The research question centers on whether financial inclusion (FI) enhances human development (HD) in low- and middle-income countries. HD, as defined by the United Nations, encompasses a richer human life beyond mere economic wealth, focusing on life expectancy, education, and standard of living. The World Bank emphasizes HD as crucial for improving lives and promoting sustainable development. Significant global disparities in HDI scores exist, with developed nations boasting scores above 0.8 and least developed countries below 0.55. These discrepancies stem from factors like economic stability, governance, and access to healthcare and education. The definition and measurement of FI remain debated, but the World Bank's definition – access to affordable and useful financial products and services – is widely used. However, the United Nations broadens this to include the quality, timeliness, and responsiveness of these services. The 2017 Global Findex survey highlights significant global disparities in financial inclusion, with high-income countries exhibiting far greater inclusion than low-income countries. Existing research on the FI-HD relationship presents mixed findings; some studies suggest a positive impact while others show negative or insignificant correlations. This study aims to address these inconsistencies by examining the relationship in low- and middle-income countries specifically, using SGMM and DTP analyses to account for potential endogeneity and non-linearity. Low- and middle-income countries were chosen because of their generally weak levels of both FI and HD, making this investigation particularly pertinent to their development challenges.

The literature extensively discusses the challenges of defining and measuring financial inclusion (FI), with no universally accepted definition. Early approaches focused on access to formal financial services, but more recent definitions, like that of Demirgüç-Kunt et al. (2017), emphasize the quality, affordability, and utilization of these services. Measuring FI also poses challenges, with varying approaches to creating composite indices. Beck et al. (2007) pioneered the use of banking service usage and accessibility, but their aggregate indicators faced limitations. Allen et al. (2016) addressed these by using micro-level survey data. Sarma (2008) and Sarma and Pais (2011) are credited with the initial creation of a FI index combining availability, accessibility, and usage data. However, subsequent research has proposed alternative indices with variations in normalization, weighting, and aggregation techniques. Several theoretical frameworks link FI and human development (HD). Sen's (1993) Capabilities Approach views FI as empowering individuals by providing resources for various life aspects. Schultz's (1961a, 1961b) human capital theory emphasizes the role of FI in investments in education and health, leading to higher productivity and well-being. The empowerment theory highlights FI's role in fostering economic independence and self-determination. The inclusive growth theory links FI to broader economic participation, and the social capital theory focuses on FI's role in strengthening community networks and trust. These theories provide a basis for understanding the complex interplay between FI and HD. Empirical studies on FI's impact on HD have yielded mixed results. While many studies report a positive association (Kamalu and Wan Ibrahim 2021; Matekenya et al. 2021; Van et al. 2021), some show negative or insignificant correlations. Chowdhury and Chowdhury (2023) found a positive impact in Bangladesh, India, and Pakistan. Barik et al. (2022) showed a positive relationship in India. Sarma and Pais (2011) observed a strong correlation between FI and HD across 49 countries, while Nanda and Kaur (2016) also found a positive association but noted its dependence on income level. Some studies explored the causal relationship, like Ofosu-Mensah Ababio et al. (2021), who found that low HD can lead to low FI, and Matekenya et al. (2021), who showed that FI positively impacts HD in Sub-Saharan Africa. Others, like Kamalu and Wan Ibrahim (2021) and Anurag et al. (2014), found positive correlations in different contexts. In contrast to direct impact studies, some research focuses on indirect effects through growth, poverty, and inequality. Studies such as those by Ali et al. (2021), Demir et al. (2022), Khan et al. (2021), Dogan et al. (2022), and Van et al. (2021) are examples of this approach. Few studies, however, have examined non-linear relationships, such as those by Kim and Lin (2011) and Abdelaziz and Helmi (2019). This study contributes to the literature by focusing specifically on low- and middle-income countries and by using both linear and non-linear models.

This study uses panel data from 79 low- and middle-income countries between 2000 and 2017 to examine the relationship between financial inclusion (FI) and human development (HD). The sample is divided into three subsamples: 27 upper-middle-income countries, 26 lower-middle-income countries, and 26 low-income countries. The Human Development Index (HDI) from UNDP reports serves as the HD measure, encompassing health (life expectancy), education (mean and expected years of schooling), and standard of living (GNI per capita). FI is measured using three approaches: 1) the access dimension (ATMs and bank branches per 100,000 adults); 2) the usage dimension (bank deposits and domestic credit to the private sector as a percentage of GDP); and 3) a composite FI index created using Principal Component Analysis (PCA) from the four aforementioned variables. The study considers several control variables, including domestic and foreign investments, trade openness, infrastructure, gross savings, gross national expenditure, and IBRD/IDA loans, to account for other factors influencing HD. The primary empirical method employed is the System Generalized Method of Moments (SGMM), suitable for panel data where the number of individuals (N) exceeds the number of time periods (T). SGMM addresses endogeneity concerns and provides more reliable estimates than OLS, fixed effects, or random effects models. Initially, a slope homogeneity test (Pesaran and Yamagata, 2008) is applied to assess the appropriateness of SGMM. Next, the Wald and Fisher linearity tests are used to determine whether a linear model is appropriate or if a dynamic panel threshold model (Kremer et al., 2013) should be used to capture potential non-linear relationships between FI and HD. The dynamic panel threshold model accounts for potential threshold effects whereby the relationship between FI and HD may vary depending on the level of FI. Three models are estimated to investigate the relationship: Model 1 assesses the impact of the access dimension on HD; Model 2 examines the impact of the usage dimension on HD; and Model 3 employs the composite FI index (from PCA) as a robustness check. Lagged dependent variables are used as instruments in the dynamic panel data models. The study employs a three-step estimation process and tests for multicollinearity to ensure the reliability of the results. The study also includes several robustness checks to ascertain the consistency of the results.

Descriptive statistics reveal significant differences in HDI and FI indicators across the three income groups, with UMICS exhibiting the highest levels of both. The multicollinearity test indicates no severe multicollinearity issues among the independent variables. The slope homogeneity test confirms the applicability of SGMM for the majority of models, except for Model 3 (using the FI index) in LICs, where fixed effects are more suitable. Linearity tests show that the relationship is non-linear for LMICS and UMICS, therefore employing the dynamic panel threshold (DTP) model for those. Results from the SGMM analysis (and fixed-effects where appropriate) show a positive and significant relationship between FI and HD for LMICS and UMICS in both access and usage dimensions. In the access dimension, ATMs significantly impact HD across all income levels, while bank branches have only a small impact on HD in LMICS. The usage dimension showed that both bank deposits and domestic credit to the private sector have a positive impact on HD, particularly in LMICS and UMICS. However, the impact of FI on HD is not significant for LICs. This is reflected in the access to financial services. Using the composite FI index as a robustness check, the positive relationship persists for LMICS and UMICS but the impact of FI on HD is still not significant for LICs. The DTP analysis reveals a threshold effect, with FI positively impacting HD both above and below the threshold levels in LMICS and UMICS. Furthermore, it reveals that investments (both foreign and domestic) have a negative and significant effect on HD for UMICS, trade openness positively impacts HD only in LICs and that savings and national expenditure positively impact HD only in UMICS. A robustness check using micro-level FI measures (account ownership, depositors, and borrowers from commercial banks) confirms the positive impact of FI on HD in LMICS and LICs.

The findings demonstrate a robust positive relationship between financial inclusion (FI) and human development (HD) in lower-middle-income and upper-middle-income countries. This supports the hypothesis that improved access to and usage of financial services contributes to better health, education, and living standards. The threshold effect suggests that the impact of FI on HD may be nonlinear, indicating the necessity to reach a certain level of financial inclusion before significant improvements in HD are realized. This highlights the importance of achieving a critical mass in financial penetration and usage for optimal impact. The negative impact of investment on HD in upper-middle-income countries is intriguing and suggests potential trade-offs. It may indicate that the benefits of investment are not translating into substantial improvements in HD indicators in these nations. The positive effect of trade openness in low-income countries aligns with the view that international trade can improve living standards through increased economic growth and opportunities. The lack of a significant FI-HD relationship in low-income countries warrants further investigation. It may be attributed to the limitations in measuring FI in this specific context, differences in access, and usage of financial services.

This study provides strong evidence supporting the positive link between financial inclusion and human development in lower-middle-income and upper-middle-income countries. The findings highlight the significance of increasing access to financial services and their responsible use in boosting HD indicators. The existence of a threshold effect underscores the need for targeted strategies to surpass critical levels of financial inclusion for maximum impact on HD. Further research should explore the qualitative aspects of FI, compare findings across all income levels, and delve deeper into the reasons for the lack of a significant relationship in low-income countries. Additionally, a deeper examination of the negative impact of investment on HD in upper-middle-income countries could provide valuable insights into policy interventions.

The study focuses primarily on the quantitative aspects of financial inclusion (access and usage), neglecting the qualitative dimension (responsible and sustainable delivery of services). The analysis is limited to low- and middle-income countries, excluding a comparison with high-income nations. The chosen FI indicators may not fully capture the nuances of financial inclusion in all contexts. Furthermore, the causal relationship between FI and HD is not explicitly addressed, only the correlation. The use of specific indicators for FI and HD may be a source of bias in interpreting the result.