Mortgage and structured products markets Report Interpretation

Overview

This mortgage and structured-products market update examines agency MBS technicals and credit-score transition effects, non-QM RMBS call activity, CMBS refinancing and maturity risks, and relative-value views across structured-credit sectors. Goldman Sachs sees near-term pressure on agency MBS spreads but identifies selective opportunities elsewhere in structured credit.

Core views

Agency MBS technicals point to greater dependence on money managers in the second half of 2026. Net issuance totaled $123 billion in January-August, versus $107 billion in the same period of 2025, and Goldman Sachs expects issuance to approach its previously published $210 billion full-year target, although higher mortgage rates create some downside risk. The report assumes $100 billion of net demand each from domestic banks and overseas investors, while judging GSE valuations unattractive for renewed portfolio net additions. As a result, money managers may need to buy more MBS in 2H2026 than in 1H2026; all else equal, this higher absorption burden should bias spreads wider. Rate sensitivity remains the immediate driver of the MBS basis. A sharp Treasury selloff lifted the 10-year yield to its highest level since October 2023 and widened current-coupon MBS Treasury OAS to 29bp, the widest level of 2026 and 3bp above the long-term average. Goldman Sachs remains neutral on the MBS basis, expecting the relationship between yields and MBS spreads to persist while upward pressure on long-end rates continues. If long-end yields stabilize after the forthcoming FOMC meeting and rate volatility falls, bank demand could improve after quarter-end. Its relative-value table adds that elevated yields, geopolitical risks and Fed-policy uncertainty skew near-term MBS spread risks wider, although it expects current-coupon OAS to finish 2026 near its long-term average if inflation normalizes and de-escalation expectations materialize. The expansion of VantageScore 4.0 acceptance is central to the report's agency-MBS credit discussion. Fannie Mae and Freddie Mac previously accepted VS4 from a limited lender set, but all lenders are now eligible to use it; FICO 10T is not yet accepted by the GSEs. VS4's share of September-to-date issuance was 4.7%, down from 6% in August, and most issuance still comes from Rocket and UWM, whose VS4 share appeared to plateau at about 22% on average in August and September. Goldman Sachs believes broader eligibility could raise adoption: its earlier framework estimated possible 35% VS4 penetration, while broader use with fewer restrictions could lift Rocket and UWM usage above 50%. The report expects only limited aggregate prepayment effects from the VS4 shift. Revised GSE loan-level price adjustment grids effectively treat a VS4 score as roughly 20 points above a classic FICO score: the adjustment for a 740 VS4 borrower matches that for a 720 classic-FICO borrower. Therefore, a borrower receiving a typical 20-point score increase should not receive a new refinancing incentive; only borrowers with improvements exceeding 20 points may obtain lower mortgage rates. Goldman Sachs expects the largest potential rate improvement among lower-FICO borrowers and steeper WALA ramps for them, driven by credit curing and the transition to VS4. It also expects GSE acceptance of FICO 10T in coming months. FHA plans to accept both VS4 and FICO 10T for case numbers assigned from January 1, 2027, potentially ahead of the GSEs. Because FHA does not use risk-adjusted pricing, credit scores are less important at origination but remain important inputs for investor delinquency and roll-rate models. The unchanged minimum score despite newer models generally producing higher scores could modestly expand the FHA credit box, although the report notes that sub-620 loans are a small share. FHA's allowance for commingling scores from multiple models could make post-January credit-score data incomparable with issuance through December 2026, complicating roll-rate modeling and buyout CPR forecasts for GN pools. In non-QM RMBS, calls remain unusually strong. About 20 deals were called from June through August, taking the estimated 2026 total through August to 43—just below the 2021 full-year high of 45—with 17 distinct issuers, a new high. The 2022 and 2023 vintages account for most activity, but their timing differs. About 85% of called 2023-vintage deals were called within three months of becoming eligible after their three-year seasoning requirement, whereas 2022-vintage deals are increasingly called at or shortly before a 100bp coupon step-up at 48 months. Goldman Sachs attributes the distinction to the lower financing cost of 2022 deals, which initially favored leaving them outstanding, until the approaching step-up changed that calculation. Higher gross weighted-average coupons and financing rates appear to be the main economic drivers of calls. Goldman Sachs expects non-QM calls to remain robust through year-end as more 2023 deals become eligible and additional 2022 deals reach their step-up dates. It estimates total 2026 calls could approach 80, with reissued balances supporting R-deal issuance in 2026 and early 2027. The report estimates 2026 R-deal issuance at roughly 96% of total called-deal balances across the same shelves between 2026 and the end of 2025, and forecasts total 2026 non-QM issuance of $120 billion. In relative value, Goldman Sachs is overweight non-QM AAAs versus the MBS basis, expecting spread differentials to narrow because of non-QM's better convexity profile and strong investor demand; it prefers higher-quality exposure because BBB-B spreads are tight versus AAA. CMBS maturity pressure remains elevated despite some improvement in past-maturity balances. The share of CMBS collateral maturing within two years is near the highest level since 2012, while loans past their original maturity date remain at the 80th percentile since 2012. Goldman Sachs frames this as a refinancing challenge because fixed-rate loans originated during 2018-22 at historically low coupons and tight spreads are now maturing in a sharply different rate environment. Performance is highly uneven by property type: more than half of industrial and multifamily past-maturity loans resolve within one year, while the share of office loans resolving within that period is near a record low. Conditional on resolution, office loans have a loss rate of nearly 17%, almost double the next-highest sector. The report sees a modest risk of agency-CMBS spread widening in coming months due to negative supply seasonals, but expects any widening to remain contained at less than 5-7bp, all else equal. Across other sectors, the report favors MBS over investment-grade corporates for diversification, prefers specified pools to TBAs near term, and favors middle loan-balance stories of $175,000-$275,000 for insulation from prepayment and policy risks. It is modestly underweight GNs, citing seasonal supply, and sees FN 6.0s as favored by carry in a steady-rate environment, while a rate rally would favor FN 5.0s and a further sharp backup could hurt FN 5.5s and 6.0s through extension risk. Goldman Sachs is neutral on CES and RPL AAAs, overweight consumer unsecured versus subprime auto and home-improvement ABS versus consumer unsecured, underweight subprime auto and solar, and overweight aircraft ABS relative to A-rated subprime auto ABS. It is neutral on digital infrastructure overall but favors hyperscaler exposure and data-center ABS trusts with long operating histories, high DSCRs and long lease terms.

Analysis framework

Goldman Sachs combines supply-demand estimates, spread and OAS comparisons, rate and volatility sensitivity, issuance and call-vintage analysis, credit-score policy changes, and property-type CMBS resolution data. It then translates these findings into relative-value views across MBS, RMBS, CMBS and ABS sectors.

Methodology notes

• Fixed Income and CreditSpread analysis

  MBS option-adjusted spread and cross-asset spread comparisons

  The report uses Treasury OAS, historical averages and relative spread levels to assess MBS valuation and compare structured-credit sectors with corporate bonds.

• Fixed Income and CreditDuration and Convexity Analysis

  Rate sensitivity, extension risk and convexity profiles

  Goldman Sachs evaluates how changes in long-end yields affect MBS spreads, coupon performance and prepayment or extension risk.

• Industry AnalysisSupply-demand framework

  Agency MBS issuance and investor-demand technicals

  The report estimates issuance and demand from banks, overseas investors, GSEs and money managers to infer likely pressure on MBS spreads.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Agency MBS

  Near-term spreads are exposed to elevated long-end yields and greater reliance on money-manager demand.

  Strengths

  Goldman Sachs expects current-coupon OAS to finish 2026 near its long-term average under inflation normalization and de-escalation expectations.

  Weaknesses

  Higher issuance, weak near-term bank demand and limited GSE portfolio demand can bias spreads wider.

  Comparison

  Preferred versus IG corporates for diversification benefits.

  Risks

  Further long-end rate increases, rate volatility, geopolitical risks and Fed-policy uncertainty.

• Non-QM RMBS

  Robust deal calls may generate R-deal issuance and support 2026 and early-2027 supply.

  Strengths

  Goldman Sachs favors non-QM AAAs versus the MBS basis because of better convexity and strong demand.

  Weaknesses

  BBB-B spreads are tight relative to AAA exposure.

  Comparison

  Expected spread differentials versus the MBS basis to narrow.

• CMBS

  A large maturity wall and high refinancing rates pressure collateral, particularly office properties.

  Strengths

  Industrial and multifamily past-maturity loans show stronger resolution rates.

  Weaknesses

  Office loans resolve slowly and have nearly 17% losses when resolved within 12 months.

  Comparison

  Goldman Sachs prefers SFR over CMBS, citing stronger single-family rent growth.

  Risks

  Near-term spread widening from negative supply seasonals and refinancing pressure.

• GNMA MBS

  Government MBS credit-score reporting may become less comparable after FHA model changes.

  Weaknesses

  Seasonals could increase GN supply more than conventional supply.

  Comparison

  Goldman Sachs prefers the GN I/FN 4.0s swap over the GN I/FN 4.5s swap.

  Risks

  Credit-score commingling could complicate roll-rate and buyout CPR modeling.

Key data

• Agency MBS net issuance$123 billionJanuary-August 2026, versus $107 billion in January-August 2025; Goldman Sachs expects full-year issuance to approach $210 billion.
• Current-coupon MBS Treasury OAS29bpHighest level of 2026 and 3bp wider than the long-term average.
• VS4 share of issuance4.7%September-to-date issuance, versus 6% in August.
• Rocket and UWM VS4 shareAbout 22%Average for August and September; usage appeared to plateau.
• Non-QM deal calls43Estimated calls through August 2026, versus the 2021 full-year high of 45.
• Potential 2026 non-QM deal callsNear 80Goldman Sachs estimate based on upcoming eligibility, rate step-ups, financing incentives and issuer call history.
• 2026 non-QM issuance forecast$120 billionCall activity and related R-deal issuance could support full-year issuance.
• Office CMBS loss rate upon timely resolutionNearly 17%Conditional on resolving past-maturity balances within the following 12 months; nearly double the next-highest sector.

Impact & implications

The report's central near-term implication is that agency MBS valuations remain vulnerable to persistent long-end yield pressure and a larger money-manager absorption requirement. Credit-score-model adoption could alter mortgage prepayment and credit-data interpretation at the margin, while robust non-QM calls may support issuance. CMBS refinancing stress remains concentrated in office collateral despite better resolution trends in industrial and multifamily loans.

Risks

• Continued upward pressure on long-end Treasury yields could widen the MBS basis.
• Higher mortgage rates could reduce agency MBS purchase issuance relative to the report's expectation.
• FHA credit-score-model commingling could make post-January 2027 credit data less comparable and complicate GN-pool modeling.
• CMBS collateral faces significant refinancing pressure as low-coupon loans mature into a much higher-rate environment.
• Office CMBS collateral has particularly weak past-maturity resolution rates and high loss severity.

What to watch

• Whether long-end Treasury yields stabilize after the next FOMC meeting and rate volatility declines.
• Domestic-bank and GSE demand for agency MBS, and the scale of money-manager purchases in 2H2026.
• The pace of broader VS4 adoption, GSE acceptance of FICO 10T and disclosure of GSE internal risk scores.
• FHA implementation of VS4 and FICO 10T from January 1, 2027 and the treatment of mixed credit-score models.
• Non-QM call activity, 2023-vintage call eligibility, 2022-vintage rate step-ups and associated R-deal issuance.
• CMBS maturity resolutions by property type, especially office losses and refinancing outcomes.