Statistics Done Right This is where I write about my research.

Working Papers

Do Closing Minority Depository Institutions Affect Credit in Their Communities

Abstract: I construct a panel dataset and apply an event study framework to estimate the effects of Minority Depository Institution (MDI) and non-MDI branch closures on credit accessibility in racially diverse local markets, defined by census tracts. My findings reveal minimal negative effects, with four notable exceptions: 1) Asian MDI branch closures lower large mortgage originations within the Asian community, 2) Hispanic MDI branch closures lower small business loan originations for small firms, 3) non-MDI branch closures lower mortgage originations in Black communities served by Black MDIs and 4) non-MDI branch closures lower small-sized mortgage originations. Surprisingly, non-MDI bank branch closures increase total small business loan originations. Using a lender level Herfindahl–Hirschman index (HHI), I show branch closures do not lead to more concentrated lending markets, rather encourage entry of non-local lenders. The results highlight the evolving role of physical bank branches in an increasingly digital banking landscape.

Who Distributed More Paycheck Protection Program (PPP) Loans? A Nonparametric Approach to Determine the Characteristics of Banks that Influenced Their Ability to Distribute PPP Loans

Abstract: The Paycheck Protection Program (PPP) was an emergency measure taken during Covid-19 pandemic to support small businesses that faced mandated business closures. Using Federal Deposit Insurance Corporation (FDIC) Call Report data from June 2020, I measure how much PPP loans/assets were given out by FDIC registered banks. I use nonparametric multivariate Kernel regressions and semiparametric smooth coefficient Kernel regressions to understand what institutional features lead to a higher amount of PPP loans/asset for every bank. I find that commercial and industrial loan commitments (C&I) to large businesses, C&I loans larger than $250,000 issued to small businesses, unused C&I loan commitments, core deposits and the status of the institutions as community banks, all positively influence the amount of PPP loans/asset that are disbursed by banks.

How Extreme Weather Events Impact the Frequency of Loan Take-up by Bangladeshi Households in Vulnerable Communities: An Analysis with Count Data Models.

Abstract: I study whether facing extreme weather events increases the likelihood of multiple loan take-up by rural Bangladeshi households. Microcredit loans are marketed to low income rural households that do not have access to commercial banks. These institutions charge higher interest rates than commercial banks and practice predatory lending. Using count data models, I show that households that face extreme weather events, households that own agricultural or non agricultural enterprises, households with members suffering from chronic illness and those who have faced sudden illness or death of an earning member are all significantly likely to make up multiple loans from microcredit institutions.