IFE Group conducted the independent analysis of the economic capital estimations of FHA’s mortgage insurance programs for more than a decade. FHA insures home mortgages for lower income households and first time home buyers. IFE Group processed the historical month-by-month performance of big data consisting of over 25 million loans; statistically estimated prepayment, default, and loss severity rates; and projected stochastic future economic scenarios and the associated prepayment and default events. The model was implemented under a Monte Carlo simulation framework to compute the economic value of the programs under normal and stressed economic scenarios. The model has been frequently verified and validated by various government agencies and the general public.

IFE Group is engaged by HUD to update, re-estimate and evaluate its Technology-Open-to-Approved-Lenders (TOTAL) scorecard, a statistically derived decision tool that works in conjunction with an AUS to arrive at logic-based automatic accept decisions for FHA loan applications.  The engagement entails baselining the TOTAL scorecard; incorporating new credit scores; developing challenger models and examining the predictive power of detailed credit attributes; establishing an evaluation framework and criteria for champion-challenger comparison; conducting fair lending impact analyses; and incorporating new modeling and data mining techniques.

IFE Group is engaged by HUD to conduct fair lending compliance review of Fannie Mae’s and Freddie Mac’s Automated Underwriting Systems to determine if the data, model, or implementation of the AUS indicate “disparate treatment or impact” on protected classes, with special attention to race, ethnicity, and gender.  IFE Group designed a methodology to track the performance of the GSE’s underwriting and identified potential improvements by building alternative models, and conducting statistical testing.

IFE Group validated the mortgage underwriting systems for international foreign government’s national housing agency. The engagement included: evaluating different approaches, such as Hedonic, Repeated Sales, Tax Proxy, and aggregation estimates from sub-models; conducting comprehensive validation, including code review, soundness of conceptual design, estimation analyses, reasonableness of model selection, and robustness of model performance; identifying issues and proposing remediation solutions; conducting stress-testing to evaluate model performance of the AVMs under hypothesized housing downturns; reviewing underwriting models and validating soundness of conceptual design and reasonableness of input variables; and presenting analyses to regulators.

IFE Group has been engaged by a foreign government agency to assess the effectiveness of their housing market analysis framework, which was designed to detect problematic housing conditions for key markets.  As a benchmark of its effectiveness and to explore ways to enhance it, we applied the framework to the U.S. national market and select CBSAs.  The engagement consisted of baselining the framework, including creating a database with U.S. housing and demographic metrics; establishing a selection criteria to select key U.S. markets and applying the framework to them; leveraging our expertise of the U.S. housing market and conducting market research; reviewing model framework for reasonableness; incorporating additional econometric techniques; reviewing framework governance; and presenting our findings and recommendations.