Ginnie Mae: Risk Management and Cost Modeling Require Continuing Attention

In response to this recommendation, as of December 2015, Ginnie Mae took the following actions: (1) Assessed available data from the Federal Housing Administration (FHA), Veterans Administration (VA) and United States Department of Agriculture's Rural Housing Service. Specifically, Ginnie Mae officials stated they reviewed loan performance models used by FHA and VA. In addition, Ginnie Mae selected a new contractor in May 2015 to assist Ginnie Mae with developing its own models. The contractor assessed the best available data for these models. With this assessment, Ginnie Mae determined that using its own loan level data in its models allowed for incorporating 10 years of historical data and alleviated timing issues in receiving data from other agencies. Moreover, Ginnie Mae used data from other agencies in its models to test data quality of loan level data received from its own issuers. (2) Procured private U.S. Stock and Treasury data to supplement data used in its models (3) Developed and documented an additional suite of econometric models, with the assistance of a contractor, to enhance its modeling capability. Within the documentation, Ginnie Mae discusses the benefits and costs related to the methodology used for modeling, which was not available at the time of our report. By taking these actions, Ginnie Mae has helped ensure that its econometric model more closely follows certain practices identified in Financial Accounting Standards Advisory Board guidance.

In response to this recommendation, Ginnie Mae awarded a new contract in May 2015 for the development of a suite of models, which includes models related to guarantee fee income, issuer default, and issuer buyout. As part of this contract, the contractor is required to conduct and document sensitivity analyses. By taking these actions, Ginnie Mae has helped ensure that its econometric model more closely follows certain practices identified in Financial Accounting Standards Advisory Board guidance.

In response to this recommendation, in December 2015, Ginnie Mae presented management assumptions used in its FY 2017 Credit Subsidy Estimate Result Review. Information was presented explaining the changes in assumptions for interest rates and foreclosure and maintenance costs, among others. In addition, methodology documents developed by Ginnie Mae and its contractor detailed how the management assumptions were determined. Moreover, Ginnie Mae and its contractor have been developing two models--issuer default and issuer buyout. Officials stated the purpose of these models is to minimize the use of management assumptions where data are available in order to conform with Federal Accounting Standards Advisory Board (FASAB) guidance. By taking these actions, Ginnie Mae has helped to enhance the transparency of its econometric model and ensure that its model more closely follows certain practices identified in FASAB guidance.

In response to this recommendation, Ginnie Mae's Economic Modeling Division and its contractor have been working to complete two models--issuer default and issuer buyout. Officials stated that the purpose of these models is to minimize the use of management assumptions where data is available in order to conform with Federal Accounting Standards Advisory Board (FASAB) guidance. By taking these actions, Ginnie Mae has helped ensure that its econometric model more closely follows certain practices identified in FASAB guidance.