Analyzing financial inclusion with explainable machine learning: Evidence from an emerging economy

This study aims to investigate the importance of factors that influence financial inclusion in emerging markets, using China as a case study. To accomplish this objective, the study collected datasets from China, encompassing macroeconomic and microeconomic factors spanning from 2014 to 2020. The authors employed various machine learning models, giving particular attention to the XGBoost model for SHapley Additive exPlanations (SHAP) feature importance explanation. The findings of the study reveal that four primary factors hold greater importance in achieving financial inclusion: urban-rural status, household income, internet coverage, and financial development. Urbanization, higher income, increased internet coverage, and enhanced financial development are likely to facilitate financial inclusion. Furthermore, financial exclusion groups are more likely to be affected by the features above. Lastly, the study identifies observed interaction effects between urbanization and other factors. Heterogeneity analyses underscore the pronounced urban-rural divide in financial inclusion and reveal region-specific vulnerabilities in Southwest China. These findings can be utilized to improve financial inclusion in emerging markets, enabling cost savings through the identification of key factors.