Poverty Alleviation through Digital and Traditional Finance: Empirical Evidence from Emerging European Economies

Introduction

The idea of sustainable development has become increasingly significant on the global level, with particular attention being paid to the UN Sustainable Development Goals (SDGs). One of these goals is the eradication of extreme poverty, a challenge that the World Bank and the UN 2030 agenda also seek to address. One of the issues that has been raised in this context is financial inclusion, with the aim of offering affordable financial services to people from marginalized groups (Arner et al. 2020). This is based on the assumption that better integrating marginalized groups into the formal financial system could sustain economic growth and alleviate poverty. However, recent global events, such as the COVID-19 pandemic and the spikes in food and energy prices, triggered for the most part by Russia’s aggression against Ukraine and climate change, have shaken the stability of the global economy. The impact on different countries has been uneven, and so, too, have the policy choices to address it. The pandemic had another, more significant consequence – it accelerated the process of digitization. For example, online banking has become a necessity to many customers, thus putting financial institutions under pressure to provide digital services. Banks have been forced to innovate, and have placed more emphasis on digital channels and mobile-based banking services, as opposed to the traditional, bank-based ones (Şendur 2022).

Poverty continues to exist. For example, some EU member states, such as the Czech Republic (12.0 %) and Slovenia (13.7 %), have lower poverty levels than others, such as Germany (21.3 %). On the other hand, some new EU countries – such as Romania (32.0 %) and Bulgaria (30.0 %) – have more serious poverty problems.^[1] The poverty indicators show that the social protection systems in the Czech Republic and Slovenia have proven to be more efficient than in neighboring countries. Meanwhile, Poland (16.3 %), Hungary (19.7 %), and Croatia (20.7 %) fall somewhere in between, with poverty rates close to or below the EU average.

Financial inclusion has traditionally been measured by the number of bank branches and accounts or number of ATMs. However, expanding physical banking infrastructure is costly and often unprofitable, especially in rural areas. It is becoming increasingly evident that mobile technology has served as a driver of many processes, especially in developing countries. For example, it allows previously “unbanked” populations to access financial services, without ever visiting a bank (Gosavi 2018). The wider adoption of mobile technology was accompanied by the growing global importance of digital financial inclusion, especially in developing regions. This paradigm shift has transformed digital platforms into key enablers of financial access worldwide. To better understand how financial inclusion has affected poverty reduction, this study divides it into two categories: traditional or conventional channels, on the one hand, and digital channels on the other. Does digital financial inclusion have the same impact on poverty reduction as traditional financial inclusion? If not, what is their proportionate impact?

Here, there is a need for a scientific response. This study aims to contribute to a better understanding of the relationship between financial inclusion and poverty alleviation. Specifically, the following questions will be addressed: Does the moderating role of EU membership change the poverty alleviation effects of digital compared to traditional finance? How does digitally driven financial inclusion affect the fight against poverty? In this study, I will use the term “emerging Europe” to refer to a set of former state socialist EU members along with candidate and aspirant countries, those being: Albania, Bosnia and Herzegovina, Bulgaria, Croatia, Czech Republic, Estonia, Hungary, North Macedonia, Kosovo, Latvia, Lithuania, Montenegro, Poland, Romania, Serbia, Slovakia, Slovenia, and Turkey.

There is currently intense scholarly interest in digital inclusive finance since it provides a transformative solution that dissolves the geographic barriers of traditional financial systems. Digital technology implementation enables universal service delivery to groups that lack access to traditional financial services and helps reduce costs for financial service institutions. This research uses empirical evidence to study whether digital financial inclusion lowers poverty levels more than traditional financial inclusion channels. This regional focus is important for two reasons. First, extensive recent research on the development of financial inclusion, mainly focusing on traditional channels (Beck, Demirgüç-Kunt, and Levine 2007; Rojas-Suarez and Amado 2014; Le et al. 2019; Poghosyan 2023; Ganić 2023) provides a solid basis for comparative analysis.^[2] Second, empirical research on the impact of digital financial inclusion on poverty reduction in this context is limited, which is precisely the research gap this study seeks to fill. Specifically, there are no studies comparing the proportionate impact of traditional and digital financial inclusion channels on poverty alleviation, including the effects of EU membership. This study thus compares the impact of traditional and digital financial inclusion channels on reducing poverty for the panel of 18 “emerging” European countries, comprising both EU and non-EU states. To the best of my knowledge, it is the first investigation to include EU membership effects on digital and traditional finance-related poverty reduction across such a wide panel of economies.

The study makes two distinct contributions to the research literature. First, it shows that financial inclusion strategies using traditional and digital methods can bring about improvements even if digital channels have less influence than more traditional financial inclusion channels. Second, the study demonstrates that traditional banking channels continue to implement methods of poverty eradication efficiently. However, policymakers still need to adopt measures to protect people from the increased poverty risks that are also inherent in digital financial services. Further, the research shows that measures aimed at boosting digital penetration can help economies with low levels of financial inclusion more broadly.

After an introductory section, which explains the research problem, the objectives, and the expected results, the study is organized as follows. A review of the pertinent literature is followed by a detailed explanation of the research models, with the fourth and fifth sections then devoted to presenting the findings and conclusions, as well as recommendations for future research.

Literature Review

Scholars such as McKinnon (1973), Shaw (1973), King and Levine (1993) showed that the financial industry provides wider access to financial services thus stimulating economic growth. In terms of allocating resources and promoting general economic development, they stressed the significance of creating a vibrant, effective, and inventive financial sector. In addition, some research revealed that, compared to other sectors such as manufacturing and hospitality, the financial industrystands out for the depth and use of information and communication technology (ICT) (Shamim 2007). In developing nations, the adoption of ICT can be crucial for achieving operational efficiency. In fact, several studies have shown that the spread of ICT can reduce the cost of information processing in the financial sector, which improves competition and inclusion, and at the same time contributes to long-term economic growth prospects (Asongu and Odhiambo 2020; Muto and Yamano 2009; Shamim 2007).

Digitalization makes the achievement of financial inclusion more feasible. In an analysis of 50 developing and developed countries, Kowalewski and Pisany (2023) demonstrated the importance of financial technologies for the economy. By showing that advanced human capital is created when a business becomes more digital, a further study, focusing on European countries, attests to the positive impact of digitalization on people’s well-being (Grigorescu and Ion 2021).

The role of mobile money in the financial landscape of developing countries is significant, with traditional institutions becoming redundant. One example is the use of mobile phones to access financial services, which has improved financial inclusion (Blancher et al. 2019) by, among other things, reducing existing inequalities in financial infrastructure (Chatterjee 2020).

Understanding the causes of poverty is essential in order to take steps to reduce it. Regional differences in culture, political environment, and socioeconomic setting determine the factors that contribute to poverty. For instance, social capital, income, ethnic inequality, local political rivalry, federal funds, and the proportion of foreign-born individuals in the local population are important factors that determine extreme poverty in the United States (Rupasingha and Goetz 2007). For EU countries, on the other hand, the size of farm holdings, productivity of resources, social protection, and domestic material consumption are key determinants of poverty (Ulman and Căutişanu 2020).

Technology is an essential component of the financial system that has received less attention in the literature, despite the fact that the impact of financial development – measured through banks and stocks – on reducing poverty has been extensively studied. Economic growth, the advancement of financial inclusion, and financial development all depend on technology (Ozili 2018). Another study concluded that it reduces poverty and vulnerability in rural areas by improving e-commerce, fostering social connections, and facilitating better access to finance.^[3] Similarly important is the effect of increasing income and financial development aimed at reducing poverty in certain areas. Demirgüç-Kunt and Klapper (2013) examined financial inclusion in 148 countries, focusing on individual and country variables linked to three traditional channels of financial inclusion: holding a bank account; saving in a bank account; and using bank loans. They implied that the degree of financial inclusion is impacted by disparities in income, both within and between nations. Furthermore, they found that people with higher incomes and levels of education typically benefit more from increased financial inclusion.

Mushtaq and Bruneau (2019) studied the role of ICT in poverty reduction, using data from 61 countries (2001–2012). They found that enhancing financial frameworks for ICT growth is key to reducing poverty and income inequality. Rewilak (2017) analyzed the impact of financial sector penetration on poverty in middle- and low-income countries (2004–2015) and found that financial depth was more influential than access to finance. Appiah-Otoo and Song (2021) examined the impact of financial technology on poverty in 31 Chinese provinces (2011–2017), focusing on third-party payments and lending. Their research showed that technology, including third-party payments and loans, has contributed to poverty reduction. Similarly, Ye, Chen, and Li (2022) confirmed the previous findings for a panel of 31 provinces in China. Their study indicated that low-income provinces experience a significantly greater impact from measures using financial technology to reduce poverty than high-income ones.

The studies conducted by Wang and He (2020) and Riley (2018) confirmed that financial technology helps to promote financial inclusion. It makes financing easier for small and medium-sized businesses, which frequently struggle to secure funding from conventional state-owned banks. Focusing on emerging nations, Asongu and Asongu (2018) investigated the relationship between mobile banking, poverty, inequality, and inclusive growth. The authors pointed to the existence of an inverse relationship between the use of mobile money and poverty not only in Latin American countries, but also in Central and Eastern Europe and the Asia–Pacific region.

To address some open questions related to financial inclusion and poverty alleviation, my empirical analysis will start by testing hypothesis H ₁: Digitally driven and traditional channels of financial inclusion contribute to poverty alleviation in emerging European countries. The rationale behind this hypothesis is the correlation between financial inclusion and poverty reduction. Indeed, access to a wide range of financial services is assumed to enable individuals to invest, thus increasing their economic opportunities; digitization is an important part of this since it remains relatively inexpensive and opens up more financial services.

The work of Kwan and Chiu (2015) and other neoclassical models of economic development and sustainable living provide the theoretical basis for this study. These models demonstrate the relationship between ICT and the engagement of marginalized communities. By utilizing the expanding influence of mobile technology, recent developments in financial technology offer remarkable potential to overcome obstacles to financial inclusion, which can lead to more balanced income development, while promoting economic growth at the same time (Gabor and Brooks 2017; Salampasis and Mention 2018; Demirgüç-Kunt et al. 2018; Sahay et al. 2020).

Based on data from 182 countries, Yilmaz and Koyuncu (2018) also found that increasing internet access significantly reduces poverty and economic disparity (2000–2013). Drawing on the implications of the empirical research presented, the following hypothesis, H ₂, was developed and will be empirically tested: Digitally driven channels of financial inclusion in “emerging” European countries have a bigger impact on poverty reduction than traditional channels of financial inclusion.

According to economic theory, inefficiencies in the financial system make it more difficult for those living in poverty to access formal financial services, which prevents them from rising out of poverty (Galor and Zeira 1993; Kim 2016). For example, formal access of financial services is restricted in sub-Saharan Africa (SSA) by variables such as cost, location, and documentation requirements (Demirgüç-Kunt and Klapper 2012). The reasons most frequently given by rural residents for not having a bank account are the absence of required documentation and the distance to banks. Inadequate infrastructure with limited communication options and strict banking regulations hinder the growth of bank branches. Galor and Zeira (1993) showed that access to credit for the poor was hampered by dysfunctional credit markets. Therefore, greater access to finance is seen as one way to reduce poverty and economic inequality (Aslan et al. 2017; Park and Mercado 2018; Ganić 2023). Similarly, in his empirical analysis of 52 African countries, Asongu (2015) found that the poor benefit from mobile phone penetration. Mobile phones helped redistribute income, and their evolution into mobile devices functioning as a sort of “pocket bank” has made financial services more widely accessible.

When it comes to EU membership, the expectation is that this leads to enhanced financial inclusion channels and thus poverty alleviation, due to the better quality of institutions, robust regulatory regimes, and easier access to EU finance and innovation. As well as enhancing institutional quality, EU membership can also (1) improve consumer protection, which stimulates trust in financial services (Allen et al. 2016; Demirgüç-Kunt et al. 2018; Ozili 2018); (2) develop digital finance services such as mobile banking, enabled by EU-wide infrastructures and policies (Beck, Degryse, and Kneer 2014);^[4] (3) encourage the integration of the banking sector within the borders of the EU’s single market; and (4) increase the competitiveness and spread of standard financial services (Claessens and van Horen 2015; Comparato 2015). Based on these positive assumptions about the role of the EU’s institutional framework, the following hypothesis, H ₃, is proposed: In EU member states, both digital and traditional financial inclusion channels have greater impact on poverty reduction than in non-EU countries.

Methodology and Data

The study uses the panel of the 18 abovementioned “emerging” European economies, selected based on the availability of consistent and complete data. The time frame of the research was decided based on the most recent data and the presence of sufficient observations to conduct robust statistical analysis. The study empirically examines whether digitally driven channels of financial inclusion have a stronger effect on poverty alleviation than traditional channels for the period between 2004 and 2021. Another primary interest is to determine the moderating role of EU membership in this relationship.

The study uses a panel data analysis to examine the three proposed hypotheses. On this basis, cross-country differences and the changes over time are considered. To estimate and measure the impact of technology-driven and traditional financial inclusion on poverty, the following general panel data regression model is applied:

The dependent variable is “poverty”, and it is measured by two different indicators used by the World Bank: PVTA1 – poverty headcount ratio, i.e. the percentage of the population living in households with a per capita consumption or income that is below the poverty line; and PVTA2 – poverty gap, which measures the depth of poverty or the average income shortfall of the total population below the poverty line.

Moreover, the research employs complex statistical approaches to provide an accurate study with real-world data. First, I use a two-way effect for country and time heterogeneity to control for country-specific elements and isolate the net impact of the explanatory variables from possible bias. Second, I avoid the cross-section random effect, which would lead to conclusions beyond the specific countries in our dataset (Baltagi 2021). The panel-corrected standard errors (PCSE) method (Beck and Katz 1995; Wooldridge 2010) is also used to address standard data problems, such as uneven variances, the unobserved relationships between groups, and heteroscedasticity (Bailey and Katz 2011; Hasan and Rahman 2019). Evidence indicates that the PCSE is a reliable tool to tackle these issues. I then cross-check the results by applying the generalized method of moments (GMM) technique to examine the model’s assumptions and control for unobserved fixed effects by considering the potential endogeneity of the explanatory variables (Baltagi 2021).

The core relationships between poverty and financial inclusion to be examined using the baseline model are expressed as follows:

Here, i presents the unit of country; t denotes the years in which poverty _it alternates between PVTA1 and PVTA2. As mentioned above, the poverty headcount ratio provides the extent or the breadth of poverty (PVTA1), while the poverty gap (PVTA2) provides information on the depth of poverty. Digitally driven channels of financial inclusion refer to the measure of financial technology by utilizing two proxy variables: DigT1 is proxied by the percentage of individuals using internet banking, while DigT2 is proxied by the percentage of using the internet for selling goods or services, sourced from the Eurostat database.^[5] γ _i refers to the country fixed/random effect; δ _t to time fixed effects; and ɛ _it is the error term. Furthermore, I introduce interaction coefficients to estimate the contribution of EU membership to enhancing the poverty elasticity of financial inclusion based on digital and traditional channels. In fact, a model that includes the moderating role of EU membership and the impact of selected channels of financial inclusion on poverty reduction serves as the baseline model (Equation 4).

Here, EU _i x TFinI1 _it refers to an interaction effect of EU membership and TFinI1; EU _i x TFinI2 _it to an interaction effect of EU membership and TFinI2; EU _i x DigT1 _it to an interaction effect of EU membership and internet banking (DigT1); and EU _i x DigT2 _it an interaction effect of EU membership and online selling (DigT2). EU takes the value of 1 when a country is an EU member, and 0 otherwise.

Amidžić, Massara, and Mialou (2014), Sarma (2015), Wang and Guan (2017), and Ganić (2023) all employed indicators of traditional channels of financial inclusion, such as the number of bank branches per 1,000 adults and the number of ATMs per 1,000 adults. Similarly, the current study uses bank accounts per 1,000 adults to serve as TFinI1, and commercial bank branches per 100,000 people as TFinI2, sourced from the World Bank database, to measure the level of traditional channels of financial inclusion. Gross domestic product per capita (GDPpC) and population (POPUL) are used as control variables as these can have an impact on poverty. When conducting a cross-country analysis, it is important to account for differences in economic activity as measured by GDP. The model therefore accounts for the GDPpC as unique to each country. Additionally, the model accounts for population growth, implicitly assuming that population growth naturally stimulates an increase in financial and economic activities that will probably have an impact on financial development (Aduba, Asgari, and Izawa 2023).

The moderating effect of EU membership in the context of digital and traditional finance is further considered by constructing the interaction term between EU and non-EU countries which have varying income levels, poverty levels, and differ in terms of their initial use of digital and traditional financial instruments. Panel analysis has become an increasingly common tool for studying three-dimensional data since it combines time series analysis and cross-sectional analysis. In order to prevent erroneous regression findings, it is critical that cross-sectional and temporal factors are taken into consideration when analyzing panel data. The choice between dynamic random effects (RE), fixed effects (FE), and first differences (FD) models is crucial. Regarding explanatory factors, RE estimation makes no assumptions. When examining possible correlations between the explanatory variables and RE, the Hausman test is frequently used. Two estimates are compared in this test: one that holds true under the alternative and null hypotheses, and the second that holds true under the null hypothesis alone. If the idiosyncratic effects are truly random, the RE estimator performs better. On the other hand, the FE estimator is entirely consistent in this scenario if the estimates of the idiosyncratic effects of the RE and FE estimators exhibit disparities in convergence.

The variables used in my research comprise channels of digitally driven financial inclusion, traditional financial inclusion, and poverty alleviation. These variables were sourced from the World Bank, Eurostat, and the Poverty and Inequality Platform (PIP) World Bank databases, respectively. The headcount and the poverty gap are the two metrics employed to track poverty. These indicators work best together. While the poverty gap more accurately gauges the depth of poverty, the headcount shows the relationship between those who live in substandard conditions and those who earn more than the poverty level. Using the poverty headcount ratio variable to ensure strong results, the study analyzes whether the development of financial inclusion affects poverty. Table 1 displays a set of variables employed in the econometric models.

Empirical Results

There are four models in panel data regression, including a two-way-effect cross-section fixed effect and period fixed effects (M1), a cross-section random effect (M2), panel estimated generalized least squares (EGLS) or panel corrected standard error (M3), and panel generalized method of moments (M4) displayed in Tables 2 and 3 for two different dependent variables. The constants in models M1–M3 are statistically significant, revealing a baseline poverty level (POVTA ₁). The persistence effect disappears in Model M4 because it uses lagged values. The coefficient −0.709448 indicates strong mean reversion, showing that poverty trends stabilize toward long-term averages.

The results of the output for the first poverty alleviation proxy variable show that digital channels of financial inclusion have mixed effects on poverty alleviation. In fact, internet banking (DigT1) has an inverse and significant effect in two of the four models (M1 and M4), suggesting that an increase in internet banking reduces poverty. The research output for online selling goods or services (DigT2) in contrast reveals a positive and significant effect in three of the four models (M1, M3, and M4), indicating that an increase in online selling widens poverty as measured by POVTA ₁. In addition, in M1 and M4, the findings reveal that poverty is reduced by increasing internet banking, while the current level of online selling use is not shown to be a significant contributor to poverty alleviation. Moreover, for traditional finance channels, bank accounts (TFinI1) and commercial bank branches (TFinI2) are statistically significant in three out of four models (M1, M2, and M4) and in all four models (M1, M2, M3, and M4), respectively, having the expected signs, at the 1 %, 5 %, and 10 % significance levels.

The research indicates that an expansion in internet banking (DigT1) has a minor impact on poverty, which is insignificant; a 10 % increase in use only reduces poverty rates by 1.1–1.4 %. Although statistically significant, this modest effect may be because the poorest households rarely have access to or face difficulties using digital banking. It implies that the growth of internet banking may alleviate poverty but not single-handedly. Policies geared toward internet banking growth should go hand in hand with other policies directly assisting vulnerable groups.

The analysis indicates that online goods sales have complex implications for poverty. Although such activities are also associated with the fact that the number of people living below the poverty line has increased slightly (+0.07 % for every 10 % increase in the number of online sellers), they are also associated with a small decrease in the depth of poverty (−0.05 %). This implies that online marketplaces could help some people to earn decent incomes but cannot easily solve the problem of the poorest. By contrast, the benefits of traditional banking services are more consistent: physical bank branches reduce poverty by two to three times more than digital tools (by 0.12–0.28 % per 100 new branches), and bank accounts primarily benefit people at or just below the poverty line, as opposed to the poorest poor. This suggests that, in “emerging Europe”, various financial services are available to various verticals of society. Moreover, EU membership enhances the power of internet banking to decrease the level of poverty (increasing its impact by −0.0009), perhaps owing to the EU’s regularized infrastructural investments, which facilitate access to digital finance.

That said, economic growth is still the greatest driver, since 1 % growth in GDP per capita cuts poverty eightfold (−0.0080) relative to even the EU-enhanced digital banking drivers. Variations in population size do not have a strong impact on poverty reduction. Improved access to finance and better economic opportunities are in fact the main driving factors. Traditional financial inclusion (such as having basic access to banks) is higher in countries that have reached higher levels of poverty reduction than in those where such inclusion is lower. Although internet banking also contributes to poverty alleviation (more in the EU, as indicated in models M1–M3), this impact is augmented by EU membership, as with traditional banking channels (TFinI1/TFinI2). Of the two control variables, population growth (Popul) is mostly insignificant, suggesting that it does not strongly influence poverty levels, while economic growth (LnGDPPc) has mixed effects, with a negative but weak significance in M2 and M4. This indicates that a higher GDP per capita may not directly translate to lower poverty levels (Table 2).

Table 2 reveals that the M1 model (two-way effect) has the highest R-squared (0.574), suggesting that it explains the greatest variation in poverty. The value of R-squared for M2 and M3 have weaker explanatory power, while M4 (GMM) does not provide an R-squared value but incorporates dynamic effects, confirming the persistence of poverty. Diagnostic tests run for the GMM model show that the AR (1) and AR (2) tests in M4 indicate no strong autocorrelation issues.

To validate its main results, the study also used the poverty gap indicator (POVTA 2) presented in Table 3. The findings determine that former poverty is a strong predictor of present poverty. Studies have demonstrated that historical poverty determines present levels. The depth of the rate of improvement of poverty, however, does not match the overall rate of change of poverty. The results of programs aimed at alleviating extreme poverty in selected “emerging Europe” countries appear to take effect more rapidly compared to larger-scale anti-poverty initiatives.

The poverty gap analysis (PVTA2) shows that although digital and traditional financial inclusion contributes to the reduction of deep poverty, their effectiveness varies. Internet banking (DigT1) exhibits mixed results (from −0.0003 in M1 to −0.0009 in M3) and the impacts are stationary in two of the four models. Online selling (DigT2), on the other hand, has stronger, albeit very small effects (−0.0002 to −0.0005), implying that a 10 % increase in online selling reduces deep poverty by just 0.2–0.5 % (Table 3). The traditional channels, especially bank branches (TFinI2), achieve greater reductions (-0.0001–-0.0009), indicating the historical value of such channels for the poorest. Furthermore, EU membership increases the importance of internet banking (EU × DigT1 = −0.0009 in M3) and diminishes the role of the traditional channel, which reflects the policy changes concerning digital solutions. Importantly, GDP growth (−0.008 in M3) is significantly more influential than financial inclusion alone, with changes in population being insignificant (Table 3). The findings highlight that tackling profound poverty will require a multipronged approach, comprising digital, physical, and economic expansion, in order to achieve a significant reduction.

Membership in the EU promotes online selling of goods and services, though so far it has only proven capable of enhancing internet banking and payments in the newer member states. While EU integration can have a positive effect on aspects that help reduce poverty by increasing the accessibility of conventional bank accounts, it can also undermine the importance of physical bank branches in certain regions. On the one hand, EU membership is associated with increased adoption of online financial instruments, paving the way for improvements in financial inclusion; on the other hand, this has been accompanied by gradual abandonment of the traditional banking system. This is not in line with findings from China, where the usage of financial technology has dramatically reduced poverty, especially in low-income areas (Appiah-Otoo and Song 2021; Ye, Chen and Li 2022). In Africa, too, financial inclusion has been stagnating or even declining as traditional banking decreased but digital finance rose (Asongu 2015; Čihak and Ratna 2020). Comparable situations have been found to exist in low-income countries such as China, India, and Nigeria, where poor, rural populations are difficult to reach through the established routes of finance (Asongu and Asongu 2018; Demirgüç-Kunt and Klapper 2012).

Two issues of vital policy interest become apparent in these modest effect sizes that proved to be statistically significant: (1) Digital financial tools alone do not have the potential to dramatically reduce poverty. (2) Traditional banking access still has an outsized impact compared to the digital tools. The findings of this study also show that online solutions need to be combined with certain traditional, in-person services. In addition, it is important to offer financial education programs and provide targeted support, such as subsidies, enabling vulnerable groups to take part in the digital economy.

Conclusion

Using panel data analysis, this study compares the impact of traditional and digitally driven channels of financial inclusion on poverty reduction from 2004 to 2021, with a focus on 18 “emerging” European economies. It shows that digital financial inclusion produces a range of results regarding poverty reduction outcomes. Poverty levels decrease when individuals use internet banking, while the use of online selling has limited and mixed results. Both traditional finance channels, that is bank accounts and bank branches, generate statistically significant poverty reduction effects in most of the tested models, which provides empirical evidence for the validity of hypothesis H ₁. The presence of physical bank infrastructure remains essential for reducing poverty.

The relationship between the outputs of traditional and digitally driven channels of financial inclusion and poverty is the reverse to that which was anticipated. The study finds that among traditional financial inclusion channels the number of bank accounts per 1,000 adults has a stronger impact on poverty alleviation than the number of commercial bank branches per 100,000 persons. In other words, a 1 % increase in traditional channels of financial inclusion led to a 0.21 % (bank accounts) and 0.16 % (bank branches) lower poverty respectively. On the contrary, the same increase in digitally driven channels yields only a 0.0259 % reduction. Empirically, traditional channels of financial inclusion have a higher impact on reducing poverty than digital channels of financial inclusion, thus hypothesis H ₂ is rejected.

Hypothesis H ₃ is partially confirmed. In fact, the moderating role of EU membership stimulates poverty reduction via both digital and traditional financial inclusion channels. Nevertheless, their respective impacts differ. Internet banking (DigT1) has the largest effect, lowering poverty significantly. Traditional financial services, on the other hand, display more mixed findings, with EU integration enhancing the role of bank accounts (TFinI1) in reducing poverty, but doing little to help poverty reduction by means of bank branches (TFinI2). Moreover, online selling (DigT2) seems to have context-specific effects. These ambiguous impacts imply that poverty alleviation requires specific policy responses that go hand in hand with the transformation of financial technology.

Although diversifying bank units is an efficient way to reduce poverty in non-EU states because it increases accessibility, in EU states, this is less important because the banking network is already available to most citizens. My research indicates that policymakers will need to focus on improving internet banking, in particular in countries that have only just embarked on the EU integration process. They also need to consider the obstacles to the adoption of online retail by underserved populations. They would be well advised to adopt a two-pronged approach to enhancing digital systems in underprivileged regions, while regulatory control should concentrate on poverty alleviation.

The empirical results provide causal evidence of EU membership intensifying the impacts of financial inclusion on poverty reduction. Integration into the EU amplifies the role of instruments such as internet banking (DigT1) and standard bank accounts (TFinI1) in the reduction of poverty by exploiting existing differences during the EU accession process. On the other hand, the findings indicate associational dynamics between financial inclusion channels and poverty alleviation in the 18 “emerging” European countries. Compared to digital channels, traditional forms of financial access (bank branches and bank accounts) have stronger and more stable negative correlations with both incidence and depth of poverty. Thus, the role of physical infrastructure in providing the poor with access to financial services cannot be ignored in the “emerging” European markets. Conversely, it is not entirely clear how digital financial tools play into the reduction of poverty in the 18-country panel, as the relationship between internet banking and the poverty reduction is very limited. Online selling can assist economically vulnerable groups, but does not help the poor themselves, which means that digital financial inclusion tools have mixed results among both low- and high-income individuals.

The findings of this study are limited due to the lack of available data, especially digital finance indicators, including mobile money and ICT-based financial services. This restricts the analysis to classic banking variables. The study does, however, contribute to a better understanding of the financial inclusion–poverty nexus. Future studies would be well advised to (1) restructure existing data to integrate additional variables introduced by financial technology-related processes; (2) use threshold regression to define inclusion thresholds in terms of poverty reduction; and (3) use a wider range of methods to measure and depict emerging digital financial environments and their diverse effects on the population.