Causal Impact of Cashback Campaigns on Post-Marketing Default Behavior in Consumer Lending

Abstract

This study examines how cashback marketing campaigns influence subsequent credit-risk outcomes. Using an AB testing framework implemented on a dataset of 640,000 loan applicants, the treatment group received a one-time cashback incentive, while the control group did not. Propensity-score weighting and double-machine-learning estimators were applied to isolate causal effects. Results show that campaign exposure increases short-term activation rates by 18.4% but also raises 90-day delinquency risk by 6.7 percentage points. Heterogeneity analysis reveals significantly larger default spillovers among new-to-bank customers (increase of 11.3 pp). A cost-risk decomposition suggests that while the campaign boosts revenue in the first month, cumulative losses surpass gains after 5.6 months. The findings highlight the risk implications of customer-acquisition incentives.