Register here for next month’s call: Thursday, February 19th, 2026, 1 p.m. ET. 

In the January Models & Markets call, our quantitative modeling team hosts their first monthly deep dive of the year into prepayment model performance, an updated analysis of second liens and HELOCs using Equifax data, and the evolving macroeconomic backdrop shaping mortgage markets. 

Here’s a quick recap in case you missed it. 

(Click here for the entire 20-minute recording or continue reading for a summary.)  

Revised HELOC and HEL Results Using Equifax ADS Data

  We performed a comprehensive analysis of second liens and HELOCs using Equifax’s Analytic Data Set (ADS), which represents a 10% anonymized sample of U.S. consumer credit data at the tradeline level. 

Following the resolution of data quality issues identified in an earlier analysis, the revised results now align much more closely with economic intuition. Prepayment speeds behave consistently across vintages, credit score bands, and refinancing regimes. 

One key takeaway holds that higher credit score borrowers tend to prepay faster, particularly during refinancing waves, while lower credit score segments remain slower. This pattern is especially evident in post-COVID vintages. Overall credit quality for HELOCs and second liens remains strong, with performance clustering closer to the highest credit score bands. 

Another notable observation is the role of seasonality in newer HELOC vintages. In a high-rate environment with limited refinancing activity, turnover-driven prepayments become more prominent. Baseline prepayment speeds for HELOCs are running around 15 CPR, higher than what is typically observed in first-lien portfolios under similar conditions. These dynamics provide useful signals for understanding how first-lien behavior may differ when second liens or HELOCs are present on the same property. 

  We plan to expand this analysis further, including deeper investigation into correlations between first- and second-lien prepayment behavior. 

Mortgage Rates Remain Likely to Stay Higher for Longer 

The broader economic outlook remains one of persistence rather than relief. Federal Reserve projections point to unemployment stabilizing around the low-4% range and real GDP growth near 2% over the medium term. Meanwhile, expectations for the fed funds rate suggest limited room for significant cuts beyond 2026. 

Longer-term rates tell a similar story. Consensus forecasts indicate the 10-year Treasury is unlikely to fall meaningfully below 4% over the next two to three years, implying mortgage rates are likely to remain near (and potentially above) the 6% level for much of the period ahead. Temporary dips tied to policy announcements or market events have proven short-lived, with rates quickly reverting back toward recent levels. 

Consumer Stress Continues to Build 

While headline spending remained strong during the most recent holiday season, the composition of that spending tells a more cautious story. Consumers increasingly favored lower-cost retailers, suggesting budget sensitivity and selective spending behavior. 

Survey data reinforces this theme. Year-over-year consumer sentiment and expectations have declined meaningfully, and perceptions of job insecurity (particularly among college-educated workers) have become more negative. These dynamics could have important implications for credit performance and housing activity as economic uncertainty persists. 

Prepayment Model Performance: v. 3.7 Continuing to Track Market Performance Well 

RiskSpan’s prepayment models continue to perform well across Agency collateral. 

RiskSpan’s Prepayment Model v3.7 continues to demonstrate strong performance across collateral types. Recent back-testing shows that model projections remain closely aligned with realized speeds, even as seasonal effects and calendar nuances influence month-to-month results. 

For conventional 30-year loans with lower coupons, December’s modest uptick in observed CPRs was largely attributable to four additional collection days relative to November. After adjusting for day count effects, actual prepayment speeds continue to trend lower, consistent with expectations in a higher-rate environment. 

Premium cohorts also remained largely stable. Despite a brief decline in mortgage rates late last year, the move was insufficient to trigger a meaningful new refinance wave. Most refinance-eligible borrowers have already acted, and the refinancing “pull-forward” effect appears largely exhausted. This dynamic is also visible in the S-curve, which has flattened back toward historical averages after October’s temporary acceleration. 

Agency collateral shows similar patterns. Ginnie Mae discount cohorts tracked model expectations closely, while premium cohorts remained flat. One area of ongoing refinement is deep in-the-money, very high-coupon Ginnie Mae loans, where actual speeds have run slightly slower than model projections as refinance incentives flatten out earlier than in prior cycles. 

Looking Ahead 

In summary: 

• RiskSpan’s Prepayment Model v3.7 continues to perform well across most collateral segments 

• HELOC and second-lien analysis using Equifax data now shows economically intuitive and stable results 

• Mortgage rates are likely to remain near 6% in the absence of a major macro shock 

• Consumer behavior is showing increasing signs of stress and caution 

• RiskSpan plans to release additional analytics later this year, including a new non-QM credit model in the first half of the year and a next-generation prepayment model in the second half. 

We continue to add additional analytics reports on the Platform. Please visit www.riskspan.com/request-access to request free access. 

As always, please feel free to contact us to discuss or learn more. 

Register here for this month’s call: Thursday, January 22nd, 2026, 1 p.m. ET. 

Just before the holidays, RiskSpan’s quantitative modeling team hosted its December Models & Markets call, offering its monthly, detailed look at prepayment model performance, evolving macroeconomic conditions, and what to expect in 2026. Led by Shane Lee and Divas Sanwal, the discussion highlighted a housing and credit market navigating elevated rates, slowing growth, and increasing consumer stress. 

Here’s a quick recap in case you missed it. 

(Click here for the entire 24-minute recording or continue reading for a summary.)  

Why Rate Cuts Aren’t Lowering Mortgage Rates 

Although the Federal Reserve delivered multiple rate cuts toward the end of 2025, the Fed Funds rate remains in the 350–375 basis point range, with futures markets expecting only gradual additional cuts in 2026. As the following charts and tables illustrate, even a move toward 300–325 bps next year leaves policy rates well above pre-pandemic norms. 

More importantly for housing, longer-term rates continue to dominate mortgage pricing. Market consensus forecasts presented on the slides show the 10-year Treasury remaining above 4% for the next two to three years, a view that has remained remarkably stable across forecasting sources. As a result, mortgage rates have been largely unchanged over recent months despite easing monetary policy. 

The implication is clear: refinance and cash-out activity remain extremely constrained and are likely to stay that way well into 2026. Any incremental increase in prepayment activity will come principally from turnover, not rate-driven refinancing. 

Home Prices: Growth Slows, Regional Divergence Emerges 

We used unadjusted Case-Shiller and FHFA data to highlight that month-over-month home prices declined across many large metro areas, even where seasonally adjusted figures appear more stable. Seasonal patterns have shifted materially in recent years, making unadjusted trends especially informative. 

The FHFA four-quarter appreciation map illustrated this growing regional dispersion. Parts of the Sun Belt, including California, Texas, and Florida, have experienced notable price declines, with the Fort Myers area standing out as a recent weak spot. At the same time, select Northeast markets continue to see positive appreciation, with areas near New York showing some of the strongest gains. 

Overall, while a broad-based housing downturn has not materialized, slowing appreciation reduces borrowers’ financial flexibility and reinforces the current lock-in environment. 

Consumers Under Pressure 

As has been a recurring theme in several of our recent monthly calls, the consumer credit environment is showing increasing signs of strain. 

Unemployment has edged higher, reaching 4.6% in November, with younger workers (ages 16–25) experiencing disproportionately higher joblessness. Inflation, while easing slightly, remains stubbornly above target, with recent CPI readings still near 2.7% year over year. 

We are also continuing to see historically high levels of consumer debt and a notable slowdown in spending growth. Unlike typical holiday-season patterns, consumer spending has not accelerated meaningfully, suggesting households are becoming more selective and cautious. 

One particularly telling trend is the rapid growth of buy now, pay later (BNPL) usage. Increasing reliance on BNPL for essential purchases points to tighter household budgets and reduced financial resilience. 

Taken together, these indicators support expectations—also shown in the Fed’s December Summary of Economic Projections—that GDP growth is likely to remain near or below 2% over the next several years, while credit performance warrants close monitoring. 

Prepayment Model Performance: Holding Up Across Collateral Types 

RiskSpan’s prepayment models continue to perform well across Agency collateral. 

For Fannie Mae and Freddie Mac pools with WACs of 5.5% and below, observed turnover speeds declined modestly month over month. As highlighted below, this softness largely reflects seasonal effects and a shorter reporting month. While the model projected slightly higher speeds, overall alignment with observed behavior remained strong. 

For higher-coupon GSE collateral (6.0% and above), December marked a normalization following unusually aggressive prepayment speeds observed in the prior month. As shown in the charts, observed speeds moderated, allowing the model to close the gap and better track realized behavior. 

A similar pattern emerged in the Ginnie Mae collateral, with both discounted and premium coupon cohorts showing improved alignment between modeled and observed speeds. In particular, the moderation in higher-coupon Ginnie Mae prepayments mirrored trends seen in the GSE universe, underscoring the consistency of borrower behavior across agency channels. 

During Q&A, the team also addressed VA loan performance. Internal loan-level analysis suggests VA loans tend to prepay faster than baseline model projections, an area RiskSpan continues to evaluate closely.  

Looking Ahead: 2025 in Review and What’s Coming in 2026 

In 2025, RiskSpan delivered several major Platform enhancements: 

• Prepayment Model v3.7, introducing an out-of-the-money (OTM) slope to better capture turnover lock-in effects 

• Prepayment Model v3.8, adding a new ARM sub-model and additional tuning controls 

• Prepayment Model v3.11, a fully redeveloped framework for non-QM collateral 

• Credit Model v7.0, featuring a full delinquency transition matrix for GSE and Ginnie Mae loans 

Looking ahead, we outlined an ambitious 2026 release schedule, including: 

• A Non-QM Credit Model v7.1 with full delinquency transitions, expected in the first half of the year 

• A broader non-agency credit model later in 2026 

• A completely new prepayment framework—currently referred to as Prepayment Model 4.0—built from the ground up 

We continue to add additional analytics reports on the Platform. Please visit www.riskspan.com/request-access to request free access. 

As always, please feel free to contact us to discuss or learn more. 

Register here for next month’s call: Thursday, December 18th, 2025, 1 p.m. ET. 

Each month, we host a Models & Markets call to offer our insights into recent model performance, emerging credit risks, and broader economic indicators. This month’s call reviewed recent prepayment performance, presented new research on identifying cash-out refinance activity in GSE data, and walked through key macroeconomic and consumer-debt indicators shaping mortgage behavior going into 2026. 

Here’s a quick recap in case you missed it. 

(Click here for the entire 24-minute recording or continue reading for a summary.)  

New Research: Estimating Cash-Out Refinance Activity Using GSE Data 

Cash-out refinance is a component of prepayment modeling that has traditionally been difficult to observe directly. Shane Lee explained how we have been getting at it using publicly available GSE performance data.

Originations vs. Prepayments: Understanding the Gaps 

Voluntary prepayments consist of turnover, rate-refinance, and cash-out refinance components. While originations include a loan-purpose indicator (“purchase,” “refinance,” “cash-out”), payoff data does not. 

Nationally, the gap between prepaid loan counts and contemporaneous originations is significant, especially in earlier years. This is driven in part by new construction, properties without existing liens, and cross-region relocations. 

To improve attribution, our team has been evaluating data at the ZIP3 level, where prepay and origination volumes show much tighter alignment. Shane presented examples, including ZIPs near Ventura, Tucson, St. Louis, Boulder, and Austin, demonstrating that refinances and cash-outs can be reasonably inferred when prepaid loan totals track closely with origination totals in the same geography. 

Where origination and prepay counts align well, origination loan-purpose shares can serve as a proxy for prepay-purpose shares, enabling estimation of the cash-out fraction among prepaid loans. 

Prepayment Model Performance: Stable Overall, With Pockets of Divergence

Guanlin Chen presented a review of our v3.7 model back-testing results. In summary: 

Low-Coupon (≤5.5%) Conventional and Ginnie Cohorts 

Actual October CPRs tracked the model closely for low-coupon pools across Fannie, Freddie, and Ginnie. October’s slight upward movement in discount speeds (which the model had projected to decline) was explained by a calendar effect: one additional collection day offset typical seasonal slowdown. 

When adjusting for day-count, both actual and projected CPRs show similar downward trends. The alignment reinforced Guanlin’s point that lock-in remains firmly intact. Despite lower rates during parts of October, borrowers with sub-4% or low-4% mortgages still show little inclination to refinance, consistent with recent months. 

High-Coupon (≥6%) Cohorts: Speeds Running Hotter Than Expected 

The premium sector told a different story. Borrowers holding 6%–7% coupons responded more aggressively to rate movements than historical incentive-matched periods would suggest. The S-curve steepened further in October, with realized CPRs meaningfully exceeding v3.7 model predictions. 

To address this, RiskSpan’s v3.8 prepayment model introduces a configurable “in-the-money multiplier” that allows users to steepen the S-curve to better capture this more responsive behavior. 

Outliers and Ongoing Calibrations 

While most premium segments prepaid faster than expected, deep-in-the-money Ginnies (WAC >7%) actually prepaid slower than v3.7 projected. We are actively evaluating updated calibration approaches for these cohorts. 

Market Indicators: Rates, Labor Markets, Home Prices, and the Fed 

Mortgage News Daily data showed a recent ~25bp increase in the 30-year fixed rate. The prevailing question on clients’ minds—“Where do rates go from here?”—was addressed via futures and FedWatch probability data: 

• Fed Funds futures suggest the policy rate will likely remain unchanged in December, despite fresh unemployment data. 

• Projections show the 10-year Treasury hovering around 4% for the next several years, implying mortgage rates likely remain above 6% through 2026. 

Labor Market Softening 

The latest (delayed) September unemployment rate rose to 4.4%. Rising unemployment, paired with persistent inflation pressures, creates a challenging backdrop for housing demand. 

Home Price Growth Slowing Nationally 

Case-Shiller data, nationally and across metros, showed: 

• A 0.3% month-over-month national decline in the latest reading. 

• Major metros increasingly showing broad-based price deterioration, with formerly resilient cities like Los Angeles slipping negative. 

While inventory is rising toward a buyer-leaning market, transaction volumes remain soft. 

Consumer Debt: Elevated, Shifting & Stress-Inducing 

Debt rose $200B quarter-over-quarter, with long-term increases far outpacing inflation and population growth in several categories: 

• Student loans: +600% since 2003 

• Mortgage balances: +165% 

• Auto loans: similarly elevated 

Inflation (+71% cumulative since 2003) and adult population growth (~6%) alone cannot explain these increases. 

Aging Households Carrying More Debt Than Ever 

A striking trend: borrowers 60+ years old have experienced 300–500% increases in total debt held. 

In 2003, the 70+ population held only 4% of total U.S. household debt. 
In 2025, that share stands at 10%. This is an extraordinary shift.

This appears to be evidence of structural strain: As people age, they are unable to pay down their debts. Also, wage growth has not kept up with inflation.

Younger households, meanwhile, face increasing difficulty obtaining new credit.

We continue to add additional analytics reports on the Platform. Please visit www.riskspan.com/request-access to request free access. 

As always, please feel free to contact us to discuss or learn more. 

Register here for next month’s call: Thursday, November 20th, 2025, 1 p.m. ET. 

Each month, we host a Models & Markets call to offer our insights into recent model performance, emerging credit risks, and broader economic indicators. This month’s call focused on the impact of the Fed rate cut, key macro indicators and a spotlight on the surging second-lien market. 

Here’s a quick recap in case you missed it. 

(Click here for the entire 23-minute recording or continue reading for a summary.)  

Rates Ease, but Headwinds Persist 

October has brought a modest reprieve in mortgage rates, with the 30-year fixed rate having fallen to approximately 6.2%, the lowest level in nearly a year. 

Affordability remains constrained, however, and long-term headwinds appear far from resolved. Specifically: 

Unemployment remains near 4.2%, and core PCE inflation continues to hover around 2.8%. While steady, this remains above the Fed’s comfort zone. 

Home price growth is slowing nationally, with several major metros posting month-over-month declines. 

Fed Funds futures suggest rates will stay elevated into 2026, with year-end 2025 expectations still in the 3.5–3.75% range. 

Together, these indicators suggest a “higher for longer” policy regime even as the market eyes rate cuts later this year. 

HELOC and Second-Lien Insights: Delinquencies on the Rise 

Leveraging the Equifax Analytic Dataset, a 10% sample of active U.S. credit borrowers with anonymized tradeline-level detail, enables us to dive deep into Home Equity Loans (HELs) and Home Equity Lines of Credit (HELOCs). These asset classes are gaining renewed investor attention as homeowners tap existing equity rather than selling into a high-rate market.

Delinquency rates are trending upward for both HELs and HELOCs, particularly among lower-credit-score borrowers. Aggregated five-year views on page 11 highlight the steady climb, with 600-score cohorts showing the sharpest deterioration. 

These findings echo broader signals of consumer strain visible across other loan products. 

Consumer Balance Sheets Under Pressure

The New York Fed’s Q2 2025 Household Debt and Credit Report underscored the strain many consumers face. Total household debt continues to climb, driven by non-housing credit categories—auto loans, student debt, and revolving balances in particular. 

Credit card and auto loan delinquencies have risen sharply, while mortgage and HELOC performance, though still comparatively solid, are trending downward. Even with stable macro indicators, consumers remain financially stretched. This dynamic is likely to influence credit performance and securitization trends into 2026.

Prepayment Model Updates 

Our prepayment models continue to align well with observed speeds across both Conventional and Ginnie collateral. Lower-coupon collateral (WAC ≤ 5.5%) experienced some deceleration versus forecasts—a function of seasonality and slower housing turnover.  

Higher-coupon cohorts (WAC ≥ 6.0%) reflected more volatility, consistent with recent refinance activity at the margins. 

We continue to add additional analytics reports on the platform. Please visit www.riskspan.com/request-access to request free access. 

Also, please feel free to contact us to discuss or learn more.

Register here for next month’s call: Thursday, October 16th, 2025, 1 p.m. ET. 

Each month, we host a Models & Markets call to offer our insights into recent model performance, emerging credit risks, and broader economic indicators. This month’s call focused on the impact of the Fed rate cut, key macro indicators and a spotlight on the surging second-lien market. 

Here’s a quick recap in case you missed it. 

(Click here to listen to the entire 27-minute recording or continue reading for a summary.)  

Market Backdrop: September 2025

Mortgage rates have hit their lowest levels in nearly a year, averaging close to six percent. The Federal Reserve delivered its first rate cut of the current cycle in September 2025, reducing the target range from 400–425 basis points to a projected 350–375 basis points by year-end. Despite this easing, markets continue to anticipate relatively high rates into 2026. 

Inflation and unemployment are holding stable, but long-term headwinds persist, including sluggish real wage growth and affordability constraints in the housing market. Longer-dated Treasury yields are the key driver of mortgage rates, making them essential for investors to keep tabs on. As bond yields set the tone for borrowing costs across the economy, their movement will be critical in shaping both origination volumes and prepayment activity in the coming quarters. 

Spotlight on Second Liens 

The second lien mortgage market continues its emergence as one of the most active areas in structured finance. Issuance and securitization of second lien products have been increasing rapidly, with no signs of slowing. This expansion is driven in part by rising homeowner demand for tapping into the accumulated equity and lenders’ interest in capturing additional credit exposure in a higher-rate environment. 

Prepayment behavior in second lien mortgages, however, differs significantly from that of first liens. This divergence makes specialized model calibration critical. RiskSpan’s Prepayment Models, calibrated against actual second lien performance, indicate that the models are capturing observed dynamics effectively. With issuance expected to continue climbing, accurate modeling of second lien prepayment risk will remain an essential tool for market participants seeking to price and manage these assets. 

Prepayment Model Updates 

Back-testing continues to show that RiskSpan’s prepayment models are tracking well against observed performance across a variety of collateral types. Recent analysis of agency MBS vintages from 2021 and 2022 revealed that higher-coupon pools, particularly those in the 6.5% range, are slowing more than originally anticipated. 

FN/FH 2021-2022 6.5s

By contrast, lower- and mid-coupon pools—those ranging from 1.5% through 5.5%—have remained steady and closely aligned with model expectations. This outcome reinforces the robustness of the models across different coupon bands and provides confidence in their ability to capture nuanced prepayment behavior. 

FN/FH 2021-2022 1.5s – 3.5s

FN/FH 2021-2022 4s – 5.5s

As we introduced during our August call, consumer credit remains a major focus of RiskSpan’s modeling enhancements. Using the Equifax Analytic Dataset, the team has constructed prepayment aging curves for both auto loans and personal loans. These analyses confirm that borrower credit score bands, measured using VantageScore 4.0, influence prepayment behavior in a manner similar to mortgage loans. For auto loans, the score sensitivity is particularly evident across borrower segments. Personal loan data show similar trends, with one notable difference: the effect of loan term is more pronounced after the first year of loan seasoning. This suggests that term structure plays a more significant role in personal loan prepayment decisions compared to auto loans. 

Auto Loan Prepayment Aging Curves

Personal Loan Prepayment Aging Curves

We are in the process of finalizing these consumer loan prepayment models and will release them shortly on the RiskSpan Platform. This will give clients the ability to incorporate a new level of borrower insight into their own portfolio analytics. 

Looking Ahead 

The integration of Equifax ADS into the construction of prepayment aging curves is just the beginning. We continue to expand our modeling capabilities and data integration in order to provide clients with deeper and more actionable insights. Credit card and student loan models are already in the pipeline, and their release will extend RiskSpan’s modeling coverage across the full spectrum of consumer credit products. 

In addition, the team is adding new analytics reports to the Platform, giving clients free access to timely updates and market intelligence. These ongoing enhancements underscore our commitment to equipping the investment management community with the tools and data needed to navigate complex and evolving credit markets. 

Contact us to discuss or learn more.

Register here for next month’s call: Thursday, September 18th, 2025, 1 p.m. ET. 

Each month, we host a Models & Markets call to offer our insights into recent model performance, emerging credit risks, and broader economic indicators. This month’s call was a wide-ranging update on new model developments, consumer credit insights, and macroeconomic trends shaping structured finance. 

Here’s a quick recap in case you missed it. 

(Click here to listen to the entire 30-minute recording or continue reading for a summary.)  

Market Outlook: August 2025

Stable employment and inflation notwithstanding, the macro backdrop remains dominated by persistent headwinds: 

• Mortgage Rates: Still above 6.5% and expected to stay above 6% for the next several years. 

• Home Prices: Case-Shiller data shows relative stability, with modest month-over-month declines and low year-over-year growth. 

• Labor & Inflation: Both unemployment and PCE inflation are holding steady. 

• Fed Policy: The Fed Funds Rate remains in the 4.25%–4.50% range, with the first cut expected in September 2025. Markets anticipate a year-end rate of 3.75%–4.00%, but long-term rates remain elevated. 

• 10Yr rates unlikely to see a significant decline over next few years, leading to a high mortgage rate environment (>~ 6%) for next 3-5 years. 

New Equifax Data Integration 

We introduced our latest research leveraging the Equifax Analytic Dataset (ADS), a borrower-level anonymized sample representing 10% of the U.S. active credit population. Using tradeline-level detail (credit scores, balances, payments, etc.), we have constructed aging curves for auto loans and personal loans segmented by credit score bands. 

Some key takeaways: 

• Auto Loan Defaults: Clear segmentation appears across credit score bands, with default curves validated against Federal Reserve data. 

• Personal Loan Defaults: Similar segmentation trends, with early results indicating significant variation across risk tiers. 

• Credit card and student loan performance curves: Coming soon. 

The final versions of these datasets will be accessible directly within the RiskSpan platform, allowing clients to benchmark their portfolios against robust national trends. 

Model Updates 

Prepayment Models (Versions 3.2 & 3.7) 

Our prepayment models continue to perform strongly against observed market behavior. The latest back-testing of agency cohorts (Fannie Mae and Freddie Mac 2021/2022 vintages across 1.5%–6.5% coupons) shows that speeds remain broadly consistent with expectations. However, higher coupon pools have recently exhibited slower-than-expected speeds, reflecting both tighter refinancing conditions and borrower credit constraints. 

1.5 to 3.5 Coupons 

6.5 Coupons 

Credit Model 7.0 

Our much-anticipated Credit Model v7 is now available in production on the RiskSpan Platform. Key features include: 

• Delinquency Transition Matrix – A granular 3-D framework tracking monthly movement of loans through delinquency buckets (30D, 60D, 90D, 120D, 150D, 180D+, Foreclosure, REO). 

• Severity & Liquidation Enhancements – Expanded severity vectors and a liquidation timeline module allow for more nuanced control of loss projections. 

• Integration with MSR Engine – Provides detailed P&I and T&I cash flow accounting that captures probabilistic delinquency transitions. 

These enhancements equip investors and risk managers with deeper tools for analyzing loss dynamics across mortgage, GSE, FHA, and VA loan cohorts. 

Contact us to learn more.

Register here for next month’s call: Thursday, August 21st, 2025, 1 p.m.

Each month, we host a Models & Markets call to offer our insights into recent model performance, emerging credit risks, and broader economic indicators. This month, as interest rates remain elevated and economic uncertainty persists, we addressed how both conventional and AI-based modeling techniques are shaping decision-making processes across agency, non-QM, and ARM products.

Here’s a quick recap in case you missed it.

(Click here to listen to the entire 30-minute recording, or continue reading for a summary.)

Model Performance: Prepayment Dynamics in Focus

RiskSpan’s prepayment model continues to perform well based on benchmarking against actuals across coupon stacks. The team noted:

• Speeds in higher coupons have slowed relative to expectations, in line with broader refinancing trends as mortgage rates remain high.
• RiskSpan’s Non-QM Prepayment Model (v3.11) shows strong back-testing performance. While most vintages perform as expected, the 2022 vintage diverged, potentially due to ambiguous underwriting guidelines in QM loans that may have led to adverse selection in the Non-QM space. One possible reason is that this reflects borrower composition differences not captured by traditional metrics.

New ARM Model Launch

An enhanced ARM Prepayment Model (v3.8) is now live in production. It exhibits refined sensitivity to rate shocks and aims to provide improved accuracy for adjustable-rate portfolios in today’s volatile environment.

Claude the Research Assistant: AI in Action

One of the highlights of the call was a deep dive into how we are testing Claude (Anthropic’s well-known LLM) as a mortgage research assistant.

Using a dataset from RiskSpan’s Snowflake instance, Claude orchestrated an end-to-end analytical workflow, including:

• Retrieving and aggregating partially pre-aggregated loan-level data
• Generating Python code for analysis and visualization
• Annotating charts and analyzing prepayment trends

Key Insights from Claude’s Analysis

Claude surfaced several noteworthy trends:

• FICO Score Sensitivity: Higher credit score bands (>750) showed dramatically higher prepayment rates than lower bands (<650), highlighting the refinancing advantage for more creditworthy borrowers.
• Loan Size Effect: A positive correlation (0.22) between loan size and prepayment rates suggests that larger loan holders are more motivated to refinance.
• Mortgage Vintage: Newer vintages (especially 2015–2016) demonstrated greater prepayment sensitivity, likely due to looser underwriting and seasoning effects.
• Interest Rate Sensitivity: Claude captured the sharp inverse relationship between rates and prepayment, particularly the COVID-era spike and the post-2022 slowdown.

Claude correctly reasoned with the provided data but could not identify some features (like “Spread at Origination”). This raises interesting questions about LLMs’ capacity to reason beyond their training corpus.

Market Outlook: Economic Signals Turning Cautionary

The macro backdrop continues to weigh on securitization and borrower behavior. Highlights from July’s indicators:

• Mortgage Rates: Remain above 6.5%, with little sign of easing before the Fed’s expected first rate cut in September.
• Fed Funds Rate: Currently 4.25–4.50%, with year-end projections settling around 3.75–4.00%.
• Home Prices: Showing stability with little YoY movement in the Case-Shiller Index.
• Labor and Inflation: Both unemployment and PCE inflation measures remain steady, but signs of economic headwinds are beginning to appear.

On the Horizon

• RiskSpan’s new credit model (v7), which includes a new delinquency transition matrix, is on track for release by the end of the month.
• Continued enhancements are being made to the Platform, including new prepayment and performance visualizations for private credit and agency MBS sectors.

Contact us to learn more.

Register here for next month’s call: Thursday, July 17th, 2025, 1 p.m.

Each month, we host a Models & Markets call to offer our insights into recent model performance, emerging credit risks, and broader economic indicators. This month, we showcased our responsiveness to shifting macroeconomic dynamics and introduced new transparency elements (i.e., back-testing tools) to our prepayment and credit modeling.

Click here to listen to the entire 23-minute recording, or continue reading for a summary.

Agency Prepay Model: Back-testing and Enhanced Control

We are launching a new loan-level prepayment back-testing tool using nearly all agency loans (FN/FH/GN) aged 10 years or less. The tool runs every month through our models with historical home prices and interest rates. Based on this data, we have an interactive dashboard that will allow users to drill down into model performance with far more granularity than currently possible.

Key Enhancements to Prepay Model v. 3.8

A soon to be released version of the prepay model will include:

• User-defined slope multipliers for both Out-of-the-Money (OTM) and In-the-Money (ITM) performance, offering finer control over refinance sensitivity and turnover behavior.
• Independent knob control across CONV 30, CONV 15, FHA, and VA loan types.

A redesigned ARM prepayment framework, derived from the fixed-rate model. The new ARM component includes:

• A realistic payment shock element that aligns prepayment spikes with rate reset events.
• Improved seasonality and aging ramp that reflects empirical loan behavior

These updates give users the ability to more precisely tune model responses under a variety of macroeconomic and borrower scenarios.

Credit Model: V7 and Delinquency Transitions

The delinquency transition matrix incorporated into our new Credit Model V7 provides users a more nuanced credit risk assessment. This model works in conjunction with the enhanced prepayment model to better simulate the joint dynamics of default and prepay behavior across economic cycles.

Macroeconomic Context: Rates and Risk in a Holding Pattern

We remain cautious in our outlook for the remainder of 2025 and into 2026:

The Fed Funds Rate is expected to remain elevated—currently in the 4.25–4.50% range—with the first rate cut likely in September. By year-end 2025, the market expects it to settle around 3.75–4.00%.

Mortgage rates remain stubbornly high, hovering above 6.5%, putting pressure on origination volumes and reinforcing the value of accurate prepayment modeling.

Home prices and broader macro indicators like unemployment and PCE inflation remain stable, suggesting a “wait-and-see” mode for both consumers and investors.

What’s Next: More Models, More Tools, More Insights

We continue to expand our Platform with new analytics, model documentation, and client-facing tools. Users can soon access the new back-testing report directly within the Platform, alongside these updated prepayment and credit models. These developments reflect our commitment to model transparency, data-driven innovation, and practical tools for real-time market adaptation.

Contact us to learn more.