Uncovering Unobserved Heterogeneity in Digital Financial Inclusion

Abstract

This study investigates how digital financial literacy (DFL) influences financial inclusion (FI) among rural Tanzanian women while addressing the critical, yet underexplored, issue of unobserved heterogeneity in behavioral finance research. It examines how individual-level differences is reflected through latent subgroups that affect the link between DFL and FI. The model includes financial confidence and financial attitude as mediators, and mobile network quality as a contextual moderator, offering an analysis of how behavioral and infrastructural enablers interact across different segments of rural women. Data were collected from 301 rural women in Tanzania’s Mbeya, Dodoma, and Kigoma regions through a cross-sectional survey using purposive sampling. The study employed Partial Least Squares Structural Equation Modeling (PLS-SEM) via SmartPLS 4.0, followed by Finite Mixture PLS (FIMIX-PLS) segmentation. This approach enabled the identification of latent segments within the population that differ significantly in their structural pathways, thus revealing hidden patterns that conventional models often obscure. While the pooled analysis confirms that DFL has a direct and positive impact on FI with financial confidence and attitude acting as significant mediators and mobile network quality as a moderator. The FIMIX-PLS uncovered three distinct behavioral segments. Segment 1 exhibited strong DFL–FI links with effective mediation and moderation; Segment 2 relied primarily on financial attitude despite moderate DFL; and Segment 3 showed constrained outcomes due to infrastructural limitations, despite adequate literacy. These findings highlight the presence of meaningful subpopulation differences in digital finance behavior. The findings emphasize the need for segment-specific interventions that go beyond one-size-fits-all models. Policymakers and practitioners should tailor financial education, confidence-building programs, and digital infrastructure investment according to the behavioral profiles of different subgroups