‘Flawed models’ are cause of false mortgage discrimination findings

“It is particularly disheartening then that lenders are often the subject of ill considered accusations regarding discrimination, accusations based upon analyses that lack statistical rigor”
Michael Fratantoni, MBA’s Vice President of Research and Economics

Dennis Norman

This week the Mortgage Bankers Association (MBA) released a paper, “A Review of Statistical Problems in the Measurement of Mortgage Market and Credit Risk” conducted by Professor Anthony M. Yezer of the George Washington University and sponsored by MBA’s Research Institute for Housing America (RIHA). This paper examines the fundamental assumptions that are often used as an analysis of whether their is discrimination present in the process of mortgage lending. “Mortgage lenders throughout the industry are committed to fair lending, and expend considerable resources to ensure that they adhere to the letter and the spirit of all applicable laws and regulations. It is particularly disheartening then that lenders are often the subject of ill considered accusations regarding discrimination, accusations based upon analyses that lack statistical rigor,” said Michael Fratantoni, MBA’s Vice President of Research and Economics. Mr. Fratantoni claims that the problems with the analysis process that leads to these false claims is well known in the academic world.

The paper by Professor Yezer supports the theory that many of the claims of discrimination in mortgage lending based upon statistical data are inaccurate due to flawed statistical models. The report is fairly lengthy and, as you would expect in a research report from a Professor, quite technical, but here are some highlights from the report:

  • Conventional models of mortgage discrimination lack robust theoretical support. Without a solid grounding in theory, the statistical techniques commonly used today to estimate these models produce biased and inconsistent results, falsely indicating discrimination where it does not exist.
  • In choosing loan terms such as, the value of the property being purchased and down payment, applicants know that they are influencing their cost of credit and their probability of rejection. Thus the probability of rejection and mortgage rates are due to choices of both the applicant and the lender.
  • False assumptions introduce systematic biases into the estimates that make the models fail in ways that are particularly troubling. Discrimination tests tend to produce false positive indications of discrimination where none exists and tests for default risk are particularly bad at detecting instances where future default rates are likely to rise significantly.
  • The serious limitations of current statistical approaches to testing for discrimination and credit risk in mortgage lending have likely contributed to recent problems in mortgage markets. If these limitations are not recognized and naïve reliance on them continues, current problems are likely to recur in the future. Alternatively, there are major gains to be made if economic analysis of mortgage market discrimination and mortgage credit risk can be improved.

Don’t miss these posts!

We don’t spam! Read our privacy policy for more info.

🤞 Don’t miss info like this!

We don’t spam! Read more in our Privacy Policy

Leave a Reply