Measuring Bank Regulations: A Text-Based Approach

Job Market Paper

Selected for Emerging Scholar Initiative (ESI) at Financial Management Association 2025 

Presentations (* scheduled):  Frontier Risks, Financial Innovation and Prudential Regulation of Banks (J. Fin. Intermediation-CFAR-UNC-Gothenburg Conference), RiskLab/Bank of Finland/ESRB Conference on  AI and Systemic Risk Analytics, FDIC 23rd Annual Bank Research Conference (Poster Session), Eastern Finance Association, Southwestern Finance Association, Issues in Financial Markets and Banking (IFMB) Conference, Five-star Asia Pacific Workshop in Finance, MoFiR Workshop on Banking, European Finance Association (Cancelled), Asian Finance Association,  RMI 18th Annual Risk Conference, * Financial Management Association, Louisiana State University (E. J. Ourso College of Business), Monash University (Monash Business School), Université du Québec à Montréal (ESG UQAM), University of Technology Sydney (UTS Business School),  Southeast Missouri State University (Harrison College of Business & Computing), University of Florida (Warrington College of Business), National University of Singapore (NUS Business School)

Abstract: I introduce a novel text-based measure of U.S. banking regulation intensity from historical newspapers spanning 1926-2023. The Bank Regulation Index tracks changes around crucial events like Glass-Steagall's introduction and repeal. Deregulation displays a boom-bust pattern: increased bank stock returns and lending in the short term, followed by higher crisis likelihood in longer horizons. Decomposing the BRI into topics shows that credit-specific regulation reliably predicts future banking distress beyond well-established leading indicators. This pattern, confirmed across five other anglophone countries, underscores how monitoring credit deregulation through text-based analysis offers policymakers an even earlier warning indicator for detecting financial instability before crises materialize. 

Was the Glass–Steagall Act Justified? Text-Based Evidence from the Pecora Hearings

with Gustavo Cortes & Marc Weidenmier

Presentations (* co-author presentations): Federal Reserve Monetary and Financial History Workshop  (Federal Reserve Bank of St. Louis), International Macro History Online Seminar (Centre for Economic Policy Research),  Midwest Macroeconomics Meeting*, Rutgers University*, Utah State University (Huntsman)*, University of Calgary

Abstract: The Glass--Steagall Act of 1933 is one of the most influential and controversial pieces of financial regulation in U.S. history. Enforced for 66 years, the legislation was designed to prevent commercial banks from engaging in speculative activities in the securities market. We examine the passage of the Act in the 1930s and its repeal in the 1990s using a novel bank-level text-based risk measure, along with hand-collected daily stock prices and balance sheet data from major U.S. banks. We leverage machine learning methods on contemporary newspaper coverage by training our language model on the transcripts of the 1932--34 Pecora Hearings to identify Great Depression-era banking risk terminology. Our difference-in-difference (DID) estimates document that the Act's enactment significantly reduced banks' idiosyncratic volatility and text-based risk measures. Affected banks improved their distance-to-default and extended more credit, possibly mitigating the Depression's credit crunch. In contrast, using a staggered DID analysis of commercial and investment bank mergers, we find that the repeal of the Act in the 1990s increased idiosyncratic volatility. The reintegration reduced mortgage lending growth and loan approval rates while decreasing shareholder dividend yields. Our findings highlight the consequences of combining commercial and investment banking for financial stability.

Partisanship in Mutual Fund Portfolios

Presentation(s):  University of Florida (Warrington College of Business), Financial Management Association 2021, American Finance Association 2022 (Poster Session)

Abstract: Partisan bias in fund portfolios is the effect of fund manager’s political affiliation on portfolio allocation decisions. I study two potential manifestations of this bias: biased expectations where managers become optimistic (pessimistic) when their party comes in (goes out of) the government, and in-group favoritism where managers choose higher holdings of politically aligned firms. I find strong evidence for the biased expectations channel, using recent data that includes the effects of the 2020 Presidential election. However, contrary to past literature, I find no evidence for in-group favoritism. I also document a partisan bias in holdings of stocks exposed to politicized topics (COVID-19 and Brexit). The COVID-19 result does not carry over to earlier pandemics (H1N1, Ebola and Zika).