The Research Of The Network Characteristics Of P2P Lending And The Effect On Default Risk

In recent years,with the rapid development of the P2 P online lending industry,the hidden credit risks are also increasing.Various runs occur frequently,which sounds the alarm for the development of the industry.Although the industry tends to be stable under the rectification and regulation of the government and regulatory agencies,the healthy,orderly and sustainable development of the P2 P lending industry still needs long-term regulatory mechanism.Among them,the identification of credit risk is particularly important.In P2 P lending,information can be classified as hard information and soft information,which are also two main aspects of scholars’ research in identifying the credit risk of P2 P lending.Unlike the information directly available above,this paper explores the default risk through the characteristics of lending behavior of participants from the perspective of the network.This paper takes RenRenDai platform as a research sample and constructs a network using the loan data and bidding behavior data in the platform.On the basis of getting the whole network,the paper analyses the node’s topology structure and its practical significance in the P2 P lending,then uses the characteristics of the nodes in the network to establish the default risk model by using Logit model,which describes and models the risk of P2 P lending from a new perspective.The study finds that a small number of participants carry out a large amount of lending activities on the platform,but most participants have a small amount of investment.At the same time,in the capital flow network,the edge weight which represents the amount of investment per investment,obeys the power law distribution.From the perspective of borrowing,the in-weight of participants obeys power law distribution.However,from the perspective of investors,the out-weight of participants does not obey the power law distribution,but concentrates on the cumulative investment of smaller amounts.In the lending network,the degree distribution of P2 P lending network obeys power law distribution,and there is a weak hierarchical structure inside.The power exponent of the network evolves with time and gradually approaches the theoretical deduced value.In general,new borrowers are higher than investors,and the two types of participants are in balance.Then we use Logit model to study the effect of network characteristics of participants on default probability in the P2 P lending.Empirical research shows that the third kind of participants’ trading behavior has dual information value and can reveal the default risk of the loan.The third kind of participants can use information advantage to identify the lower risk loans,and the third kind of participants themselves will publish lower risk loans.There exists the behavior that both borrowers and lenders invest in the same loan in the P2 P lending.This behavior will not reveal the lower risk of the investment target,and may even indicate a higher default probability.The borrower’s out-clustering coefficient has a positive impact on default probability only in the sub-group issued by the borrower with AA,A and D rating or in the sub-group with the amount between 3000 and 5000.The larger the in-degree,the more investors that the borrower can attract,and the lower the default risk of the loan.To sum up,this paper describes the lending behavior in the P2 P lending platform from a new perspective,and establishes a default risk model,which is constructive for investors,platforms and the regulators.