Banks operate in one of the nation’s most heavily regulated industries, where policy intervention aims for wide-ranging goals that include limiting risk, protecting consumers, and ensuring fair treatment of individuals through equal access to credit. The Community Reinvestment Act (CRA), enacted in 1977, is a much-studied example, though unlike most banking regulations, which restrict banks, the CRA encourages them to extend credit to targeted groups within certain communities. Though a number of studies (e.g., Agarwal, Benmelech, Bergman, and Seru (2012), Saadi (2020)) have examined whether the CRA encourages risky lending, our new paper looks at a different potential consequence of the law: banks’ willingness to take costly actions to avoid CRA oversight.
In our paper, we focus on a feature embedded in a 1995 CRA revision that increased regulatory intensity and monitoring for banks with assets greater than $250 million – large banks – while in some ways reducing regulation of other banks, small banks. Among other changes, small banks faced a streamlined and narrower evaluation process and were evaluated at five-year intervals rather than the two-year intervals required of large banks. On the other hand, large banks faced a more comprehensive lending evaluation on top of the investment and service tests. The results from these evaluations have had consequences, as regulatory agencies now consider a bank’s past CRA performance when evaluating applications for new branch openings and bank mergers and acquisitions.
A discrete increase in regulatory requirements tied to the $250-million threshold suggests a perceived cost of complying with the CRA. Our study exploits this threshold by examining the tendency of a bank to strategically manage its assets to fall just shy of the threshold. As the strategic slowing of a bank’s growth has a cost, our approach is based on a revealed-preference argument, whereby banks bunching below $250 million view the increase in regulatory cost triggered upon crossing the threshold as greater than the cost of slowing growth.
Consistent with this idea, we begin by documenting significant bunching of banks at the $250 million asset threshold from 1996 to 2004. In contrast, we find no evidence of bunching in the pre-reform period (1986-1993) nor signs of bunching at other salient asset values ($150 million and $350 million), which were not tied to CRA regulation. In addition, we confirm bunching using “excess mass” techniques from public finance (e.g., Kleven and Waseem (2013), DeFusco, Johnson, and Mondragon (2020)).
To identify how a bank circumvents the more rigorous CRA assessment and estimate possible effects on the area it serves, we turn to a difference-in-differences approach. Here, the first differential captures the difference in outcomes for banks with pre-reform assets falling just below the threshold (treated) relative to similarly sized banks (assets less than $350 million), while the second differential captures how this difference changes following the 1995 reform (post). We find that banks with 1994-measured assets between $200 million and $250 million (i.e., banks with the greatest incentive to bunch below the threshold once the reform was in place) experienced a 4.4pp reduction in post-reform asset growth. Reassuringly, this estimate is stable across various lower bounds for the treated group (e.g., $220 million). Moreover, we find no evidence of a pre-trend, with the effect immediately realized in 1995. This result is consistent with banks strategically managing asset growth to retain the reduction in regulatory costs associated with a small bank classification.
Subsequently, we shift our attention to the potential real effects of the CRA on areas served by treated banks. First, we consider the potential heterogeneity in a bunching bank’s rejection rate for loans that qualify for CRA credit (“low- and moderate-income” or LMI loans) compared with other loans. We find that banks with 1994-measured assets just below the $250 million threshold experience a 1.3pp to 2.2pp increase in rejection rates for LMI-qualifying loans. Importantly, these estimates control for local economic conditions with county-year fixed effects and include bank-LMI and LMI-year fixed effects to account for general differences across banks and loans over time. This result is particularly interesting, suggesting that the increase in regulatory costs is associated with a reduction in credit offered to a specific group of potential borrowers targeted by the CRA.
We also examine the equilibrium effects on areas served by bunching banks. We begin by examining the composition of business establishments across counties. Literature has documented that small firms rely more on small banks (e.g., Berger, Saunders, Scalise, and Udell (1998), Weston and Strahan (1996)). Therefore, we would expect any negative effect on business establishments to be concentrated in smaller firms. Consistent with this idea, the post-CRA reform share of small businesses (less than 20 employees) decreases in counties served by bunching banks. We also consider the potential effects on independent innovation. To the extent that independent inventors behave like small firms and rely disproportionately on lending from smaller banks (Babina, Bernstein, and Mezzanotti (2020)), slower bank growth may hamper innovation. Consistent with this idea, we find a reduction in independent innovation in treated counties, with a post-reform decrease in the rate of individually-assigned patents of 4.9 percent to 7.1 percent.
Taken together, these results suggest that, rather than promoting lending and economic growth in areas they serve, banks instead elected to reduce their economic footprint to avoid an increase in regulatory oversight. This strategy was costly for the avoiding banks but also for borrowers whom the act was designed to benefit, because it limited their economic growth. Our results are also relevant in light of recent efforts to modernize and expand the CRA to institutions currently not obligated to comply. Our findings caution against the potential negative consequences of this action if the newly regulated institutions take similar steps to avoid regulatory oversight.
Agarwal, Sumit, Efraim Benmelech, Nittai Bergman, and Amit Seru, “Did the Community Reinvestment Act (CRA) lead to risky lending?”, Technical Report, Working Paper, National Bureau of Economic Research 2012.
Babina, Tania, Asaf Bernstein, and Filippo Mezzanotti, “Crisis Innovation”, Technical Report, Working Paper, National Bureau of Economic Research 2020.
Berger, Allen N, Anthony Saunders, Joseph M Scalise, and Gregory F Udell, “The effects of bank mergers and acquisitions on small business lending”, Journal of Financial Economics, 1998,50(2), 187–229.
DeFusco, Anthony A, Stephanie Johnson, and John Mondragon, “Regulating household leverage”, The Review of Economic Studies, 2020, 87 (2), 914–958.
Kleven, Henrik J and Mazhar Waseem, “Using notches to uncover optimization frictions and structural elasticities: Theory and evidence from Pakistan”, The Quarterly Journal of Economics, 2013, 128 (2), 669–723.
Saadi, Vahid, “Role of the Community Reinvestment Act in mortgage supply and the US housing boom”, The Review of Financial Studies, 2020, 33 (11), 5288–5332.
Weston, James and Philip E Strahan, “Small business lending and bank consolidation: Is there cause for concern?”, Current issues in Economics and Finance, 1996, 2 (3).
This post comes to us from professors Jacelly Cespedes at the University of Minnesota’s Carlson School of Management, Jordan Nickerson at the University of Washington, and Carlos Parra at the Pontifical Catholic University of Chile’s School of Business. It is based on their recent article, “Strategically Staying Small: Regulatory Avoidance and the CRA,” available here.