Get Started
Category: Article

Overcoming Data Limitations (and Inertia) to Factor Climate into Credit Risk Modeling

With each passing year, it is becoming increasingly clear to mortgage credit investors that climate change is emerging as a non-trivial risk factor that must be accounted for. Questions around how precisely to account for this risk, however, and who should ultimately bear it, remain unanswered. 

Current market dynamics further complicate these questionsLate last year, Politico published this special report laying out the issues surrounding climate risk as it relates to mortgage finance. Even though almost everyone agrees that underinsured natural disaster risk is a problem, the Politico report outlines several forces that make it difficult for anyone to do anything about it. The massive undertaking of bringing old flood zone maps up to date is just one exampleAs Politico puts it: 

The result, many current and former federal housing officials acknowledge, is a peculiar kind of stasis — a crisis that everyone sees coming but no one feels empowered to prevent, even as banks and investors grow far savvier about assessing climate risk. 

At some point, however, we will reach a tipping point – perhaps a particularly devastating event (or series of events) triggering significant losses. As homeowners, the GSEs, and other mortgage credit investors point fingers at one another (and inevitably at the federal government) a major policy update will become necessary to identify who ultimately bears the brunt of mispriced climate risk in the marketOnce quantified and properly assigned, the GSEs will price in climate risk in the same way they bake in other contributors to credit risk — via higher guarantee fees. For non-GSE (and CRT) loans, losses will continue to be borne by whoever holds the credit risk 

Recognizing that such an event may not be far off, the GSEs, their regulator, and everyone else with credit exposure are beginning to appreciate the importance of understanding the impact of climate events on mortgage performance. This is not easily inferred from the historical data record, however. And those assessing risk need to make informed assumptions about how historically observed impacts will change in the future. 

The first step in constructing these assumptions is to compile a robust historical dataset. To this end, RIskSpan began exploring the impact of certain hurricanes a few years ago. This initial analysis revealed a significant impact on short-term mortgage delinquency rates (not surprisingly), but less of an impact on default rates. In other words, affected borrowers encountered hardship but ultimately recovered. 

This research is preliminary, however, and more data will be necessary to build scenario assumptions as climate events become more severe and widespread. As more data covering more events—including wildfires—becomes available, RiskSpan is engaged in ongoing research to tease out the impact each of these events has on mortgage performance.  

It goes without saying that climate scenario assumptions need to be grounded in reality to be useful to credit investors. Because time-series data relationships are not always detectable using conventional means, especially when data is sparse, ware beginning to see promise in leveraging various machine learning techniques to this endWe believe this historical, machine-learning-based research will provide the backbone for an approach that merges historical effects of events with inputs about the increasing frequency and severity of these events as they become better understood and more quantifiable. 

Precise forecasting of severe climate events by zip code in any given year is not here yet. But an increasingly reliable framework for gauging the likely impact of these events on mortgage performance is on the horizon.  

RiskSpan’s Edge Platform Wins 2021 Buy-Side Market Risk Management Product of the Year

RiskSpan, a leading SaaS provider of risk management, data and analytics has been awarded Buy-Side Market Risk Management Product of the Year for its Edge Platform at’s 2021 Risk Markets Technology Awards. The honor marks Edge’s second major industry award in 2021, having also been named the winner of Chartis Research’s Risk-as-a-Service category.

Licensed by some of the largest asset managers and Insurance companies in the U.S., a significant component of the Edge Platform’s value is derived from its ability to serve as a one-stop shop for research, pre-trade analytics, pricing and risk quantification, and reporting. Edge’s cloud-native infrastructure allows RiskSpan clients to scale as needs change and is supported by RiskSpan’s unparalleled team of domain experts — seasoned practitioners who know the needs and pain points of the industry firsthand

Adjudicators cited the platform’s “strong data management and overall technology” and “best-practice quant design for MBS, structured products and loans” as key factors in the designation.


Edge’s flexible configurability enables users to create custom views of their portfolio or potential trades at any level of granularity and down to the loan level. The platform enables researchers and analysts to integrate conventional and alternative data from an impressive array of sources to identify impacts that might otherwise go overlooked.

For clients requiring a fully supported risk-analytics-as-a-service offering, the Edge Platform provides a comprehensive data analysis, predictive modeling, portfolio benchmarking and reporting solution tailored to individual client needs.

An optional studio-level tier incorporates machine learning and data scientist support in order to leverage unstructured and alternative datasets in the analysis.

Contact us to learn how Edge’s capabilities can transform your mortgage and structured product analytics. 

Learn more about Edge at 

The NRI: An Emerging Tool for Quantifying Climate Risk in Mortgage Credit

Climate change is affecting investment across virtually every sector in a growing number of mostly secondary ways. Its impact on mortgage credit investors, however, is beginning to be felt more directly.

Mortgage credit investors are investors in housing. Because housing is subject to climate risk and borrowers whose houses are destroyed by natural disasters are unlikely to continue paying their mortgages, credit investors have a vested interest in quantifying the risk of these disasters.

To this end, RiskSpan is engaged in leveraging the National Risk Index (NRI) to assess the natural disaster and climate risk exposure of mortgage portfolios.

This post introduces the NRI data in the context of mortgage portfolio analysis (loans or mortgage-backed securities), including what the data contain and key considerations when putting together an analysis. A future post will outline an approach for integrating this data into a framework for scenario analysis that combines this data with traditional mortgage credit models.

The National Risk Index

The National Risk Index (NRI) was released in October 2020 through a collaboration led by FEMA. It provides a wealth of new geographically specific data on natural hazard risks across the country. The index and its underlying data were designed to help local governments and emergency planners to better understand these risks and to plan and prepare for the future.

The NRI provides information on both the frequency and severity of natural risk events. The level of detailed underlying data it provides is astounding. The NRI focuses on 18 natural risks (discussed below) and provides detailed underlying components for each. The severity of an event is broken out by damage to buildings, agriculture, and loss of life. This breakdown lets us focus on the severity of events relative to buildings. While the definition of building here includes all types of real estate—houses, commercial, rental, etc.—having the breakdown provides an extra level of granularity to help inform our analysis of mortgages.

The key fields that provide important information for a mortgage portfolio analysis are bulleted below. The NRI provides these data points for each of the 18 natural hazards and each geography they include in their analysis.

  • Annualized Event Frequency
  • Exposure to Buildings: Total dollar amount of exposed buildings
  • Historical Loss Ratio for Buildings (Bayesian methods to derive this estimate, such that every geography is covered for its relevant risks)
  • Expected Annual Loss for Buildings
  • Population estimates (helpful for geography weighting)

Grouping Natural Disaster Risks for Mortgage Analysis

The NRI data covers 18 natural hazards, which pose varying degrees of risk to housing. We have found the framework below to be helpful when considering which risks to include in an analysis. We group the 18 risks along two axes:

1) The extent to which an event is impacted by climate change, and

2) An event’s potential to completely destroy a home.

Earthquakes, for example, have significant destructive potential, but climate change is not a major contributor to earthquakes. Conversely, heat waves and droughts wrought by climate change generally do not pose significant risk to housing structures.

When assessing climate risk, RiskSpan typically focuses on the five natural hazard risks in the top right quadrant below.

Immediate Event Risk versus Cumulative Event Risk

Two related but distinct risks inform climate risk analysis.

  1. Immediate Event Analysis: The risk of mortgage delinquency and default resulting directly from a natural disaster eventhome severely damaged or destroyed by a hurricane, for example.  
  2. Cumulative Event Risk: Less direct than immediate event risk, this is the risk of widespread home price declines across an entire area communities because of increasing natural hazard risk brought on by climate changeThese secondary effects include: 
    • Heightened homebuyer awareness or perception of increasing natural hazard risk,
    • Property insurance premium increases or areas becoming ‘self-insured, 
    • Government policy impacts (e.g., potential flood zone remapping), and 
    • Potential policy changes related to insurance from key players in the mortgage market (i.e., Fannie Mae, Freddie Mac, FHFA, etc.). 

NRI data provides an indication of the probability of immediate event occurrence and its historic severity in terms of property losses. We can also empirically observe historical mortgage performance in the wake of previous natural disaster events. Data covering several hurricane and wildfire events are available.

Cumulative event risk is less observable. A few academic papers attempt to tease out these impacts, but the risk of broader home price declines typically needs to be incorporated into a risk assessment framework through transparent scenario overlays. Examples of such scenarios include home price declines of as much as 20% in newly flood-exposed areas of South Florida. There is also research suggesting that there are often long term impacts to consumer credit following a natural disaster 

Geography Normalization

Linking to the NRI is simple when detailed loan pool geographic data are available. Analysts can merge by census tract or county code. Census tract is the more geographically granular measure and provides a more detailed analysis.

For many capital markets participants, however, that level of geographic specific detail is not available. At best, an investor may have a 5-digit or 3-digit zip code. Zip codes do not directly match to a given county or census tract and can potentially span across those distinctions.

There is no perfect way to perform the data link when zip code is the only available geographic marker. We take an approach that leverages the other data on housing stock by census tract to weight mortgage portfolio data when multiple census tracts map to a zip code.

Other Data Limitations

The loss information available represents a simple historical average loss rate given an event. But hurricanes (and hurricane seasons) are not all created equal. The same is true of other natural disasters. Relying on averages may work over long time horizons but could significantly underpredict or overpredict loss in a particular year. Further, the frequency of events is rising so that what used to be considered 100 year event may be closer to a 10 or 20 year event. Lacking data about what losses might look like under extreme scenarios makes modeling such events problematic.

The data also make it difficult to take correlation into account. Hurricanes and coastal flooding are independent events in the dataset but are obviously highly correlated with one another. The impact of a large storm on one geographic area is likely to be correlated with that of nearby areas (such as when a hurricane makes its way up the Eastern Seaboard).

The workarounds for these limitations have limitations of their own. But one solution involves designing transparent assumptions and scenarios related to the probability, severity, and correlation of stress events. We can model outlier events by assuming that losses for a particular peril follow a normal distribution with set standard deviations. Other assumptions can be made about correlations between perils and geographies. Using these assumptions, stress scenarios can be derived by picking a particular percentile along the loss distribution.

A Promising New Credit Analysis Tool for Mortgages

Notwithstanding its limitations, the new NRI data is a rich source of information that can be leveraged to help augment credit risk analysis of mortgage and mortgage-backed security portfolios. The data holds great promise as a starting point (and perhaps more) for risk teams starting to put together climate risk and other ESG analysis frameworks.

EDGE: An Update on GNMA Buyout Efficiency

In July, we examined buyouts of delinquent GNMA loans, with special focus on the buyout efficiency for bank servicers. At that time, several banks were 98% to 99% efficient at buying out delinquent loans, where efficiency is defined as the percentage of 90+ days delinquent loans that are repurchased. In this short note, we update the buyout efficiency of major bank and non-bank servicers. 

Buyout efficiency varies widely among banks. While the most efficient banks repurchase nearly 100% of eligible loansothers, including Flagstar and Citizens Bank, opt to leave virtually all the 90+ day delinquent loans they service in securities. In the table below, we show the dollar-weighted buyout efficiencies for top banks, as well as the UPB of each bank’s unpurchased 90+ day delinquent loans, as of the January 2021 factor date.

Buyout efficiency for 90+ day delinquent loans, data as of January 2021. 

Servicers listed by total UPB serviced.

The overhang of seriously delinquent loans serviced by Flagstar and Citizens is spread across several GN2 Multi-lender sectors, with concentrations of delinquent loans rising to just 1% of the total current face of 2018 4% and 2018 4.5% cohorts. If Flagstar and Citizens were to repurchase all of their delinquent loans in a single month, it would add roughly 11-12 CPR to these cohorts. This represents the upper limit in involuntary speed, and actual speeds would likely be much slower with repurchases spread over several months.

The markedly lower buyout efficiency among GNMA non-bank servicers has created involuntary prepay overhang that is potentially much more daunting. The following table summarizes top non-bank servicers, their buyout efficiency over the past two quarters, and their current overhang of 90+ day delinquent loans.

Buyout efficiency for 90+ day delinquent loans, data as of January 2021.

Servicers listed by total UPB serviced.

Both Penny Mac and Lakeview have improved their buyout efficiency over the last quarter and may continue to do so, as more investors begin to embrace the GNMA EBO trade. The multi-lender cohorts with the most exposure to 90+ day DQ loans serviced by Penny Mac or Lakeview include 2020 3.5s as well as 2017-19 production 3.5s and 4s, with each cohort ranging between 4% to 5% of its current face.

This final table, below, illustrates the impact of forbearance on buyout activity among non-banks. While forbearance status seems to pose no impediment to buyouts for banks — in fact, banks with the highest buyout efficiency seem to favor repurchasing loans that are in COVID-forbearance over loans that are “naturally” delinquent – non-bank behavior is more nuanced.

Of the top five non-bank servicers, only Lakeview has generated significant repurchases of loans in COVID forbearance, repurchasing 10% of eligible loans in Q4. In the table below, we separate the 90+ day delinquent loans by their forbearance status and then compute each servicer’s buyout efficiency across these sub-cohorts.

Buyout efficiency for 90+ day delinquent loans, data as of January 2021.

Lakeview’s buyout behavior suggests that forbearance is not an impediment to non-bank repurchases. If we see continued improvements in buyout efficiency over the next few months, involuntary speeds in GNMA securities have the potential to rise significantly.

Contact us if you are interested in seeing variations on this theme. Using Edge, we can examine any loan characteristic and generate a S-curve, aging curve, or time series.

Nearly $8 Trillion in Senior Home Equity Pushes Reverse Mortgage Market Index Upward

The NRMLA/RiskSpan Reverse Mortgage Market Index (RMMI) rose to 280.99 during the third quarter of 2020, an all-time high. This reflects a 1.6% increase in senior home equity, which now stands at an estimated $7.82 trillion. Growth in senior homeowner’s wealth was largely attributable to an estimated 1.6% (or $149 billion) increase in senior housing value, offset by 1.6% (or $28 billion) increase of senior-held mortgage debt.

The National Reverse Mortgage Lenders Association (NRMLA) and RiskSpan have published the Reverse Mortgage Market Index (RMMI) since the beginning of 2000. The RMMI provides a trending measure of home equity among U.S. homeowners age 62 and older.

The RMMI defines senior home equity as the difference between the aggregate value of homes owned and occupied by seniors and the aggregate mortgage balance secured by those homes. This measure enables NRMLA to help gauge the potential market size of those who may be qualified for a reverse mortgage product. The chart above illustrates the steady increase in this index since the end of the 2008 recession.

Increasing house prices drive the index’s upward trend, mitigated to some extent by a corresponding modest increase in mortgage debt held by seniors. The most recent RMMI report (reflecting data as of the end of Q3 20202) was published last week on NRMLA’s website.

Note on the Limitations of RMMI

To calculate the RMMI, an econometric tool is developed to estimate senior housing value, senior mortgage level, and senior equity using data gathered from various public resources such as American Community Survey (ACS), Federal Reserve Flow of Funds (Z.1), and FHFA housing price indexes (HPI). The RMMI is simply the senior equity level at time of measure relative to that of the base quarter in 2000.[1]  The main limitation of RMMI is non-consecutive data, such as census population. We use a smoothing approach to estimate data in between the observable periods and continue to look for ways to improve our methodology and find more robust data to improve the precision of the results. Until then, the RMMI and its relative metrics (values, mortgages, home equities) are best analyzed at a trending macro level, rather than at more granular levels, such as MSA.

[1] There was a change in RMMI methodology in Q3 2015 mainly to calibrate senior homeowner population and senior housing values observed in 2013 American Community Survey (ACS).

Cash-out Refis, Investment Properties Contribute to Uptick in Agency Mortgage Risk Profile

RiskSpan’s Vintage Quality Index is a monthly measure of the relative risk profile of Agency mortgages. Higher VQI levels are associated with mortgage vintages containing higher-than-average percentages of loans with one or more “risk layers.”

These risk layers, summarized below, reflect the percentage of loans with low FICO scores (below 660), high loan-to-value ratios (above 80%), high debt-to-income ratios (above 45%), adjustable rate features, subordinate financing, cash-out refis, investment properties, multi-unit properties, and loans with only one borrower.

The RiskSpan VQI rose 4.2 points at the end of 2020, reflecting a modest increase in the risk profile of loans originated during the fourth quarter relative to the early stages of the pandemic.

The first rise in the index since February was driven by modest increases across several risk layers. These included cash-out refinances (up 2.5% to a 20.2% share in December), single borrower loans (up 1.8% to 52.0%) and investor loans (up 1.4% to 6.0%). Still, the December VQI sits more than 13 points below its local high in February 2020, and more than 28 points below a peak seen in January 2019.

While the share of cash-out refinances has risen some from these highs, the risk layers that have driven most of the downward trend in the overall VQI – percentage of loans with low FICO scores and high LTV and DTI ratios – remain relatively low. These layers have been trending downward for a number of years now, reflecting a tighter credit box, and the pandemic has only exacerbated tightening.

Population assumptions:

  • Monthly data for Fannie Mae and Freddie
  • Loans originated more than three months prior to issuance are excluded because the index is meant to reflect current market
  • Loans likely to have been originated through the HARP program, as identified by LTV, MI coverage percentage, and loan purpose, are also These loans do not represent credit availability in the market as they likely would not have been originated today but for the existence of HARP.

Data assumptions:

  • Freddie Mac data goes back to 12/2005. Fannie Mae only back to 12/2014.
  • Certain fields for Freddie Mac data were missing prior to 6/2008.
  • GSE historical loan performance data release in support of GSE Risk Transfer activities was used to help back-fill data where it was missing.

This analysis is developed using RiskSpan’s Edge Platform. To learn more or see a free, no-obligation demo of Edge’s unique data and modeling capabilities, please contact us.

EDGE: GNMA Forbearance End Date Distribution

With 2021 underway and the first wave of pandemic-related FHA forbearances set to begin hitting their 12-month caps as early as March, now seems like a good time to summarize where things stand. Forbearance in mortgages backing GNMA securities continues to significantly outpace forbearance in GSE-backed loans, with 7.6% of GNMA loans in forbearance compared to 3.5% for Fannie and Freddie borrowers.[1] Both statistics have slowly declined over the past few months.

Notably, the share of forbearance varies greatly amongst GNMA cohorts, with some cohorts having more than 15% of their loans in forbearance. In the table below, we show the percentage of loans in forbearance for significant cohorts of GN2 30yr Multi-lender pools.

Percent of Loans in Forbearance for GNMA2 30yr Multi-lender Pools:

Cohorts larger than $25 billion. Forbearance as of December 2020 factor date.

Not surprisingly, newer production tends to experience much lower levels of forbearance. Those cohorts are dominated by newly refinanced loans and are comprised mostly of borrowers that have not struggled to make mortgage payments. Conversely, 2017-2019 vintage 3s through 4.5s show much higher forbearance, most likely due to survivor bias – loans in forbearance tend not to refinance and are left behind in the pool. The survivor bias also becomes apparent when you move up the coupon stack within a vintage. Higher coupons tend to see more refinancing activity, and that activity leaves behind a higher proportion of borrowers who cannot refinance due to the very same economic hardships that are requiring their loans to be in forbearance.

GNMA also reports the forbearance end date and length of the forbearance period for each loan. The table below summarizes the distribution of forbearance end dates across all GNMA production. This date is the last month of the currently requested forbearance period.[2]

For loans with forbearance ending in December 2020 (last month), half have taken a total of 9 months of forbearance, with most of the remaining loans taking either three or six months of forbearance.

For loans whose forbearance period rolls in January and February 2021, the total months of forbearance is evenly distributed between 3, 6, and 9 months. Among loans with a forbearance end date of March 2021, more than half will have taken their maximum twelve months of forbearance.[3]

In the chart below, we illustrate how things would look if every Ginnie Mae loan currently in forbearance extended to its full twelve-month maximum. As this analysis shows, a plurality of these mortgages – more than 25 percent — would have a forbearance end date of March 2021, with the remaining forbearance periods expiring later in 2021.

A successful vaccination program is expected to stabilize the economy and (hopefully) end the need for wide-scale forbearance programs. The timing of this economic normalization is unclear, however, and the distribution of current end dates, as illustrated above, suggests that the existing forbearance period may need to be extended for some borrowers in order to forestall a potentially catastrophic credit-driven prepayment spike in GNMA securities.

Contact us if you interested in seeing variations on this theme. Using Edge, we can examine any loan characteristic and generate a S-curve, aging curve, or time series.

[1] As of the December 2020 factor date, using the data reported by the GSEs and GNMA. This data may differ marginally from the Mortgage Bankers Association survey, which is a weekly survey of mortgage servicers.

[2] Data as of the December 2020 factor date.

[3] Charts of January, February and March 2021 rolls are omitted for brevity. See RiskSpan for a copy of these charts.

EDGE: COVID Forbearance and Non-Bank Buyouts

November saw a significant jump in GNMA buyouts for loans serviced by Lakeview. Initially, we suspected that Lakeview was catching up from nearly zero buyout activity in the prior months, and that perhaps the servicer was doing this to keep in front of GNMA’s requirement to keep seriously delinquent loans below the5% of UPB threshold. [1]


Buyout rates for some major non-bank servicers.

Using EDGE to dig further, we noticed that Lakeview’s buyouts affected both multi-lender and custom pools in similar proportions and were evenly split between loans with an active COVID forbearance and loans that were “naturally” delinquent.

The month-on-month jump in Lakeview buyouts on forborne loans is notable. The graph below plots Lakeview’s buyout rate (CBR) for loans that are 90-days+ delinquent.

Further, the buyouts were skewed towards premium coupons. Given this, it is plausible that the buyouts are economically driven [2] and that Lakeview is now starting to repurchase and warehouse delinquent loans, something that non-banks have struggled with due to balance sheet and funding constraints.

Where do the current exposures lie? The table below summarizes Lakeview’s 60-day+ delinquencies for loans in GN2 multi-lender pools, for coupons and vintages where Lakeview services a significant portion of the cohort. Not surprisingly, the greatest exposure lies in recent-vintage 4s through 5s.

To lend some perspective, in June 2020 Wells serviced around one-third of 2012-13 vintage 3.5s and approximately 8% of its loans were 60-days delinquent, all non-COVID related.

This analysis does not include other non-bank servicers. As a group, non-bank servicers now service more than 80% of recent-vintage GN2 loans in multi-lender pools. The Lakeview example reflects mounting evidence that COVID forbearance is not an impediment to repurchasing delinquent loans.

If you interested in seeing variations on this theme, contact us. Using Edge, we can examine any loan characteristic and generate a S-curve, aging curve, or time series.

[1] Large servicers are required to keep 90-day+ delinquencies below 5% of their overall UPB. GNMA has exempted loans that are in COVID forbearance from this tally.

[2] Servicers can repurchase GNMA loans that have missed 3 or more payments at par. If these loans cure, either naturally or due to modification, the servicer can deliver them into a new security. Given that nearly all GNMA passthroughs trade at a significant premium to par, this redelivery can create a substantial arbitrage opportunity, even after accounting for the trial period for the modification.

Chart of the Month: Fed Impact on Credit ETF Performance

On March 23rd, The Fed announced that its Secondary Market Corporate Credit Facility (SMCCF) would begin purchasing investment-grade corporate bonds in the secondary market, first through ETFs and directly in a later phase. 

In June, we charted the impact of this announcement on the credit spreads of various corporate bonds. This month we are charting its impact on ETF performance.

This month’s chart plots the price of ETFs relative to their price as of March 23rd 2020 (i.e., all ETF prices are set to 1.00 as of that date). Data runs from Feb 24th to Nov 16th 2020.

EDGE: Unexplained Prepayments on HFAs — An Update

In early October, we highlighted a large buyout event for FNMA pools serviced by Idaho HFA, the largest servicer of HFA loans. On October 28, FNMA officially announced that there were 544 base-pools with erroneous prepayments due to servicer reporting error. The announcement doesn’t mention the servicer of the affected pools, but when we look at pools that are single-servicer, every one of those pools is serviced by Idaho HFA.

FNMA reports the “September 2020 Impacted Principal Paydown” at $133MM. The September reported prepayment for FNMA Idaho HFA pools was 43 CPR on a total of just over $6B UPB. If we add back the principal from the impacted paydown, the speed should have been 26 CPR, which is closer to the Freddie-reported 25 CPR.

FNMA provides an announcement here and list of pools here. According to the announcement, FNMA will not be reversing the buyout but instead recommends that affected investors start a claims process. We note that Idaho HFA prepayment speeds will continue to show these erroneous buyouts in the October factor date.

Contact us to try Edge for free.

Get Started