Beyond Account Ownership: Digital Payments and Formal Savings as Robust Correlates of Financial Inclusion, 2011–2024

Technical Summary: Beyond Account Ownership: Digital Payments and Formal Savings as Robust Correlates of Financial Inclusion, 2011–2024

Problem Statement
While financial inclusion is a central pillar of global development policy, cross-country evidence regarding the drivers of account ownership remains fragmented across survey waves and shifting indicator definitions. A critical gap exists in distinguishing between mere access (account ownership) and active usage (savings, payments, borrowing). Prior literature often relies on single-wave analyses or treats account ownership as a static outcome, failing to isolate the specific behavioral dimensions that robustly correlate with inclusion when controlling for persistent cross-country heterogeneity. Furthermore, the extent to which technology-driven access translates into sustained usage behaviors, and how these relationships evolve over time, requires a longitudinal perspective that accounts for both country-specific institutional factors and global time trends.

Methodology
This study utilizes a country-year panel dataset derived from six waves (2011, 2014, 2017, 2021, 2022, and 2024) of the World Bank Global Findex Database. The analysis covers up to 162 economies, with the core regression sample restricted to 294 complete observations across 162 countries for the years 2014, 2017, 2021, and 2024, where data on all covariates were available.

The empirical strategy employs a two-way fixed-effects panel regression model. The dependent variable is the share of adults (aged 15+) with an account at a financial institution or mobile-money provider. The independent variables are four usage-dimension covariates: digital payment use, mobile money account ownership, formal savings, and formal borrowing.

The estimation proceeds in three nested specifications:

1. Pooled Ordinary Least Squares (OLS) with heteroskedasticity-robust standard errors.
2. OLS augmented with survey-wave fixed effects and country-clustered standard errors.
3. A two-way fixed-effects specification adding country fixed effects to absorb all time-invariant country characteristics (e.g., institutional quality, financial sector structure).

As a robustness check, a median (quantile) regression with wave fixed effects is estimated on the same complete-case sample to assess the influence of outlier countries. The study explicitly treats the regression findings as descriptive correlates rather than causal effects, acknowledging the potential for reverse causality and unobserved within-country developments.

Key Results

• Trends in Ownership: Global average account ownership rose from 50.6% in 2011 to 78.7% in 2024. The fastest absolute gains occurred in South Asia (35.2% to 83.9%) and Sub-Saharan Africa (23.3% to 58.2%). However, the Middle East, North Africa, Afghanistan, and Pakistan aggregate showed non-monotonic progress, including a decline between 2017 and 2021.
• Inequality Gaps: The global gender gap in account ownership narrowed only modestly, moving from 14.1 percentage points in 2011 to 11.7 in 2024. The within-country income gap (richest 60% vs. poorest 40%) also narrowed slightly from 14.1 to 11.7 points but remains substantial. The urban–rural gap was observable only in the 2024 wave, standing at 6.7 percentage points.
• Determinants of Ownership: In the two-way fixed-effects model (controlling for country and year), digital payment use ($\beta = 0.552, p < 0.01$) and formal savings behavior ($\beta = 0.418, p < 0.01$) emerged as robust, statistically significant correlates of account ownership.
• Non-Robust Correlates: Conversely, the coefficients for mobile money account ownership and formal borrowing shrank toward zero and lost statistical significance once country fixed effects were absorbed. This suggests that the raw cross-sectional association between mobile money/credit and account ownership largely reflects stable, time-invariant country-level factors (such as overall financial sector development) rather than a dynamic within-country relationship over time.
• Model Fit: The two-way fixed-effects model yielded an $R^2$ of 0.972, a value the authors attribute to the inclusion of a large number of country fixed effects relative to the sample size, cautioning against interpreting this as strong out-of-sample predictive power.

Significance and Claims
The study claims originality in tracing these relationships consistently across the full 2011–2024 Findex series, rather than relying on single-wave snapshots. Its primary significance lies in isolating within-country correlates net of persistent cross-country differences—a distinction largely absent from prior work.

The findings challenge the assumption that mobile money or credit access alone drives account ownership dynamics over time. Instead, the results suggest that digital payment infrastructure and the formalization of savings are the most robust behavioral drivers associated with account ownership. The paper posits that policy efforts focused solely on expanding account ownership or credit access may be insufficient; rather, investments in digital payment ecosystems and mechanisms that encourage formal savings are likely more effective in closing remaining inclusion gaps.

The authors maintain a modest stance, emphasizing that their results describe associations rather than causal mechanisms. They note that the persistence of the gender gap despite broad access growth implies that structural and behavioral barriers require targeted interventions beyond generic access expansion. Similarly, the weak link between formal borrowing and account ownership in the fixed-effects model suggests that credit market development requires distinct policy instruments (e.g., credit information infrastructure) separate from account ownership campaigns.