Can causal machine learning reveal individual bid responses of bank customers? - A study on mortgage loan applications in Belgium

Christopher Bockel-Rickermann, Sam Verboven, Tim Verdonck, Wouter Verbeke

Abstract: Highlights•The study highlights the role of causal inference in loan pricing.•We frame bid response modeling within the Neyman–Rubin potential outcomes framework.•We analyze factors leading to observed cofounding in loan pricing data.•We assess the impact of observed confounding on various machine learning methods.•Real-life mortgage loan data from a Belgian bank is used for empirical evaluation.