited affordable purchase options. In fact, Realtor.com reports that about 80% of mortgagors enjoy sub-5% fixed-rate mortgage (FRM) loans, which contributes to the dramatically slow sales activity in the existing home market. While it is unlikely that the 30-year FRM will revert to the historic lows of 2020, anticipated rate cuts by the Federal Reserve (Fed) and a contraction of the spread between the 30-year FRM and the 10-year U.S. Treasury note yield in 2024 should generate momentum in home inventory levels and an increase in mortgage originations.

Renting May Be Cheaper Than Buying, But…

The lack of existing homes for sale in the U.S. has fueled the rental market during the past several years. Moreover, population growth rates of 0.2%, 0.4%, and 0.5% (according to the U.S. Census Bureau) in 2021, 2022, and 2023, respectively, continue to bode well for occupancy rates across the U.S.

Also, our economic forecast calls for an average 30-year FRM of 6.8% in 2024, which, along with 2023, would be the highest annual rate in over 20 years. The consequent impact on affordability in the housing market, coupled with population growth and the time-lag for single-family housing completions, suggests that rental market fundamentals will likely be supportive in 2024. Nevertheless, the inventory of single-family investment properties for acquisition will continue to be subdued until some of the existing inventory frees up.

Falling rental vacancy rates in the South benefits single-family-rental market

Vacancy rates last spiked during the Global Financial Crisis (GFC), when the housing market was overbuilt and there were historically high foreclosure rates. The recent population movement away from the West Coast has benefitted the single-family-rental market, as well as owner-occupied activity, in the South (see chart 1); so while the Southern region has generally demonstrated the highest vacancy rates historically, it has fallen more in line with the overall U.S. rate in recent years. We believe this is due to the population migration trends during the past five years as well as the growing popularity of single-family rentals in the Sunbelt States. Regional variation is also evident in home prices, as we discussed in a recent publication (see "Accounting For Regional Volatility When Assessing Home Price Appreciation: A Portfolio Management Theory Approach," published Nov. 22, 2023).

Chart 1

Multifamily starts expected to decline, but inventory could still rise, which may lead to downward pressure on rents

The increase in financing costs in the commercial real estate sector (particularly at banks that may be more involved in multifamily lending) has likely contributed to the pullback in multifamily starts. Despite the expected decline in multifamily permits in 2024, inventory of multifamily properties would likely rise during 2024 due to completions already underway and entering the market. The additional supply could put downward pressure on rents in that sector. The same may not necessarily be true for the single-family-rental sector because rent growth is typically less volatile relative to multifamily rates, and single-family-home renters tend to be longer-term tenants. Single-family starts have fared relatively well (notwithstanding mortgage rates near 8%), with 2023 levels hovering around one million units, which is just below the 50-year average. Our 2024 forecast for total starts (which includes multifamily starts) is for slightly more than 1.3 million units, down marginally from 2023.

Chart 2 depicts the historical relative sizes of age cohorts. The consistency of the "25-44" bracket indicates that this cohort will continue to create demand, at least in the short to medium term.

Chart 2

Homeownership rates are healthy

During first-half 2023, the homeownership rate in the U.S. was 66%, roughly the same as the average level in 2022. This is down slightly from 67% at the start of the pandemic, but still well above the post-GFC low of 63% during first-half 2016. These healthy levels are consistent with population growth as well as the demographic shift of homebuyers: there are now more Millennials than there are Baby boomers and those in Generation X. Moreover, Millennials are in their prime home-buying years and are forming families. We expect this trend to continue, at least in the near term, with housing demand remaining robust in 2024 and beyond. This also points to a positive outlook for single-family-rental occupancy rates, because the still-stubborn spread of the 30-year FRM over the 10-year Treasury note yield continues to restrict purchasing power. Multifamily occupancy rates may be more muted as supply picks up in the near term.

Idiosyncratic trends and risks at the regional level

While housing at the national level can be informative, it does mask idiosyncratic trends and risks at the regional level. Focusing on metropolitan statistical areas (MSAs), we expect pandemic-related population migration dynamics to stabilize in 2024, as effects of the pandemic fade into the rearview mirror. While population migration started prior to 2020, the pandemic exacerbated movement away from expensive MSAs into tax-friendly and lower-cost retirement havens at a time when remote work was encouraged or even required. Assuming that hybrid work arrangements took on a near-final form for most institutions in 2023, population migration will be driven not only by economic factors present pre-pandemic, but also the newfound flexibility in the workforce. Chart 3 shows MSA-level population growth from third-quarter 2022 to third-quarter 2023 (based on data from the U.S. Census Bureau).

Chart 3

Chart 3 suggests that migration patterns between third-quarters 2022 and 2023 resemble those pre-pandemic. Interestingly, even after rapid, pandemic-related migration into certain areas--such as Tampa, Florida; Boise, Idaho; and Austin, Texas--the trend continued into 2023, but it remains to be seen whether or not this is a secular shift. Secondary cities are also likely to benefit from the spillover from overpriced coastal areas, especially now that remote/hybrid work have largely stabilized.

Spread Between FRM And U.S. Treasury Likely Won't Normalize In 2024

The greatest annual increase in mortgage rates in over 40 years occurred in 2022, an event few anticipated. The Fed's monetary policy certainly contributed to the increase in FRM; however, the increase of 500 basis points (bps) in the 30-year-fixed rate was driven in part by the widening spread to the 10-year Treasury note yield. The spread, normally about 170 bps, captures a mixture of investor concerns about residential mortgage-backed securities (RMBS), including duration and liquidity risk, as well as inherent frictions and other costs. During 2023, the spread widened to over 300 bps and has remained in that range to date.

Despite the major challenge to affordability, the housing market performed reasonably well since the third-quarter 2022, when prices were down nationally for the first and only time in a decade (as measured by the FHFA seasonally adjusted quarterly purchase-only index). This resilience suggests sustained performance in 2024 as we forecast steady declines in the 30-year FRM to 7.3%, 7.0%, 6.6%, and 6.2%, during each of the four respective quarters. Our forecasts for the 10-year Treasury over the same four quarters are: 4.7%, 4.4%, 4.2%, and 3.9%. This implies a spread tightening of 60 bps to a level of 230 bps, still well above the historical norm of about 170 bps, but clearly easing as the path of interest rates becomes more certain. We do not, however, expect normalization for several years (see chart 4).

The currently wide spread of the 30-year FRM to the 10-year Treasury note yield was last seen in the years immediately after the GFC, when markets were gripped by flight-to-quality behavior and investors generally shunned risky fixed-income products. An informative study by the Brookings Institute analyzed the components of this spread and determined that much of the widening is due to prepayment risk. The widening of the spread post-GFC, on the other hand, was largely driven by liquidity concerns. While liquidity risks remain in today's RMBS markets, the dominant risks appear to be interest-rate-related.

Chart 4

Momentum For Originations And Securitizations In 2024

Mortgage originations and RMBS issuance were muted in 2023

Both mortgage originations and RMBS issuance were muted in 2023. Non-agency volume was roughly $78 billion, down from $140 billion in 2022 and $190 billion in 2021. Fannie Mae expects total origination volume (purchases and refinances) of over $1.5 trillion in 2023, which is down 36% from 2022 and 67% from 2021. To appreciate the decline, after controlling for inflation, one can divide the origination volumes by the conforming loan limits in respective years. Doing so leads to 2.1 million units originated in 2023, which was the lowest reading in the past decade at merely 45% of the 10-year average. The next lowest reading was in 2014 at 3.1 million units.

Origination and securitization volume are expected to increase in 2024

With an anticipated reduction in mortgage rates this year, one would expect an increase in origination and securitization volume. Indeed, Fannie Mae's December housing forecast points to an approximate 25% increase in total originations in 2024, which should drive RMBS issuance, most of which is in the agency space. Our non-agency issuance forecast calls for a 30% increase in volume to $100 billion, proportionately more than origination volume. Our view on the greater growth in non-agency securitization reflects the likelihood of better financing prospects as the expected downward path of interest rates comes into sharper focus. It also reflects the fact that some subsectors (e.g., single-family rental and seasoned re-performing) are less dependent on origination volume and asset pool representation. See chart 5 for a breakout of our forecasts by subsector.

Chart 5

Non-qualified mortgage (non-QM) RMBS volume expected to increase

We expect non-QM RMBS volume to increase about 25% to approximately $37.5 billion, leading non-agency volume. The strong expectations reflect the broadening investor base as well as the fact that originators have become aware of, and are more comfortable with, the non-QM subsector. These originators are also likely to have excess capacity. We expect re-performing/non-performing RMBS volume to increase to about $17.5 billion as 2021 vintage non-performing loan transactions enter their callable phase. For the prime 2.0 (jumbo/conforming) sector, we expect a slight uptick as mortgage rate uncertainty recedes. Nevertheless, uncertainty regarding changing capital rules continues to influence the sector. Credit-risk transfer (CRT) and single-family rental are forecast to have issuance activity in line with that of 2023 at around $10 billion and $5 billion, respectively. We are projecting some growth in the "other" category, which includes second liens/home equity lines of credit (HELOCs), reverse mortgages, residential transition loans, and shared appreciation assets.

Second lien/HELOC subsector will be sensitive to mortgage rates and housing inventory

The forecast for the second lien/HELOC subsector is particularly sensitive to the path of mortgage rates and housing inventory. Assuming a downward trajectory in rates, currently high levels of homeowner equity could drive issuance activity. Although unlikely, falling rates could mean the economic benefits of a cash-out refinance on a first-lien mortgage would outweigh the benefits of adding a second-lien mortgage at a rate several hundred bps over the conforming rate. To illustrate when a blended rate would be out-of-the-money relative to a first-lien cash-out refinance, we consider various combined loan-to-value (LTV) ratios and interest rate combinations (see table 1). This sensitivity analysis assumes the homeowner has a first-lien mortgage balance of $300,000 with a 60% LTV ratio and obtained a second-lien mortgage with an assumed rate of 11%.

Table 1

We are forecasting a decline in the 30-year FRM to a 4.8% average in 2027 from a 6.8% average in 2024. For perspective, the 30-year FRM has averaged about 7.7% over the past 50 years, although this mean is heavily influenced by the high interest rates of the 1980s. Over the past 30 years, the average rate is 5.6%. Table 1 above shows different combinations of locked-in first-lien rates and second-lien balances to illustrate the blended rate hurdle (with the bottom area depicting the 2027 forecast).

Borrowers with larger second-lien amounts have higher breakeven blended rates as opposed to smaller second-lien amounts, and it would take less interest rate reduction to create an economically viable refinance scenario. Furthermore, higher first-lien interest rates are more likely to create an economically viable refinance scenario sooner (as opposed to lower first-lien interest rates) because less downward rate movements are needed to create an economically viable refinance scenario. For example, a borrower with a 5% rate on a first lien and a $125,000 amount on a second lien would need prevailing interest rates to drop below 6.76% to create an economically viable refinance scenario. On the other hand, a borrower with a 5% rate on a first lien and a $25,000 amount on a second lien would need prevailing interest rates to drop below 5.46% to create an economically viable refinance scenario.

Perhaps the most pressing question about second-lien mortgages concerns prepayment speeds. Historically, based on data from the GFC, this subsector was dominated by leveraged and high combined LTV (CLTV) purchases, such that much of the prepayment activity was involuntary (i.e., represented defaults). In 2023, second-lien origination was much different, with lower CLTV cash-out refinancings and strong underwriting, and we expect a similar scenario in 2024. A meaningful decrease in mortgage rates would therefore create an economically viable refinance scenario. It would also impact prepayment speeds of first-lien mortgages with subordinated debt because of the financial incentive to quickly pay down the second-lien mortgage and refinance into a single first lien, albeit with a higher rate than the original first lien, but one that is lower than the former blended rate. Additional factors to consider would be:

• Refinance-related expenses (both monetary and time investment based);
• Amortization reset on the first lien if wrapping a first and second lien with a refinance;
• Differing voluntary curtailment behavior from borrowers with first liens given the higher interest rate on the second lien;
• Tax deductibility, which may or may not be a factor related to the effective interest rate the borrower pays;
• Servicer recapture motivation and proactiveness to identify and direct borrowers to economically wrap a first and second lien;
• General mortgage origination pertaining to first-lien purchase origination volumes and the corresponding focus commitment on second-lien wrapped refinances;
• Psychological effects (i.e., even if the blended rate is lower, the willingness to give up the lower contractual rate on the first lien needs to occur); and
• Shorter amortization profiles on second liens compared with typical 30-year-amortization terms on first liens, thus watering down the stationary effect of the blended rates in the table 1.

Outlook For RMBS Credit Performance

In 2023, we reverted our outlook for RMBS to benign in the guidance to our criteria, "RMBS: Methodology And Assumptions For Rating U.S. RMBS Issued 2009 And Later," published April 17, 2020. So far, the housing and economic market fundamentals have supported our view. Strong home equity among borrowers, lack of meaningful housing price correction, low unemployment levels, nominal wage growth, and the fact that most borrowers have locked in low FRMs have buoyed performance, and we have observed only mild delinquency upticks in our rated portfolio (some of which might be due to seasonal factors). For 2024, we are forecasting an unemployment rate of 4.3% and a real GDP growth of 1.5%. As expected, prepayments continue to drag and hover at or below historical turnover/mobility rates (see chart 6). The one-month conditional prepayment rate (CPR) for prime cohorts (Prime 2.0 and CRT) were approximately 5%, which is lower than the historical turnover/mobility CPR of roughly 6%, based on agency loan behavior in rising interest rate periods. The slow speeds are no doubt due primarily to borrowers' unwillingness to give up ultra-low mortgage rates (sometimes referred to as the "golden handcuffs"). The secular shift in remote working norms may have also contributed to muted turnover, as individuals don't necessarily have to move when their job relocates to a different city.

Chart 6

Potential credit impact of adjustable-rate mortgages (ARMs)

Over 90% of mortgages in the U.S. are fixed rate and serve as a hedge against inflation, although the origination share may be lower as ARMs are being used more frequently now to manage costs. However, a portion of our outstanding RMBS portfolio has exposure to ARMs, which typically take on a hybrid structure with a fixed-period of three, five, seven, or 10 years at a rate slightly below the FRM, followed by a floating rate for the remainder of the 30-year term. Because many of these ARMs were originated in a lower-rate environment, they are exposed to the risk of interest rate shocks, which could affect performance of related RMBS. To determine the potential impact, we examined outstanding ARM loans in our portfolio (predominantly non-QM/debt service coverage ratio loans) with initial fixed-period resets between August 2022 and March 2023. We compared performance (by loan count) before the initial fixed-payment reset to the performance after the payment reset. Before the initial fixed-payment reset, the average loan coupon was 6.21% and, after the payment reset, the average loan coupon increased to 8.03%, representing a reset of over 180 bps. Chart 7 shows that the reset did result in observable performance differences, as average 30-day-plus delinquencies jumped to around 6.2% post reset from around 4.6% prior to the reset.

Chart 7

While ARMs make up less than 10% of the mortgages outstanding in the U.S. (far less than prior to the GFC), they tend to be more popular in the non-QM subsector. Of the ARMs within our non-QM rated portfolio, the majority have fixed periods of five or seven years. This means that the 2017 and 2019 vintages will be subject to resets in 2024. However, these mortgages tend to feature periodic (typically annual) caps that limit the size of the rate increase in a given period (see chart 7). The typical rate cap is 200 bps, although about one-third of these ARMs have caps of 100 bps, and a small portion have higher caps. The net effect is that upon reset, the average rate increase was about 175 bps. Note that in chart 7, we include only loans that survived with pre- and post-reset six-month data. Chart 8 depicts upcoming resets in 2024 for the given population we assessed. Chart 8 also compares post-reset debt to income (DTI) to a hypothetical reference DTI of 36%.

Chart 8

A Jumpstart To The New Year

The markets for housing and mortgages will likely be invigorated in 2024 to the extent the Fed begins to cut its benchmark rate, benefitting borrowers as mortgage rates follow suit. It remains to be seen how much rates need to fall for a meaningful number of borrowers to break free from their golden handcuffs and free up some of the existing home supply. Although the impact of rate movements on home prices is not always obvious, some supply/demand fundamentals typically drive home price appreciation (HPA). The recent and sudden jump in rates dampened demand and caused short-term negative HPA. However, it also led to the golden handcuffs phenomenon that resulted in the dampened supply, which subsequently supported HPA. While falling rates could boost demand and put upward pressure on house prices, they could also lead to an increase in inventory levels. Therefore, the supply side needs to be considered as well. Although HPA is viewed as credit-positive, excessive HPA could make an already unaffordable market even more problematic for the prospective home buyer. In a recent commentary (see "Residential Mortgage Performance: 'Like A Rock' And 'Still the Same'," published Nov. 3, 2023), we showed that each decline of 25 bps in the FRM would boost purchasing power by 2% to 3%, which would likely manifest as positive HPA. As affordability continues to erode, the rental market will likely remain active, with vacancies at low levels. However, the rental market will also depend on the developments in single-family-rental and multifamily markets over the next few years.

As we move into the new year, we will also be keeping an eye on the FICO update from the Federal Housing Finance Agency and developments related to the 2023 brokerage commission case.