Financial Development, Financial Inclusion and Economic
Welfare
By: Jorrit van Veen
s2251191
Supervised by: Prof. Dr. B.W. Lensink
19-01-2018
Submitted for the degrees of: MSc Economics MSc Finance
1 Abstract
This study investigates the influence of financial inclusion and financial development on economic welfare with a cross-country analysis. This study partly explores how these factors apply to the Sub-Saharan African region. Based on a sample of 214 countries, evidence shows that the long-run growth rates for GDP per capita over the period 2011-2017, and the growth rate for welfare-relevant TFP for the period 2011-2014, are unaffected by financial development and financial inclusion. Robustness analysis strengthens these findings. Only control variables are robustly related to economic welfare. Some evidence is found for a negative relationship between financial development and GDP per capita growth for the Sub-Saharan African region. However, when robustness is taken into account this significance disappears.
Keywords: Economic Welfare; Financial Development; Financial Inclusion; Sub-Saharan Africa
3 1 Introduction
Financial inclusion is essential for economic welfare given the social and economic implications. It can be reasoned that financial inclusion belongs to basic needs such as safe water, education and health services (Peachey & Roe, 2004). Among advancing financial development, the World Bank (2015) announced the goal of Universal Financial Access (UFA) for 2020. Financial inclusion is considered key to reducing poverty and boosting prosperity. One of the initiatives to improve financial inclusion is the Brookings Financial and Digital Inclusion Project (FDIP), which includes 26 countries. What is apparent is that most these countries are located in the Sub-Saharan African (SSA) region. According to the Global Findex, 24% of adults in the SSA region have an account at a financial institution compared to an average of 51% worldwide. Besides the heightened attention for financial inclusion, financial development is widely recognized as a relevant driver for economic growth with the potential to improve poverty alleviation (Beck, Demirgüç-Kunt & Levine, 2004). The Global Financial Development Report reports the SSA region to have a financial depth of 20.9 compared to 107.4 for the developed countries, where depth measures private credit by deposit money banks to GDP in percentages (World Bank, 2014). Although financial development and financial inclusion improved over recent years for the SSA region, a vast disparity exists compared to the developed world. Access to financial products is essential as they enable individuals to diversify risks and smooth consumption. Moreover, lack of access induces inefficient forms of savings and insurances. Financial inclusion has potential to reduce poverty and ensure inclusive economic growth (Aportela, 1999; Demirgüç-Kunt & Klapper, 2012). Moreover, financial development indicators demonstrate that a financial structure can support and improve financial transactions. The economy benefits from an efficient allocation of resources. Therefore, it is not surprising these topics receive considerable attention by policymakers and are of interest to researchers.
4
Numerous policies are geared to address financial market imperfections as these are particularly binding on agents that lack collateral, credit history or networks. Addressing these constraints opens up financial opportunities for the poor (Galor & Zeira, 1993). However, given the recent literature, policymakers have to take into consideration that an oversized financial system possibly increases the fragility of the economic system. Notably, the most recent global financial crisis of 2007-2008 comes to mind which has emphasized the risks arising from excessive financial deepening (Arcand, Berkes & Panizza, 2015; Bezemer, Grydaki & Zhang, 2016).
5
available for the SSA region on financial inclusion is limited. Data collection on financial inclusion has been initiated by the World Bank (2011) in 2011 with the Global Findex, and is only collected every three years. This complicates time-series analysis for the SSA region, and therefore this study mainly focuses on exploring the linkages between the financial sector and economic welfare worldwide, based on a cross-country country analysis. With the current limits, the SSA region will be explored by including a dummy variable for the SSA region.
Based on a sample of 214 countries, this study attempts to answer the question how financial inclusion and financial development interact with economic welfare with a cross-country regression analysis. The growth rate for GDP per capita for the period 2011-2017, and growth rate for welfare-relevant TFP are used as indicators for economic welfare. A large body of empirical studies on economic growth determinants exist. The literature is not very explicit which variables belong in the “true” regression. Therefore, robustness analysis is used to explore the significance of the explanatory variables. The regression analysis provides no evidence, which suggests that the growth rate of GDP per capita is affected by financial development and financial inclusion. Similar findings are observed for welfare-relevant TFP. Robustness analysis reinforces these findings as only control variables are robustly related to economic welfare. The analysis for the SSA region provides some evidence for a relationship between financial development and GDP per capita growth. The results suggest financial development negatively influences economic growth. However, when robustness is taken into account this significance disappears. Additionally, none of results for the SSA region indicates that financial inclusion strengthens economic welfare.
6
2 Theory
The literature on economic welfare, financial inclusion, and financial development are discussed in this chapter, which provides the theory for the hypotheses in the last section of this chapter. Furthermore, financial inclusion and financial development are discussed for the SSA region.
2.1 Economic welfare
Economic welfare is concerned with the utility individuals or a group of persons attain due to an extensive range of factors such as income, wealth, health and literacy. It is a general concept, which complicates a simple definition. Samuelson & Nordhaus (2004) define economic welfare as: “The level of prosperity and standard of living of either an individual or a group of persons”. They argue that in the field of economics this comes down to utility gained from achievement or material goods through economic activity. Another definition is given by Hueting (2011), who defines welfare as: ”The satisfaction of wants derived from our dealings with scarce goods”. The emphasis of scarcity is evident and widely used in definitions of economics. Accomplishing economic welfare is limited by for example time, income, and availability of resources, and winds down to optimizing these factors to attain maximum satisfaction.
2.2 Financial development
7
2.2.1 Economic growth and financial development
9 growth in credit increases more rapidly relative to GDP. Recent literature thus suggest that the finance-growth relationship found until the 1990s does not hold for the 1990s and 2000s.
2.2.2 Financial development and inequality
10 2.2.3 Financial development for Sub-Saharan Africa
Because of a disappointing history of economic performance, the financial systems in the Sub-Saharan African (SSA) region has faced several transformations of the banking systems. Especially following a persistent slowdown in economic growth induced financial reforms during the 1980s. Research on the effectiveness of these reforms, however, produces ambiguous (Fowowe, 2013). Some researchers argue the financial systems have benefited from the reforms (Ghirmay, 2004; Mugume, 2007; Tornell, Westermann & Martinez, 2004), however, the majority of the studies observes mixed findings for these reforms on financial deepening and efficiency of banking (Ahmed, 2013; Musonda ,2008). The economies of the SSA region vary widely across several development indicators such as political environment, legal structure, population density, and size. Moreover, high heterogeneity exists for their financial systems in terms of sophistication and depth. The 48 SSA economies are highly underdeveloped, relatively small, and bank dominated with the exception of South Africa and the surrounding countries constituting the Southern African Customs Union and Mauritius (Čihák et al., 2013; Mlachila, Park & Yabara, 2013). The SSA region ranks lower compared to developed countries and most developing countries on nearly all indicators of financial development (Čihák et al., 2013). Figure 1 presents a brief overview of the financial depth of SSA region compared to other regions.
Figure 1: Financial development
This figure presents levels of financial development for the different regions in the world. The levels are computed by taking domestic credit to the private sector (% of GDP) for each region.
0 50 100 150 200
11 Whereas financial depth has improved over recent years, the majority share of assets across all country groups can be attributed to foreign-owned subsidiaries account or assets from state-owned banks. However, these branches do not necessarily contribute greatly to economic growth. Moreover, as financial depth has increased, the SSA regions have failed to catch up relative to other developing regions such as the Middle East and North Africa (Mlachila et al., 2016). Furthermore, the region is characterized by large bank penetration for urban areas, with banks having expanded in other activities such as securities markets, fund management, insurances, which troubles to identify different financial institutions.
2.3 Financial inclusion
12 2.3.1 Financial inclusion for Sub-Saharan Africa
13 Figure 2: Financial inclusion
This figure presents levels of financial inclusion for different dimensions. It compares the SSA region with the world average. Description of ACCESS, USAGE and BARRIER can be found in Appendix 1.
2.4 Hypothesis development
2.4.1 Financial development and economic growth
The literature provides broadly two avenues for which financial development affects economic growth (Ghirmay, 2004). From a savings rate perspective, financial development affects the savings rate, thereby increasing resources available for investment, hence increasing economic growth. This theory stems mainly from the research by McKinnon (1973) and Shaw (1973). The McKinnon-Shaw model stresses the importance of the role for financial institutions to mobilize and aggregate savings. Without institutions, the task of mobilizing savings would be troublesome if not impossible. Connecting savers and investors would be costly due to information asymmetries and transaction costs. Financial institutions provide the service of mobilizing savings on a large scale reducing costs and information asymmetries, enabling confidence with savers to provide their funds. Secondly, from a financial efficiency perspective financial development should increase efficiency of allocation of resources, spurring productivity of investments. This theory has its roots in the studies by Schumpeter (1982), and Goldsmith (1969). The economy not only benefits from volume effects of savings, Goldsmith (1969) argues differentiated allocation of capital expenditures can be considered as an additional advantage of financial development. Moreover, with financial development, the productivity of investments benefits in two ways. Allocation of resources improves due to
14 advancement in processing information relevant to evaluate alternative investment projects and increased opportunities to hedge and diversify risks (Ghirmay, 2004). Although some debates exist on the channels that affect economic growth, more fundamentally no consensus exist whether financial development actually influences economic growth. Whereas some argue financial development is essential for economic growth due to the reasons mentioned above, others argue financial development can be considered a side effect to economic growth, as the economy grows demand for financial products increases. Furthermore, financial development has possible implications for other welfare indicators. Literature exists on the link between inequality and financial development. The literature is inconsistent concerning the direction of this relationship. Greenwood & Jovanovic (1990) theorize a relationship, like the Kuznets curve, between finance and inequality. Inequality increases first as only a few agents have access to finance, and therefore benefit from early stages of development in the financial market. As the development matures, more agents gain access, and benefit from financial development, which decreases inequality. The literature suggests a threshold exists beyond which financial development can disrupt economic growth. The financial deepening possibly increases volatility (Sahay et al., 2015). However, given the relatively low levels of financial development for the Sub-Saharan region, suggests sufficient advancement can be made before reaching such a threshold. For the developed world, more finance not necessarily implies growth. In addition, although caveats lie within the direction of the finance-growth relationship, we expect according to the theories suggested above financial development to have a positive impact on economic growth. Moreover, the study will look into the relationship between financial development and inequality. The following hypothesis will be central to the investigation of the finance-growth relationship:
H1: An increase in financial development positively influences economic growth.
2.4.2 Financial inclusion and economic welfare
15 projects are forgone, which in turn slows down economic growth. Increasing financial inclusion alleviates these inefficiencies (Beck & Demirgüç-Kunt, 2006). Moreover, financial development does not necessarily resolve financing constraints; financial inclusion provides a measure of accessibility to finance for firms (Claessens & Tzioumis, 2006). Secondly, the human-capital theory holds the idea that human capital development is of fundamental importance for economic growth. Human capital is a combination of knowledge and skill accrued due to investment in these particulars, which have economic significance in a similar way as capital (Schulz, 1961). Reducing financial constraints therefore not only affects capital accumulation, access to finance opens up possibilities for households to invest in human capital through education, and effectively become more productive resulting in securing better-paying jobs. Like capital accumulation, human capital accumulation provides agents the tools to be productive. Furthermore, from a firm-behavior perspective financial inclusion affects the economy through positive externalities. Reducing total cost of capital through the reduced necessity of, for example, intermediaries is likely to increase production and expand employment opportunities (Ndlovu, 2017). Besides the effects for incomes, financial inclusion has possible implications for other welfare factors. From a consumption smoothing perspective individuals who have access to finance can smooth consumption over time, according to their time preference, due to the possibility of saving and loaning. Moreover, improved financial inclusion opens up opportunities for individuals to invest in healthcare, education and insurances (Dupas & Robinson, 2012). Financial inclusion also has ramifications from an inequality perspective. Although the literature has been limited on this topic, the World Bank Development Indicators (World Bank 2006, 2008) suggest financial inclusion is potentially related to poverty and inequality. Whereas developed countries show high levels of financial inclusion and low levels of poverty, developing countries have considerably lower levels of financial inclusion and higher poverty levels. Moreover, financial exclusion tends to be high for low-income households (Dymski, 2005). Given the theory above, this study expects financial inclusion to have a positive effect on economic welfare. Moreover, the theory is broad in providing arguments for financial inclusion to impact GDP positively. Besides affecting incomes, financial inclusion is theorized to have implications for other welfare indicators such as inequality, education, and health. Therefore, the following hypothesis will be central to the investigation of the relationship between financial inclusion and economic welfare:
16 3 Methodology
3.1 Sample and data selection
The dataset for this study consists of 214 countries. The starting values are obtained for the year 2011. This is mainly due to the limited data available for the demand side of financial inclusion, which is tracked by the Global Findex since 2011. The dataset on demand side of financial inclusion is updated every three years. Data for the supply side of financial inclusion is obtained from International Monetary Fund’s (IMF) Financial Access Survey (FAS) database (IMF, 2017b). The data on the demand side of financial inclusion is retrieved from the Global Findex (World Bank Group, 2015). The World Bank provides the financial development data with Global Financial Development Database (GFDD), which keeps track of an extensive range of financial development indicators, observing financial depth, which is the size of the financial sector, financial stability and financial efficiency of the intermediaries. The growth rate for GDP per capita is based on the data obtained from the GFDD (World Bank, 2017). Additionally, the Penn World Table (Feenstra, Inklaar & Timmer, 2015) provides a welfare indicator.
3.1.1 Independent variable: financial development
17
can be overlooked. The concerns about allocation of credits are addressed by King & Levine (1993). They make distinctions between financial institutions, which is limited to the distinction between deposit banks and the central bank, due to the data available. The indicator BANK examines to which degree the commercial banks allocates credit compared to the central bank. According to King & Levine (1993), this measure is relevant as banks are more likely to provide investment information services and risk diversification compared to the central bank. Beck et al. (2000) are able to distinguish between three financial institutions based on new datasets and add the group “other financial institutions”. These institutions generally do not incur liabilities, which can function as a means of payment, while providing a role as financial intermediary. This additional distinction provides opportunities for new measures. Secondly, a measure of relative amount of credit going to the private sector is constructed to alleviate the allocation concerns. King & Levine (1993) propose the indicator PRIVY, which is the ratio of claims on the nonfinancial private sector to GDP. This indicator can be thought of credit issued solely to the private sector, while government, government agencies, and public enterprises are not involved. Similarly, Levine (1997) suggest the indicator PRIVATE, which measures the claims on the non-financial private sector to domestic credit. These measures are relevant assuming that financial systems, who provide comparable more credit to private firms, are more engaged in researching firms, pressuring corporate control, offer risk management services, and facilitating transactions.
18 direct role in allocating credit in many countries. Additionally, commercial banks are not the only financial intermediaries; hence, this measure possibly omits considerable cross-country variation in financial development. Ghirmay (2004) argues some weaknesses exist for the measure PRIVATECREDIT. Notably, it is a narrow measure of financial development as it solely focuses on financial development in the banking system. However, for developing countries, financial development has mainly occurred in the banking system. Therefore, PRIVATECREDIT is the preferred choice.
3.1.2 Independent variable: financial inclusion
Financial inclusion is a multidimensional phenomenon, which complicates direct measurement. Widely accepted is the approach by the World Bank (2008) which suggest financial inclusion can be largely captured by access, usage and quality of financial services. Measurement of financial inclusion has been primarily focused on density indicators. The focus of these measures are predominantly on the supply side of financial inclusion, with measures that include number of banks or ATMs per capita, number of bank branches among others. High levels of formal financial services provide an indication of accessibility to formal financial systems. However, these measures provide a limited view on financial inclusion (Cámara & Tuesta, 2014). This issue is addressed by Demirgüç-Kunt & Klapper (2012) with the Global Financial Inclusion Index (Global Findex), introduced by the World Bank in 2011. The database provides information, on the demand side of the financial sector, how people save, borrow, make payments, and manage risk.
19 expensive. Distance measures number of unbanked due to access points, which are perceived by the individuals as too far away. Lastly, documents measures the number of unbanked due to the lack of necessary documents (Cámara & Tuesta, 2014). All the barrier indicators are measured in terms of total population, and are provided by the Global Findex. Based on availability of the variables for a wide range of countries two variables are chosen for every dimension. The description of the variables are provided in table.1
Table 1: Financial inclusion variables
This table contains two variables for each of the financial inclusion dimensions access, usage, and barrier.
Variables Description
ACCESS Automated Teller Machines (ATMs) per 100,000 adults ACCESS1 Automated Teller Machines (ATMs) per 1,000 km2 USAGE Account at a financial institution
USAGE1 Borrowed from a financial institution
BARRIER Borrowed from financial institution, income, poorest 40% BARIER1 Account, income, poorest 40%
Some concerns consists for the variables to measure the same element. Table 2 provides the correlation matrix for the different variables. For each dimension, one variable is chosen which is used for the regression analysis. Based on collinearity concerns the variables ACCESS1, USAGE1 AND BARRIER1 are dropped, and ACCESS, USAGE, and BARRIER will be used as indicators of financial inclusion.
Table 2: Correlation matrix financial inclusion
This table shows the correlations between the different variables for financial inclusion.
20
3.1.3 Dependent variable: economic welfare
21
of the index causes research on intercountry development levels to generate the same outcomes. Hence, the composite index provides no additional findings compared to each of the individual components. Hence, it can be argued that taking solely an index on income, think of index like Purchasing Power Parity (PPP), provides sufficient information on how well a country is doing. A more recent study by Jones & Klenow (2016) attempts to compose a more reliable indicator for economic welfare. They base their index on consumption, leisure and inequality, rather than income and literacy. Compared to the HDI, which is of arbitrary nature, the indicator by Jones & Klenow (2016) is more explicitly grounded in the economic theory of utility maximization. Notably, the study by Jones & Klenow (2016) find that cross-country inequality in welfare is greater compared to income inequality. While Europe is closer to the United States in terms of economic welfare, developing countries lack behind even more.
22 3.1.4 Control variables
23
3.2 The model
This study makes use of cross-sectional data, which comprises a large range of countries. The methodology is based on a vast literature, which investigates economic growth determinants based on cross-country regressions. The literature on cross-country regressions is very broad, and over 50 variables have been found to be related to growth (Levine & Renelt, 1992). The basic methodology consists of the following equation:
ɣ = α + β1x1 + β2x2 + … + βnxn + ϵ (1)
The ɣ is the vector with rates of growth in GDP per capita and welfare-relevant TFP, and vectors x1,...., xn are explanatory variables. The first part of the regression analysis will make
use of this model. Moreover, some of the independent financial variables will be rotated excluding the other financial variables to assess multicollinearity. A typical issue faced by empirical growth economists is which explanatory variables belong in the “true” regression, given that given that they know how the “true” model looks like. Many studies exist on different explanatory variables, and even may find some of these explanatory variables to be significant when some of the control variables of the regression are changed. Studies by Sala-i-Martin (1997) and Leven & Renalt (1992) address this issue by investigating the robustness of the explanatory variables. They use the following model:
ɣ = αj + βyjy + βzjƶ + βxjxj + ϵ (2)
The equation differs somewhat to the extent that y is a vector of variables which appears in all of the regressions. These are considered fixed variables, which have been widely used in the literature and are robustly related to the dependent variable. The variable ƶ is the variable of interest, and xj is a vector of up to three variables taken from a pool of control variables (𝒳).
The subscript j indicates the different models, with βzj the estimates of interest to this research.
In an attempt to carefully investigate the relationship between financial determinants and economic welfare, regressions are run for several different control variables. The model specification (2) will be used for robustness checks, in which model j estimates regressions that combine variables, which are fixed for all the regressions (y), the variable of interest (ƶ), and additional control variables (xj) selected from the pool 𝒳. The argument for keeping three
24
has at least seven variables on the right-hand-side of the regression (Sala-i-Martin, 1997). The previous section has discussed several control variables frequently mentioned in the literature and elaborated on the theory behind these variables. They provide the basis for the regression analysis. It is important that the fixed variables are robustly related to the dependent variable, meaning that they systematically matter in regressions run in earlier studies and that they are available for the beginning of the period to avoid endogeneity. For the fixed variables Levine & Renelt(1992) use initial level of income, the investment rate, the secondary school enrollment rate, and the rate of population growth, whereas Sala-i-Martin (1997) chose initial level of income, life expectancy, and primary-school enrollment rate. Since most studies on the growth literature include the variable initial level of income and find it to be significant, it is fairly straightforward to select this variable as one of the fixed variables. Additionally, life expectancy and primary and secondary enrollment rate are chosen as fixed variables, as these variables are widely used in the literature for measuring initial stock of human capital. The variables that have been discussed in section 3.1.4., and not have been chosen for vector y are included in the pool 𝒳. These are the variables government size, openness of trade, inflation and investment ratio, and will be used to check for the robustness of the variables of interest. The variables of interest are the financial inclusion dimensions access, barrier and usage, and the variable privatecredit which measures financial development.
Whereas Levine & Renelt (1992) use an extreme-bound test to identify the robustness of the empirical relations in the economic growth literature, Sala-i-Martin (1997) assigns confidence levels to the variables. The extreme-bound test for, for example, variable z relies on the extreme bounds, in which the upper bound is defined as the largest value βzj + 2𝜎zj , and the lower
bound is defined as the lowest value βzj - 2𝜎zj . When the lower bound is negative and the upper
25
This results in a total of 32 regressions. The estimates for the different regressions are used to assign confidence to the different variables. The average mean (𝛽̂) and standard deviation (𝜎𝑧 ̂) 𝑧2 are computed for M models, with weights (𝜔𝑧𝑗), according to equation (3) and (4). These averages are used to calculate the cumulative distribution function [CDF(0)] using the standard normal distribution. Note that zero divides the area under the density in two, and for simplicity, the larger of the two areas will be called CDF(0). The unweighted averages are chosen over the weighted averages as models with endogenous variables have a better fit, which raises corresponding weights, and therefore are susceptible of dominating the estimates (Sala-i-Martin, 1997). The unweighted averages do not suffer from this problem and are, therefore, preferred to compute the confidence levels.
𝛽̂ = ∑𝑧 𝑀𝑗=1𝜔𝑧𝑗𝛽𝑧𝑗 (3) 𝜎̂ = ∑ 𝜔𝑧𝑗𝑧2 𝜎
𝑧𝑗2 𝑀
26 4 Empirical results
4.1 Descriptive statistics
Table 3 provides the descriptive statistics of the variables in terms of mean, median, minimum, maximum, standard deviation and skewness. The variables are winsorized to limit extreme variables and avoid possible spurious outliers. Like initial level of GDP per capita, other variables are transformed when skewness suggest non-normality, where the interval [-2,2] is a frequently used indicator for normality (Trochim & Donnelly, 2006). Although no assumptions are made for linear models on the distribution of the independent variables, the transformations aid to reduce skewness and clarify the relationship with the dependent variable, which should benefit the predictive ability of the model. The variables for investment, government size, inflation and openness of trade experience positive skewness, and are subjected to the logarithmic transformation. The variable for human capital is negatively skewed and is transformed by a power function. The resulting variables fall within the interval [-2,2]. Appendix 3 provides the descriptive for the non-transformed variables.
The average growth level of GDP per capita for the whole sample over the period 2011-2017 amounts to 1.57%, whereas the growth for welfare-relevant TFP, for the period 2011-2014, is negative with -0.088%. The variation for both dependent variables is comparable with a similar minimum, maximum and standard deviation. The financial inclusion dimensions access (41.34), usage (46.09) and barrier (8.459) are displaying large differences, and likewise, the financial development (53.03) vary widely with a minimum of 4.359 and a maximum of 202.3. The control variables and initial level of GDP per capita (8.584) vary less, with the exception of the investment ratio (24.72) which has a relatively high standard deviation compared to the mean.
27
GDP, life expectancy, financial development the financial inclusion dimension access. Additionally, the correlations between life expectancy and initial GDP per capita, and the financial dimension access and initial GDP per capita exceeds the threshold of 0.7. To address these multicollinearity concerns additional regressions are run to check for the robustness of the estimates.
Table 3: Descriptive statistics
This table contains descriptive statistics for the dependent variables GDPC and WTFP, the independent variables of interest ACCESS, USAGE, BARRIER and PRIVATECREDIT, and the control variables. Appendix 1 provides an explanation of the variables.
VARIABLES N mean min max sd kurtosis skewness
28 Table 4: Correlation matrix
This table contains the correlation coefficients between the major variables, namely the dependent variables GDPC and WTFP, the variables explanatory variables PRIVATECREDIT, ACCESS, USAGE, BARRIER, and the control variables.
GDPC WTFP IGDP EDUCA TION HEALT H PRIVAT ECRE. ACCES S USAGE BARRI ER GOVER NMENT INVEST MENT INFLAT ION OPENN ESS GDPC 1.00 WTFP 0.36 1.00 IGDP -0.23 -0.12 1.00 EDUCATION 0.21 0.20 0.48 1.00 HEALTH -0.03 -0.08 0.81 0.58 1.00 PRIVATECREDIT 0.01 -0.09 0.68 0.39 0.67 1.00 ACCESS -0.14 -0.11 0.76 0.34 0.68 0.68 1.00 USAGE -0.09 -0.09 0.86 0.44 0.73 0.75 0.76 1.00 BARRIER 0.10 0.05 0.42 0.37 0.48 0.44 0.37 0.49 1.00 GOVERNMENT -0.10 0.01 0.28 0.22 0.22 0.25 0.27 0.45 0.04 1.00 INVESTMENT 0.26 -0.17 -0.06 0.21 0.00 -0.04 -0.09 -0.05 0.05 0.05 1.00 INFLATION 0.08 0.03 -0.43 -0.21 -0.32 -0.40 -0.34 -0.36 -0.07 -0.26 0.11 1.00 OPENNESS -0.01 0.12 0.30 0.19 0.22 0.21 0.09 0.30 0.14 0.15 0.13 -0.14 1.00 4.2 Regression results
This section provides the results from the regressions used for evaluating the interplay between economic welfare and financial determinants. Firstly, the results for the cross-country analysis will be discussed. After this, the robustness of these results will be analyzed. Lastly, in section 4.2.3 the focus is redirected to the results for the SSA region.
4.2.1 Cross-country Results
29
What is immediately observed from table 5 is that initial GDP per capita, human capital, and investment ratio are significant for all models when taking a 10% level. For some models, they are significant at 5% and the 1% level. The sign for initial GDP per capita is negative which is in accordance with the convergence theory. Furthermore, the signs of the estimates for human capital and the investment ratio are in accordance with theory, that they positively impact economic growth. What is apparent from the different models is that from the financial inclusion dimensions, not a single estimate suggest that financial inclusion positively affects economic growth. Moreover, the estimated coefficients are close to zero and negative for some of the estimates of financial access and financial barriers. Similarly, the estimate for financial development is not significant for any of the regressions and close to zero. Furthermore, life expectancy, government size, and openness of trade are significant for some of the models. This is in line with the observation of Sala-i-Martin (1997) that changing some explanatory variables results in different explanatory variables to be significant. The only control variable, which is not significant for any of the models, is inflation. In the next section the robustness is explored and confidence is assigned to the results. The goodness of fit estimates of the models are similar, with model (3) having the lowest R-squared. The models are also subject to the Jarque-Bera test for normality, to check whether the residuals are normally distributed. The results for most models suggest this is no concern. Only model (2) suggest non-normality as the null-hypothesis of normality is rejected at the 5% level. Overall, the results from models 1-5 do not support hypothesis 1 that financial development improves economic welfare. Likewise, no evidence is found in support of hypothesis 2 that financial inclusion positively impacts economic welfare.
30
financial inclusion estimates fail to support the theory that financial inclusion positively impact economic welfare.
4.2.2 Robustness of the cross-country results
31
estimates for financial inclusion and financial development are not robustly related to growth in welfare-relevant TFP.
Summarizing, whereas there is evidence in support of the relationship between economic welfare and the control variables, this is not the case for financial inclusion and financial development. This suggests that financial inclusion and financial development are not important channels to improve economic welfare. However, this need not to imply no effects exist. The different dimensions of financial inclusion are captured by separate variables; however, the variables chosen may lack explanatory power. The Global Findex provides a comprehensive list of variables; however, the measure for financial inclusion is restricted as not all variables are observed for all countries. The literature provides some guidance to which variables to incorporate (Cámara & Tuesta, 2014); however, attempts to compose a multidimensional index are scarce or incomplete. In addition, another weakness is that the measure for financial development is predominantly bank based (Ghirmay, 2004). This measure possibly omits essential information on financial development. Similarly, the measures used for economic welfare focuses mainly on the income part of economic welfare. Whereas welfare-relevant TFP attempts to account for other factors. However, different indicators for welfare are potentially superior.
4.2.3 Results Sub-Saharan Africa
Tables 8 and 9 presents the results of the regressions, which investigate whether differences exist for the SSA region, compared to the rest of the world. The dummy variable captures the effects for the SSA region. Models 17-21 provide the results with GDP per capita growth as the dependent variable, and models 22-26 investigate the results with welfare-relevant TFP growth as the dependent variable. Models (17) and (22) includes all the financial variables, control variables, the dummy variable for the SSA region, the interactions for the SSA region, and the financial variables. For models 18-21 and 23-26 the financial variables are rotated.
32
financial development actually negatively affects economic growth for the SSA region. However, when financial development is regressed without the financial inclusion
dimensions, this interaction is not significant. Additionally, the financial inclusion estimates are low and none of the interactions with the SSA region are significant. The results for the goodness of fit of the models slightly improve when compared to regressions excluding the dummy variables. However, this increase is limited and likely a result of adding variables to the regression, which will tend to increase the R-squared. To examine whether the residuals of the regressions are normally distributed the Jarque-Bera test is computed. The only model that raises some concerns is model (18), which reject the null hypothesis of normality at the 5% level.
Table 9 presents the regression results with welfare-relevant TFP growth as the dependent variable. Similar to the results in table 6 initial level of GDP per capita negatively impacts the welfare-relevant TFP growth for models 22-26. Additionally, human capital is positively related to the dependent variable for models 23-26. Other factors seem not to matter, as there is no significant effect on the dependent variable. Both the estimates for financial inclusion and financial development are low and not significantly related to welfare-relevant TFP growth. Additionally, the estimates seem to be unaffected when taking into account the SSA region. The goodness of fit is slightly higher compared to the regressions that do not take into account the SSA region; however, the increase can be attributed to the added variables that are likely to increase the R-squared. All the models (22)-(26) comply with null-hypothesis of Jarque-Bera test for normality of the residuals; hence, the estimates can be expected to be approximately correct.
Summarizing, no evidence is found for financial inclusion to positively impact economic welfare for the SSA region, as none of the financial inclusion dimensions are significantly related to the economic welfare variables. Moreover, weak evidence suggests financial development impacts economic welfare for the SSA region. However, the findings suggest that the effects are negative, and not robust considering the fact that only one of the
33 Table 5: Cross-Country regression results GDP per capita growth
This table shows the estimates for the cross-country OLS regressions of the financial variables and the control variables on the growth rate of GDP per capita. Columns 2-5 rotate the financial variables. The standard errors are presented in the parentheses. The symbols ***, **, * denote statistical significant at the 1%, 5% and 10% levels, respectively. Appendix 1 provides an explanation of the variables.
(1) (2) (3) (4) (5) Dependent var. GDPC GDPC GDPC GDPC GDPC IGDP -0.00697* -0.0103*** -0.00925*** -0.00719* -0.00544** (0.00390) (0.00285) (0.00308) (0.00372) (0.00273) EDUCATION 0.0405** 0.0532*** 0.0499*** 0.0379** 0.0366** (0.0178) (0.0176) (0.0178) (0.0172) (0.0174) HEALTH 0.000651 0.000831* 0.00108** 0.000549 0.000541 (0.000467) (0.000476) (0.000488) (0.000451) (0.000457)
PRIVATECREDIT 4.08e-05 8.50e-05
(7.74e-05) (6.60e-05)
ACCESS -6.87e-05 -4.18e-05
34 Table 6: Cross-country regression results welfare-relevant TFP growth
This table shows the estimates for the cross-country OLS regressions of the financial variables and the control variables on the growth rate of welfare-relevant TFP. Columns 7-10 rotate the financial variables. The standard errors are presented in the parentheses. The symbols ***, **, * denote statistical significant at the 1%, 5% and 10% levels, respectively. Appendix 1 provides an explanation of the variables. (6) (7) (8) (9) (10) Dependent var. WTFP WTFP WTFP WTFP WTFP IGDP -0.00984* -0.00924** -0.0108*** -0.0113** -0.0131*** (0.00535) (0.00378) (0.00389) (0.00512) (0.00358) EDUCATION 0.0663** 0.0601** 0.0536* 0.0646** 0.0654** (0.0308) (0.0301) (0.0301) (0.0298) (0.0291) HEALTH 0.00123** 0.00106* 0.00111* 0.00124** 0.00115** (0.000574) (0.000568) (0.000572) (0.000552) (0.000554)
PRIVATECREDIT -6.43e-05 -5.26e-05
(8.99e-05) (7.88e-05)
ACCESS -2.46e-05 6.51e-06
(9.01e-05) (7.62e-05)
USAGE -8.83e-05 -4.94e-05
35 Table 7: Main results of the regressions
This table provides the results for the robustness analysis. Columns (11) and (14) present the average estimates. Columns (12) and (14) display the average standard deviations. The last columns (13) and (16) presents the confidence estimates.
36 Table 8: SSA dummy regression results GDP per capita growth
This table shows the estimates for the cross-country OLS regressions for growth rate of GDP per capita, with a dummy for the SSA region. Columns 18-21 rotate the financial variables. The standard errors are presented in the parentheses. The symbols ***, **, * denote statistical significant at the 1%, 5% and 10% levels, respectively. Appendix 1 provides an explanation of the variables.
(17) (18) (19) (20) (21) VARIABLES GDPC GDPC GDPC GDPC GDPC IGDP -0.00934** -0.0102*** -0.00925*** -0.00740* -0.00569** (0.00435) (0.00289) (0.00319) (0.00378) (0.00276) EDUCATION 0.0402** 0.0568*** 0.0501*** 0.0328* 0.0323* (0.0189) (0.0186) (0.0189) (0.0178) (0.0179) HEALTH 0.00144* 0.000895 0.00134** 0.000856 0.000820 (0.000747) (0.000642) (0.000639) (0.000641) (0.000637) DSSA 0.0143 0.00534 0.00600 -0.00119 0.000653 (0.0148) (0.00991) (0.00959) (0.0115) (0.0122)
PRIVATECREDIT 4.98e-05 9.37e-05
(8.26e-05) (6.87e-05)
ACCESS -6.26e-05 -4.69e-05
(7.88e-05) (7.18e-05)
USSAGE 9.30e-05 6.96e-05
37 Table 9: SSA dummy regression results welfare-relevant TFP growth
This table shows the estimates for the cross-country OLS regressions for growth rate of welfare-relevant TFP, with a dummy for the SSA region. Columns 18-21 rotate the financial variables. The standard errors are presented in the parentheses. The symbols ***, **, * denote statistical significant at the 1%, 5% and 10% levels, respectively. Appendix 1 provides an explanation of the variables.
(22) (23) (24) (25) (26) VARIABLES WTFP WTFP WTFP WTFP WTFP IGDP -0.00753 -0.00899** -0.0101** -0.0114** -0.0122*** (0.00615) (0.00384) (0.00406) (0.00509) (0.00360) EDUCATION 0.0522 0.0571* 0.0543 0.0561* 0.0665** (0.0363) (0.0339) (0.0336) (0.0336) (0.0313) HEALTH 0.000282 0.000668 0.000544 0.000220 0.000244 (0.000989) (0.000819) (0.000819) (0.000788) (0.000799) DSSA -0.0255 -0.00931 -0.0108 -0.0282 -0.0185 (0.0234) (0.0154) (0.0140) (0.0169) (0.0185)
PRIVATECREDIT -5.64e-05 -4.68e-05
(9.73e-05) (8.31e-05)
ACCESS -2.58e-05 1.05e-05
38 5 Conclusion
Summarizing, the attention for financial inclusion is increasing. The goal of Universal Financial Access (UFA) for 2020 set by the World Bank (2015) demonstrates that financial inclusion is increasingly recognized as an important driver for economic welfare. To a certain extend financial inclusion belongs to basic needs such as safe water, education and health services. Financial inclusion enables individuals to diversify risks and smooth consumption through financial products such as savings and insurances. Besides financial inclusion, financial development is widely believed to be a driver of economic prosperity. Well-developed financial systems have higher savings rates. More resources are available as institutions mobilize savings, and connect investor and savers. Additionally, financial development support efficiency of allocation of resources to the most productive projects. Contrary to research in the 1990s, the recent literature fails to discover a relationship between financial development and economic growth. In addition, the research on the interaction between financial inclusion and aggregates such as GDP is limited. This research explores these areas, by first exploring the interplay between financial development and economic growth as well as welfare-relevant TFP, and secondly investigate how financial inclusion affects these aggregates.
39
40 6 References
Abu‐Bader, S., & Abu‐Qarn, A. S. (2008). Financial development and economic growth: empirical evidence from six MENA countries. Review of Development Economics, 12(4), 803-817.
Agbetsiafa, D. (2004). The finance growth-nexus: evidence from Sub-Saharan Africa. Savings and
Development, 271-288.
Ahmed, A. D. (2013). Effects of financial liberalization on financial market development and economic performance of the SSA region: An empirical assessment. Economic Modelling, 30, 261-273.
Aportela, F. (1999). Effects of financial access on savings by low-income people.
Arcand, J. L., Berkes, E., & Panizza, U. (2015). Too much finance?. Journal of Economic Growth, 20(2), 105-148.
Baliamoune-Lutz, M. (2008). Financial development and income in North Africa. International
Advances in Economic Research, 14(4), 422-432.
Banerjee, A. V., & Newman, A. F. (1993). Occupational choice and the process of development.
Journal of political economy, 101(2), 274-298.
Beck, T., & Demirguc-Kunt, A. (2006). Small and medium-size enterprises: Access to finance as a growth constraint. Journal of Banking & finance, 30(11), 2931-2943.
Beck, T., Demirgüç-Kunt, A., & Levine, R. (2004). Finance, inequality, and poverty: cross-country evidence (No. w10979). National Bureau of Economic Research.
Beck, T., Demirgüç-Kunt, A., & Levine, R. (2007). Finance, inequality and the poor. Journal of
economic growth, 12(1), 27-49.
Beck, T., Levine, R., & Loayza, N. (2000). Finance and the Sources of Growth. Journal of financial
economics, 58(1), 261-300.
Becker, G. S. (1994). Human capital revisited. In Human Capital: A Theoretical and Empirical Analysis
with Special Reference to Education (3rd Edition) (pp. 15-28). The University of Chicago Press.
Bellemare, M. F., Masaki, T., & Pepinsky, T. B. (2017). Lagged Explanatory Variables and the Estimation of Causal Effect. The Journal of Politics, 79(3), 000-000.
Bennett, R. L. (1963). Financial innovation and structural change in the early stages of industrialization: Mexico, 1945–59. The Journal of Finance, 18(4), 666-683.
Bezemer, D., Grydaki, M., & Zhang, L. (2016). More mortgages, lower growth?. Economic Inquiry, 54(1), 652-674.
Bloom, D. E., Canning, D., & Sevilla, J. (2004). The effect of health on economic growth: a production function approach. World development, 32(1), 1-13.
Bolt, J., Lensink, B.W. & Raster, T. (2017). Structural transformation and inequality in Africa: an investigation into the Kuznets curve.
Bruhn, M., & Love, I. (2014). The real impact of improved access to finance: Evidence from Mexico. The Journal of Finance, 69(3), 1347-1376.
41 Čihák, M., Demirgüč-Kunt, A., Feyen, E., & Levine, R. (2013). Financial development in 205 economies, 1960 to 2010 (No. w18946). National Bureau of Economic Research.
Claessens, S., & Tzioumis, K. (2006). Measuring firms’ access to finance. World Bank, 1-25.
Dev, S. M. (2006). Financial inclusion: Issues and challenges. Economic and political weekly, 4310-4313.
Demirgüç-Kunt, A., & Klapper, L. F. (2012). Measuring financial inclusion: The global findex database.
Demirgüç-Kunt, A., & Klapper, L. (2013). Measuring financial inclusion: Explaining variation in use of financial services across and within countries. Brookings Papers on Economic Activity, 2013(1), 279-340.
Dormann, C. F., Elith, J., Bacher, S., Buchmann, C., Carl, G., Carré, G., & Münkemüller, T. (2013). Collinearity: a review of methods to deal with it and a simulation study evaluating their performance.
Ecography, 36(1), 27-46.
Dupas, P., & Robinson, J. (2012). The (hidden) costs of political instability: Evidence from Kenya's 2007 election crisis. Journal of Development Economics, 99(2), 314-329.
Dymski, G. A. (2005). Financial globalization, social exclusion and financial crisis. International
Review of Applied Economics, 19(4), 439-457.
Edward, S. (1973). Financial deepening in economic development. Business and Economics Review. Eryılmaz, F., Bakır, H., & Mercan, M. (2014). Financial Development and Economic Growth: Panel Data Analysis. Handbook of Research on Strategic Developments and Regulatory Practice in Global
Finance, 233.
Feenstra, R. C., Inklaar, R., & Timmer, M. P. (2015). The next generation of the Penn World Table.
The American Economic Review, 105(10), 3150-3182, available for download at www.ggdc.net/pwt.
Accessed on 12-12-2017
Fowowe, B. (2013). Financial Liberalization in Sub‐Saharan Africa: What Do We Know?. Journal of
Economic Surveys, 27(1), 1-37.
Galor, O., & Zeira, J. (1993). Income distribution and macroeconomics. The review of economic studies, 60(1), 35-52.
Ghirmay, T. (2004). Financial Development and Economic Growth in Sub‐Saharan African Countries: Evidence from Time Series Analysis. African Development Review, 16(3), 415-432.
Goldsmith, R. W.(1969). Financial Structure and Development. New Haven, CN: Yale University Press.
Greenwood, J., & Jovanovic, B. (1990). Financial development, growth, and the distribution of income.
Journal of political Economy, 98(5, Part 1), 1076-1107.
42 Hueting, R. (2011). The future of the Environmentally sustainable national income. Ökologisches
Wirtschaften-Fachzeitschrift, 25(4), 30.
IMF. (2016). Regional Economic Outlook: Sub-Saharan Africa Multispeed Growth. Washington, D.C: International Monetary Fund, Publication Services.
IMF. (2017a). World Economic Outlook Update: A Shifting Global Economic Landscape. Washington, D . C .: International Monetary Fund.
IMF. (2017b). Financial Access Survey. Retrieved from: http://data.imf.org/?sk=E5DCAB7E-A5CA-4892-A6EA-598B5463A34C. Accessed on 15-12-2017.
Jones, C. I., & Klenow, P. J. (2016). Beyond GDP? Welfare across countries and time. The American
Economic Review, 106(9), 2426-2457.
Jung, W. S. (1986). Financial development and economic growth: international evidence. Economic
Development and cultural change, 34(2), 333-346.
Levine, R. (1997). Financial development and economic growth: views and agenda. Journal of economic literature, 35(2), 688-726.
Levine, R. (2005). Finance and growth: theory and evidence. Handbook of economic growth, 1, 865-934.
Levine, R., Loayza, N., & Beck, T. (2000). Financial intermediation and growth: Causality and causes. Journal of monetary Economics, 46(1), 31-77.
Levine, R., & Renelt, D. (1992). A sensitivity analysis of cross-country growth regressions. The
American economic review, 942-963.
Kindleberger, C. P. (1978), Manias, Panics, and Crashes: A History of Financial Crises, Basic Books, New York.
King, R. G., & Levine, R. (1993). Finance and growth: Schumpeter might be right. The quarterly journal of economics, 108(3), 717-737.
Kuznets, S. (1937). Introduction to" National Income and Capital Formation, 1919-1935". In National
Income and Capital Formation, 1919-1935 (pp. 1-2). NBER.
McGillivray, M. (1991). The human development index: yet another redundant composite development indicator?. World Development, 19(10), 1461-1468.
McKinnon, R. I. (1973). Money and Capital in Economic Development. Washington, DC: Brookings Institution, 1973.
Meier, G., & Seers, D(1984). Pioneers in development. New York, NY: Oxford U. Press.
Minsky, H. P. (1974), "The modeling of financial instability: An introduction," in Modelling and
Simulation, Vol. 5, Proceedings of the Fifth Annual Pittsburgh Conference, Instruments Society of
America, pp. 267–72
43 Mlachila, M., Cui, L., Jidoud, A., Newiak, M., Radzewicz-Bak, B., Takebe, M., & Zhang, J. (2016).
Financial Development in Sub-Saharan Africa: Promoting Inclusive and Sustainable Growth.
International Monetary Fund.
Mugume, A. (2007). Market structure and performance in Uganda's banking industry. African
Econometrics Society.
Ndlovu, G. (2017) Access to financial services: Towards an understanding of the role and impact of financial exclusion in Sub-Saharan Africa.
Nordhaus, W. D., & Tobin, J. (1972). Is growth obsolete?. In Economic Research: Retrospect and
Prospect, Volume 5, Economic Growth (pp. 1-80). Nber.
Odhiambo, N. M. (2007). Supply‐leading versus demand‐following hypothesis: Empirical evidence from three SSA countries. African Development Review, 19(2), 257-280.
Reed, W. R. (2013). A note on the practice of lagging variables to avoid simultaneity (No. 13/32). Ram, R. (1986). Government size and economic growth: A new framework and some evidence from cross-section and time-series data. The American Economic Review, 76(1), 191-203.
Rioja, F., & Valev, N. (2004). Does one size fit all?: a reexamination of the finance and growth relationship. Journal of Development economics, 74(2), 429-447.
Robinson, J. (1952). The Generalization of the General Theory. The rate of interest, and other essays. London: Macmillan, pp. 67-142.
Rousseau, P. L., & Wachtel, P. (2011). What is happening to the impact of financial deepening on economic growth?. Economic Inquiry, 49(1), 276-288.
Sahay, R., Čihák, M., N'Diaye, P., & Barajas, A. (2015). Rethinking financial deepening: Stability and growth in emerging markets. Revista de Economía Institucional, 17(33), 73-107.
Sala-i-Martin, X. X. (1997). I just ran two million regressions. The American Economic Review, 178-183.
Samuelson, P. A., & Nordhaus, W. D. (2004). Economics, chapter 18. Protecting the Environment, McGraw-Hill.
Schultz, T. W. (1961). Investment in human capital. The American economic review, 51(1), 1-17. Schumpeter, J. A., (1982). The Theory of Economic Development: An Inquiry into Profits, Capital, Credit, Interest, and the Business Cycle (1912/1934). Transaction Publishers.–1982.–January, 1, 244. (Original work published in 1911)
Shaw, E.S. (1973) Financial Deepening in Economic Development. New York: Oxford University Press.
Solow, R. M. (1957). Technical change and the aggregate production function. The review of Economics
and Statistics, 312-320.
44 Trochim, W., & Donnelly, J. (2006). The research knowledge methods base. Cincinnati, OH: Atomic
Dog Publishing.
UNDP (1990). Human Development Report 1990. New York, NY: Oxford U. Press. Verbeek, M. (2008). A guide to modern econometrics. John Wiley & Sons.
World Bank. 2014. Global Financial Development Report 2014: Financial Inclusion. Washington, DC: World Bank.
World Bank. (2015, April 17). Spring Meetings Flagship Event: Universal Financial Access 2020. Retrieved from: http://www.worldbank.org/en/events/2015/04/06/universal-financial-access-2020. Accessed on 10-11-2017
World Bank. (2017). Global Financial Development Database. Retrieved from:
http://databank.worldbank.org/data/reports.aspx?source=global-financial-development. Accessed on 12-12-2017
World Bank Group. (2015). Global Financial Inclusion (Global Findex) Database.
Retrieved from: http://datatopics.worldbank.org/financialinclusion/. Accessed on 10-12-2017
45 APPENDIX 1: Variable overview
This table gives an overview of the variables included in the analysis. Variables Description
GDPC Long-run growth rate GDP per capita (constant 2010 US$) period 2011-2017 WTFP Long-run growth rate welfare-relevant productivity(constant , 2011 value =1
for all countries) period 2011-2014
IGDP Level of initial GDP per capita, year 2011
EDUCATION School enrollment, primary and secondary as % of the population HEALTH Life expectancy at birth, total (years)
PRIVATECREDIT Domestic credit to private sector (% of GDP)
ACCESS Automated Teller Machines (ATMs) per 100,000 adults USAGE Account at a financial institution
BARRIER Borrowed from financial institution, income, poorest 40%
GOVERNMENT General government final consumption expenditure (% of GDP) INVESTMENT Gross capital formation (% of GDP)
INFLATION Inflation, consumer prices (annual %) OPENNESS Import plus Export (% of GDP)
46 APPENDIX 2: Country overview
This table present the countries included in the analysis.
Country List
Afghanistan Cayman Islands Gibraltar Lesotho Norway Sudan Albania
Central African
Republic Greece Liberia Oman Suriname
Algeria Chad Greenland Libya Pakistan Swaziland
American Samoa Chile Grenada Liechtenstein Palau Sweden
Andorra China Guam Lithuania Panama Switzerland
Angola Colombia Guatemala Luxembourg Papua New Guinea Syrian Arab Rep. Antigua and
Barbuda Comoros Guinea Macao SAR, China Paraguay Taiwan Argentina Congo, Dem. Rep. Guinea-Bissau Macedonia, FYR Peru Tajikistan Armenia Congo, Rep. Guyana Madagascar Philippines Tanzania
Aruba Costa Rica Haiti Malawi Poland Thailand
Australia Cote d'Ivoire Honduras Malaysia Portugal Timor-Leste Austria Croatia
Hong Kong SAR,
China Maldives Puerto Rico Togo
Azerbaijan Cuba Hungary Mali Qatar Tonga
Bahamas, The Cyprus Iceland Malta Romania Trinidad and Tobago Bahrain Czech Republic India Marshall Islands Russian Federation Tunisia
Bangladesh Denmark Indonesia Mauritania Rwanda Turkey
Barbados Djibouti Iran, Islamic Rep. Mauritius Samoa Turkmenistan Belarus Dominica Iraq Mexico San Marino Turks and Caicos Is. Belgium Dominican Republic Ireland Micronesia Sao Tome Tuvalu
Belize Ecuador Isle of Man Moldova Saudi Arabia Uganda
Benin Egypt, Arab Rep. Israel Monaco Senegal Ukraine
Bermuda El Salvador Italy Mongolia Serbia
United Arab Emirates Bhutan Equatorial Guinea Jamaica Montenegro Seychelles United Kingdom Bolivia Eritrea Japan Morocco Sierra Leone United States Bosnia Herzegovina Estonia Jordan Mozambique Singapore Uruguay Botswana Ethiopia Kazakhstan Myanmar Slovak Republic Uzbekistan
Brazil Faroe Islands Kenya Namibia Slovenia Vanuatu
British Virgin
Islands Fiji Kiribati Nauru Solomon Islands Venezuela, RB Brunei Darussalam Finland
Korea, Dem.
People’s Nepal Somalia Vietnam
Bulgaria France Korea, Rep. Netherlands South Africa Virgin Islands (U.S.) Burkina Faso French Polynesia Kosovo New Caledonia South Sudan West Bank and Gaza
Burundi Gabon Kuwait New Zealand Spain Yemen, Rep.
Cabo Verde Gambia, The Kyrgyz Republic Nicaragua Sri Lanka Zambia Cambodia Georgia Lao PDR Niger St. Kitts and Nevis Zimbabwe
Cameroon Germany Latvia Nigeria St. Lucia
Canada
Ghana Lebanon
47 APPENDIX 3: Non-transformed descriptives
This table contains non-transformed descriptive statistics for the dependent variables GDPC and WTFP, the independent variables of interest ACCESS, USAGE, BARRIER and PRIVATECREDIT, and the control variables. Appendix 1 provides an explanation of the variables.
VARIABLES N mean min max sd kurtosis skewness
GDPC 199 0.0157 -0.0938 0.0689 0.0257 6.198 -0.949 WTFP 115 -0.000883 -0.0936 0.0659 0.0252 5.631 -0.233 IGDP+ 201 13,882 342.9 86,001 19,022 6.216 1.920 EDUCATION+ 146 0.969 0.647 1.094 0.0838 7.328 -2.011 HEALTH 199 70.68 49.02 83.32 8.589 2.539 -0.645 PRIVATECREDIT 178 53.04 4.359 202.3 44.16 4.441 1.364 ACCESS 182 41.34 0.377 207.6 41.66 5.821 1.573 USAGE 146 46.09 1.522 99.65 32.01 1.699 0.361 BARRIER 146 8.459 0.205 27.82 6.574 3.620 1.104 GOVERNMENT+ 178 16.56 5.911 51.18 7.117 10.17 2.047 INVESTMENT 178 24.72 5.389 67.65 10.15 7.544 1.662 INFLATION+ 178 6.897 -0.364 47.31 6.623 18.41 3.353 OPENESS+ 193 98.52 23.70 421.9 58.33 15.34 2.934 Transformations: IGDP = ln*(IGDP+) EDUCATION = (EDUCATION+)2 GOVERNMENT = ln*(GOVERNMENT+) INFLATION= ln*(INFLATION+) OPENNESS = ln*(OPENNESS+)
APPENDIX 4: Calculation Growth Rates
The average growth rate for GDP per capita, for period 2011-2016, and welfare-relevant TFP, for period 2011-2014) are computed by the following functions:
48 APPENDIX 5: Example Robustness regressions variable ACCESS
This table shows the results for the variable ACCESS regressed with different control variables, to test for robustness of the variable ACCESS. A similar approach is used for the variables PRIVATECREDIT, USAGE, and BARRIER. Additionally, this process is repeated for the dependent variable of welfare-relevant TFP. The standard errors are presented in the parentheses. The symbols ***, **, * denote statistical significant at the 1%, 5% and 10% levels, respectively. Appendix 1 provides an explanation of the variables.
(a1) (a2) (a3) (a4)
VARIABLES GDPC GDPC GDPC GDPC IGDP -0.00987*** -0.0108*** -0.00839*** -0.00849*** (0.00305) (0.00300) (0.00304) (0.00305) EDUCATION 0.0496*** 0.0601*** 0.0502*** 0.0513*** (0.0179) (0.0172) (0.0176) (0.0179) HEALTH 0.00117** 0.00122** 0.00101** 0.00108** (0.000484) (0.000488) (0.000483) (0.000490)
ACCESS -4.88e-05 -4.71e-05 -2.64e-05 -5.72e-05
(7.09e-05) (7.16e-05) (7.21e-05) (7.05e-05)
