Bankruptcy Prediction and Risk Contagion Analysis of China’s P2P Lending Platforms: A Theory-Driven Design and Macro Level Analysis

Description

In recent years, online peer-to-peer (P2P) lending platform, a new channel of financing through which lenders are matched with borrowers, has become increasingly prevalent. Because of the low overhead and efficient electronic distribution of funds, P2P lendingcan lead to higher returns for both borrowers and lenders. However, there could also be significant risks associated with such practice involving borrower defaulting, fraud, platform insolvency and bankruptcy. According to statistics from wdzj.com, the number of problematic P2P platforms totalled 2,856 in August 2019. Ezubao, a P2P lending platform for example, allegedly attracted about 50 billion yuan ($7.6 billion) in funds from 900,000 investors and was shut down in 2016 due to fraudulent activities. These problematic P2P platforms, if undetected, could cause immeasurable damage to the investor confidence and the overall economy.In this research, we aim to examine the risks associated with P2P lending at a macro level. Specifically, we seek to (1) identify factors that help predict the failure and bankruptcy risk of a P2P lending platform, (2) evaluate how does the risk of one P2P lending platform propagate to other platforms, and (3) analyze how the government regulatory policy influence the development of the P2P financial industry. In order to answer the above questions, we propose a novel analytic framework to extract structural characteristics from the P2P network, which will be developed using both external news information and internal registry associations. We will then train machine learning algorithms to predict the probability of platform bankruptcy using data from multiple sources. We will further use probability-generating function to analyze the risk contagion mechanism of platform bankruptcy based on the P2P network. Finally, we will conduct rigorous empirical analysis to study the effect of three influential Chinese policy events on the risk-taking activity (operational risk, credit risk, and users’ perceived risk) of the P2P lending platforms. Findings from this research project will have significant theoretical as well as practical implications. From a theoretical perspective, we believe that an effective and efficientalgorithm to evaluate risk will mitigate the issue of adverse selection commonly observed in online marketplaces due to information asymmetry. From a practical perspective, this study can help government regulators and investors to achieve their investmentobjectives goals while protecting the public interest and the overall health of the financial industry. We also expect that this project will provide additional insights for the development of financial technology industry in Hong Kong.