Alfred M. Pollard, General Counsel
Federal Housing Finance Agency
400 Seventh St. NW, 8th Floor
Washington, DC 20219
Attention: Comments/ RIN 2590-AA98: Validation and Approval of Credit Score Models by Fannie Mae and Freddie Mac
Dear Mr. Pollard,
On behalf of National Taxpayers Union, R Street Institute, Citizens Against Government Waste, Institute for Liberty and Taxpayers Protection Alliance (“the undersigned”), we respectfully submit these comments to the Federal Housing Finance Agency (FHFA) concerning its Notice of Proposed Rulemaking for the validation and approval of credit score models. The undersigned are pleased to comment in favor of the proposed rule, which we believe represents a fair and reasonable interpretation of section 310 of the “Economic Growth, Regulatory Relief, and Consumer Protection Act of 2018.” Our organizations have long been involved in financial services issues and are prominent voices on housing finance matters. Importantly, we all follow the same housing policy fundamentals: a system that promotes broad access to credit for qualified borrowers, a significant private capital buffer, and administrative actions that promote competitive markets and protect taxpayers.
Since Fannie Mae and Freddie Mac (the GSEs) were placed into conservatorship more than a decade ago, our organizations have called for Congressional and administrative actions to mitigate taxpayer exposure to risky activity. Taxpayers have expressed concern about greatly expanded GSEs activities in the private market over the past few years while enjoying the unique competitive advantages of government backing. With the GSEs
having more than $5 trillion of mortgage risk on their extremely leveraged balance sheets, which are by extension underwritten by taxpayers, the federal government holds a considerable level of risk.
Economists widely agree that the significant increase in housing foreclosures that fueled the 2008 financial crisis and subsequent recession was a result of a weakening of GSE mortgage standards through affordable housing goals. These goals required the GSEs meet annual quotas of low- and moderate-income mortgages. As time went on and the number of prime borrowers dried up, the GSEs had to expand operations in the subprime market in order to meet their annual quotas. As the market shrank, the GSEs found it harder and harder to find creditworthy borrowers causing them to lower their standards to meet their affordable housing goals. This involved either reducing the accepted credit score, lowering the required down payment, raising the debt-to-income ratio, or accepting low or no documentation.
Accepting lower credit standards certainly expanded the number of people who were eligible for a mortgage, but it allowed a greater number of under-qualified borrowers to obtain a loan who would have otherwise been denied such a large line of credit. Once defaults skyrocketed and the housing bubble burst, the GSEs were wired more than $190 billion from taxpayers to keep them afloat and were placed into conservatorship where they remain to this day.
If there is one lesson from the 2008 housing crisis that should have been learned, it is that overly ambitious affordable housing goals and the rush to qualify numerous borrowers by any means can put the economy, and taxpayers, at great risk. GSEs utilize credit scores in several ways including benchmarks for risk fees, loan eligibility guidelines, and (for Freddie Mac) one of many attributes in making a credit assessment. They are also used internally to balance counterparty risk, an often-overlooked but very important role. Thus, allowing new credit score models into the GSE framework could have major consequences for their operations, their risk, and in turn taxpayer liabilities.
Such consequences would also reverberate throughout the private sector, as lenders, loan servicers, mortgage insurers, and other parts of the industry would face all manner compliance and implementation costs. New credit scoring methods in the GSEs could also eventually spill over into taxpayer-backed lending programs at the Federal Housing Administration, the Small Business Administration, and other agencies. In an environment where GSEs and FHA appear to be more heavily weighting their portfolios with higher-risk loans, the introduction of new credit scores could even affect the overall systemic risk calculation at an especially delicate point in financial markets. These factors are discussed in greater detail in a Policy Paper that National Taxpayers Union filed separately with FHFA.
It is of particular concern to free market, limited government groups to see how the “Credit Score Competition Act,” included as Section 310 of S. 2155, “the Economic Growth, Regulatory Relief, and Consumer Protection Act” that passed in 2018, will be implemented. Section 310 directs FHFA to create a process for evaluating new credit scoring models for use by the GSEs, but does not mandate they accept more just one type of credit score. We believe FHFA interpreted the legislative text in a careful and thoughtful manner that complies with legislative intent. The proposed rule issues standards for compliance, which sets forth several factors that must be considered in the validation and approval process, including the credit score model’s integrity, reliability, and accuracy, its historical record of predicting borrower and credit behaviors, and consistency of any model with GSE safety and soundness.
Further, FHFA rightly notes in the proposed rule that alternative scores may immediately gain a competitive advantage in the market. As such, the rule specifically “prohibits an Enterprise from approving any credit score model developed by a company that is related to a consumer data provider through any common ownership or control, of any type or amount.” In addition, the proposed rule not only calls for sound cost-benefit analysis in evaluating new models, it also builds in conflict-of-interest guardrails (which are standard in other regulatory spheres) to ensure that those models compete on a level playing field. This holds promise for creating a true market-driven competitive environment with an opportunity for innovation.
Additionally, we are supportive of the straightforward, four-step process which the Enterprises evaluate and implement alternative credit-scoring models. The process is summarized below:
1) Solicitation of applications from credit score model developers;
● Proposes that solicitation for new applications occur at least every seven years, or as determined necessary by FHFA.
2) Initial review of submitted applications;
● Each GSE would obtain the data from the data provider on behalf of the applicant.
3) Credit score assessment;
● During this assessment phase, each credit score model would be assessed for accuracy, reliability, and integrity.
● Approaches for assessing accuracy include:
1) Comparison-based. This approach will not require the applicant’s credit score to be more accurate than the existing credit score in use by the GSEs. This approach would be more subjective and indicate reasonableness of the credit score’s accuracy.
2) Champion-Challenger. The applicant’s credit score must be more accurate than the existing credit score in use by the GSEs. This would be a bright line test.
3) Benchmark-Based, an absolute statistical standard would be established and all scores would have to surpass the standard. For example, a K-S or Gini score could be established that must be surpassed.
4) Transitional approach. This approach would allow one of the other approaches be applied for the initial credit score assessment and a possible different approach for subsequent credit score evaluations.
4) Enterprise business assessment;
● During this phase, a GSE would assess the credit score model in conjunction with the GSEs business systems and processes.
● In addition, the GSE must consider impacts on the mortgage finance industry, assess competitive effects, conduct a third-party vendor review, and any other evaluations established by the GSE.
The validation and approval process, which produces the resulting approved credit score model, must meet these five statutory requirements:
(i) satisfy minimum requirements of integrity, reliability, and accuracy; (ii) have a historical record of measuring and predicting default rates and other credit behaviors; (iii) be consistent with the safe and sound operation of the corporation; (iv) comply with any standards and criteria established by the Director of the Federal Housing Finance Agency under section 1328(1) of the Federal Housing Enterprises Financial Safety and Soundness Act of 1992; and (v) satisfy any other requirements, as determined by the corporation.
Taxpayers have a significant stake in the housing finance market and FHFA must do everything in its power to ensure that taxpayer risk is mitigated to its fullest extent. The proposed rule will undoubtedly help to protect from a potential “race to the bottom” effect to qualify as many possible borrowers as possible through political manipulation of tools that are supposed to be reliable predictors of risk. Significant innovation in the credit scoring space is already occurring through the advent of refinements and expansions to existing standard tools, which themselves are being subjected to rigorous testing. We believe these modernizations can be balanced with the benefits of a stable, predictable system of lending and finance that measures and protects against risk, not only to borrowers and lenders, but also to taxpayers.
Thank you for the opportunity to offer our views on this proposed rule. We urge FHFA to adopt the rule as is and implement it in a timely manner.
Pete Sepp, President
National Taxpayers Union
Alex J. Pollock, Distinguished Senior Fellow
R Street Institute
Tom Schatz, President
Citizens Against Govenrment Waste
David Williams, President
Taxpayers Protection Alliance
Andrew Langer, President
Institute for Liberty