Algorithmic credit allocation and the rise of financial inequality

Evidence from U.S. fintech platforms

Bibliographic Data

ID9403933
AuthorsTouitou Mohammed (0000-0002-4442-7888, University of Algiers 3, corresponding author)
Year2025
Volume75
Issue4
Pages599-629
Publication date2025-11-14
Open AccessNo
TypeARTICLE
VenueActa Oeconomica (JOURNAL)
Journal identifiersISSN: 0001-6373 • E-ISSN: 1588-2659
PublisherAkadémiai Kiadó (PUBLISHER • HU)
DOI10.1556/032.2025.00134
OpenAlexW4416415513
LanguageEN
References cited38

Algorithmic credit systems used by fintech platforms are increasingly transforming access to financial services across the United States. While machine learning–driven models improve credit availability, they also raise concerns about fairness, transparency, and structural bias. This study presents a hybrid empirical framework—combining causal inference techniques with explainable AI audits—to evaluate the distributional impacts of fintech lending from 2010 to 2023 at the ZIP-code level. Using difference-in-differences estimations and Causal Forest models, we find that fintech entry increases approval rates by 6–9 percentage points, particularly in low-income and majority-minority areas. However, these gains are accompanied by higher interest rates and spatially embedded proxy biases, especially in ZIP code–linked features. SHAP-based interpretability and fairness diagnostics reveal significant disparities and fairness decay under demographic drift. To address these challenges, we propose a novel Algorithmic Accountability Index (AAI) that quantifies disparities in transparency, proxy risk, and model equity. These findings provide policy-relevant insights for digital financial governance, algorithmic audit standards, and the equitable design of AI-based credit infrastructures

Tools

Open DOI