Modeling Binary Causal Effects of Related Repayments: Causal Inference Techniques
摘要
The long-tail customers in retail banking represent a vast demographic with a relatively low individual value. This segment, however, holds considerable potential and can be a significant asset. Nevertheless, leveraging this potential necessitates the use of technological advancements rather than relying solely on manual efforts. Enhancing long-tail customer engagement translates into increasing the reliance of the general populace on a banking institution, decreasing customer attrition, and bolstering assets under management (AUM) along with the number of monthly active users (MAU). The term “dual-card customers” refers to retail clients who possess both a credit card and a debit card. Typically, banks encourage valuable customers to link their credit and debit cards, allowing the system to automatically deduct repayment amounts from the debit card’s current account on the credit card’s due date. The business rationale for this practice is multifaceted: it requires customers to maintain sufficient funds in their debit card current accounts, thereby increasing deposits and the usage rate of debit cards. It also consolidates idle funds within the designated bank, simultaneously enhancing AUM and MAU while mitigating the risk of inadvertent credit card payment delays. Moreover, the linkage of credit and debit cards fosters greater customer loyalty. For customers, failing to link the cards necessitates either transferring repayment funds from another bank or using services like Alipay or WeChat, both of which involve additional transaction fees and a higher likelihood of incurring overdue interest due to missed payments, making this approach less advantageous.