Herding behaviour in P2P lending markets

We explore individual lender behaviour on Renrendai.com, a leading Chinese peer-to-peer (P2P) crowdlending platform. Using a sample of roughly 5 million investor-loan-hour observations and applying a high-dimension fixed effect estimator, we establish evidence of herding behaviour: the investors in our sample tend to prefer assets that had attracted strong interest in previous periods. The herding behaviour relates to both the experience of the investor and the length of time of each investment session. The results show that herding happens mostly in the first or final hour of long sessions. Herding behaviour is further confirmed by estimates at the listing-hour data.