The dataset was compiled from a Chinese online lending platform for the research project titled `Peer-To-Peer Lending Market in China: Pricing, Risk Management and Regulation,’ Economic and Social Research Council (ESRC) grant ES/P004741/1. The data were in public domain. The dataset includes about 7.5 million listings posted over the period between October 2010 and October 2018. The project follows General Data Protection Regulations and conforms to data confidentiality and anonymisation. Several variables that could result in the identification of borrowers or lenders have been deleted.The research project focuses on P2P loan applications to a major Chinese lender from its start-up in 2010 to December 2016 (some 800,000 in total), collecting all information provided by the borrowers in support of their applications. This includes both quantitative and qualitative data about their employment status, income, personal details, their credit histories and loan purposes. The data will enable us to examine how borrowers present themselves and the effect of this on lender behaviour - i.e. the conditions of loans offered. Using this data, the project examines: a) The factors that influence lenders' decisions - who to lend to, what conditions to impose. How do lenders discriminate between borrowers? b) How soft information about borrowers affects lenders' decisions. Does it improve their assessment of borrower risk? c) Whether there is an incentive for borrowers to misrepresent their income in order to improve their chances of obtaining a loan or the terms offered. The project will examine whether misreporting by borrowers affects lender decisions and loan outcomes. d) Whether lenders are affected by other lenders' decisions, and if so whether they are likely to operate rationally. e) How information about external events (e.g. publicity given to frauds on lenders) as exogenous shocks affects borrower and lender behaviours. f) How major regulation changes in 2015 have affected borrower and lender behaviours.
The data were collected by a Research Assistant using python web-scraping codes from http://www.renrendai.com. The downloaded HTML files were parsed and converted into csv files. Further, the data quality checks were applied to remove any potential problems. Finally, all information that could help to identify a person has been removed.
