A Decomposition Approach to Analyze the Development Ranking Score of Online Peer to Peer (P2P) Lending Platforms

基於分解法的針對P2P 借貸平臺發展評級指數之研究

Student thesis: Doctoral Thesis

View graph of relations


Related Research Unit(s)


Awarding Institution
Award date28 Aug 2017



Online Peer-to-Peer (P2P) lending industry has grown rapidly in the past few years in China and this has significantly changed the way people lend and borrow money. It has also promoted marketization of interest rates and forced banks to address the needs of previously disregarded customers. However, without proper regulation, the wild growth of the P2P lending industry might hinder its long-term growth. Unlike banks, which are under more stringent regulations and with better risk management practices, P2P lending industry has suffered high default rates, especially after the adjustment of the securities markets in the second half of 2015. Because of such widespread defaults, investors have suffered significant losses. However, two questions are always asked: Why weren’t any warnings provided by P2P lending platforms to help investors avoid such losses? Was the development ranking score of P2P lending platforms useful and reliable in predicting such defaults?

In order to search for an answer and find a solution to improve the current situation, I conducted an in-depth analysis of the development ranking scores (the Development Score) created and used by the most comprehensive and influential P2P lending portal - Wangdaizhijia (WDZJ). Before my study, there was no research or study specifically addressing such issue.

My study aims at creating a better understanding of implications of the Development Score and to answer the question: What are the implications of the Development Score? Does the Development Score contain any hard information - the common economic factors or soft information - WDZJ’s own judgement on the values of a set of factors? What are the relationships between the Development Score and both the hard and soft information? Is the Development Score useful and reliable in predicting default of P2P lending platforms? WDZJ is chosen as the subject for this study for two reasons:
1. It provides the most comprehensive data since 2013; and
2. It is currently one of the most important and frequently visited P2P lending portals in China.

To identify the relationships between the Development Score and both Hard and Soft information, a Decomposition Model was adopted in this study. Firstly, I partition the factors that affect the Development Score into Hard information and Soft information. Secondly, I define the relationship between the Development Score and the Hard information (common economic factors), and the relationship between the Development Score and Soft information (self judgement of WDZJ). After these relationships were identified, I further conducted an Event Study to identify whether key factors were available to predict the possibility of default for P2P lending platforms.

By using decomposition approach to conduct the Event Study, some key factors were identified. The final findings confirmed positive correlations between the Development Score and both Hard information and Soft information. In addition, some key variables were identified and recommended to be closely monitored as they can be used to predict the possibility of P2P lending platform defaults.

Based on the results of this study, recommendations can be formulated for WDZJ to improve the composition of the ranking score. For example, the magnitudes of changes of variables can be included to alert investors of possible defaults of P2P platforms. Above all, this study provides a new perspective with better understanding of the development ranking score information for P2P lending industry stakeholders including regulators, investors, borrowers and other interested parties. Future studies can be extended to test the sensitivity and reliability of the development ranking model when more data sets are available.

    Research areas

  • P2P lending, Wangdaizhijia, development ranking score, decomposition method