Working Papers

[1]The Social Costs of the Current Expected Credit Loss Approach: Evidence from Mortgage Lending


I examine the impact of the current expected credit loss (CECL) approach on geographic disparities in mortgage credit access between low- and moderate-income (LMI) and non-LMI census tracts. Banks are subject to the Community Reinvestment Act (CRA), a key policy to reduce credit access inequality for borrowers in LMI census tracts. Using a triple-difference design, I find that following CECL adoption, banks reduce mortgage approval rates and increase mortgage interest rates in LMI census tracts, suggesting a decrease in mortgage credit supply by CECL-adopting banks in these areas. Non-bank lenders mitigate the reduction in bank credit partially. I show that following CECL adoption, non-bank lenders increase mortgage approval rates in LMI census tracts with greater exposure to CECL-adopting banks, and non-bank lenders also slightly loosen their underwriting standards, but do not charge higher interest rates when extending mortgage credit in these areas. The aggregate quantity of mortgage credit in LMI census tracts also decreases following CECL adoption. Findings highlight that the CECL approach has social costs by reducing banks' promotion of equitable credit access under the CRA. Results also inform the policy debate on extending CRA lending requirements to non-bank lenders.

[2]Fraud at a Distance? How Remote Work Shapes Financial Misconduct (NBER WP 34417) with John M. Barrios and Jessie Guo
Financial misconduct is often a team activity, facilitated by face-to-face interactions, shared norms, and trust. We exploit the sudden shift to remote work during COVID-19 to examine how workplace organization shapes collusion and financial misconduct. Using a novel firm-level measure of work-from-home feasibility, we find that firms that are more able to operate remotely experienced large post-2020 declines in misconduct. This decline is found across multiple misreporting proxies and is robust to various alternative measures of remote work. Cross-sectional tests indicate stronger declines in teamwork-intensive firms, firms with effective internal controls, and firms with weaker pre-COVID employee perceptions of culture and leadership, consistent with environments more reliant on dense in-person relational coordination. Overall, our findings reflect that financial misconduct is a team activity, sensitive to the organizational structure of the firm, with important implications for governance and organizational design.