Can financial inclusion enhance human development? Evidence from low- and middle-income countries

The study investigates whether financial inclusion (FI) enhances human development (HD) in low-, lower-middle-, and upper-middle-income countries. HD, grounded in health, education, and standard of living, varies widely across countries, with lower-income contexts exhibiting low HDI scores. FI—access to affordable, appropriate financial services—has expanded globally but remains uneven, especially in poorer countries. Prior literature offers mixed findings on the FI–HD link, motivating this paper’s core question: does increased FI improve HD in low and middle-income countries, and does this relationship exhibit nonlinearity? The study’s significance lies in informing policy to leverage FI for better health, education, and living standards in countries where both FI and HD are relatively weak.

The literature highlights conceptual and measurement disagreements around FI, with evolving definitions emphasizing not only access but also quality, cost, and usage of financial services. Composite FI indices (e.g., Sarma 2008; Sarma and Pais 2011; Mialou et al. 2017) capture multiple dimensions (access, usage), though methodological choices differ. Theoretical frameworks linking FI to HD include Sen’s Capabilities Approach, human capital theory, empowerment theory, inclusive growth, and social capital, all suggesting FI can expand capabilities, enhance productivity, and improve well-being. Empirical evidence largely reports positive FI–HD associations across contexts (India, SSA, OIC, cross-country studies), though some findings are mixed and point to potential nonlinearity and threshold effects. Studies also document FI’s indirect benefits via growth, poverty reduction, and inequality mitigation, with effects often stronger in low-FI and lower-income settings. A noted gap is limited focus on low- and middle-income countries and on disentangling FI’s access versus usage dimensions; this study addresses both and explores nonlinear dynamics.

Data: Panel of 79 countries classified as low-income (LI), lower-middle-income (LMI), and upper-middle-income (UMI), observed from 2000–2017 (initially 94 countries; 15 dropped due to data gaps). Sources: World Development Indicators (WDI) for FI and controls; UNDP for Human Development Index (HDI). The sample is split into three sub-samples: 27 UMI, 26 LMI, 26 LI countries. Outcome: Human Development Index (HDI), a [0,1] composite of life expectancy, education (mean years and expected years of schooling), and gross national income per capita, using UNDP methodology. Advantages and limitations of HDI are discussed (e.g., responsiveness, omitted welfare indicators). Financial inclusion measures: (1) Access dimension: ATMs per 100,000 adults (ATM) and bank branches per 100,000 adults (BRAN). (2) Usage dimension: bank deposits to GDP (DEPO) and domestic credit to the private sector (% of GDP, DCPS). (3) Composite FI index (IFI) built via Principal Component Analysis (PCA) from ATM, BRAN, DEPO, DCPS as a robustness measure. Controls: Gross fixed capital formation (INVES), FDI inflows (% GDP), trade openness (% GDP), infrastructure (individuals using the Internet, %), gross savings (% GDP, GSAV), gross national expenditure (% GDP, GNE), and World Bank loans/credits (log IBRD/IDA, LOANS). Additional micro-level FI measures for robustness (2011–2021 window): account ownership (ACC), number of depositors (DEPOS), and borrowers (BORR) per 1000 adults. Empirical approach: Three-step strategy. 1) Access model: HDI_it = f(ATM, BRAN, controls) estimated by System GMM (SGMM) or fixed effects depending on homogeneity tests. 2) Usage model: HDI_it = f(DEPO, DCPS, controls) via SGMM. 3) FI index model: HDI_it = f(IFI, controls) as robustness. Dynamic structure includes lagged HDI as an endogenous regressor; lags of HDI used as instruments. Diagnostics include Arellano–Bond AR(1)/AR(2) and Sargan tests. Slope homogeneity (Pesaran–Yamagata, 2008) and linearity tests (Wald/Fisher) guide model selection. Nonlinearity: Dynamic Threshold Panel (DTP) model (Kremer et al., 2013) applied for UMI and LMI groups where homogeneity and linearity tests warrant; IFI serves as both threshold variable and regime-dependent regressor with regime-specific slopes and intercepts. Estimation selects the threshold minimizing residual sum of squares; GMM used with appropriate instruments. For LI countries, fixed effects are used where DTP is not supported. Multicollinearity: Checked via correlation matrices and variance inflation factors (VIFs), showing no severe multicollinearity.

- Descriptives: Average HDI by group: UMI ≈ 0.72, LMI ≈ 0.61, LI ≈ 0.46. Access indicators (ATMs/branches) and usage indicators (DEPO/DCPS) are markedly lower in LI countries; PCA-based IFI averages: UMI 0.38, LMI 0.14, LI −0.62. - Access dimension (Table 5): - ATMs per 100,000 adults positively and significantly increase HDI in all three groups (UMI, LMI, LI) using SGMM/fixed effects as appropriate. - Bank branches are generally insignificant; exception: positive and significant for LMI countries. - Investments (INVES) and FDI exhibit negative and significant associations with HDI in UMI and LMI; mostly insignificant in LI. Suggested channels include environmental and health externalities from non-clean energy use and profit repatriation. - Trade openness positively and significantly relates to HDI in LI in the access model; other controls (GSAV, GNE) are positive and significant primarily in UMI. - Diagnostics: AR(2) and Sargan tests support correct specification. - Usage dimension (Table 6): - UMI: Both DEPO and DCPS are positive and significant; LMI: DCPS positive and significant (DEPO not robust); LI: usage measures are not significant. - TRADE, INFRA, GSAV, GNE often positively associated with HDI in UMI/LMI; effects not robust in LI. - Diagnostics (AR(2), Sargan) acceptable across models. - Threshold effects (Table 7 and linearity tests Table 9): - Linearity rejected; DTP supports a significant threshold in IFI for UMI and LMI groups. Estimated IFI thresholds are around 0.20 (reported near y ≈ 0.20 for both groups). Below and above the threshold, FI significantly increases HDI, with a larger marginal effect below the threshold (stronger β below-threshold than above-threshold), indicating diminishing marginal returns as FI deepens. - Robustness with composite IFI and micro-level FI measures (Table 8): - IFI positively and significantly relates to HDI; INFRA strongly positive; LOANS positive in LMI and negative in UMI in the DTP specification. - Micro FI (2011–2021): LI countries show significant positive effects of account ownership (ACC) and borrowers (BORR) on HDI; LMI shows positive effect of depositors (DEPOS); UMI effects are generally insignificant. Illustratively, a 1% increase in account ownership is associated with a 7.16% increase in HD for LI (based on coefficient magnitude). - Overall: SGMM results confirm that FI enhances HD in LMI and UMI; the effect is not statistically discernible for LI in aggregate access/usage models, though micro-level FI measures indicate benefits in LI and LMI.

The findings answer the research question by showing that improving financial inclusion elevates human development in lower- and upper-middle-income countries, with evidence of a nonlinear (threshold) relationship. FI appears to operate through increased access and usage of financial services—particularly ATM availability and private credit/deposits—which likely improve consumption smoothing, investment in education and health, and job creation. The absence of significant effects for LI in aggregate usage measures may reflect measurement issues (aggregate ratios capturing financial development more than inclusion) or insufficient financial infrastructure/quality. The threshold results suggest that at low levels of FI, marginal gains for HD are larger; as FI deepens beyond the threshold, gains persist but at a reduced marginal rate, underscoring the importance of early-stage inclusion policies and complementary investments (infrastructure, literacy). Trade openness, savings, and national expenditure also contribute positively to HD, while investment and FDI can correlate negatively with HD in some contexts, potentially due to environmental and health externalities, emphasizing the need for responsible, clean, and locally beneficial investment strategies.

This study contributes by focusing on low- and middle-income countries, decomposing FI into access and usage, and employing both SGMM and dynamic threshold panel methods. It shows that FI significantly enhances HD in LMI and UMI countries, with robust evidence of threshold effects around an IFI level of approximately 0.20. Results are robust to a PCA-based FI index and to micro-level inclusion measures, which are especially relevant for LI and LMI. Policy recommendations include: expanding FI via fintech, digitization, and innovation; strengthening infrastructure and financial literacy; maintaining supportive regulatory frameworks; and leveraging trade and savings to further boost HD. Given evidence of negative associations of investment/FDI with HD in some groups, policymakers should promote environmentally sustainable, health-safe, and locally beneficial investments. Future research should incorporate qualitative aspects of FI (responsibility, sustainability), extend datasets beyond 2017 where possible, and conduct comparative analyses including high-income countries.

- FI measurement focuses on quantitative dimensions (access and usage), excluding qualitative aspects (e.g., consumer protection, responsible lending). - FI data constraints limit the panel to 2000–2017 (and 2011–2021 for micro-level measures), potentially missing recent dynamics. - High-income countries are excluded, limiting generalizability and cross-income comparisons. - HDI, while comprehensive, omits some welfare aspects (e.g., distribution, gender equality, environmental degradation) and may be less responsive to short-run changes. - Some FI proxies (e.g., DCPS, DEPO) may reflect financial development rather than inclusion, especially in LI contexts.