The Home Mortgage Disclosure Act (HMDA) is the nation’s most complete record of mortgage origination activity.
The HMDA helps industry experts and researchers gauge market activity by lender, geography, and race or ethnicity. This helps them assess credit availability, the riskiness of the market, potential discrimination, and the need for public and private investment in housing.
HMDA data are free, public, and highly reliable. Their availability and dissemination deepens our collective understanding of the market, making it more efficient and enabling informed policymaking.
In 2015, the Consumer Financial Protection Bureau (CFPB) published a rule implementing HMDA reforms required by the Dodd-Frank Act (PDF). But earlier this month, the CFPB released two items that threaten to substantially curtail the availability of these data:
- a Notice of Proposed Rulemaking (NPRM) (PDF) that would reduce the number of reporting institutions by raising the number of loans an institution must make to be required to report to anyone, including regulators, under the HMDA
- an Advance Notice of Proposed Rulemaking (ANPR) (PDF) that opens the possibility of eliminating much of the data specifically added to the regulation in response to the abuses and subsequent disaster of the early 2000s
Both actions have unnecessarily short comment periods: just 30 days from appearance in the Federal Register for the NPRM (comments due June 12) and 60 days for the ANPR (comments due July 8).
NPRM proposals would allow at least 759 institutions to go dark on mortgage lending
The NPRM raises the threshold for HMDA reporting on closed-end loans from 25 loans in either of the previous two years to either 50 or 100 loans (alternative proposals), effective January 1, 2020. This would apply to banks and nonbanks.
For open-ended loans, such as home equity lines of credit, the 500-loan threshold for inclusion would be extended through 2021 and then reduced to 200 loans in 2022, rather than to 100 loans, as in the 2015 HMDA rule. The CFPB estimates that average savings for the 100-loan, closed-end proposal would be less than $5,000 per year per institution.
According to the NPRM, these threshold increases would allow 759 institutions (under a 50-loan threshold) or 1,718 institutions (under a 100-loan threshold) to go dark on mortgage origination reporting. By 2022, 681 out of 1,014 institutions would go dark on open-ended loans. This is a serious loss of insight on a market vital to the health of the economy; regulators failed to fully understand the importance of these loans and the role they played leading up to the 2008 financial crisis.
These steps could substantially weaken the effects of the 2015 HMDA rule, which expanded the number of data points collected to track the abuses that triggered the Great Recession. It’s difficult to see how the loss of this information, and the resulting loss of market efficiency and policymaking awareness, is worth the small reduction in cost for those newly exempted.
Additionally, the process that the CFPB has used to roll out this NPRM is concerning. Banks have just implemented the 2015 HMDA rule—why change the rules again now?
Responsible rulemaking requires that the CFPB institute a longer comment period and show cause for the proposed change. Instead, the CFPB’s short comment period puts the burden on the public to demonstrate why the rule shouldn’t be changed. As the CFPB itself admits, the HMDA is a public disclosure statute. If disclosure is the purpose of the statute, the burden of proof must be on those who wish to curtail it.
The ANPR proposals would curtail the collection of critical data on manufactured housing, loan denials, and multifamily lending
The ANPR proposals, which would apply to all HMDA reporters, are perhaps more troubling. We will analyze these in more detail, but three issues immediately stand out.
- Curtailing manufactured housing data
The CFPB says that some stakeholders found it burdensome to report information on whether the borrower owns or leases the land where a manufactured home is located. Though not required in HMDA reporting before the 2015 rule, this information has potentially large public benefit.
Urban Institute research shows that as many as 60–70 percent of borrowers who purchased a manufactured home using a chattel loan (a loan not secured by real property) might have been eligible for a mortgage or real estate loan (a loan secured by the structure and the land). The latter is substantially less expensive and has exponentially more consumer protections.
But our numbers were approximations. We could more reliably study this topic with HMDA data on exactly how many borrowers were affected.
Under the Duty to Serve rule, Fannie Mae and Freddie Mac are also exploring expanded work in manufactured housing. The importance of chattel lending in this sector underscores the value of new HMDA data in enabling the government-sponsored enterprises to meet this obligation.
Given the very tight supply of affordable housing, expanding options for lower payments on manufactured housing could spur more buying and the production of more units of this affordable housing type. And fully understanding those benefits requires the data the CFPB proposes to eliminate.
- Curtailing the “open-ended” response to why a loan is denied
The ANPR also proposes, based on anonymous “stakeholder feedback,” to eliminate the open-format description of why a loan was denied. This contradicts the 2015 rule, which required a reason for a denial from reporters but permitted an open-format description of the reason if none of the standard reasons fit. Reasons for loan denial are critically important.
The CFPB can reduce the number of instances when lenders must use the open-format field by expanding the choices of reasons for denial. This can easily be achieved by analyzing the 2018 HMDA filings and including the most common free-form text responses as choices instead.
- Curtailing data on multifamily lending
The ANPR also asks for comments on whether the CFPB should exclude coverage of business or commercial purpose loans made to a nonnatural person (a business entity) and secured by a multifamily dwelling on the argument that it is unnecessary to fulfill the purposes of the HMDA and that the burden of reporting does not outweigh the benefits.
We can’t think of anything more central to the purpose of the HMDA than multifamily lending, mostly for rental properties. Using HMDA data, we recently showed that multifamily lending contributes disproportionately to banks’ requirements under the Community Reinvestment Act (CRA).
We also showed that multifamily lending is much more concentrated than single family lending, so monitoring whether these lenders are serving low- and moderate-income tracts, as well as better-off tracts, is critically important.
The ANPR essentially proposes to do what the CFPB refused to do in the 2015 rule—namely, exclude reporting of the very large number of multifamily loans that are made to nonnatural persons, such as limited liability corporations.
Although much of the data required under the 2015 rule is collected (although not in a standard format) by banks for the CRA and other regulatory purposes, the lack of HMDA inclusion would make it more difficult for examiners to see how one institution compares with similar institutions. And the public would not be able to view individual institutions nor make comparisons between institutions as to their relative activity in providing loans to low- to moderate-income areas or otherwise serving their communities. To the extent that there are data fields (beyond those the CFPB has already exempted) that are irrelevant or excessively burdensome for multifamily lenders, the solution is to exempt additional fields, not the entire category of loans.
The ANPR proposal is a step backward, and its adoption would reduce transparency in the multifamily loan market, which is critical to supporting both subsidized and naturally occurring affordable rental housing.
HMDA reporting can and should be improved, but the answer isn’t to collect less information, it is to collect information more intelligently. Reducing the information available to regulators, policymakers, and the public will reduce our collective understanding of this critical market, making it less efficient and making policymaking more speculative. No one should want that outcome.