Essays on an online marketplace with asymmetric information : empirical examination on herding and leaderboard effects in P2P lending

Abstract

Peer-to-peer (P2P) lending allows individual borrowers to raise funding from multiple individual lenders. P2P lending has recently become a popular IS research field due to information asymmetry issues that exist and result in investors having a limited amount of information with which to distinguish between good and bad borrowers. The intermediary, in general, publicly provides the funding status for listings such as the percent funded in real time, thus aiming to help the decisions of subsequent investors. This is likely to lead to peer influence. Irrational herding happens when lenders mimic the choices of others by referring to their decisions and following well-funded listings. Whereas, rational herding occurs when lenders make decisions based on observational learning. We use a unique panel data set that tracks the funding dynamics of borrower listings on Popfunding.com, the largest P2P lending site in Korea. We implement the econometric model for panel data in order to analyze if herding influences lending decisions on Popfunding.com. In addition, if so, we have developed a model to identify whether such herding is irrational or rational. We focus on the bid amount and the number of bids as the measure of how a request attracts investors. We find evidence of herding among lenders, which echoes the results of previous research on P2P lending. However, our results show that the existence of rational herding is ambiguous, a finding different from the results of Zhang and Liu (2012). The moderating effects of the total amount on herding are identified significant while the effects of the interest rate are not consistent. Most presumably, the investors in Popfunding.com have limited information regarding the borrowers. In this line of investor behavior, the information provided by the intermediary is very crucial in that investors may make their decisions easily based on the impression received from the basic information provided on the website. The role of information provided by the intermediary, specifically the leaderboard, grows in importance in this setting. In this P2P lending context, the leaderboard, where popular loan requests are displayed on the websites front page, provides information for lenders to use when evaluating the requests. This is similar to how many e-commerce websites display products based on their popularity, in other words, ranking items according to previous sales records. Information related to popularity such as a leaderboard will serve as an indicator of the choices previously made by people who decided on a particular purchase. We empirically examine the effects of leaderboard information regarding the most popular existing loan requests in the perspective of both the likelihood of being financed and loan performance. Our results show that the leaderboard information works ex ante in attracting additional bids to get loan requests successfully financed. However, it does not work ex post in improving the performance, thus there is less potential for default.