PEER TO PEER LENDING, DEFAULT PREDICTION-EVIDENCE FROM LENDING CLUB

Sriharsha Reddy () and Krishna Gopalaraman
Additional contact information
Sriharsha Reddy: Associate Professor, Institute of Management Technology (IMT), Hyderabad, India
Krishna Gopalaraman: Managing Consultant, Enkeyed Consulting and Analytics, Hyderabad, India

Journal of Internet Banking and Commerce, 2016, vol. 21, issue 03, 01-19

Abstract: Purpose– The objective of this paper is to outline an approach towards a Classification Problem using R. The focus is on two problem statements as stated below: 1. To combine the data on loans issued and loans declined and build model that replicates Lending Club Algorithm closely 2. Using Lending Club’s published data on loans issued and its various attributes, build model that can accurately predict probability of delinquency. Design/methodology/approach– In order to build a model which replicates lending club algorithm closely various classification techniques such as Logistic Regression, Basic Classification Trees, Generalized Linear Model with Penalization, Ensemble of Decision Trees and Boosted Trees were used using R. Boosted Trees classification method is deployed to build model that can accurately predict probability of delinquency. Findings– Risk Score variable figures as the top of the variable importance list followed by length of employment as one of the more important variables in determining whether loans where eventually issued. Risk Score (at Origination) figures as the top of the variable importance list. This is followed by Amount Paid as a % of Loan Amount as one of the more important variables in determining whether loans would turn delinquent. The performance (accuracy) on training as well as test set is best given using the xgboost model at 99%. Practical implication– The paper includes implications for the borrowers to understand the factors influencing the decisions of issuance of loan and for the investors to understand the reasons for delinquency in peer to peer lending. Originality/value– This paper fulfills an identified need to build a model to predict probability of success in getting loans with identification of reasons for issuance of loans at Lending Club. Similarly, it also attempts to build a model to predict probability of delinquency and reasons contributing to delinquency to benefit investor’s community at Lending Club.

Keywords: Lending Club; Probability of Default; Peer To Peer Lending (search for similar items in EconPapers)
JEL-codes: A11 (search for similar items in EconPapers)
Date: 2016
References: Add references at CitEc
Citations:

Downloads: (external link)
https://www.icommercecentral.com/open-access/peer- ... g-club.php?aid=81766 Full text (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:ris:joibac:0198

Access Statistics for this article

Journal of Internet Banking and Commerce is currently edited by Vijaya Lakshmi, Nahum Goldmann and Dale Pinto

More articles in Journal of Internet Banking and Commerce
Bibliographic data for series maintained by Dale Pinto ().