Lessons learned in the private-label market in 2024

Chris Helwig | December 20, 2024

This document is intended for institutional investors and is not subject to all of the independence and disclosure standards applicable to debt research reports prepared for retail investors. This material does not constitute research.

As 2024 comes to a close, RMBS investors have likely gleaned some meaningful takeaways on both fundamentals and technical drivers of private-label performance. Originations of non-conforming loans were resilient in the face of higher rates and grew increasingly diverse. From a fundamental perspective, credit concerns remain largely focused on near-prime borrowers that are performing more like subprime than prime borrowers, while in prepayments, investors are looking to ARMs and other collateral attributes that may offset the pronounced effect of lock-in and provide upside to pulling to par on discount bonds. From a valuation perspective, RMBS spreads have breached agency MBS relative value benchmarks but certain products may have additional room to tighten further.

Supply shrugs off higher rates, grows and stretches

Supply of private-label RMBS likely surprised to the upside this year. With roughly $150 billion in total issuance across all flavors of non-agency RMBS, this year’s issuance will be the most since 2021 when the market had much more favorable tailwinds in the form of substantially lower rates and concerted efforts to push more conforming loans to private channels. Securitization issuance doesn’t paint the entire picture though. While securitizations of non-QM loans will be the largest contributor to total issuance, the likely more meaningful evolution this year was the growing presence of insurance balance sheet capital available to fund production of non-QM loans. Insurance capital has proven to be both a stable, permanent capital home for the asset class while also providing somewhat of a governor on non-QM whole loan spreads given a far less volatile cost-of-funds than securitization. The sustained presence of insurance capital has made originators less reliant on the need to reverse engineer loan pricing to securitization economics, subsequently fueling an uptick in originations.

While securitization rates of non-QM whole loans dipped this year given the inverted yield curve and resultant lower cost-of-funds associated with longer dated annuity liabilities than shorter-date securitization ones, supply was buttressed by loans against borrower equity and other, more novel exposures. Securitizations backed by closed-end second liens (CES) witnessed an uptick in supply, increased acceptance amongst investors, particularly at the top of the capital structure and a subsequent tightening across the spread curve. The CES product should continue to grow next year as it should offer both strategic and economic benefits to the originator community. Originators who have the ability to tap their existing MSR book to source CES lending opportunities can reduce customer acquisition cost to originate, protect the value of the existing MSR as it will be less susceptible to cash-out refinancing and garner attractive gain-on-sale margins against the backdrop of subdued first lien origination volumes. Equity extraction was a consistent tailwind to private-label originations broadly as HELOCs, Home Equity Investment (HEI) and cash out refinancings of non-QM investor and DSCR loans all contributed in varying degrees to total supply.

ARMs and other collateral attributes have helped offset the lock in effect

The promulgation of second lien lending served to exacerbate the existing lock-in effect. Investors looking for upside to relatively elevated prepayments on discounted private-label MBS were able to side step the lock in effect to some degree given the presence of certain collateral attributes, although that opportunity appears to be somewhat of a melting ice cube, particularly when it comes to ARM collateral backing non-QM trusts.

The re-steepening of the yield curve over the past 12 months has driven elevated prepayment rates across ARM loans collateralizing Non-QM trusts. To date, adjustable-rate loans have made up more than half of total Non-QM prepayments despite making up less than 30% of over $300 billion of Non-QM collateral analyzed. ARM collateral comprises just over 15% of outstanding Non-QM collateral, making it increasingly difficult, and likely more costly, to gain meaningful exposure (Exhibit 1). Prepayment rates on outstanding hybrid loans appear likely to remain elevated as the WAC on the outstanding population of ARM loans is roughly 40 bp greater than the population of prepaid loans.

Exhibit 1: Non-QM Collateral Attributes – Prepaid versus Outstanding

Source: Santander US Capital Markets, CoreLogic LP

Screening for other collateral attributes that may provide upside to prepayments, loans with higher balances have exhibited faster out-of-the-money speeds than smaller ones. Controlling the population for fixed-rate observations, loans with balances greater than $1 million and -125 bp of refinancing incentive prepaid anywhere from 4.5 to 6.5 CPR faster than other cohorts. Faster observed out-of-the-money speeds on large balance loans are somewhat of an anomaly as they have historically exhibited steeper prepayment S-curves and slower-out-of- the-money speeds than smaller balance loans. (Exhibit 2)

Exhibit 2: Larger loans exhibit faster-out-of-the-money speeds

Source: Santander US Capital Markets, CoreLogic LP, Observations limited to fixed-rate collateral only with a two-year observation period of June 2022 through November 2024. Refi incentive is SATO adjusted to normalize for risk-based pricing at origination.

Elevated deep out-of-the-money speeds were more pronounced for loans collateralized by investment properties and second homes where there is likely less lock-in effect as investors may look to extract equity associated with home price appreciation and second homeowners tend to exhibit higher rates to mobility and turnover. From a credit perspective, deep credit investors may be able to add large loan exposure, particularly in lower LTV investment properties or second homes where compensating credit characteristics may help to offset exposure to large loan risk.

Near prime looks more like subprime

Weaker performance across consumer exposures has been primarily localized to low-to-moderate credit quality borrowers where pandemic and post-pandemic fiscal policies including mortgage payment forbearance and direct fiscal stimulus translated to upward drift in borrowers’ credit scores. Per studies conducted by the Federal Reserve, the share of borrowers with subprime credit scores fell to the lowest observed levels since the late 1990s, with the share of those borrowers falling from 23% to 18% of total.* The Fed’s study defines subprime borrowers as having credit scores of less than 620 with near prime borrowers maintaining a credit score between 620 and 719. Based on this taxonomy, FICO inflation likely had the most meaningful impact in the near prime category where previously subprime consumers had been re-classified.

As the effects of fiscal policies abated and the broader US economy has weakened, fundamental borrower performance in near prime credits has begun to deteriorate. Looking at both serious delinquency rates and rolls of performing loans to early-stage delinquency, lower-tier near prime loans are performing markedly worse than stronger credits. Consistent with the Fed’s analysis, loans in Non-QM trusts with FICOs less then 720 are rolling to early stage delinquencies at markedly higher rates than those with scores greater than 720. A similar phenomenon is apparent in 60+ day delinquency rates as well (Exhibit 3).

Exhibit 3: Roll and seriously delinquent rates elevated on near prime Non-QM loans

Source: Santander US Capital Markets, CoreLogic LP Analysis Exclusive to 2019-2023 vintage Non-QM transactions

Digging a bit deeper into performance of near prime loans in Non-QM trusts shows dispersion in performance based on certain loan attributes. Owner occupied loans exhibit higher roll and delinquency rates than both investor loans and second homes. Somewhat surprisingly, LTV underwritten loans in the form of both no ratio and low DSCR loans exhibit the highest roll and serious delinquency rates relative to other forms of underwriting such as full or alternative documentation loans. While it’s surprising that these loans are exhibiting higher delinquency rates, it is potentially supportive of no loss abatements of these loans given their generally lower ingoing original LTVs relative to other cohorts.

Non-agency spreads breach agency benchmarks but some have room to tighten further

This year was more of a reminder than a lesson for certain non-agency investors. And that reminder was   spreads on certain private-label exposures can and sometimes do trade through those of agency MBS, a phenomenon that was last evidenced in a meaningful way in 2021. As covered in the private-label MBS buyer’s guide for 2025, top tier non-QM ‘AAA’s are currently trading through current coupon MBS nominal spreads. Certain private-label exposures’ ability to trade through MBS spreads is generally a function of the market perceived duration of these assets, allowing them to de-couple from an MBS benchmark to a short duration corporate benchmark.

* The Effects of Credit Score Migration on Subprime Auto Loan and Credit Card Delinquencies Jan. 12, 2024