This paper investigates the impact of introducing junior unsecured loans (i.e., FinTech loans) into the small business lending market. Using French administrative data, we find that firms experience a 13% increase in bank credit after receiving a FinTech loan. We use propensity score matching procedures and a shift-share instrument to account for credit demand. The credit increase only occurs when FinTech borrowers invest in new assets, and Fintech borrowers are subsequently more likely to pledge collateral to banks. This suggests that firms use FinTech loans to acquire assets that they then pledge to banks, thereby increasing total borrowing capacity

Borrowers' use of cashless payments improves their access to capital from FinTech lenders and predicts a lower probability of default. These relationships are stronger for cashless technologies providing more precise information, and for outflows. Cashless payment usage complements other signals of borrower quality. We rationalize these empirical findings using a framework in which borrowers signal their lower likelihood of diverting cash flows through payment technology choice, and screening accuracy is further strengthened by informational complementarities. The informational synergy we uncover provides a rationale for the joint rise of cashless payments and FinTech lending, as well as for open banking

We study the nature and effects of cultural biases in choice under risk and uncertainty by comparing peer-to-peer loans the same individuals ( lenders ) make alone and after observing robo-advised suggestions. When unassisted, lenders are more likely to choose co-ethnic borrowers, facing 8% higher defaults and 7.3pp lower returns. Robo-advising does not affect diversification but reduces lending to high-risk co-ethnic borrowers. Lenders in locations with high inter-ethnic animus drive the results, even when borrowers reside elsewhere. Biased beliefs explain these results better than a conscious taste for discrimination: lenders rarely override robo-advised matches to ethnicities they discriminated against when unassisted.

We study gender differences in on-time loan payment responsiveness to collection mechanisms using randomized dunning text messages sent to 17,545 FinTech borrowers. Reminder text messages significantly reduce delinquency rates relative to a no-message control, with messages incorporating social or financial incentives proving more effective. Women are more responsive to social pressure, while men are more sensitive to financial incentives. These results are robust to observable control variables and matching methods. Channel analyses indicate that gender differences in responsiveness to the existence and size of the incentive explain the observed gender differences under the social incentive treatment. However, only gender differences in sensitivity to the existence of incentives explain the gender difference in financial incentive treatments. These findings inform practitioners and policymakers that some seemingly gender-neutral practices may create unintended gender disparities in financial markets

Fraud indicators within the Paycheck Protection Program (PPP), a major COVID relief program, are highly geographically concentrated. ZIP codes and counties with high rates of suspicious PPP loans are strongly socially connected, with evidence that fraud spreads spatially over time through social networks. Individuals in suspicious social media groups have higher rates of PPP fraud, and socially connected ZIP codes frequently use the same specific FinTech lenders, consistent with social networks influencing detailed loan decisions. Our findings suggest that more proactive data analysis is needed for fraud prevention, detection, and prosecution to prevent the social spread of fraudulent schemes

This study analyzes the evolution of climate-risk discourse in banking using 4,887 news articles (2008–2024) collected from ProQuest. We apply Natural Language Processing and add two novel layers: (i) an event-alignment analysis that links coverage dynamics to dated policy and supervisory milestones, and (ii) a discourse-network analysis connecting banks and regulators. We document a marked post-2020 shift, with ESG emerging as the dominant framing (7,860 mentions) alongside persistent geographic asymmetries (U.S.-led coverage) and uneven sectoral engagement (Risk Management highest salience; Fintech lowest). Sentiment skews positive (≈4,000 positive vs. ≈1,500 negative), and topic modeling identifies eight stable thematic clusters spanning operations, ratings, ESG assessment, disclosures, and market instruments. Event alignment shows media attention is typically anticipatory (median peak two months before an anchor), with COP26 producing a sustained level shift (+100% within a ±6-month window) and the Bank of England’s CBES results generating the largest single spike (210 articles), whereas some 2022 rule-making announcements (e.g., SEC climate-disclosure proposal) exhibit sharper but less durable attention. The discourse network centers on two regulatory hubs (the Federal Reserve and the ECB) with key banks (e.g., Citigroup, JPMorgan, UBS) bridging into supervisory narratives. Collectively, the findings show climate risk becoming embedded in core banking practice while revealing structural, regional, and functional asymmetries that matter for policy design and implementation. • Provides the first longitudinal NLP-based analysis of climate risk discourse in the banking sector • Reveals how climate risk integration in banking has evolved across regulatory, operational, and market dimensions • Identifies distinct thematic domains shaping climate risk narratives in banking over time • Shows that climate risk discourse is predominantly anticipatory around major policy and supervisory milestones • Maps the institutional structure of climate risk governance by linking banks and regulators within a discourse network