Heterogeneous effects of Unified Payment Interface readiness on digital payment adoption in India: Evidence from causal forest models

Type of the article: Research ArticleAbstractThe rapid expansion of the digital payment system has transformed financial inclusion in India; however, it has not eliminated persistent socioeconomic and spatial disparities in digital financial participation. The Unified Payments Interface (UPI) adoption has been marked by highly uneven adoption across socioeconomic groups, which limits its direct benefits. Using the World Bank’s Global Findex 2025 microdata, the machine learning-based causal forest model analyzes the heterogeneous effects of UPI readiness on individual-level digital payment adoption. The results show that UPI readiness has a positive average treatment effect of 0.161 on the digital payment index (0-1 scale). Substantial heterogeneity persists across urban–rural location, income, education, and gender, confirming entrenched digital divides. The most pronounced heterogeneity emerges along an inclusion-exclusion spectrum: individuals facing the greatest affordability and usability constraints experience treatment effects up to 3.7 times larger than the most digitally included groups. The findings point to the need to develop targeted digital inclusion policies for socially vulnerable population groups, aimed at strengthening their financial resilience and expanding their digital engagement. The study also provides novel empirical evidence on the effective use of machine learning methods to inform public finance policies aimed at fostering inclusive digital ecosystems.