Logistic Regression Model for Predicting Buy Now Pay Later Default Risk Among Generation Z Users

Sari

consumption patterns among Generation Z, a demographic characterized by high digital adoption but comparatively limited financial experience. This study develops a binary logistic regression model to predict BNPL default risk among Generation Z users, incorporating financial literacy, financial self-control, impulsive buying tendency, income level, digital transaction frequency, and peer influence as predictors. Using an illustrative dataset structured to mirror a survey of 300 Generation Z BNPL users (aged 17–27). The model was evaluated using odds ratios, the Hosmer Lemeshow goodness-of-fit test, Nagelkerke R², classification accuracy, and the ROC AUC. In the illustrative run, financial literacy and financial self-control were associated with lower odds of default, while impulsive buying tendency, transaction frequency, and peer influence were associated with higher odds of default, consistent with the hypothesized directions; income level was not statistically significant. The model demonstrated adequate discriminative ability (illustrative AUC = 0.72) and good calibration (Hosmer-Lemeshow p > 0.05). These findings, once confirmed with primary survey data, are expected to offer practical implications for credit risk management, responsible lending policy, and product design in fintech lending platforms.
Keywords: buy now pay later; logistic regression; default risk; generation Z; fintech; credit risk management; machine learning.