Now Covering the Full Spectrum of Mortgage Risk
Built on the industry’s largest and most robust loan performance datasets, CoreLogic® RiskModel forecasts mortgage prepayments, defaults, losses, and loan-level cash flows for portfolios and securities’ underlying loan pools.
We recently added RiskModel AGENCY to the RiskModel suite of mortgage behavioral models. This agency-focused credit and prepayment model is as effective in predicting performance for the higher-credit-quality asset class as our non-agency models.
Drawing upon servicer-contributed agency-type loans from the vast CoreLogic Loan-Level Market Analytics (LLMA) mortgage performance dataset, RiskModel AGENCY extends the RiskModel non-agency transition framework to predict voluntary and involuntary prepayment risk and all stages of delinquency for agency loans. It consists of separate sub-models for fixed and adjustable rate mortgages and direct modeling treatments for loan modifications.
The RiskModel AGENCY development dataset covers twelve years of origination and performance observations with sample validations occurring over this same time period. To further validate the model, out-of-sample back-tests were conducted using performance data recently released by each of the GSEs.
RiskModel Can Perform a Number of Critical Tasks
Analysis of structured agency credit risk (STACR) bonds – providing investors with new insight into the propensity and magnitude of future credit events associated with STACR reference pools.
Stress testing and accounting of bank-owned and other entities’ holdings
Mortgage servicing rights (MSR) valuations – given its granular breakout of delinquency states and focus on prime conforming collateral, RiskModel AGENCY is a more appropriate choice for MSR valuation than traditional prepayment models
Banks can stress-test non-agency RMBS holdings and whole loan portfolios, measure risks and expected performance of their holdings to facilitate OTTI and GAP distressed-asset requirements
Broker-dealers and investors can measure risks on current RMBS holdings, identify arbitrage opportunities, and exit positions that will ultimately yield undesirable ROIs.
Originators can realistically access available whole-loan transaction options and evaluate likely future outcomes if they pursue securitization.
Accurate mortgage risk determinations depend on the accuracy of current and future home price assessments. RiskModel is the only commercially available predictive analytics program that integrates CoreLogic HPI—used by both the Federal Reserve and OCC. CoreLogic HPI integration allows more than 80% of the loans in the model’s development dataset to be estimated at the ZIP-code level.
RiskModel also offers the innovative and proprietary Refinance Qualification Index (RFI), an automated assessment of the likelihood a prospective borrower will qualify for refinancing—helping to predict prepayment sensitivity with greater accuracy.
By combining rich data and modeling with proven stochastic analytics, RiskModel forecasts:
- Prepayments, defaults, losses, cash flows
- Changes in mortgage status including current, all stages of delinquency
- Terminal events including voluntary prepayment and two forms of liquidation
- Probability of default (PD)
- Loss given default (LGD)
- Loss severity
Employed by leading mortgage-backed securities broker-dealers, traders, investors, banks and regulators, the CoreLogic RiskModel is a broadly trusted solution for understanding residential mortgage risk now and into the future.
Post-Crisis Predictive Power and Effective Solutions for New Regulations
Leveraging the full scope of CoreLogic data resources, RiskModel can predict the risk associated with whole loan portfolios and formulate accurate assumptions to support the valuations of both agency and non-agency residential mortgage-backed securities.
Drawing from a wealth of historical loan-performance data, RiskModel’s sophisticated age curves enable it to capture the effect of credit burnout on more seasoned loans that continue to perform despite the presence of negative equity.
Both comprehensive and cost-effective, RiskModel can play a critical role in:
- Fed-mandated stress tests
- OTTI estimates
- Setting bank reserve levels
- Projecting and reporting asset impairments now required by GAAP
Dialing for Model Sensitivities
Model sensitivities can be quickly and easily modified to reflect the behavioral effect of loan characteristics, such as:
- Loan modification type
- Product type
- Delinquency status
- Payment history
By fine-tuning a model's sensitivity to such variables based on experience, RiskModel users can improve estimates of future performance as it relates to a portfolio or a specific security.
This adjustable framework also lets users create dials that adjust for supplemental data—enabling the factoring in of borrower credit, public record, servicer, origination channel, and other relevant data.
RiskModel draws on the mortgage industry’s most comprehensive databases—with more than 70 million active and historical prime, alt-A, and subprime loans—including:
- The CoreLogic Loan-Level Market Analytics (LLMA) database—covering 42 million active mortgages, 76% of residential mortgages, worth $7.6 trillion
- The CoreLogic securities database—covering 97% of all non-agency MBS and ABS securities, worth more than $1.13 trillion
- The CoreLogic LLMA-derived agency residential mortgage dataset—covering more than seven million servicer-contributed active agency loans
- CoreLogic HPI—the most comprehensive and timely U.S. housing market price index.
CoreLogic HPI incorporates more than 30 years of repeat sales transactions, representing more than 65 million observations sourced from CoreLogic’s industry-leading property record, securities and servicing databases.
CoreLogic HPI provides the most comprehensive and timely set of monthly home price indices and median sales prices available covering:
- 600+ ZIP codes (58 percent of total U.S. population)
- 640 Core Based Statistical Areas (86 percent of total U.S. population)
- 1200+ counties (84 percent of total U.S. population) located in all 50 states and the District of Columbia
RiskModel uses ZIP-code level indices to approximate the current property value for every loan it analyzes. Future home prices are simulated at the CBSA level based on assumptions for expected growth and volatility that are easily modified by the user.
To learn more about the CoreLogic HPI, click here.
How to Access RiskModel
CoreLogic offers multiple ways to access RiskModel’s power:
- Desktop graphical user interface (GUI)—RiskModel’s Windows-based GUI allows users to leverage its sophisticated analytics through a simple graphical user interface, providing access and flexibility right on the user’s desktop
- RiskModelDIRECT—this non-agency, MBS-vectors data service is available directly via SFTP, generated by RiskModel with future home price simulations guided by CoreLogic HPI Forecasts™. Users can subscribe to the entire MBS/ABS dataset or customize to match their portfolio’s focus. Vectors can be integrated with structured finance modeling tools—such as INTEXdesktop™ and INTEXcalc™ —or through select third-party partners
- Advisory and Professional Services engagements—our advisory and professional services experts first understand users’ needs then tailor their efforts to provide appropriate RiskModel-based analysis, a process that often saves both time and money. Professional Services staff can also liaise with CoreLogic analytics teams to develop client-directed custom calibrations for unique portfolios
- Application programming interface (API)—clients can also incorporate a Windows-based, customizable API that will automate analytics research and reporting by integrating it directly into the company’s existing systems infrastructure.