How can we learn from a borrower’s online behaviors? The signal effect of a borrower’s platform involvement on its credit risk

Abstract

Internet consumer credit services are defined as the provision of consumer credit services through internet platforms. While these services have benefited the public, they also present numerous credit management challenges, particularly in credit scoring. As a result, platforms are seeking ways to develop a more accurate and effective credit scoring system for internet consumer credit services. Using a credit application dataset from a large e-commerce platform in China, our study addresses this issue. Drawing on signaling theory, we explore how the signal of a borrower’s platform involvement intensity reflects its credit risk. We discover that a borrower with high involvement intensity pose lower credit risk, whereas that with low involvement intensity exhibit the opposite trend. Furthermore, our research reveals that a borrower's personal characteristics influence the signal effect of its involvement intensity. This is possibly due to platforms adjusting their perception of the signal effect based on prior expectations of borrower groups. Specifically, the signal effect of a borrower’s involvement intensity is stronger when it has higher involvement stability and a better credit history, and weaker when it comes from an economically developed region. Our study not only enhances understanding of credit scoring but also offers practical recommendations for developing credit assessment strategies for platforms.