Financial inclusion, education, and employment: empirical evidence from 101 countries

The paper investigates whether and how financial inclusion promotes employment and whether the level of educational development moderates this relationship. It situates the problem within global gaps in access to finance, persistent gender disparities in financial usage, and uneven development of inclusive finance across countries, especially with the rise of digital financial services. Against the backdrop of economic shocks (e.g., COVID‑19) and the need for resilient employment, the study asks: (1) Does financial inclusion raise employment? (2) Does education directly enhance employment? (3) Does education amplify financial inclusion’s positive effect on employment? The authors posit three hypotheses: H1, financial inclusion is positively associated with employment; H2, the level of educational development is positively associated with employment; and H3, education positively moderates the effect of financial inclusion on employment.

Prior work links financial inclusion to employment through improved access to savings, credit, and digital services, which can directly and indirectly expand jobs and optimize employment structure. Studies highlight that savings and formal accounts enhance household financial flexibility; inclusive finance supports SMEs, entrepreneurship, and can reduce unemployment rates, with stronger effects often in higher‑income settings. Digitalisation can improve equity in employment opportunities. Separately, education influences employment via human capital formation, skills matching, and reduced unemployment risk; although classic theories suggest wage adjustments can offset productivity gains in the long run, many studies document positive effects of education investment on employment and earnings, particularly where public sector demand is high. Gaps identified include limited comprehensive measures of financial inclusion that combine traditional and digital dimensions, few analyses of education as a moderator in the finance–employment link, and limited heterogeneity analyses using gender as the grouping basis for external employment effects of financial inclusion.

Data: Panel data for 101 countries across 2011, 2014, 2017, and 2021 from World Bank datasets, including Global Findex (Financial Inclusion), Financial Access Survey, Doing Business, Education Statistics, and related sources. Due to missingness, country counts by year were 96 (2011), 99 (2014), 100 (2017), and 92 (2021); the basic panel comprises 384 observations. Some missing values were treated via interpolation; countries lacking sufficient data across years (e.g., 17 countries including Guatemala, Haiti, Jamaica, Tunisia) were removed from trend comparisons. Variables: - Dependent: Employment level (employed population / working‑age population), chosen over unemployment to better reflect inclusive employment coverage. - Core explanatory: Financial Inclusion Index (IFI). Constructed following Sarma (2012) and G20 inclusive finance taxonomy across four dimensions: usage, availability, quality, and benefits to specific groups. Multiple secondary indicators populate each dimension. Weights are based on coefficients of variation. Dimension indices are synthesized and then aggregated to a total IFI. - Moderating: Education development level (Education), proxied by government expenditure on education as a share of GDP. - Controls: GDP per capita (log, real), urbanisation rate, industrial value added/GDP, household and NPISH final consumption/GDP, general government final consumption/GDP, trade openness ((imports+exports)/GDP), net FDI inflows/GDP. - Heterogeneity: gender-specific employment levels (Maleemploy, Femaleemploy) and World Bank income groups (low, lower‑middle, upper‑middle, high income). - Robustness variable: Unemployment rate. IFI construction (outline): 1) Compute coefficient of variation for each indicator within each dimension; derive weights proportional to variation. 2) Standardise indicators using min–max bounds. 3) Aggregate weighted indicators to obtain each dimension index. 4) Combine dimension indices into the overall IFI using analogous aggregation. For 2021, the system adds digitalisation‑focused indicators (e.g., mobile/Internet bill payments) and subjective financial experience measures (e.g., financial stress from medical expenses, use of savings for emergencies). Modeling: - Baseline two‑way fixed‑effects panel model (country and year FE) after testing: LM/FE test (Prob>F=0.0000) favours FE over pooled OLS; Hausman test p=0.0000 supports FE over RE. Baseline model: Employment = a + β1·IFI + β2·GDPper + β3·Urbanisation + β4·Industry + β5·Consumption + β6·Government + β7·Openup + β8·Foreigninvest + ε. - Moderation model adds Education and the interaction Education*IFI: Employment = α + β1·IFI + β2·Education + β3·(Education*IFI) + controls + FE + ε. Significance and sign consistency of β1 and β3 indicate positive moderation. - Endogeneity: Addressed using an instrumental variables (2SLS) FE approach with Mobile (log mobile cellular subscriptions per 100 people) as instrument for IFI. First‑stage strength and over‑identification/unidentifiable tests are reported. - Robustness: (i) Replace Employment with Unemployment as dependent variable; (ii) winsorise IFI at 10% tails; (iii) IV estimation; (iv) heterogeneity regressions by gender and income groups. Descriptive statistics (selected): Mean IFI=0.619 (SD=0.107), Employment=0.568 (SD=0.113), Maleemploy=0.666, Femaleemploy=0.470; Education mean share of GDP=0.144. IFI trends suggest non‑monotonic changes with increased digital/user‑experience indicators in 2021 possibly lowering measured IFI amid pandemic disruptions. Estimation uses standard errors reported with significance levels; models include year and country fixed effects.

- Baseline effect (H1): IFI positively associated with employment. Without controls, β(IFI)=0.184, p<0.01. With controls, β(IFI)=0.119, p<0.05 (Table 5). This indicates higher financial inclusion correlates with higher employment levels. - Controls: Except for GDP per capita (positive), several controls (government consumption, urbanisation, industry share, etc.) show negative associations with employment in this sample and specification, potentially reflecting structural shifts, competition, or period/country composition. - Robustness: Results hold when (i) winsorising IFI (β≈0.117, p<0.05); and (ii) substituting unemployment as the outcome: β(IFI) on Unemployment = −0.115, p<0.05, consistent with employment gains. - Endogeneity/IV: Using Mobile as instrument, first stage shows strong relevance (F≈10.49, p<0.05). Second stage: IFI remains positively associated with employment (p<0.05). Exogeneity test p=0.001 supports the need for IV; under‑identification test p=0.020 rejected, indicating identification. - Education main effect (H2): β(Education)=0.500, p<0.01, confirming a positive link between education spending and employment. - Moderation (H3): Interaction β(Education*IFI)=1.445, p<0.01, indicating education amplifies the employment‑promoting effect of financial inclusion. - Gender heterogeneity: IFI positively associated with both male and female employment (p<0.05); the coefficient is larger for females, implying stronger gains for women. - Income‑group heterogeneity: Positive effects across income categories, with the largest effect in lower‑middle‑income countries, followed by low‑income, upper‑middle‑income, then high‑income groups. Significance weak in the lowest income group but significant elsewhere, suggesting inclusive finance requires a minimal economic/infra foundation yet yields relatively larger marginal benefits in poorer settings. - Descriptive insight: Global IFI development remains uneven; average employment rates among working‑age populations indicate substantial room for improvement. The 2021 IFI dip aligns with greater weight on digital/user‑experience indicators and pandemic-era disruptions.

The findings support the central hypothesis that greater financial inclusion fosters higher employment by expanding access to credit and payments, lowering financing frictions for SMEs and households, and enabling digital entrepreneurship. Education not only directly raises employment through human capital and better job matching but also strengthens the ability of individuals and firms to leverage inclusive financial services; thus, it amplifies finance’s effect on employment. Gender results suggest inclusive finance particularly benefits women’s employment, consistent with its focus on previously underserved groups. Income‑group results indicate that while some economic and institutional capacity is necessary to extract benefits from inclusive finance, the marginal employment gains are strongest outside the highest‑income settings, highlighting inclusive finance’s role in reducing disparities. These results reinforce the need for policies that co‑develop financial infrastructure and education (including financial literacy), promote digital financial services, and target underserved populations to maximise employment impacts.

This study constructs a digitally informed financial inclusion index and uses two‑way fixed‑effects and IV panel regressions on 101 countries (2011–2021) to show: (1) financial inclusion significantly promotes employment; (2) education directly increases employment; and (3) education positively moderates the finance–employment link. Benefits are more pronounced for women and in lower‑income settings (notwithstanding some insignificance in the lowest income tier), implying inclusive finance can narrow employment gaps when supported by adequate economic and regulatory foundations. Policy recommendations include: coordinating development of inclusive finance across regions, accelerating digital transformation and user‑centric service quality, expanding targeted microcredit and SME financing, enhancing financial literacy within education systems, and protecting vulnerable groups (e.g., women) in labour markets. Future research should extend coverage as data availability improves, refine digital/user‑experience components of IFI, explore causal channels (e.g., entrepreneurship, firm dynamics), and examine moderated effects within subgroups (e.g., education’s moderation by gender or age) as data permit.

- Data constraints led to exclusion of some countries/years and reliance on interpolation for missing values, potentially affecting generalisability and precision. - The moderating role of education could not be explored within gender-stratified heterogeneity regressions due to data limitations. - IFI measurement varied across years, with added digital and subjective indicators in 2021 that may affect comparability over time. - Potential residual endogeneity may remain despite IV strategy; instrument strength borders conventional thresholds. - Control variable interpretations may reflect period- and country-specific factors; causal mechanisms (e.g., entrepreneurship, firm investment) are inferred rather than directly observed.