Research interest: Banking, Fintech, Corporate Finance, Household Finance, Financial Innovation, Applied Artificial Intelligence
Abstract: As artificial intelligence (AI) plays an increasingly pivotal role in financial decision-making, a key question arises: does AI complement or substitute human judgment in credit markets? This study examines the performance of two Fintech lending strategies, Human+AI and Algorithm+AI, using instrumental variable (IV) and difference-in-differences (DiD) methodologies. The results demonstrate that combining human judgment with AI significantly reduces loan charge-offs, especially during economic downturns, with human input proving more valuable under high information asymmetry. Loan screening intensity emerges as a critical behavioral mechanism through which human judgment improves outcomes. However, the platform’s fixed-interval loan listing policy undermines human performance due to heightened competition and time constraints. These findings underscore the complementary role of human judgment in enhancing AI-driven credit decisions, offering valuable insights for optimizing hybrid decision-making systems in Fintech lending.
Abstract: This paper examines the impact of AI adoption on bank loan performance, focusing on non-performing loans (NPLs). Using a human capital-based measure of AI implementation and an instrumental variable approach, we find that AI adoption reduces NPL ratios. The effect is stronger in banks with greater human interaction, suggesting complementarity between AI and human judgment. Gains are limited for banks with complex, commercially oriented portfolios, reflecting AI’s limitations in handling non-standardized data. Banks with lower cost efficiency also face challenges in realizing AI’s potential. Our findings offer insights on how institutional characteristics influence AI’s effectiveness in supporting financial stability.
Abstract: Credit supply plays a crucial role in the economy, especially during periods of economic and financial crises. As emerging and rapidly growing fintech lenders, marketplace lenders now hold a significant share of the credit supply in the economy, This paper aims to examine marketplace lenders’ credit supply reliability relative to traditional banks amidst the COVID-19 pandemic. Employing the Difference-in-Differences (DID) framework alongside the instrumental variable methodology, our findings reveal that marketplace lenders decrease their credit supply by 0.55% more than large banks and 0.33% more than small banks in the post-COVID period. Furthermore, we explore the fundamental reason behind the greater vulnerability of marketplace lenders’ credit supply compared to banks. Our analysis indicates that the reliance on individual investors makes the funding sources of marketplace lenders less stable relative to those of banks. Additionally, we find a statistically significant indication that individual investors reduce their credit supply by 0.05% and 0.06% compared to passive and active institutional investors, respectively, in response to the COVID-19 crisis. Overall, the findings offer valuable insights for policymakers in designing strategies to strengthen credit supply stability and bolster the resilience of marketplace lenders in times of economic crises.