Dynamic Financial Inclusion in ASEAN 8: Do Macroeconomics and Financial Technology Matter?

This study aims to estimate the effects of macroeconomic indicators and financial technology on financial inclusion in ASEAN 8 during 2010-2018. There are three financial inclusion indicators, which include debit card ownership (Model 1), credit card ownership (Model 2), and domestic credit to GDP ratio (Model 3). Furthermore, the dynamic panel is applied to demonstrate dynamic financial inclusion models. The findings show that the domestic credit to GDP ratio is influenced by the unemployment rate, inflation, and financial technology. In addition, Model 1 and 2 show that the FEM is a robust model, while Model 3 indicates that REM is a robust model. This study encourages governments in ASEAN 8 to manage macroeconomic indicators progressively and stably to expand equal financial inclusion for the community.


Introduction
Financial inclusion encourages and facilitates all individuals to engage in a broad and integrated financial system (Berry, 2015;Appleyard, Rowlingson, & Gardner, 2016;Salignac, Muir, & Wong, 2016). In general, the definitions of financial inclusion tend to vary or are not universal. Put simply, Lenka and Barik (2018) described that financial inclusion is identical to the process of providing various financial products and services such as deposit and credit facilities, check services, mobile/internet banking and insurance facilities for poor and low-income households at affordable costs. At the high-level conference held in Seoul, South Korea in November 2010, financial inclusion became one of the nine main pillars of the Global Development Agenda (GPFI, 2011). The access to finance through financial inclusion will improve savings among people who are not familiar with formal finance such as farmers, so that they can manage their expenses. Demirguc-Kunt, Klapper and Singer (2017) said that this access is important for people living in the poor category, because the financial inclusion will help them reduce inequality and poverty.
Financial inclusion can be measured by several indicators. The latest financial inclusion indicators have been published by the World Bank since  However, the increase in credit card ownership and domestic credit to GDP ratio is relatively slow. This means that there is a business risk and relatively low financial transactions in ASEAN 8 which become obstacles to the acceleration and expansion of financial inclusion. Similarly, the condition has happened in the development of financial technology. Meanwhile, developments illustrated in macroeconomic indicators show an upward trend. A significant decrease occur in the poverty rate. The development of economic growth, inflation, and unemployment rates tend to be stable.
Based on GDP growth data (annual in %) published by the World Bank, ASEAN countries experience accelerated growth, with an average of 6.02% during 2010 -2018. Le, Chuc, and Taghizadech-Hesary (2019) described that Asia has strong growth. Therefore, policymakers should improve the poor access to financial services to ensure normal growth. On 7-8 April 2016, the Asian Development Bank Institute, the APEC Business Advisory Council, and the Foundation for Development Cooperation held a forum to discuss the issue of Financial Inclusion in the Digital Age. The forum discussed the importance of accessing financial services to individuals and groups to be able to get benefit from broad and integrated financial service products.
Many empirical studies estimate the financial inclusion models both at the level of a country and cross-country analysis. In addition, various approaches or methods have been used to estimate the determinants of financial inclusion. Chikalipah (2017) identified several factors that influenced financial inclusion in Sub-Saharan Africa (SSA) in 2014 using the OLS method. Those factors were literacy, GDP growth rate, population density, infrastructure index, and GNI per capita. Meanwhile, Inoue (2018) employed independent variables such as poverty, real GDP per capita, inflation, and trade openness. Specifically, Lenka and Barik (2018) found that changes in financial inclusion in India did not produce significant growth in rural areas compared to cities. There is a gap of financial inclusion in rural and urban areas which can be caused by a large number of multinational companies in urban areas that drive financial services in urban areas to be more adequate. The development of the financial inclusion index (FII) was conducted by Goel and Sharma (2017) that a value of 0<IFI≤0.4 indicates low financial inclusion, 0.4<FII≤0.6 describes medium financial inclusion, and 0.6<FII≤1 describe high financial inclusion. The findings of the previous empirical study motivated proofing of the influence of macroeconomic and financial technology on financial inclusion in ASEAN 8. This study focuses on dynamic panel modeling of financial inclusion.
This study aims to estimate the effects of macroeconomic indicators and financial technology on financial inclusion in ASEAN 8 during 2010-2018. There are financial indicators which become financial inclusion proxies, namely: debit card ownership, credit card ownership, and domestic credit to GDP ratio. Meanwhile, the selection of the eight ASEAN countries is based on relatively similar developments in economic indicators. Many ASEAN countries are still experiencing problems in accelerating financial inclusion due to limited financial system accessibilities and capabilities at the level of the banking industry and society. Furthermore, 2010 is chosen as an indication of the efforts of ASEAN countries to strengthen the implementation of the banking (financial) integration framework.
This empirical study contributes to the existence of literature in several ways. First, the financial inclusion models consist of three-panel models because it uses three dependent variable or indicators so that it will provide information on the impact of macroeconomic indicators and financial technology on each model properly. Second, dynamic panels are used to estimate the effect of lagged financial inclusion indicators on each estimation model. Third, the financial technology observer is used to determine the impact of the number of users and seekers of financial technology transactions on financial inclusion in ASEAN 8 during 2010-2018.

Previous Empirical Studies
An empirical study of financial inclusion focusing on the relationship between economic growth and the increase of complexity of the financial system was introduced by Goldsmith (1969). More broadly, this empirical study has developed in the issue of financial inclusion (Le et al., 2019). Berry (2015) argued that financial inclusion takes place in response to financialization, by increasing the participation and involvement of individuals in the financial system. Salignac et al. (2016) explained that access to financial services rises the concept of supply and demand for financial inclusion. Basic formal financial services include credit, savings, insurance, payment, and money transfer facilities. Without these services, individuals often use informal financial sources or financial exclusions, which may harm individuals (Inoue, 2018). Furthermore, formal financial inclusion begins with having a deposit account at a bank or other financial service provider, to make and receive payments and saving money (Demirguc-Kunt et al., 2017). Ellis (2007) added that financial market liberalization is not enough to ensure financial inclusion in individuals who are still limited to financial access services. In addition, financial inclusion problems also occur in micro, small, and medium enterprises due to limited access to credit and domestic and foreign markets (ADBI, 2016;Hunter, 2016).
The relation between financial development and economic growth has been widely analyzed in several literatures in economic field (Goldsmith, 1969;Gleb, 1989;King & Levine, 1993;and Fry, 1997). Empirically, research on the relation between the two variables has been done by Sharma (2016), which found that financial inclusion stimulates economic growth in India. The same finding has been explained by Iqbal and Sami (2017). Therefore, financial sector policy reform and innovation need to be carried out appropriately and progressively in India. Furthermore, Lenka and Barik (2018) explained that the expansion of financial inclusion in India was also supported by financial inclusion policies at both the city and village levels. However, financial inclusion growth in cities is higher than in villages. Thus, a significant increase or decrease in economic growth will have significant implications for financial inclusion (Anarfo, Abor, & Osei, 2019). A different thing was found by Chikalipah (2017) that good financial literacy will encourage an increase and expansion of financial inclusion compared to macroeconomic indicators such as economic growth and GDP per capita in Sub-Saharan Africa countries. Empirical development carried out in this study is the selection of GDP growth as one of the determinants of three financial inclusion indicators in ASEAN 8 during 2010-2018 using dynamic panels. In addition, inflation is also a determining factor in financial inclusion in the dynamic panel model.
Meanwhile, the link between financial inclusion and poverty has been made by Lal (2017). His findings showed that financial inclusion through cooperative businesses has a Pandhit Dynamic Financial Inclusion in ASEAN 8: … significant impact on reducing poverty levels in India. Furthermore, Inoue (2018) revealed that financial inclusion and financial depth have a negative and significant effect on poverty in India. This empirical study uses poverty and unemployment rates in the ASEAN 8 under a dynamic panel model.
Empirically, the link between financial inclusion and financial technology has been made by Lashitew, Tulder, and Liasse (2018), Mushtaq and Bruneau (2019)

Dataset
This study selected three indicators as proxies for financial inclusion in ASEAN 8 during 2010-2018 obtained from financial index publications. These three indicators are indicators that are commonly used in the empirical analysis of financial inclusion. Meanwhile, macroeconomic indicators consist of economic growth, inflation, poverty rate, and unemployment rate. The data was obtained from the World Bank publication. Furthermore, this study establishes the indicator of financial technology observer as a proxy for financial technology obtained from Google Trend. ASEAN countries that became the study sample were Indonesia, Malaysia, Thailand, Cambodia, the Philippines, Vietnam, Lao PDR, and Myanmar. These countries have similar characteristic bents in the development of economic and financial indicators. Table 1 informs the research variables. Dependent variables consist of debit card ownership or DC (Model 1), credit card ownership or CC (Model 2), and domestic credit to GDP ratio or CRD (Model 3). Meanwhile, independent variables include economic growth (GDPG), recovery rate (POVR), inflation (INF), unemployment rate (UER), and financial technology observer (FTO). The α0 is the intercept while β1, β2, β3, β4, and β5 are the parameters/slopes of the equation. The values of β1, β2, and β5 are > 0, while β3 and β4 are < 0. Furthermore, the i is the cross-section of ASEAN 8 countries.
Model 3 will estimate the effect of macroeconomic and financial technology (FTO) on the dynamic domestic credit to GDP ratio (CRD). This model is also utilized as a robustness test for Models 1 and 2. Macroeconomic indicators used to consist of economic growth (GDPG), unemployment rate (UER), and inflation (INF). The dynamic panel model to be estimated is as follows: CRDit = α0 + β1CRDit-1 + β2GDPGit + β3UERit + β4INFit + β5FTOit + εit (3a)

Result and Discussion
This empirical study estimates the effects of macroeconomic indicators and financial technology on financial inclusion in ASEAN 8. Financial inclusion is proxied by three indicators namely debit card ownership (DC), credit card ownership (CC), and domestic credit to GDP ratio (CRD). The mean DC value is 572.3225%. It means that on average, an Pandhit Dynamic Financial Inclusion in ASEAN 8: … ASEAN community member has more than 1 debit card. The countries that have relatively low percentages of debit card ownership are Cambodia, Myanmar, and Lao PDR while the countries that have high percentages of debit card ownership are Vietnam, Thailand, the Philippines, Malaysia, and Indonesia. From 2010-2018, the mean CC value in ASEAN 8 was 121.4883%. Countries that have a relatively low CC percentage are Cambodia, Myanmar, and Lao PDR. Meanwhile, the mean CRD value is 67.9301%. Countries that have relatively high CRD percentages are Vietnam, Thailand, and Malaysia. Thus, the development of financial inclusion in Vietnam, Thailand, and Malaysia tend to be more progressive than the five ASEAN countries. Thus, the three countries are relatively difficult to absorb labor in the domestic market compared to other countries in ASEAN 8. Table 3 describes the estimated results of dynamic debit card ownership (DC). DC is an indicator of financial inclusion in the dynamic panel model (Model 1). Based on the Hausman test results, it can be seen that FEM is the best panel model. FEM estimation results show that the DC lag has a significant and positive effect on DC. It means that the development of debit card ownership in ASEAN 8 is currently closely related to the dynamics of debit card ownership in previous periods. Meanwhile, macroeconomic and financial technology indicators have no significant effect. This finding is different from the Pooled OLS and REM estimation results which indicate that macroeconomic indicators (such as economic growth and inflation) and financial technology have a significant effect. However, an increase in economic growth and financial technology led to a decrease in debit card ownership in ASEAN 8. Simply stated, this condition indicates that economic growth and financial technology achieved have not been able to encourage significant and evenly distributed public savings activities for all people. Other indications show that an increase in inflation causes an increase in debit card ownership (Pooled OLS and Random Effects estimation results). It means that people tend to reduce the risk of monetary value at the time of inflation by saving with the hope that they can obtain the appropriate interest rate on savings/deposits. The R-square of the Fixed Effects Model (FEM) is about 0.7959 (within-group). It means that 79.59% of the dependent variable is influenced by variations in the dependent variable. Furthermore, the R-square of cross-sectional estimation is 99.78% (between groups). Meanwhile, the R-square of the overall estimation is 98.60%. Source: The authors' estimation Note: () denotes standard Error [ ] denotes Z-statistics ***, ** and * denote significant levels at 1%, 5% and 10%, respectively A dynamic panel of financial inclusion estimates is carried out on Model 2 to obtain robust estimation results (Table 4). The Hausman test shows that FEM is the right panel model. FEM estimation results describe that credit card ownership (CC) is significantly influenced by lagged of CC and financial technology. The number of credit card ownership (CC) in the previous period led to an increase in the current CC period. Furthermore, financial technology (FTO) has a significant and positive effect on CC. This finding is following the hypothesis developed in Model 2. It means that the higher the community seeks and utilizes financial technology, it will encourage an increase in financial inclusion in ASEAN 8. The parameter of constant also indicates a significant and positive influence. Thus, FEM estimation results are better than Pooled OLS and REM estimation results.
The R-square of FEM is 74.09% (within-group). It means that 74.09% of the dependent variable is influenced by variations in the independent variable. Besides, the R-square of cross-sectional estimation is 99.49% (between groups). Meanwhile, the R-square of the overall estimation is 98.99%.
Previous empirical studies indicate that macroeconomic indicators such as economic growth, poverty (unemployment), and inflation have a significant effect on financial inclusion. Based on Tables 3 and 4, it can be seen that the results of the dynamic panel estimation show that economic growth, poverty, and inflation have no significant effect. For this reason, this empirical study carries out an estimation in Model 3 and obtains a more robust estimation model. Source: The authors' estimation Note: () denotes standard Error [ ] denotes Z-statictics ***, ** and * denote significant levels at 1%, 5% and 10%, respectively Table 5 shows the estimated results of dynamic domestic credit to GDP ratio as one indicator of financial inclusion. The LM test indicates that the results of REM estimation are correct. REM estimation results inform that the domestic credit to GDP ratio (CRD) is influenced by the lag of CRD, unemployment rate, inflation, and financial technology. Moreover, the constant parameter of estimation also has a significant and positive effect. An increase in the domestic credit to GDP ratio (CRD) in the previous period was able to stimulate an increase in the CRD of the current period. This condition indicates the expansion of credit transactions in each ASEAN 8. Furthermore, the increase in the unemployment rate and inflation will have implications for the reduction in CRD. These results are in line with the hypothesis developed in Model 3. For this reason, ASEAN 8 governments are expected to be careful in formulating macroeconomic policies to control the amount of unemployment as well as low and stable inflation rates. However, an increase in financial technology led to a decrease in CRD. This needs to be explored indepth on how people utilize financial technology so that domestic credit transactions are reduced. Furthermore, the estimation results are not following the hypothesis formulated in Model 3 that financial technology will encourage an increase in domestic credit.
The R-square of REM is 89.63% (within-group). It means that 89.63% of the dependent variable is influenced by variations in the independent variable. Furthermore, the Rsquare of cross-sectional estimation is 99.78% (between groups). Meanwhile, the Rsquare of the overall estimation is 99.19%.
Previous empirical studies conducted by Sharma (2016)

Acknowledgement
Many thanks to Mr. Malik Cahyadin from Universitas Sebelas Maret for comments, analytical supports and motivations to improve quality of the study.

Conclusion
Financial inclusion can stimulate more efficient economy, deepen financial markets, and broaden banking activities in the community. This paper estimates the impact of macroeconomic and financial technology on financial inclusion in ASEAN 8. The dynamic panel method is chosen to identify past financial inclusion interactions in the current financial inclusion period. In addition, three financial inclusion indicators are used including debit card ownership, credit card ownership, and domestic credit to GDP ratio. The selection of these indicators is already relevant to previous empirical research. The empirical development that has been carried out is the use of dynamic panel methods and financial inclusion variables.
Model 1 shows that the fixed effect model is more appropriate. The estimation results explain that debit card ownership is significantly influenced by the lag of debit card ownership. Besides, Constant also has a significant effect. Meanwhile, macroeconomic and financial technology variables have no significant effect. This means that the independent variable does not have implications for dynamic financial inclusion in ASEAN 8 during the study period.
Model 2 describes that the fixed effects model is more appropriate. Financial inclusion is a proxy by credit card ownership indicators. Credit card ownership is significantly influenced by lagged credit card ownership, financial technology observer, and constant. The results of this estimation provide a better illustration than Model 1. It means that financial technology has significant implications for dynamic financial inclusion in ASEAN 8 during 2010-2018.
The final model is the random effects model as a more appropriate panel model. Domestic credit to GDP ratio is influenced by the lagged of domestic credit to GDP ratio, unemployment rate, inflation, financial technology observer, and constant. This means that macroeconomic indicators and financial technology have significant implications for financial inclusion in ASEAN 8 during the study period.
This empirical study provides inputs to economic and financial policymakers in ASEAN 8 to keep inflation rates low and stable. Furthermore, governments in the ASEAN 8 region should encourage and facilitate the expansion and acceleration of access to financial technology to the public to accelerate the implementation of financial inclusion on a massive and equitable basis for the wider community. However, this study has limitations in identifying non-economic factors that have significant implications for financial inclusion in ASEAN 8. Thus, further empirical research is expected to develop a model for estimating financial inclusion under non-economic factors. Furthermore, the selection of more appropriate dynamic models can also be used.