Publication Date
Spring 4-16-2025
Presentation Length
Poster/Gallery presentation
College
Jack C. Massey College of Business
Department
Economics and Finance
Student Level
Undergraduate
SPARK Category
Research
Faculty Advisor
Rudolph Bedeley
WELL Core Type
Financial Wellness
SPARK Session
We would like to be in the 2:00-3:00 session.
Presentation Type
Poster
Summary
Peer-to-peer (P2P) lending has transformed consumer credit markets by providing an alternative to traditional banking institutions. LendingClub, a pioneer in this space, facilitates lending between individual investors and borrowers through a data-driven risk assessment model. Our research aims to enhance loan default prediction by developing a more precise classification model based on LendingClub’s historical loan data, ultimately improving risk assessment for investors.
Utilizing a dataset of approximately 650,000 loans from 2007 to 2015, we construct a predictive model to classify loan default risk. Our approach focuses on key financial indicators, including interest rates, borrower grades, debt-to-income ratio, and delinquency history, while introducing a novel binary variable, is_bad_loan, which categorizes loans based on repayment outcomes. By analyzing these features, we assess the effectiveness of LendingClub’s internal grading system and explore alternative risk assessment features beyond traditional credit scoring methods we see traditionally in the United States.
Our findings provide deeper insights into the factors influencing loan defaults, offering comparable results of different predictive models. Ultimately, providing a more nuanced understanding of lender risk in P2P lending as they pose as non-traditional investors in this space.
Recommended Citation
Gilani, Shumaila and Kliimand, Viktoria Kleer, "Data-Driven Default Prediction: Insights from Lending Club" (2025). SPARK Symposium Presentations. 280.
https://repository.belmont.edu/spark_presentations/280
Included in
Business Analytics Commons, Finance and Financial Management Commons, Management Information Systems Commons
