EBA欧洲银行-Paper-Session-3.-Olena-Havrylchyk_29页_2mb

Summary of "What drives the expansion of the peer-to-peer lending?"

Core Content

This paper investigates the main drivers of the expansion of peer-to-peer (P2P) lending platforms in the United States, focusing on Prosper and Lending Club. The study explores three hypotheses: competition-related, crisis-related, and internet-related, to understand the factors influencing the geographic spread of P2P lending at the county level.

Main Hypotheses

1. Competition-related Hypothesis

   • Suggests that P2P lending expands in areas with lower bank concentration and weaker brand loyalty, as these conditions indicate less competition and easier access to alternative financial services.
   • Market structure variables, such as branch density and concentration indices, are used to test this hypothesis.

2. Crisis-related Hypothesis

   • Proposes that the expansion of P2P lending is linked to the financial crisis and bank failures.
   • Measures include the share of deposits held by failed banks and solvency ratios derived from bank capital data.
   • If confirmed, these variables would show a positive impact from crisis-related factors.

3. Internet-related Hypothesis

   • Argues that the spread of the Internet and digital technologies facilitates P2P lending growth.
   • Variables include internet access rates (broadband, mobile wireless, upload speed), and openness to innovation (patents per capita).
   • These variables are expected to have a positive correlation with P2P lending expansion.

Key Findings

• The competition-related hypothesis is supported by the data, as P2P lending expanded more in areas with lower branch density and lower bank concentration, interpreted as weaker brand loyalty.
• Spatial correlation is an important factor in the diffusion of P2P lending, as counties near major financial hubs (e.g., California, New York, Florida) show higher P2P lending per capita.
• The internet-related hypothesis is also supported by the positive correlation between internet access rates and P2P lending growth.
• Socio-economic and demographic characteristics (e.g., education, population density, poverty levels, race) are found to influence the expansion of P2P lending.
• Racial minorities may be more likely to use P2P lending due to the removal of racial identification features on platforms, reducing discrimination.

Methodology

• The study uses county-level data from Prosper and Lending Club, covering the period from 2006 to 2013.
• It introduces a spatial autoregressive model (SARAR) to account for spatial correlation both in the dependent variable and error terms.
• The model includes:
  • Market structure variables: branch density, bank concentration indices (HHI and C3).
  • Crisis variables: share of deposits from failed banks, solvency ratios.
  • Innovation and internet variables: patents per capita, internet access rates.
  • Socio-economic and demographic variables: age, education, poverty, race, population density.
• The spatial lag is incorporated to capture the influence of neighboring counties on P2P lending adoption.

Data and Variables

• Loan data from Prosper and Lending Club is used to measure P2P lending diffusion.
• County-level variables are constructed based on:
  • Loan volume and number per capita.
  • Bank branch data and FDIC records.
  • Internet access data from the NTIA’s State Broadband Initiative.
  • Socio-economic data from the U.S. Census Bureau.
• Data limitations:
  • Missing city names in later years led to data loss.
  • Some counties had no P2P loans, which were excluded from the analysis.
  • State-level dummies are used to account for differences in regulation and other state-specific factors.

Conclusion

• The study finds that competition and internet access are the main drivers of P2P lending expansion.
• Spatial effects are also significant, as P2P lending tends to spread from more developed areas to neighboring regions.
• Socio-economic factors play a role, with education and population density being positively correlated with P2P lending.
• The removal of racial identification features on P2P platforms may have increased access for racial minorities.
• The findings suggest that P2P lending could be a disruptive technology in the financial sector, especially in markets with weak traditional banking structures.

Summary Statistics and Variables

• The study includes 3,059 counties and county equivalents, with 313 counties having no P2P loans.
• Table 2 provides a detailed list of variables used in the analysis.
• Table 3 includes summary statistics for all variables, highlighting their distribution and significance.

Implications

• The results contribute to the emerging literature on P2P lending by analyzing its geographic expansion.
• They suggest that FinTech innovations can complement traditional banking, especially in areas where competition is weak and trust in technology is high.
• The spatial diffusion of P2P lending highlights the importance of network effects and local adoption dynamics in financial technology spread.