SSRN-无抵押金融科技贷款对抵押贷款市场的溢出效应（英）_68页_6mb

Summary

The paper examines the impact of unsecured FinTech lending on the mortgage market outcomes in the US using LendingClub loan data and HMDA mortgage data. It provides causal evidence that an increase in unsecured FinTech lending leads to positive spillover effects in the mortgage market. This effect is attributed to the alleviation of information frictions by FinTech lenders, which help improve traditional credit metrics such as repayment history and debt-to-income ratios. Using a quasi-exogenous variation from LendingClub's 2015 partnership with the BancAlliance network, the study shows that FinTech lending increases mortgage applications and loan originations, with a multiplier effect. Key findings include a positive impact on new home purchase applications and loans, and a significant increase in mortgage access for middle-income borrowers. Delinquency rates did not rise due to the increased mortgage lending, indicating effective creditworthiness screening by FinTech platforms. The results highlight the role of FinTech in improving credit market efficiency, particularly for borrowers with information frictions.

Key Findings

1. Spillover Effects: Unsecured FinTech lending positively affects mortgage market activity due to improved credit metrics from debt consolidation.
2. Causal Identification: The BancAlliance partnership provides quasi-exogenous variation, enabling a credible DID/DDIV identification strategy.
3. Marginal Borrowers: Middle-income borrowers show greater responsiveness to FinTech lending, likely due to higher information frictions.
4. New Home Purchases: Activities with higher information asymmetry (new home purchases) are more sensitive to FinTech lending.
5. Pricing Effects: FinTech lending leads to a shift in the credit supply curve, resulting in lower mortgage rates.
6. Robustness: Findings hold even when excluding FinTech mortgage lenders or using alternative mortgage datasets.

Methodology

• Data: LendingClub loan data and HMDA/Fannie Mae-Freddie Mac mortgage data aggregated at the zip3 level.
• Identification: DID with BancAlliance partnership as an instrument, addressing endogeneity concerns.
• Mechanism tests: Triple DID and subgroup analyses to explore heterogeneity and information asymmetry.