Essays on the Peer-to-Peer Lending Business

Peer-to-peer (P2P) lending is a new way of originating loans where borrowers and lenders are directly matched via an online lending platform. From its inception in 2005, P2P platforms have been among the fastest growing segments in the financial services space. Transparency Market Research suggests that the opportunity in the global P2P market will be worth $897.85 billion by the year 2024, up from $26.16 billion in 2015. This work investigates the P2P lending market from both the platforms’ and investors’ perspectives. First, since P2P lending is a relatively new market, the fundamentals of the market keep changing over time. Therefore, treating all historical data equally may lead to a seriously misspecified model for the current market. This problem is referred to as concept drift in statistics and machine learning. Using a rich micro-level dataset from Prosper Marketplace, which is one of the largest online P2P lending platforms in the U.S., we show evidence that borrower and lender behavior is changing over time. Second, by leveraging certain machine learning techniques (naive Bayes classifier, ensemble modeling, and hidden Markov model) and structural modeling, we find evidence that Prosper Marketplace recognizes the concept drift problem when developing their model to assess borrowers’ risk and set interest rates. More specifically, we find that Prosper assigns different weights to data from different economic environments to train its model. In the counterfactual, we demonstrate that we can improve Prosper’s revenue by changing the way Prosper uses data. Third, we show that investors on Prosper Marketplace are more likely to invest in riskier loans during the end-of-week or end-of-month periods. The end-of-period effect is in line with the mental accounting theory. The end of a mental accounting period propels people to feel less connected with their present selves. This feeling of disconnection from their present selves makes people more likely to make riskier decisions since, even if the outcomes from those riskier investments turn out to be negative, they will soon enter a new mental accounting period and can easily blame their imperfect past selves for the wrong decisions.Ph.D