Financial inclusion, aiming to provide affordable and accessible financial services to all segments of society, contrasts with traditional finance's bias towards the wealthy. Digital financial services have expanded access, particularly benefiting marginalized groups (Allen et al., 2016; Fu, 2020). Despite progress, significant gaps remain, with nearly one-third of adults unbanked in 2021 (World Bank) and a persistent gender gap (Kim, 2022; Asfaw et al., 2009). Inclusive finance addresses the shortcomings of the traditional 'rational economic man' hypothesis (Renzhia and Baek, 2020), requiring a stable macroeconomic environment and regulatory support (Johnson, 2013). The global pandemic exacerbated unemployment, highlighting the need for improved employment quality and the role of financial inclusion in providing liquidity, capital, and flexible working modes for businesses (Lai et al., 2022). This study investigates whether financial inclusion enhances job creation by expanding financing channels and fostering entrepreneurship.

Existing research explores the relationship between financial inclusion and employment, and education and employment. Studies demonstrate the positive role of financial inclusion in increasing employment and optimizing employment structure (Sakyi-Nyarko et al., 2022; Zhechen and Guosheng, 2021; Romero-Castro et al., 2023; Mais et al., 2020). Research on education's impact on employment level and quality reveals that education investment improves social quality and sustainable production (Ravi, 2021; Blanchard and Katz, 1997). However, existing research lacks comprehensive evaluations of traditional and digital finance under digital empowerment and its impact on employment. Studies on gender differences in financial inclusion's impact on employment are limited, and the moderating role of education on financial inclusion's impact on employment remains underexplored.

This study uses panel data from 101 countries (2011, 2014, 2017, 2021) from the World Bank's Financial Inclusion and Financial Availability Survey databases. A new financial inclusion evaluation index system is constructed, incorporating digital empowerment and both traditional and digital finance. A dual fixed-effect panel data regression model is employed to examine the impact of financial inclusion on employment, with education as a moderating variable. The financial inclusion index (IFI) is calculated in five steps using the coefficient of variation of multiple indicators across four dimensions: degree of use, benefit to specific groups, quality, and availability (see Table 1 for details). The employment level is measured as the proportion of employed individuals in the working-age population. The education development level is measured as the proportion of educational financial expenditure to GDP. Control variables include GDP per capita, urbanization rate, industrial added value, resident consumption level, fiscal expenditure, degree of openness, and foreign investment (see Table 2). The Hausman test justifies the use of a fixed-effects model. Models (5) and (6) show the basic regression model and the regulatory effect model, respectively. A robustness test is conducted by replacing the dependent variable with the unemployment rate and removing extreme values. An instrumental variable method is employed to address potential endogeneity using mobile cell subscriptions as an instrument. Heterogeneity analysis is performed by classifying countries based on income levels and gender.

Descriptive statistics reveal an unbalanced global development of financial inclusion (Table 3). The average IFI was 0.619, with a standard deviation of 0.107, and showed a fluctuating trend over time (Table 4). Basic regression analysis (Table 5) confirms a positive relationship between the financial inclusion index and employment level (H1), even after controlling for other factors. The robustness test, using unemployment rate as the dependent variable and removing extreme values, supports this finding. Instrumental variable analysis (Table 6) addresses endogeneity and reinforces the positive association between financial inclusion and employment. The results further support H2 (education level positively correlates with employment level) and H3 (education positively moderates the effect of financial inclusion on employment). Heterogeneity analysis (Table 7) reveals a stronger positive effect of financial inclusion on female employment compared to male employment. The positive impact of financial inclusion on employment is most significant in low- and middle-income countries, suggesting that a certain economic foundation is necessary for financial inclusion to fully realize its potential.

The findings highlight the crucial role of financial inclusion in promoting employment, particularly in developing countries. The positive impact is strengthened by higher levels of education, suggesting that policies promoting both financial inclusion and education can synergistically boost employment. The greater impact on female employment underscores the importance of gender-sensitive policies to ensure equitable access to financial services and opportunities. The stronger effect in lower-income countries suggests that financial inclusion initiatives should prioritize resource allocation to these areas to maximize their development potential.

This study contributes to the literature by comprehensively assessing the impact of financial inclusion on employment, considering the moderating role of education and accounting for heterogeneity across countries and genders. Policy implications include promoting coordinated regional development of financial inclusion, accelerating digital transformation, establishing direct channels to link financial inclusion and employment, and investing in education to amplify the positive effects of financial inclusion. Future research could explore the detailed mechanisms linking financial inclusion, education, and employment in more specific contexts.

The study's limitations include data limitations resulting in the exclusion of some countries and the inability to explore the moderating effect of education in gender-based heterogeneity analysis. Future research could address these limitations by expanding the dataset and conducting more detailed analyses.