Quietly but steadily, the public and policymakers are losing access to important data about the mortgage market and the state of credit for low- and moderate-income borrowers and communities. Three current rulemakings from the Consumer Financial Protection Bureau (CFPB) and bank regulators—one final, one proposed, and one at an earlier stage—promise to exacerbate the situation.

With COVID-19 raising the potential for another credit crisis, this pullback in public data is especially troubling. How did we get here?

In 2015, the CFPB published a final rule implementing provisions of the Dodd-Frank Act that were designed to increase the amount, type, and quality of mortgage market data available to the public under the Home Mortgage Disclosure Act (HMDA). The goal was to establish a better early-warning system for the kinds of problems in the mortgage market that led to the bubble and bust of the century’s first decade.

As part of that rulemaking (PDF), the CFPB exempted lenders making fewer than 25 mortgage loans a year from reporting. The result: 1,400 depository institutions, or 22 percent of those that had been reporting, became exempt.          

Then, in 2018, Congress passed the Economic Growth, Regulatory Relief, and Consumer Protection Act, which exempted depository institutions making fewer than 500 mortgage loans a year from reporting almost all of the new data that had been required under Dodd-Frank and the CFPB’s 2015 rule. The result: much less information from 3,250 insured banks and credit unions. That’s 67 percent of the 4,860 financial institutions currently required to report, with a disproportionate impact on rural and low-income areas and on information about small loans.

The CFPB’s new rule (PDF), issued April 16, 2020, raises the threshold for reporting any data from 25 loans (per the 2015 rule) to 100 loans, exempting 1,640 depositories and 60 nondepositories from reporting—on top of the 1,400 institutions exempted in 2015.

In the preamble to the rule, the CFPB stated it “is unable to readily quantify the loss of some of the HMDA benefits to consumers”—never mind policymakers—“with precision.” But consistent with a comment we submitted on the proposal, the CFPB noted that the effects of reduced data about mortgage availability would disproportionately affect rural and low-income census tracts.

Most importantly, the CFPB estimated the public would lose access to information about 13 percent of all multifamily loan originations. Both the structure of multifamily lending and its importance to housing low-income families makes losing these data especially troubling.

But this is not the only CFPB action to limit data on multifamily lending. In May 2019, the CFPB issued an advance notice of proposed rulemaking (PDF) to inquire about four pressure points in reporting the new fields required under its 2015 rule implementing the Dodd-Frank Act.

As we pointed out in a comment letter, the most damaging of the inquiries would be the elimination of HMDA reporting of multifamily loans to “nonnatural persons”; that is, business entities. These business entities have always been part of HMDA reporting.

Using 2018 HMDA data, the first reported pursuant to the 2015 rule, we were able to understand the importance of these entities for the first time. Using the narrowest definition of nonnatural persons, we found that exempting multifamily loans to nonnatural persons would eliminate HMDA reporting on 65.6 percent of all multifamily loans (structures with five units or more) and more than 80 percent of loans on structures with 50 units or more.

Overall, we would lose data about loans on 80.5 percent of units in multifamily structures. In other words, it would render the multifamily HMDA data useless. No decision has been reached on the fate of these loans.

Finally, in January 2020, the Office of the Comptroller of the Currency and Federal Deposit Insurance Corporation issued a notice of proposed rulemaking to modernize the Community Reinvestment Act (CRA). In addition to making major changes in how banks’ performance under the CRA will be evaluated (see our comment letter), the agencies propose to completely rewrite the public data provisions of the agencies’ CRA regulations.

Under the proposal—and notwithstanding that the proposal would significantly increase the reporting burden on banks (PDF)—the public would have access to much less, and much less granular, geographic information about small business, small farm, and community development lending. Publicly released data would be aggregated at the bank level and for all banks at the county level.

When combined with the CFPB’s cutback in mortgage lending data, it will become more difficult—and in some geographies, close to impossible—to determine how well each bank is serving its communities.

Public data are important. They allow us to look at credit availability across the market. They allow us to look at credit flowing to low- and moderate-income communities versus their affluent counterparts. They allow us to see what individual banks are doing in their lending to specific areas. We are likely to lose all of that, one regulation at a time, notwithstanding these datasets’ importance to evidence-based policymaking. Good policymaking cannot tolerate the continuous chipping away of this important resource.