From literacy to action: econometric and machine learning evidence on digital financial behaviour in India

Purpose This study examines the impact of digital financial literacy (DFL) on digital financial practices (DFP) in India by combining causal inference and machine learning approaches. It investigates whether higher DFL translates into greater engagement with digital financial services and whether this effect varies by gender. Design/methodology/approach Using primary survey data from 1,215 adults, the study applies covariate-adjusted ordinary least squares (OLS), propensity score matching (PSM) and overlap weighting to estimate the effect of DFL on DFP. Gender heterogeneity is analysed through interaction models and stratified matching. A Random Forest model is used to assess predictive importance. Findings Results consistently show that higher DFL is associated with significantly greater DFP across all estimators, with effect sizes ranging from 0.86 to 1.41 (p < 0.001). The effect is stronger among women, indicating a potential equalising role of literacy in digital finance adoption. Machine learning results further identify DFL as the most important predictor of DFP, explaining a substantial share of variation in outcomes. Research limitations/implications The cross-sectional design limits causal interpretation, and findings are specific to the Indian context. Future research can extend this work using longitudinal data and comparative frameworks. Practical implications The findings highlight the need to shift from infrastructure-led to capability-driven financial inclusion strategies. Targeted DFL interventions, particularly for women and low-income groups, can improve adoption, trust and responsible use of digital financial services. Originality/value This study integrates causal inference and machine learning to provide robust evidence on the role of DFL in shaping digital financial behaviour, offering both methodological and policy-relevant insights for advancing digital financial inclusion.