Nexus between financial inclusion, workers’ remittances, and unemployment rate in Asian economies

The study explores whether and how financial inclusion and workers’ remittances affect unemployment in Asian economies. Financial inclusion—ensuring access to affordable and appropriate financial services such as payments, credit, savings, insurance, and pensions—has grown in prominence since the early 2000s, with 1.7 billion adults still unbanked as of 2017. A well-functioning financial system supports growth via resource allocation and intermediation, potentially influencing labor markets by easing credit constraints, promoting SME growth, and fostering innovation. The post-COVID context and the SDGs underscore the urgency of understanding finance–labor market linkages. Remittances represent large and stable external inflows in many developing countries and can affect employment through consumption demand, investment, entrepreneurship, and human capital channels. Despite this, limited research examines the joint effects of financial inclusion and remittances on unemployment across Asian subregions, often with methodological limitations (e.g., endogeneity, non-normality). This study addresses these gaps by employing 2SLS, GMM, and panel quantile regression on a multi-country Asian panel to identify region-specific effects and offer policy-relevant insights.

Financial inclusion and unemployment: Theory suggests that financial development eases access to credit for households and firms, lowering external finance costs, stimulating investment, firm entry, and job creation, thereby reducing unemployment (e.g., Bernanke et al., Monacelli et al.). However, potential downsides from financialization emphasize short-termism, weakened labor bargaining power, and reduced physical investment, potentially dampening employment (Krippner; Tori & Onaran). Empirical evidence is mixed but often finds that financial liberalization and development reduce unemployment (Lin et al.; Feldmann; Borsi). Recent studies highlight the importance of inclusive finance specifically, linking financial inclusion to growth, poverty reduction, and potentially lower unemployment through enhanced financial development and economic activity (Pal & Bandyopadhyay; Koomson et al.; Mushtaq & Bruneau). Remittances and unemployment: Neoclassical frameworks imply remittances may raise reservation wages and reduce labor supply (potentially increasing unemployment), but they can also finance consumption, investment, entrepreneurship, and education, thus stimulating demand and job creation and enhancing employability (Azizi; Furlanetto & Seneca; Airola; Vadean et al.). The net effect is ambiguous a priori and context-dependent, motivating empirical investigation in Asian settings. The literature reveals gaps regarding region-specific dynamics, endogeneity, and non-normality that this study seeks to address.

Model: Baseline static and dynamic panel specifications relate unemployment (UNE) to financial inclusion and remittances, controlling for macro and openness factors: (1) UNE_it = π0 + π1 ATM_it + π2 REM_it + π3 GDP_it + π4 INT_it + π5 FG_it + π6 Inflation_it + α_i + ε_it; (2) UNE_it = π0 + λ UNE_{t-1} + π1 ATM_it + π2 REM_it + π3 GDP_it + π4 INT_it + π5 FG_it + π6 Inflation_it + α_i + ε_it. Data: Annual panel for 39 Asian economies, 2004–2021. Variables: UNE (total unemployment, % labor force, WDI); ATM (ATMs per 100,000 adults, GFDD) as a proxy for financial inclusion; REM (personal remittances received, % of GDP, WDI); GDP (GDP per capita, constant 2015 US$, WDI); INT (individuals using the Internet, % population, WDI); FG (financial globalization index, KOF); Inflation (consumer prices, annual %, WDI). Estimation: To address endogeneity, heteroskedasticity, and serial correlation, the study employs 2SLS and Arellano–Bond GMM. GMM uses lagged levels/differences as instruments, with preference for two-step estimation in unbalanced panels for efficiency and robustness to autocorrelation and heteroskedasticity. Robustness: Panel Quantile Regression (PQR) assesses effects across the unemployment distribution and accommodates non-normality, corroborated by Jarque–Bera tests indicating non-normal series. Diagnostics: Variance Inflation Factors (VIFs) indicate no serious multicollinearity (all VIF < 5; mean VIF ~2.08; tolerances > 0.1). Descriptives: Means (UNE 6.158; ATM 3.164; REM 5.740; INT 1.362; GDP 8.433; FG 1.737; Inflation 6.073) and standard deviations reported; all series reject normality.

• 2SLS results across Asia and subregions (Central Asia, East Asia, Southeast Asia, West & Middle East Asia, South Asia) show that higher ATM density (financial inclusion), remittances (REM), internet usage (INT), GDP per capita, and financial globalization (FG) are generally associated with lower unemployment; inflation’s effect is mixed across regions.
• Illustrative 2SLS patterns: Negative and significant associations for ATM, REM, INT, GDP, and (often) FG with unemployment across most regions; inflation tends to raise unemployment in Asia and West & Middle East Asia but is insignificant or negative in some subregions.
• GMM results confirm the main findings with dynamic controls: A 1% increase in ATM, REM, and INT is associated with reductions in unemployment by approximately 0.213%, 0.042%, and 0.587% in Asia; 0.128%, 0.008%, and 0.242% in Central Asia; 0.263%, 0.254%, and 1.819% in East Asia; 0.120%, 0.025%, and 1.007% in Southeast Asia; 0.126%, 0.035%, and 1.211% in West & Middle East Asia; and (ATM insignif.), 0.028%, and 0.829% in South Asia, respectively.
• GDP per capita and FG typically reduce unemployment in GMM (e.g., GDP: −0.131% in Asia, −0.427% in Central Asia, −1.245% in East Asia; FG: −0.611% in Asia, −0.991% in Southeast Asia), though some coefficients are regionally insignificant.
• Inflation raises unemployment in Asia (e.g., +0.011%) but is significantly negative in Central Asia, East Asia, and South Asia, indicating heterogeneous macro price effects.
• The lagged unemployment term in GMM is positive and highly significant, indicating persistence in unemployment dynamics.
• Panel Quantile Regression (PQR) shows heterogeneity across the unemployment distribution: ATM effects turn negative and significant at higher quantiles (70th–95th), REM and GDP exhibit consistently negative effects across most quantiles (except the lowest), INT reduces unemployment at upper quantiles (80th–95th), FG is mostly negative (10th–95th), and inflation increases unemployment from the 20th–95th quantiles.
• Together, results across methods underscore robust negative associations of financial inclusion, remittances, internet usage, GDP, and financial globalization with unemployment in Asian economies, with inflation effects varying by region and unemployment quantile.

The findings directly address the research question by showing that greater financial inclusion and higher remittance inflows are linked to lower unemployment in Asia and its subregions. Access to financial services (proxied by ATM density) likely eases credit constraints, fosters entrepreneurship and SME expansion, and facilitates risk management, thereby supporting job creation. Remittances boost aggregate demand and can finance investment, entrepreneurship, and human capital, which translate into improved employment outcomes. The negative associations for internet use indicate that digital connectivity supports labor market participation and enables access to digital financial services. GDP per capita’s negative relation with unemployment reflects expected growth-employment linkages, while financial globalization’s negative effect suggests beneficial integration channels (capital access, financial development). Inflation exhibits heterogeneous impacts, consistent with differing regional macroeconomic structures and policy regimes. Robustness across 2SLS, GMM, and PQR suggests these relationships hold under endogeneity concerns and across the unemployment distribution, though magnitudes and significance vary by region and quantile.

This paper contributes by jointly examining financial inclusion and remittances as determinants of unemployment across Asian economies using 2SLS, dynamic GMM, and panel quantile regression on a large multi-country panel (2004–2021). Consistent evidence indicates that financial inclusion (ATM density), remittances, internet penetration, higher GDP per capita, and financial globalization are associated with lower unemployment, while inflation’s effects are region-specific. Policy implications include: prioritizing inclusive finance (expand access to accounts, credit, insurance; promote financial literacy; leverage digital finance), reducing remittance costs and promoting formal channels while encouraging productive use (e.g., SMEs, infrastructure), investing in skills and entrepreneurship programs, streamlining regulations and supporting SMEs to improve the business environment, enhancing cross-border cooperation to facilitate remittances and regional integration, and instituting regular monitoring and evaluation of inclusion and remittance policies. Future research should deepen analysis of mechanisms, account for country heterogeneity, and broaden financial inclusion metrics beyond ATMs.

• Heterogeneity across Asian economies (financial systems, labor markets, institutions) may lead to varying effects; region- and country-specific factors are not fully explored.
• Mechanisms/channels (e.g., entrepreneurship, human capital, sectoral shifts) are not explicitly identified and warrant deeper investigation.
• Financial inclusion is proxied primarily by ATM density; broader, multidimensional indicators could provide a more comprehensive assessment.
• Potential data limitations from an unbalanced panel and measurement issues may affect estimates despite robustness checks.
• Comparative analyses across countries and subregions could further clarify contextual influences and policy effectiveness.