Loss Reserve Forecasting and Risk Management at the National Credit Union Administration

Challenge: The National Credit Union Administration (NCUA) faces potential losses when an insured credit union (CU) fails, and the National Credit Union Share Insurance Fund (NCUSIF) is a contingent liability account administered by the NCUA for insurance losses as a result of CU failures. The NCUSIF insures deposits of nearly 92 million account holders and assets of more than $1 trillion, so proper maintenance of NCUSIF reserve levels is critical to not only a large population of private consumers, but also to large scale economic stability. An essential component of administering the NCUSIF is forecasting the cash requirements of the fund to ensure solvency. NCUA required the design and construction of an econometric model that estimates the default risk of the portfolio of its insured CUs to more accurately forecast the NCUSIF requirements in a volatile economy.

Solution: Summit was retained to construct an econometrically enhanced model to forecast NCUSIF requirements through prediction of the individual default risk of each of NCUA’s insured CUs. Summit conducted a broad literature review and collected a diversity of financial, macroeconomic, microeconomic, and industry data on millions of CU observations in preparation for model development. Summit then constructed a suite of cutting-edge Double Hurdle and Logistic models to forecast NCUSIF reserve needs. The models incorporate proprietary NCUA examiner data, specifically Capital Adequacy, Asset Quality, Management Quality, Earnings Ability, and Liquidity Position (CAMEL) ratings of a CU, and CU-level financial (NCUA 5300 Call Report) data, as well as microeconomic and macroeconomic indicators. Summit also developed a user-friendly model interface, systems documentation, and standard operating procedures tailored to NCUA’s needs. In addition to reserve requirement forecasts, the set of econometric models that Summit produced dually generates highly informative CU examiner tools, including institution-level weighted risk measures and failure probabilities, which had effective prediction success rates of more than 92 percent. Throughout the model development and deployment process, Summit directly collaborated with and advised NCUA senior management, including the Chief Economist, Office of the Chief Economist, and the Division of Risk Management. Summit currently provides ongoing audit, forecasting, and model maintenance support to the NCUA.