Skip to content

Recent Research Highlights Critical Points for Policymakers to Consider in CRA Modernization

Written by Low Income Investment Fund

As the federal banking regulators continue to consider potential changes to Community Reinvestment Act (CRA) regulations, community development stakeholders have been participating in their own evaluation of CRA, including highlighting the many ways in which the law has benefited the industry and discussing opportunities to further strengthen the regulations. Recent research on CRA, highlighted by the Penn Institute for Urban Research (Penn IUR), presents particularly relevant information for stakeholders to consider as the regulatory process advances over the coming months.

This blog discusses a series of main takeaways from three recent working papers on CRA. The papers are the result of a February 2019 research symposium convened by the Federal Reserve Board, Federal Reserve Bank of Philadelphia, and Penn IUR to consider the future of CRA. LIIF CEO Dan Nissenbaum presented on a panel “Effectiveness of the Community Reinvestment Act: Past, Present, and Future” at the research symposium. The papers include:

  • Quantitative Performance Metrics for CRA: How Much “Reinvestment” is Enough? by Carolina Reid. Reid, a professor at the University of California, Berkley’s Terner Center for Housing Innovation, analyzes data from CRA performance evaluations released in 2011 and 2016 for all banks in California to illustrate what banks are currently doing to meet their CRA requirements.
  • The Community Reinvestment Act: What Do We Know, and What Do We Need to Know? by Laurie Goodman, Jun Zhu, and John Walsh. The Urban Institute’s Goodman, Zhu and Walsh analyze CRA lending data – including mortgage lending and data on small businesses, small farms, and community development lending – to inform what we currently know about CRA and what data is necessary to develop a more robust and accurate understanding of the law’s impacts.
  • Updating CRA Geography: It’s Not Just About Assessment Areas by Mark Willis. The NYU Furman Center’s Mark Willis articulates a vision for a new geographic framework that enables internet-based banks to allocate their CRA activity to more low- and moderate-income (LMI) communities.

Three themes emerged throughout the papers: who does and should benefit from CRA; the need to appropriately account for the growing number of branchless banks; and the need for more and better data.

I. Who Benefits from CRA?

Reid and Goodman, Zhu and Walsh detail patterns of banks receiving CRA credit for lending to higher-income borrowers in LMI areas, leading to potentially concerning implications for gentrification and displacement.

In general, banks do less single-family LMI lending than non-banks largely due to banks’ movement away from Federal Housing Administration (FHA) loans in recent years (Goodman, Zhu and Walsh, p. 9). What’s interesting, though, is that there is no difference in the share of bank loans to LMI borrowers inside and outside of assessment areas, which Goodman, Zhu and Walsh note “raises questions about the CRA’s effectiveness as currently structured for single-family lending” (p. 11).

For instance, Reid’s analysis in California found that “more than three-quarters of loans in LMI tracts—those that the banks are claiming as CRA activity—are going to middle- and upper-income households”. This trend was amplified in high cost areas, with 92 percent of loans in LMI tracts in San Francisco going to middle- and upper-income households in 2017 (p. 12).

Goodman, Zhu and Walsh found similar patterns in their analysis of bank single-family mortgages by income-level in LMI areas. Given that CRA is intended to serve LMI communities, the authors question whether the practice of providing CRA credit to any borrower in an LMI community should continue, especially given that over 60 percent of the dollar volume of bank loans in LMI census tracts are not going to LMI borrowers (p. 15). Goodman, Zhu and Walsh conclude that “examiners should make sure institutions are not solely skimming large, more profitable loans in gentrifying areas to count toward CRA requirements” (p. 15).

II. Accounting for Branchless Banks

The geographic framework of CRA is central to the modernization conversation given the advent of online and nationwide banking and financial products since CRA was enacted in 1977 and regulations were last updated in 1995. Current CRA regulations require banks to identify at least one assessment area (AA), even for banks without any physical deposit-gathering facility. However, the regulations for wholesale and limited purpose banks acknowledge that the bank’s main office is not necessarily where the bank engages in most of its activity. In addition to their designated AA(s), wholesale and limited purpose banks are evaluated on the community development (CD) test, which offers full credit for CD activities in any LMI market nationwide.

The increase in large retail banks that have an internet-based presence beyond physical branch locations necessitates a modernized approach to their evaluation process to ensure that they are not deterred from conducting CD activity outside of their AA(s). Willis proposes an evaluation process that separately evaluates a bank’s CRA performance first within its AA(s) and then beyond its AA(s).

Willis recommends the following evaluation process for large retail internet banks:

1.     Evaluate the bank’s CRA performance within its AA(s)

  • Existing AAs retain their importance under Willis’ proposal because a bank must adequately serve it’s AA(s) before it can proceed to the next step.

2.     Evaluate the bank’s CRA performance beyond its AA(s). This evaluation should include a:

  • Retail Products Test that tests for home mortgages, small business loans, and small farm loans at the local level. This test evaluates the bank’s retail presence in a particular area regardless of the presence of a physical location.
  • Community Development Test that is identical to the CD test already applied to wholesale and limited purpose banks that do not have deposit-taking locations. Banks would be eligible for full credit (both quantitative and qualitative) for CD activities in any of the LMI markets across the nation.

3.     Blend the results of the AA evaluation and the outside-AA evaluation for an overall bank rating.

  • Willis suggests that such a blended rating would more appropriately reflect the scope of these banks’ activity.

The current regulatory framework creates disincentives for internet-based and branchless banks to invest in LMI communities nationwide, despite these banks serving borrowers across the country. Willis’ approach seeks to remove these regulatory disincentives and encourage banks to serve more LMI communities.

III. The Need for More and Better Data

The importance of data was prevalent across the papers and raised serious implications about the feasibility – both substantively and logistically – of implementing a metric-based evaluation in CRA. The most notable barrier that Goodman, Zhu and Walsh encounter in their analysis is the lack of available data on CRA activities. In fact, the authors could only analyze CRA lending data because it is the only category for which data is available.

Reid also emphasizes the inconsistencies in how CRA data is collected and reported, noting that even when the information is included in a bank’s performance evaluation, there are a number of inconsistencies that make it difficult to conduct an evaluation. Inconsistencies include variations in data by regulator, by examiner, and by format – i.e. buried in a narrative or entered into a table – in addition to irregularities in the definition of certain activities.

Both papers highlight the need for significantly better data collection and reporting procedures. At a minimum, addressing these inconsistencies is important solely for the sake of evaluating CRA’s impact in communities. But as the regulators consider moving to a metric-based system that is implicitly linked to a baseline of current CRA activity, the necessity of having a consistent and effective infrastructure for data collection cannot be overstated.


Each of these papers offers a valuable contribution to the CRA modernization debate. The three themes identified above in no way capture the totality of takeaways from the research, but do indicate the breadth of the modernization challenge, as well as the importance of ensuring that CRA regulations are appropriately calibrated to the needs of communities and the shifting characteristics of financial institutions.

LIIF will continue to monitor the CRA modernization debate and provide updates as the process unfolds. We strongly encourage all community development stakeholders to stay engaged in the conversation and join our effort to ensure CRA remains an effective tool for community development nationwide.