The just-released non-agency performance data (from November 2025) grabbed more than a few headlines.
Non-QM loans saw a notable jump in early-stage delinquencies, raising understandable questions around the office (ours and others) about whether this move reflects emerging credit stress or something more benign – like, say, bad data.
We ultimately concluded that the increase, while real, is likely temporary. The most plausible explanation for November’s spike in Non-QM delinquencies points to a calendar effect tied to the month ending on a Sunday.
Benign, indeed.
So what happened in November?
In the latest data release from Cotality (formerly CoreLogic), delinquency rates rose meaningfully across the non-agency universe, driven almost entirely by a surge in the 30–59 days past due bucket.
For the servicing month ending November:
- Total Non-Agency 30-day DQs increased 48 basis points, from 3.07% to 3.55%
- Non-QM 30-day DQs increased 41 basis points, from 1.66% to 2.07%
For Non-QM loans, this one-month increase represents the largest jump in early-stage delinquencies since the COVID-related shock in April 2020, when these rates surged from 2.99% to 12.51%. For the broader non-agency universe, the increase was the largest since June 2024.
These figures appear alarming. But a closer examination reveals that, in this case, the calendar may be doing most of the work.
The Sunday Payment effect
November ended on a Sunday (not just that, but on a Sunday that was, for many folks, the end of a four-day holiday weekend). When the final day of the month falls on a weekend, payments made on that day typically do not post until the following business day (in this instance, Monday, Dec. 1). As a result, loans that were paid “on time” (or less than 30 days late at least) can be temporarily classified as 30+ days delinquent for November reporting purposes, even though the borrower ultimately made the scheduled payment.
This “Sunday month-end effect” is well documented and understood. And both internal discussions and external market commentary point to this being the primary driver of November’s delinquency spike. Among external commentators, ICE’s Andy Walden may have summarized it most succinctly: “While the topline delinquency numbers show a sharp increase, we’ve seen comparable spikes in prior years when November ended on a Sunday and scheduled payments didn’t post until early December.”
The effect appears to be amplified with Sunday-ending Novembers in particular (perhaps because of the four-day weekend effect). As noted in the ICE piece, this has most recently happened in 2014, 2008, and 2003, when delinquency rates spiked by 61 bp, 112 bp, and 57 bp, respectively. All of those increases exceeded this year’s roughly 50 bp shift.
Approach 1: A History Lesson
To test whether November’s increase fits a broader historical pattern, we examined the relationship between month-over-month delinquency changes and the day on which the month ended.
Since 2006, there have been 33 months that ended on a Sunday. Over that nearly 20-year period, overall non-agency delinquency levels are broadly unchanged. And yet, those Sunday-ending months consistently exhibit upward pressure on reported 30-day DQ rates.
Key observations:
- Across those 33 Sunday-ending months, non-agency 30-day DQ rates increased by an average of 37 bp
- 30-day DQ rates declined in only 4 of those months
- In the remaining 29 months, delinquency rates increased
- Importantly, 27 of those 29 increases were at least partially reversed in the subsequent month
In other words, when months end on a Sunday, reported delinquencies tend to rise mechanically, only to then fall back once payments post and reporting normalizes.
Chart 1: Month-over-Month Change in Non-Agency 30-day DQ rates (Sunday Month-Ends Highlighted, with green indicating a decline, and red indicating an increase)
Approach 2: Agency Data as a Leading Indicator
Non-agency delinquency data are reported with a one-month lag relative to Agency MBS. As a result, we can use Agency performance as a sort of real-time proxy for how non-Agency data may evolve in the following release.
For the December factor date (corresponding to payments due November 30):
- Fannie/Freddie D30 jumped 19 bp, from 0.73% to 0.92%
- GNMA D30 jumped 41 bp, from 3.84% to 4.25%
Crucially, both measures recovered sharply in December, declining back toward their October levels:
- Fannie/Freddie D30 fell to 0.78%
- GNMA D30 fell to 3.89%
That represents a recovery of 74% and 88%, respectively, of the November spike.
If Non-Agency and Non-QM delinquencies follow a similar pattern, a comparable recovery would imply:
- Non-QM D30 falling back to roughly 1.77%
- Total Non-Agency D30 falling back to roughly 3.20%
These levels would be broadly consistent with pre-November trends and inconsistent with a narrative of accelerating credit stress.
Chart 2: Agency vs. Non-Agency 30-day DQ Rate Changes and Subsequent Recovery (the dashed green and blue lines for December 2025 represent extrapolated D30 rates if Non-agency mortgages see similar recoveries to those experienced by Fannie/Freddie mortgages)
Conclusion
November’s spike in Non-QM delinquencies looks dramatic, but the weight of evidence points to a calendar artifact, not a structural shift in credit performance. Similar spikes usually occur when any month ends on a Sunday and are particularly pronounced when November does. History suggests 2025’s anomaly will be largely reversed in December.
Investors should continue to monitor delinquency trends closely, and we will revisit this analysis when the next Cotality data are released in early February. For now, the data argue for caution, not alarm.
This post provides an update on delinquency rate trends observed in the Non-Agency mortgage market with a deep dive on different vintages and credit segments of the Non-QM market. All of the figures in this post are based on queries of historical CoreLogic Non-Agency data from the most recent factor date (December, 2025) via our proprietary RiskSpan Edge Historical Performance module.
December delinquency rates continue to decline from their post-Covid highs in May 2025:
- As shown in Figures 1 and 2, the 60+ delinquency rate for Private Label Securities 2.0 (loans originated after 2010) is 1.98% as of December, 2025, down from 2.21% in August. The DQ rate for Legacy products (originated prior to 2010) dropped to 9.32%.
- Prime Jumbo mortgages continue to demonstrate the strongest performance from a credit perspective, with delinquency rates at 0.53%.
- 2nd Lien loans, comprising HELOCs and closed end mortgages, had a delinquency rate of 0.91% in Decemeber, down from 1.0% in August
- Non-QM loans delinquency rates declined to 2.68% in December, down from 3.0% in August
Figure 1.
Figure 2.
Figures 3 through 5 show the relative delinquency performance of mortgages across 4 segments of the Non-QM population, which comprises the largest portion of the PLS 2.0 market. While loans with full documentation represent the largest segment of this market from a total outstanding balance perspective, originations have been shifting towards DSCR/Investor and Bank statement loans since 2022.
- Fully documented loans have the lowest 60+ delinquency rate at 0.76%, though this DQ rate is higher than the post-COVID lows of 0.39% seen in October 2022.
- Delinquency rates for DSCR/Investor and Bank Statement loans fell in December to 2.92% and 3.99% respectively.
- Non-QM delinquency rates vary significantly by vintages
- DQ rates are lowest for the 2021 Vintage at 1.94%, driven in part by the much higher proportion of Full Doc loans in this vintage (54%, compared to 29% for the Non-QM population as a whole)
- DQ rates are highest for the 2023 Vintage at 6.02%. This is partially explained by the low proportion of Full Doc loans in this vintage (only 14%). But even when controlling for documentation type, the DQ rates are higher for the 2023 vintage, as shown in Figure 5. This could in part be explained by adverse selection through refinancing, where the borrowers with stronger credit have refinanced into rates that are lower than the 2023 peaks.
Figure 3.
Figure 4.
Figure 5.
Non-QM delinquency rates are highly differentiated by credit quality, but performance is still highly differentiated by documentation type when controlling for credit quality:
- As shown in Figure 6, the 640-680 FICO bucket for the full Non-QM universe has a 60+ delinquency rate that is 10x the rate for the 760+ FICO bucket (8.35% vs, 0.80%). On a relative basis, the delinquency rate is even more differentiated for the Full Doc population, where the 640-680 FICO bucket has a 6.37% delinquency rate compared to a 0.19% delinquency rate for the 760+ cohort.
- As observed in Figure 1, the Full Doc Non-QM loans have a significantly higher FICO score than the DSCR and Bank Statement Non-QM loans (763 vs. 744 and 737 respectively). However, this higher FICO score does not fully explain the lower delinquency rates for the Full Doc loans. Figure 7 shows that delinquency rates for Fully Documented loans are significantly lower than those for the DSCR and Bank Statement loans even within the same FICO bucket.
Figure 6.
Figure 7.
Figures 8 and 9 show the relative delinquency performance of Non-QM mortgages by year of origination. For these charts, vintages prior to 2021 are excluded to avoid the distorting impact of the COVID delinquency shock.
- Figure 8 shows the 60+ delinquency rate for each vintage by factor date.
- After eclipsing the delinquency rate of the 2022 vintage in July, the delinquency rate for the 2023 vintage continued to increase, hitting 6.02% in December
- The 2021 vintage’s 1.94% DQ rate is significantly lower than subsequent vintages in spite of being the most seasoned. This is in part due to the disproportionately high share of full documentation loans in this first post-COVID cohort of Non-QM loans.
- Figure 9 shows the 60+ delinquency rate for each vintage by loan age
- Consistent with the trends observed in Figure 8, the 2023 vintage DQ rates ramp up faster than any of the other vintages.
- The delinquency rates for the 2024 and 2025 vintages are tracking with the 2022 vintages.
Figure 8.
Figure 9.
Given the elevated delinquency rates of Non-QM mortgages relative to Agency and Prime Jumbo mortgages, particularly in the Bank Statement and DSCR/Investor and segments and in the lower FICO ranges, it is important for investors to monitor their portfolios that have Non-QM exposure. Our credit models at RiskSpan model these delinquency roll rates directly, and our modeling team calibrates our suite of models to capture both the overall trends and the differentiated performance across loan and product types. These models are just one component of our scaled analytics solutions to help our clients evaluate risk and make investment decisions.
This post provides an update on delinquency rate trends observed in the Non-Agency mortgage market with a deep dive on different segments of the fast growing Non-QM mortgage market. All of the figures in this post are based on queries of historical CoreLogic Non-Agency data via our proprietary RiskSpan Edge Historical module.
After reaching post-Covid highs in May 2025, delinquency rates have stabilized at slightly lower levels in August 2025, the most recent factor date available from CoreLogic:
- As shown in figures 1 and 2, the 60+ delinquency rate for Private Label Securities 2.0 (loans originated after 2010) is 2.21%, while the DQ rate for Legacy products (originated prior to 2010) continues to fall below the 10% threshold, hitting a post-COVID low of 9.61%
- Prime Jumbo mortgages continue to demonstrate the strongest performance from a credit perspective, with delinquency rates at 0.57%.
- 2nd Lien loans, comprising HELOCs and closed end mortgages, had a delinquency rate of 1.01%.
- Non-QM loans saw delinquency rates remain stable at 3.05%, slightly below the post-COVID peak of 3.17% in May.
Figure 1.
Figure 2.
Figure 3 shows the relative delinquency performance of mortgages across 4 segments of the Non-QM population, which represents the largest portion of the PLS 2.0 market. While loans with full documentation represent the largest segment of this market from a total outstanding balance perspective, originations have been shifting towards DSCR/Investor and Bank statement loans since 2022 (see Figure 4). In 2025, the combined volume of originations in the DSCR/Investor and Bank statement segments was about four times the volume of loans originated with full documentation.
- Fully documented loans have the lowest 60+ delinquency rate at 0.89%, though as this segment seasons, the DQ rate continues to creep up from the post-COVID lows of 0.39% seen in October 2022.
- Delinquency rates for DSCR/Investor and Bank Statement loans stabilized in August at 3.34% and 4.41% respectively, slightly lower than their post-COVID peaks seen in May 2025
Figure 3.
Figure 4.
Figures 5 and 6 show the relative delinquency performance of Non-QM mortgages by year of origination. For these charts, we exclude vintages prior to 2021 to avoid the distorting impact of the COVID delinquency shock.
Figure 5 shows the 60+ delinquency rate for each vintage by factor date.
- The delinquency rate for the 2023 vintage hit 5.25% in August, surpassing 2022 as the vintage with the highest delinquency rate.
- In spite of being the most seasoned, the 2021 vintage’s 2.04% DQ rate was significantly lower than the subsequent 2022 and 2023 vintage. This is largely due to the disproportionately high share of full documentation loans in this first post-COVID cohort of Non-QM rates, which can be seen in Figure 4. By contrast, the 2022 and 2023 vintages are composed primarily of the higher risk DSCR and Bank Statement originations.
Figure 6 shows the 60+ delinquency rate for each vintage by loan age.
- Consistent with the trends observed in Figure 5, we see the 2023 vintage DQ rates ramp up faster than any of the other vintages.
- The 2024 vintage is tracking between the 2022 and 2023 vintages.
- While there are only a few months of observations available the 2025 vintage, its delinquency ramp-up is tracking with the other post-2021 vintages
Figure 5.
Figure 6.
Given the elevated delinquency rates of Non-QM mortgages relative to Agency and Prime Jumbo mortgages and the backdrop of housing and macroeconomic uncertainty, it is important for investors to monitor their portfolios that have Non-QM exposure. Our credit models at RiskSpan model these delinquency roll rates directly, and our modeling team calibrates our suite of models to capture both the overall trends and the differentiated performance across loan and product types. These models are just one component of our scaled analytics solutions to help our clients evaluate risk and make investment decisions.
Arlington, VA – February 18, 2025 – RiskSpan, a leading provider of innovative trading, risk management and data analytics for loans, securities and private credit, has announced the release of its latest Non-QM Prepayment Model (Version 3.11), incorporating CoreLogic’s loan-level non-QM performance data. This update significantly enhances prepayment forecasting accuracy for non-QM loans and mortgage-backed securities by leveraging a robust, segmented modeling approach.
RiskSpan’s new non-QM prepayment model introduces a two-component framework that improves the precision of prepayment predictions:
- The first component is a Unified Turnover Model, designed to capture base prepayment trends.
- The second component, a Refinance Model Categorized by Documentation Type, is capable of distinguishing among and modeling behavioral characteristics specific to bank statement, debt service coverage ratio/investor, full documentation, and other documentation types
The model is built on loan performance data spanning October 2019 to March 2024 and intelligently incorporates long-term prepayment behavior with conventional loans, addressing the challenge of limited non-QM data history. Key enhancements include:
- Sensitivity to SATO (Spread at Origination) and Burnout Effects, refining prepayment behavior projections.
- DSCR-Specific Adjustments, incorporating prepayment penalty terms and amounts to refine refinance calculations.
By integrating granular loan-level insights from CoreLogic, this release enhances market participants’ ability to accurately assess non-QM prepayment risk, optimize portfolio strategies, and improve secondary market pricing.
“Our latest model delivers a more precise view of non-QM borrower behavior, equipping market participants with the insights needed to manage risk effectively,” said Divas Sanwal, Senior Managing Director and RiskSpan’s Head of Modeling. “By leveraging CoreLogic’s expansive dataset and an expansive GSE dataset, we’re enabling investors to better anticipate prepayment trends and make more informed decisions.” The new model is now available for integration into RiskSpan’s Platform.
The new model is now available for integration into RiskSpan’s Platform.
About RiskSpan
RiskSpan delivers a single analytics solution for structured finance and private credit investors of any size to confidently make faster, more precise trading and portfolio risk decisions and meet reporting requirements with fewer resources, and less time spent managing multiple vendors and internal solutions. Learn more at www.riskspan.com.