Bachelor Thesis in Economics, 15 credits Economics C100:2
Autumn term 2020
Governance and Foreign Direct Investments
A panel gravity approach on emerging markets
Simona Semenas
This page is left intentionally blank
Acknowledgements
My deep gratitude goes to Professor Gauthier Lanot, my research supervisor, for his patient guidance, enthusiastic encouragement and useful critiques of this research work. My appreciation also extents to Magnus Andersson at Malmö Borgarskola, my previous teacher, who introduced me the topic and inspired me to pursue a degree in Economics. I would also like to thank and recognize the dedication of the other professors at the Faculty of Economics at Umeå University and exceptional staff across the university. Finally, I wish to thank my parents and brother for their wise words and constant support throughout my studies.
Sincerely,
Simona Semenas
This page is left intentionally blank
Abstract
It is widely argued that a country’s economic performance is determined by its institutional quality. This study utilizes six governance indicators to examine what role they play in attracting foreign direct investments (FDI) net inflows into emerging markets. Using a static and dynamic panel gravity model, a data set of 26 host countries are investigated during the time period of 2002– 2019 with the aim of identifying which institutional factors are the main drivers of FDI into emerging markets. This paper provides evidence that Voice and Accountability, Political stability and absence of violence, Regulation quality and Control of corruption have a statistically significant positive effect of FDI. Government effectiveness and Rule of law were found to be significantly negative linked to FDI. In addition to that, this study examines further macroeconomic determinates that were likely to have an impact on FDI. It was found that trade openness, developed infrastructure as well as the sum of natural resources rents as a percentage of GDP created a desirable business conditions for multinational corporations to invest in. Lastly it was noted that source countries preferred to invest in wealthier nations. These results create good basis for understanding that it requires policymakers to improve and implement sound laws and regulations as well as a stable business environment in order to attract FDI into emerging markets.
Table of Contents
1. Introduction ... 1
2. Literature review ... 3
3. Theoretical framework ... 5
4. Data and model ... 7
4.1 Data and model ... 7
4.2 Estimations and its limitations ... 10
5. Results ... 13
5.1 Descriptive statistics ... 13
5.2 Regression results & discussion ... 14
6. Conclusion and policy implications ... 19
7. References ... 20
8. Appendix ... 23
1
1. Introduction
With the increased global economic activity driven by international trade and capital flows, foreign direct investments (FDI) have become an important source of private external finance for countries to spur their economic development. It allows nations to convey positive effects in technology, capital, production and knowledge through direct and indirect spillover channels.
Since 1990, competition for attracting FDI has enhanced among less- and more economically developed countries. The main reason being attracting funds to finance economic growth.
(OECD, 2002).
The analysis of the determinants of FDI flows is widely studied among academic researchers.
Several studies argue that a country’s economic performance over time is to a great extent determined by its political, institutional and legal environment. A well-established government plays an active role in defining its investment environment, hence enabling favorable conditions for economic growth (Butkiewicz and Yanikkaya, 2006). The World Bank (2020) has published six indicator measures for good governance and ranked countries on where they fall.
These play an important role in shaping behavior of economic actors and stakeholders when assessing the risk of investments. Acemoglu and Robinson (2012) state that the various levels of wealth between nations can be explained through the different incentives, citizens, politicians and companies have received from their institutions. They argue that innovation and the emergence of new technologies can boost economic growth. However, for technologies to become implemented and advanced, countries must be able to rely on quality economic and political institutions. As a consequence, this will provide the population incentives to educate themselves which will result in new innovations. High institutional quality incentivizes economic activity by promoting competition, efficiency, property rights and investments enabling an upward spiral of economic growth and development. On the other hand, Sachs (2005) argue that poor institutions are prospect to remain poor as they lack resources for investing and being productive. They may discourage FDI as poor institutions serve to bring further cost to foreign direct investments i.e. higher transaction cost and political instability which results in higher uncertainty with increased economic risks (Buchanan et al., 2012).
2
Emerging markets are especially interesting to study since institutional quality may be one of the biggest differentiators in what countries may end up successful. Therefore, understanding what role institutional quality plays in attracting FDI net inflows and identifying which of the institutional factors have a considerable impact on FDI is an issue of great importance for emerging countries. Essentially, they tend to simultaneously establish a solid government as well as attract capital to finance economic development.
The purpose of this study is to investigate the impact of institutional quality determinants on Foreign Direct Investment net inflows in 26 emerging countries by using a panel data set during the time period of 2002– 2019. This paper contributes to existing literature in several ways.
Firstly, various measures of governance are adapted for the purpose of identifying which aspects contribute to the attractiveness of FDI inflows. While that is the primary focus, additional macroeconomic determinants are added considering that they may have a significant impact on FDI. Moreover, a static and dynamic panel data gravity model is applied to generate robust estimates as well as account for potential heterogeneity and endogeneity problems.
Lastly, the empirical results are estimated, and policymakers are encouraged to undertake effective measures with the aim of improving their institutional quality and FDI inflows.
3
2. Literature review
Previous studies have proceeded to identify and explain the relation between FDI net inflows and its various determinants. Some macroeconomic determinants that have been looked at are the size and growth of an economic market, economic stability, trade openness, income level and the quality of institutions. This paper, in particular, dive deeper and measures a variety of institutional factors as well as some macroeconomic variables that are expected to have a greater impact of FDI. Past research has collectively given mixed results. They may either look at the big picture and explore the various determinants as previously mentioned, focus on a specific institutional quality i.e. corruption or explore multiple institutional factors over a given time period.
As an example, Neuhaus (2006) discusses whether FDI have a positive effect on economic growth within Central and Eastern Europe. With the help of endogenous growth models, he investigated 13 countries over every stage of development of a country during the period 1991- 2002. He substitutes the “Human Capital” variable of Mankiw, Romer and Weil’s (1992) growth model with “FDI” variable and incorporate additional explanatory variables i.e. trade openness, inflation volatility, domestic investment, lagged per capita income, government balance and government consumption. Estimations show that developing countries undergo strong capital accumulation and technology transfer through FDI, whereas developed countries mainly benefit from FDI as a vehicle of “Global technology diffusion”. He concludes that “FDI had a significant positive impact on the rate of economic growth in Central and Eastern Europe Countries” (Neuhaus, 2006). In addition, he states that FDI is a significant determinant for economic growth for transition economies.
Several previous studies argue that FDI are mainly directed at developed countries due to their favorable economic environment due to e.g. developed infrastructure and stable economic environments. Advanced open economics with an educated workforce and advanced financial markets is where FDI has its largest impact. Carbonell and Werner (2018) contradict those studies by investigating Spain, one of the largest receivers of FDI. They claim that “favorable Spanish circumstances yield no evidence for FDI to stimulate economic growth”. Johnson (2006) discusses whether FDI have a positive effect on economic growth as a result of technological spillovers and capital inflows. He investigates 90 countries during the period
4
1980-2002 by performing a cross-section and a panel data analysis and finds that “FDI inflows enhance economic growth in developing economies but not in developed economies “.
The results differ according to the method of analysis that researchers employ and selected sample countries for the analysis. Due to the mixed results, the focus has shifted to other possible determinants such as the effect of institutional quality on FDI. One part of the literature focus on assessing the one specific institutional quality such as corruption or political stability on FDI, whereas another part of the literature examines various institutional indicators.
Habib and Zurawicki (2002) have found that corruption has a significant negative impact on FDI location. That contradicts Wheeler and Mody’s (1992) earlier study that investigated the location of US foreign multinationals, which failed to find a significant negative relationship between corruption and FDI. Hines (1995) confirmed and did not either find a relationship between corruption and FDI. Globerman and Shapiro (2002) utilized newly developed indices to explore the effects of government infrastructure on FDI inflows and outflows for a broad sample of developed and developing countries. They found that sound government infrastructure is an important factor of FDI inflows and outflows.
Bénassy-Quéré et al. (2007) study the role of institutional quality in 52 host and source countries with the use of a Panel Gravity Model. They conclude that inward FDI is significantly influenced by the, transparency and lack of corruption, security of property rights, tax system efficiency, law enforcement and the ease of opening and running a business. Moreover, they claim that the wide range of institutions have an impact on FDI independently of GDP per capita.
Busse and Hefeker (2007) consider 83 developing countries during the period 1984 – 2003 when examining the relationship between political risk, institutions and FDI inflows. They identify the most significant indicators multinational corporations consider and find that multiple sub-categories of political risk i.e. government stability, quality of bureaucracy, law and order, democratic accountability, corruption and ethnic tensions, and internal and external conflict are highly significant determinant of FDI inflows.
Mengistu and Adhikary (2011) investigates the effect of six components of good governance on FDI in 15 Asian countries between 1997-2007 using a fixed effects model for panel data.
5
The results reveal that political stability and absence of violence, government effectiveness, rule of law and control of corruption are the key determinants of FDI inflows. The study, however, does not find any significance evidence with voice and accountability and regulatory quality on FDI inflows. Furthermore, they find that additional variables such as human capital, infrastructure, lending rate and GDP growth rate have a significant influence on FDI inflows.
Similarly, Gangi et al. (2012) examined the effects of the same government indicators (political stability, control of corruption, rule of law, regularity quality, voice and accountability and government effectiveness) on (FDI) flows to 50 African countries. By using a fixed and random effects model, they find that voice and accountability, government effectiveness and rule of law are statistically significant. The remaining indicators were found to be statistically insignificant.
Bellos and Subasat (2012a & 2012b) suggest that poor governance in fact attracts FDI in selected transition Latin American countries. With the use of the panel gravity model, Bellos and Subasat (2013) further investigate the impact of institutional factors of FDI for 18 Latin American countries. Their empirical results suggest that poor governance enhances FDI not only in the transition countries but also in Latin America.
This study aims to further contribute to the research by diving deeper into emerging markets who are transitioning from low income, less developed markets towards a more modern and industrialized economy.
3. Theoretical framework
The theory of “comparative advantage” which was introduced by David Ricardo (1817) can essentially be linked to international capital flows. The fundamental assumption of the theory argues that international transactions are more liable to take place in countries that hold lower relative production costs. The theory additionally coincides with the overall theories of cost- benefit analysis and financial rate of return in terms of a rational investor being interested in providing capital solely if he finds reasonable returns on the cost of his investments. The underlaying assumption on the above-mentioned theories assume risk neutrality, however if the assumption is withdrawn, risk becomes an essential factor when considering making FDI decisions.
6
Tobin (1958) along with Markwitz (1959) theory of “portfolio management” explains the reasons for FDI inflows to a particular country. The theory refers to strategically identifying and constructing an investment portfolio to optimize the expected yield of returns considering a market risk (Chen, 2020).
The theories on “agency cost” (Jensen and Meckling 1976), “transaction cost” (Williamson 1975), “modern property rights” (Coase 1960) and “information asymmetry” (Akerlof 1970) contribute to the explanation on why economic actors tend to undertake business activities in countries that offer solid property rights protection by minimizing the various transaction cost of the multinational firm. Property rights are thought to reduce transaction cost by, firstly, enabling investors to have security over their assets which incentivize them to become more risk tolerant. Second, Coase theorem argues that when there is a conflict of property rights, the involved parties can efficiently negotiate terms of trade that accurately will be allocated to their highest valued use, which will result in the most efficient outcome. Lastly, property rights can reduce the transaction cost by diminishing the incentives for freeriding (Sheednile, 2019).
Having access to international markets leaves multinational corporations with a larger amount of assets to invest in. By being integrated in an international market one can reduce the average transaction cost through FDI.
Dunning’s (1993) “eclectic theory” seeks to evaluate the significance of ownership, location and internalization variables when considering if it is beneficial to pursue an FDI. The theory concludes that “FDI flows across countries due to location and ownership advantages, as well as power, to internalize transaction costs.” Despite that, with an increased global integration due to globalization the theory fails to explain why some countries attract more FDI compared to others. (Mengistu & Adhikary, 2011)
North (1990) argues that institutions a set of formal and informal rules that are created, evolved and altered by humans. Formal rules such as taxes, laws and regulations, insurance and government policies along with informal norms of behaviors such as habits and traditions affect the economic actors’ willingness to invest in a foreign country. In general, good quality institutions help reduce the transaction cost and increase profitability. Poor institutions, on the other hand, are inefficient as they take up more time and resources. Moreover, poorly protected property rights increase the risk premium and lower the economic activity. Institutions influence international economic actors’ engagement in cross boarder investments. Lucas
7
(1993) adds to the theory and suggest that in emerging economies, institutional factors play a significant role attracting FDI inflows. (Sabir,S et al., 2019)
In a similar manner, Brewer (1993), King and Levine (1993), La Porta et al. (1997, 1998) find that well enforced rules, regulations and government policies stimulate capital market development and FDI inflows. Therefore, this paper aims to explore this matter further.
4. Data and Method 4.1 Data & method
A panel data set of 26 emerging host countries and 10 source countries with the highest FDI outflows over the period 2002-2019 was employed to investigate the effects of governance on foreign direct investment net inflows. The sample includes the MSCI emerging market index with the following countries; Argentina, Brazil, Chile, China, Colombia, Czech Republic, Egypt, Greece, Hungary, India, Indonesia, Korea, Kuwait, Malaysia, Mexico, Pakistan, Peru, Philippines, Poland, Qatar, Russia, Saudi Arabia, South Africa, Thailand, Turkey and United Arab Emirates. Taiwan is part of the list, despite that, it has been excluded due to a significant lack of data. These host countries are examined together with 10 source countries; Canada, Cayman Islands, China, France, Germany, Hong Kong, Ireland, Japan, Switzerland and United Kingdom. This paper will to a great extent follow Bechtini and Younsi (2019) method. The following log-log function is applied:
𝑙𝑛𝐹𝐷𝐼𝑖𝑗𝑡 = 𝛽0+ 𝛽1𝑙𝑛𝐺𝐷𝑃𝑖𝑡 + 𝛽2𝑙𝑛𝐺𝐷𝑃𝑗𝑡+ 𝛽3𝑙𝑛𝐷𝐼𝑆𝑇𝑖𝑗 + 𝛽4𝑙𝑛𝐺𝐷𝑃𝐷𝐼𝐹𝑖𝑗𝑡 + 𝛽5𝑉𝑂𝐴𝑗𝑡+ 𝛽6𝑃𝑆𝑇𝐴𝐵𝑗𝑡+ 𝛽7𝐺𝑂𝑉𝐸𝑗𝑡+ 𝛽8𝑅𝐸𝑄𝑗𝑡+ 𝛽9𝐶𝑂𝐶𝑗𝑡 + 𝛽10𝑅𝑂𝐿𝑗𝑡+ 𝛽11𝐼𝑁𝐹𝐿𝑖𝑗+ 𝛽12𝑂𝑃𝐸𝑁𝑖𝑗 + 𝛽13𝐿𝐹𝑃𝑅𝑖𝑗
+ 𝛽14𝐼𝑁𝐹𝑅𝐴𝑖𝑗+ 𝛽15𝑅𝑀𝐴𝑇𝑖𝑗 + 𝜇𝑖𝑗𝑡
(1)
Where, i = 1,…,10 source countries; j = 1,…, 26 emerging host countries; t = 2002,…,2019 time period measured; 𝜇 is the error term and 𝛽0− 𝛽15 are the parameters to be estimated. The dependent variable examined in the analysis is the foreign direct investment net inflows as a share of GDP based on the UNCTAD FDI database. It enables us to consider the relative size of a country’s economy. Parameter 𝛽5 − 𝛽10 is the set of Worldwide Governance Indicators collected from World Bank database. It consists of 6 indicators namely, Voice and Accountability, Political Stability and Absence of violence, Government Effectiveness,
8
Regulatory Quality, Rule of Law and Control of Corruption. These collectively measure the quality of governance which according to the World Bank is defined as “a set of traditions and institutions by which authority in a country is exercised. This includes the process by which governments are selected, monitored and replaced; the capacity of the government to effectively formulate and implement sound policies; and the respect of citizens and the state for the institutions that govern economic and social interactions among them” (World Bank, 2020).
These indicators rage from approximately -2.5, which reflects weak institutional quality, to +2.5 which considers a strong institutional quality.
In addition to Governance indicators, this study has included a set of controlled variables, which potentially may influence the inflows of foreign direct investments in emerging countries.
Firstly, Gross Domestic Product (GDP) is included to measure a county’s market size. A large and growing market may affect the willingness to invests since a larger GDP is likely to generate a higher return on investment. This study has also chosen to include the differences in GDP per capita in thousands of US$ (GDPDIF) between the host and the source countries. This is employed as a measure to examine whether the difference of individuals’ living standards has any effect on why source countries choose to invest in host countries. It captures the level of development of a country differently compared to the general economic growth of a country.
(DIST) represents distance which is one of the core variables of the gravity model. Longer distances between countries are assumed to discourage FDI flows as there are countries close by which one could invest in resulting in reduced operational cost. Inflation (INFL) measures the macroeconomic stability of the host country. High inflation give rise to uncertainty because planning for long term carries out a greater risk. It becomes more difficult to budget and predict what is going to occur in the future. High inflation increases prices and restrain exports from the host countries to foreign buyers. Trade openness (OPEN) is the exports plus imports as a percentage of GDP. It is assumed that a high trade openness is preferred if multinational corporations have the objective of exporting the finished product. To manufacture a product there may be a need to import certain goods as well and the trade openness demonstrates the ease of cooperating a business in the host country. Labor force participation rate (LFPR) is the proportion of the population aged 15-64 that supply their labor to the production of goods and services. The availability of low-cost labor decreases the production cost and may be a reason to invest in a host country. Some emerging markets have restricted laws on working hours which may benefit the companies as well. INFRA is the infrastructure measurement of a country which is measured by the number of internet users in the host country. Well established
9
infrastructure can reduce cost of production, transportation and distribution as well as increase the effective labor hours. Lastly, since there is a resource abundance in many emerging markets this study controls for natural resource endowments with the use of the sum of natural resources rents as a percentage of GDP (RMAT). All controlled variables have been taken from the World bank database, except inflation and GDP per capita which have been taken from the International Monetary Fund. (Ross 2019)
The model chosen does not control for all the potential variables that may have a significant influence of FDI. Previous studies often encountered problems with endogeneity where the explanatory variable is correlated with the error term, omitted variable bias where the statistical model leaves out at least one relevant variable and heteroscedasticity where the standard deviation is unequal across the range of values measures of a second variables that predicts it.
Hence, to control for these potential statistical issues along with capturing the effect of lagged influences, this study addresses the problem by applying a dynamic panel data model with the use of systems Generalized Method of Moments (GMM), which is written as follows:
𝑙𝑛𝐹𝐷𝐼𝑖𝑗𝑡 = 𝜆𝑙𝑛𝐹𝐷𝐼𝑖𝑗,𝑡−1+ 𝛽1𝑙𝑛𝐺𝐷𝑃𝑖𝑡+ 𝛽2𝑙𝑛𝐺𝐷𝑃𝑗𝑡+ 𝛽3𝑙𝑛𝐷𝐼𝑆𝑇𝑖𝑗 + 𝛽4𝑙𝑛𝐺𝐷𝑃𝐷𝐼𝐹𝑖𝑗𝑡+ 𝛽5𝑉𝑂𝐴𝑗𝑡+ 𝛽6𝑃𝑆𝑇𝐴𝐵𝑗𝑡+ 𝛽7𝐺𝑂𝑉𝐸𝑗𝑡 + 𝛽8𝑅𝐸𝑄𝑗𝑡+ 𝛽9𝐶𝑂𝐶𝑗𝑡+ 𝛽10𝑅𝑂𝐿𝑗𝑡+ 𝛽11𝐼𝑁𝐹𝐿𝑖𝑗 + 𝛽12𝑂𝑃𝐸𝑁𝑖𝑗 + 𝛽13𝐿𝐹𝑃𝑅𝑖𝑗+ 𝛽14𝐼𝑁𝐹𝑅𝐴𝑖𝑗+ 𝛽15𝑅𝑀𝐴𝑇𝑖𝑗+ 𝜇𝑖𝑗𝑡
(2)
Equation (2) is a dynamic form of equation (1), which can be seen from the lagged dependent variable 𝜆𝑙𝑛𝐹𝐷𝐼𝑖𝑗,𝑡−1. 𝜆 is the AR(1) coefficient which represents the persistence of how long the impact of a change in X is going to last in future values of the dependent variable. The variables have been transformed by taking the first difference to eliminate the individual effect or in other words, past values of X and Y have been subtracted as well as the unobserved effect from the error term. For our instrumental variable, 𝜆𝑙𝑛𝐹𝐷𝐼𝑖𝑗,𝑡−1, one more period is subtracted to break the correlation with the error term. There is no longer an overlapping time period to cause an endogeneity bias.
Note that in this study, natural logs for variables on both sides of the econometric specification are used creating a so-called log-log model. The coefficients estimated can be interpreted as the elasticity parameters of the respective independent variables. In other words, the coefficient is
10
the estimated percentage change in the dependent variables for a percentage change in the independent variable.
4.2 Estimation and its limitations
Due to facing econometric problems such as endogenous regressor, time-invariant variable, weak instrument bias and omitted variable bias we begin with using equation (2) to estimate fixed effects, random effects and perform the Hausman test. Fixed effects is a method that controls for omitted variables in panel data that vary across entities but do not change over time.
This could be country specific variables such as the language spoken within a country, religion and other variables that have not been accounted for. In other words, fixed effects estimate the coefficients of the explanatory variables on how they affect foreign direct investment inflows holding constant unobserved variables that do not change over time (Stock & Watson, 2014).
The benefits with using fixed effects is that, regardless of the correlation between the explanatory and country-specific effects, it provides consistent and unbiased estimates.
However, fixed effects is considered as insufficient for panel data gravity model as this study includes variables that do not vary over time, such as distance. This is a concern given that distance is a main component of the gravity model (Bechtini & Younsi, 2019). All time invariant variables are thus dropped from the fixed effects regression model.
Instead, the random effects model is considered as it takes into account time-invariant variables and reduces the variance of estimates of the coefficients. Unlike the fixed effects model which assumes that independent variables are fixed such that they represent the entire population, the random effects consider that the values of an independent variable are drawn at random from a larger population of possibilities. This increases the amount of biasness in the model (Clark &
Linzer, 2015).
In order to know what model the data supports, fixed effects and random effects are estimated and a Hausman test is conducted to determine which model is the most appropriate for the given data. The test evaluates the consistency of an estimator when compared to an alternative. If the p-value of the test is greater than 5% we accept the null hypothesis which implies that the random effects model is consistent and there is no systematic difference in coefficients of fixed effects and random effects. If the p-value is less than 5% we reject the null hypothesis and
11
suggest that there is a systematic difference in the coefficients and that fixed effects is the more suitable model. (Torres-Reyna, 2007)
A substantial amount of previous studies conducted are based on the above-mentioned static models. By disregarding time invariant variables, the model creates biased and misleading findings. In addition to that, the models make a strong assumption that there is homogeneity of effect across the panel. By creating a scatterplot of the data sample which can be found in Appendix 1, the results reveal that there indeed is heterogeneity across countries and years for attracting FDI. This implies that the countries have various characteristics that differentiate them which results in different degrees of FDI inflows. If homogeneity of variance is violated, the F statistic will be bias resulting in an under- or overestimated the significance level of the parameters depending on the sample size of the dataset. This paper adapts recent study methods by estimating a dynamic panel data model using generalized model of moments (GMM). The model controls for endogeneity of a lagged dependent variable when there is correlation between the explanatory variable and the error term in the model, omitted variable bias, unobserved panel heterogeneity and measurement errors. There are two GMM estimators: the difference GMM and the systems GMM. The difference GMM corrects endogeneity by transforming all regressors through differencing and eliminate the fixed effects and time invariant components.
Assume a static linear unobserved effects model for N observations and T time periods.
𝑦𝑖𝑡 = 𝛽0+ 𝛽1𝑥𝑖𝑡+ 𝛽2𝑥𝑖𝑡+ 𝛼𝑖 + 𝑢𝑖𝑡
𝑦𝑖𝑡 is the dependent variable observed for individual i at time t, 𝑥𝑖𝑡 is the time variant regressor matrix, 𝛼𝑖 is the unobserved time invariant individual effect, and 𝑢𝑖𝑡 is the error term. Put differently 𝛼𝑖 represents the explanatory variables that are unobserved and are statistically difficult to take into account for e.g. individual preferences and institutional culture. Moreover, the unobserved value is correlated with the independent variables.
The difference GMM subtracts the previous observation from the contemporaneous one eliminating the unobserved specific individual effect.
𝑦𝑖𝑡− 𝑦𝑖𝑡−1 = 𝛽0+ 𝛽1(𝑥𝑖𝑡−1− 𝑥𝑖𝑡−2) + 𝛽2(𝑦𝑖𝑡−1− 𝑦𝑖𝑡−2) + (𝑢𝑖𝑡 − 𝑢𝑖𝑡−1)
12
When we have the lagged dependent variable as a predictor, things become more complicated.
Since the error term is a part of lagged dependent variable, we have an endogeneity problem.
The error term is correlated with the difference. Hence, first difference is used to eliminate the unobserved value 𝛼𝑖. Thereafter, one we reach back for one more period for the lagged dependent variable and use ∆𝑌𝑖𝑡−2 = (𝑌𝑖𝑡−2− 𝑌𝑖𝑡−3) as an instrument variable for (𝑦𝑖𝑡−1− 𝑦𝑖𝑡−2). The correlation with the error term is now broken. (Roodman, 2009)
Blundell and Bond (1998) argued that when using difference GMM, lagged levels can be rather weak instruments for the first difference variables. Commonly, it can be seen when the variables present high autocorrelation. Instead, Blundell and Bond propose to use the original equation in combination with the transformed one called the systems GMM. The Systems GMM model incorporates lagged differences in combination with lagged levels of the 𝑦𝑖𝑡 into the matrix of instruments. As a result, when incorporating further information contained in the lagged difference of 𝑦𝑖𝑡, along with making the assumption that first differences of instrument variables are not correlated with fixed effects, it enables one to increase the efficiency of the estimator.
This study applies the Sargan-Hansen test to test the overall validity of the GMM instruments.
(Baum, 2013).
13
5. Results
5.1 Descriptive statistics
The descriptive statistics found in table 1 indicates that foreign direct investment net inflows are rather volatile in the sample period. With a minimum of 1.29 million dollars to a maximum of 11.86 million dollars suggest a contrasted performance of FDI attractiveness for the emerging countries studied.
Table 1
Descriptive Statistics
Variable Observation Mean Std. Dev Min Max
FDI 4590 8.85 1.40 1.29 11.86
GDPj 4680 12.91 1.03 10.60 16.47
GDPi 4680 13.64 2.06 8.32 16.47
GDPDIF 4680 -0.73 2.29 -4.81 7.88
VOA 4680 -0.09 0.82 -1.91 1.29
PSTAB 4680 -0.31 0.89 -2.81 1.26
GOVE 4680 0.29 0.56 -0.88 1.51
REQ 4680 0.26 0.59 -1.07 1.54
COC 4680 -0.01 0.63 -1.14 1.59
ROL 4680 0.09 0.64 -0.97 1.43
INFL 4660 5.53 5.08 2.81 31.17
OPEN 4660 4.18 0.51 3.10 5.35
LFPR 4680 4.12 0.15 3.81 4.48
INFRA 4560 3.45 0.90 0.43 4.60
RMAT 4420 1.24 1.75 -4.61 4.13
Note. All variables are in logarithmic form, except the governance indicators.
In contrast to Butkiewicz & Yanikkaya, (2006), who found a great variety in trade openness and labour force participation rate, this paper reveals a decreased variance in those variables and demonstrate a smaller gap between the emerging markets. As for infrastructure, there remains a rather high difference between the countries in our dataset. Similarly, inflation, which measure macroeconomic stability within a country, demonstrate a high fluctuation that varies between 2.81 and 31.17 during our sample period.
14
As for the governance indicators, there is a large variability between the countries. Governance effectiveness has a minimum of -0.88 and a maximum of 1.51 and varies the least among the six indicators. This insinuates that emerging countries exhibit slightly similar characteristics when it comes to the degree of its independence from political pressures, providing quality public and civil service, formulating and implementing sound policies and the government’s credibility to the implemented policies. Political stability and the absence of violence differs the most with a minimum of -2.81 and a maximum of 1.26. Most countries have a rather average political stability, while Colombia, Pakistan, turkey and Indonesia earlier on experienced a challenging political instability. Regulatory quality, Control of corruption and Rule of law fluctuate from minimum of -1.07, -1.14 and -0.97 to maximum of 1.54, 1.59, 1.43 respectively.
This implies that emerging countries in our dataset are diverse when it comes to governments’
ability to formulate and implement sound policies, the extent to which countries practice corruption, and the degree to which countries abide by the rules of society. Similarly, voice and accountability vary greatly between the emerging countries measured with a minimum of -1.91 and a maximum of 1.29.
5.2 Regression results & discussion
The results of the static and dynamic panel gravity model are presented in table 2. Column 1 and 2 presents the fixed and random effects model, whilst column 3 and 4 display the difference and systems GMM, respectively.
First and foremost, Pesaran’s (2004) test was conducted with the purpose of testing for cross sectional dependence in panel data analysis. It was found that at 1% significance level, the null hypothesis of no cross-sectional dependence (=45.879, Pr=0) was strongly rejected. This may be due to unobserved factors, which are omitted from the model, being correlated with the explanatory variables. There is no evidence for autocorrelation suggesting that Fixed effects and Random effects estimators are inconsistent, hence a dynamic model is needed. However, this paper choses to the fixed effects and random effects model as they are a simpler version to account for the differences in estimations. Hausman test (1978) was conducted to determine whether Fixed effects or Random effects was the most appropriate model for the analysis. The results rejected the null hypothesis revealing that the Fixed Effect model is the better fitting in comparison to Random effects. Despite the fact that Fixed effects hold more consistent estimators, it does not take into account time-invariant variables i.e. distance which is one of
15
the main determining factors in the gravity model, hence both of the models are included in this study. Fixed effect estimates the R-squared to be 0.3506, while random effects has a higher R- squared at 0.6022. F-statistics for both fixed and random effects are statistically significant at 1% significance level.
Since the gravity model examines both time series and cross-sectional data, the Wald test for heteroscedasticity is performed. As the null hypothesis is rejected, we conclude the presence of heteroscedasticity in the model which suggest that the variance of the errors is not constant across observations. In order to account for the issue, robust standard errors are derived. As for the GMM model, the two steps systems GMM is used as it is more efficient and robust to heteroscedasticity and autocorrelation compared to the one-step systems GMM. As one can see in table 2, when performing the Arellando-Bond test the first order serial correlation indicates a significant value suggesting the use of a dynamic panel data for the dataset used. The second order serial correlation accepts the null hypothesis claiming that there is no second order serial correlation and that instrument used is valid.
Subsequently, regression results are estimated and reported in table 2. The gravity model has two core variables, distance and GDP. Generally, it is assumed that countries closer to each other are expected to trade more due lower transportation cost. Both Random effects and Sys- GMM give evidence for the estimated coefficients to be insignificant. This implies that distance is not a highly determining factor for choosing what country to invest in for the dataset measured. Moreover, GDP which measures a country’s economic size, showed a positive highly significant effect for both the Fixed effects, Random effects, Diff-GMM and Sys-GMM.
1% increase in the real GDP leads to a significant increase in FDI inflows towards the emerging countries by 0.438%, 1.092%, 0.451% and 1.188%, respectively. This suggests that the richer a country is GDP wise, the greater amount of FDI it attracts.
The governance indicators yield some unanticipated interesting results. Voice and Accountability is statistically positively significant for the Fixed effects and Random effects model. 1% increase in efforts of improving freedom of expression and participation leads to a significant increase by 0.167% and 0.382% in FDI inflows to emerging markets, respectively.
This indicates that voice and accountability is an essential indicator for increasing the inflow of FDI into the emerging countries investigated. This can in particular be important with rapid development in communication technology with the use of internet. Diff-GMM suggested a
16
significant negative impact on voice and accountability, implying that current poor institutional quality with weakened policies hinders FDI inflows.
Similarly, the Political stability and absence of violence indicator is also statistically significant and positively correlated with FDI for Fixed effects, Random Effects and Diff-GMM. Sys- GMM have a statistically significant and negatively linked coefficient. A positive relationship means that a 1% increase in attempting to improve political stability and reducing politically motivated violence leads to an FDI enhancement for emerging countries by 0.072%, 0.083%
and 0.827%, respectively. The results suggest that politically motivated violence and terrorism cause a threat to business stability and may lead to discourage investors. Sys-GMM was found to have a significant negative effect on FDI inflows. Ross (2019) claim that the results can potentially be interpreted as what is more important is the confidence in a wider business environment and the accessibility of doing business there as a consequence of policy implementation and modification, rather than transformation in government and political stability. The Political stability and absence of violence indicators yield notably diverse results in various studies.
Government effectiveness revealed interesting results for all the model specifications. There is a significantly negative impact of government effectiveness on FDI inflows to emerging countries in our dataset. As a matter of fact, a 1% increase in steps to strengthen government effectiveness leads to a decrease in Fixed Effects, Random Effects, Diff-GMM and Sys-GMM by -0.584, -0.373, -3.482 and -0.735, respectively. As can be seen in the descriptive data, government effectiveness exhibit slightly similar characteristics as it does not vary significantly between countries. Overall government effectiveness is rather high indicating that a well- functioning market do not necessarily stimulate FDI inflows into emerging countries.
Regulatory quality captures the ability of governments to formulate and implement sound policies and regulations that promote private sector development. The results demonstrate a positive and statistically significant effect of FDI for all the model specifications indicating that a 1% increase in improved government regulatory quality leads to an increase in FDI by 0.279%, 0.507%, 1.728% and 1.061% in the Fixed effects, Random effects and Diff-GMM and Sys-GMM model. The results provide evidence for enhanced government regulatory quality being a significant indicator in serving to create a favorable business and investment environment.
17
Our findings for Control of corruption show a highly significant positive relationship for Diff- GMM and Sys-GMM, while for the other model specifications it is found to be insignificant.
Diff-GMM and Sys-GMM estimate the coefficients to be 2.356% and 0.976%, respectively.
An effort for improving the extent the public power exercises for private gain, increases the FDI inflows. Less corrupt countries hence have it easier to attract FDI inflows.
Lastly, out of the governance indicators, Rule of Law which captures how the different agents have confidence and abide by the rules of society was found to have a significantly negative relationship for Random effects and Sys-GMM and a significantly positive relationship for Diff-GMM. The results for Random effect and Sys-GMM estimations demonstrated that a 1%
increase in enhancement of improving the Rule of Law resulted in a decrease in FDI inflows to emerging countries by -0.429 and -0.539, respectively. This indicates that poor quality of contract enforcements, property rights, police and courts tend to attract FDI inflows. These results correspond with previous findings of Butkiewicz & Yanikkaya (2006) and Golderman and Shapiro (2002) who claim that poor protection of property rights enable opportunities for corruption.
While this study has the primary focus of investigating what effect governance quality has on attracting FDI, the empirical findings reveal an additional set of macroeconomic determinants that significantly effect FDI. Inflation was found to have insignificant results. However, trade openness had a significant impact in increasing FDI inflows to emerging countries. Fixed effects, Random effects, Diff-GMM and Sys-GMM indicate that a 1% increase in trade openness boost FDI by 0.436%, 0.505%, 0.969% and 0.366% respectively. A business may want to manufacture a product with imported parts or be able to export the finished product internationally. Nevertheless, trade openness illustrates the ease of operating a business in an international environment.
Table 2
Results of static and dynamic panel data gravity model estimates
FE RE Diff-GMM Sys-GMM
Coeff. t-stats Coeff. z-stats Coeff. t-stats Coeff. t-stats
FDIij(t-1) 0.238*** 7.49 0.274*** 8.18
GDPi 0.438** 2.44 1.092*** 32.42 0.451*** -1.73 1.188*** 23.16
GDPj 0.008 0.06 0.039** 2.54 0.043 0.20 0.744*** 3.67
GDPDIF -5.640** -2.57 -7.69*** -0.81 -0.000*** -4.04 -0.000*** -6.11
DIST - - -0.029 -6.10 - - 0.026 -0.69
VOA 0.167** 2.43 0.383*** 8.38 -2.139*** -0.78 0.095 0.88
PSTAB 0.272*** 5.35 0.083** 2.08 0.827*** 6.37 -0.241** -2.07
GOVE -0.584*** -4.39 -0.373*** -3.93 -3.482*** -13.48 -0.735*** -2.88
REQ 0.279** 2.04 0.507*** 6.47 1.728*** 8.29 1.061*** 4.61
COC 0.069 0.40 -0.012 -0.16 2.356*** 8.93 0.976*** 9.19
ROL -0.143 -0.89 -0.429*** -4.54 0.703** 2.44 -0.539** -2.49
INFL 0.092 0.90 0.004 0.76 0.141 1.03 0.003 0.66
OPEN 0.436*** 3.76 0.505*** 7.62 0.969*** 4.71 0.366*** 3.12
LFPR 0.173 0.40 -0.771*** -3.77 1.320 1.37 -0.536*** -2.87
INFRA 0.574*** 8.43 0.399*** 20.92 0.869*** 8.88 0.327*** 10.82
RMAT 0.508*** 14.30 0.272*** 14.78 0.512*** 12.94 0.170*** 5.80
Constant -2.483 -0.85 -6.54*** -6.51 -5.346*** -1.08 -6.25*** -6.25
R-squared 0.3506 0.6022
F-test/
Wald test
93.49*** 3012.57*** 1793.58*** 19332.41***
Hausman Test
265.76***
AR(2) test p-value
0.139 0.878
Note. *** one percent significance level, ** five percent significance level, * ten percent significance level. Fixed effects and Random effects accounts for heteroscedasticity. Two step Difference GMM and two step Systems GMM are estimated using robust standard errors.
18
Furthermore, our empirical evidence suggests that the difference in GDP per capita has a significant negative effect on FDI inflows to emerging markets. An increase in the difference in GDP per capita results in a decrease in FDI to emerging markets implying that the source countries prefer to invest in high-income host countries. An explanation for this could be that source countries put empathies on nations where there is higher purchasing power since it is a proxy for its market size.
The Labor force participation rate measures the proportion of the population that is of working age and is either working or actively looking for a job. The results reveal that Random effects and Sys-GMM have a significant negative effect on FDI, whilst the other two model specifications have insignificant results. A 1% increase in Labor force leads to a -0.771% and -0.536% decrease in FDI inflows to emerging markets. This result suggests that the production does not necessarily occur in countries where they are the most labor intensive. Investors may instead choose countries with low cost labor where labor cost do not necessarily reduce the level of productivity.
Empirical evidence shows that that FDI is positively and significantly influenced by quality infrastructure. Fixed effects, Random effects, Diff-GMM and Sys-GMM demonstrate that a 1%
improvement in the quality of infrastructure will increase FDI by 0.574%, 0.399%, 0.869% and 0.327% respectively. This is linked with previous studies arguing that more advanced and reliable infrastructure minimized costs for transportation, production and distribution.
Lastly, natural resource rents measure the difference between the price of a commodity and the average cost of producing it. It was reveled that natural commodity rents i.e. oil, natural gas, coal, mineral and forest rents are positively linked to FDI and highly statistically significant.
This means that exploitation of natural resources is unlikely the key factor when choosing where to invest in. Table 2 summarizes the main findings of the various model specifications and their tests.
Conclusion and policy implications
It is an essential for emerging countries to understand the what role institutional quality plays in attracting foreign direct investment and to identify what determinants stimulate foreign direct investments inflows in order to be able to address financial concerns during their substantial
19
economic development period. This study investigates the impact institutional quality has on foreign direct investment inflows amongst emerging markets for the sample period 2002-2019 using a static and dynamic panel data gravity model. To achieve this, six governance indicators are considered as a measure of institutional quality. In addition to that, macroeconomic factors that significantly effect FDI, i.e. GDP, difference in GDP per capita, inflation, trade openness, labor force participation rate, infrastructure and natural resources rents are examined as control variables. The empirical results revealed that the governance indicators as well as the macroeconomic factors had a significant impact on foreign direct investment inflows to emerging countries. In particular, it was found that Voice and Accountability, Political stability and absence of violence and Regulatory quality and Control of corruption had significant positive finding indicating that they play a crucial role in facilitating FDI inflow. Government effectiveness, on the other hand, was negatively linked but statistically significant which indicated that quality of public service, civil service, and independence from political pressure did not foster FDI inflows. Rule of law generated negative results which validated several other studies claiming that poor protection of property rights may create opportunities for corruption.
Following these empirical findings, Emerging countries are recommended to improve the quality of their institutions. Allowing citizens to express themselves by written form and speech in media and letting them participate in selecting a government enables countries to crease a functioning business climate communication wise. By operating fairly with integrity, having a politically stable government, and implementing sound policies and regulations that benefit the private sector offers an attractive business climate for FDI. In addition to that, governments should opt for trade openness and invest in infrastructure to generate an increased amount of FD. Lastly, emerging markets are recommended to undertake effective measures with the aim of improving law and regulations, hence limiting evading legislations and reinforcing sound policies.
20
References
Acemoglu, D. & Robinson, J . (2012). The Origins of Power, Prosperity, and Poverty. Why nations fail London. Profile Books, 529 pp
Adhikary, B. K., & Mengistu, A. A. (2011). Does good governance matter for FDI inflows?
Evidence from Asian economies. Taylor & Francis.
https://www.tandfonline.com/doi/abs/10.1080/13602381003755765
Akerlof, G. A. (1970). The Market for “Lemons”: Quality Uncertainty and the Market Mechanism. The Quarterly Journal of Economics.
https://academic.oup.com/qje/article-
abstract/84/3/488/1896241?redirectedFrom=fulltext
Baum, C. F. (2013). Dynamic Panel Data estimators [Slides]. Boston College.
http://fmwww.bc.edu/ec-c/s2013/823/EC823.S2013.nn05.slides.pdf
Bechtini, M., & Younsi, M. (2019). Does good governance matter for FDI? New evidence from emerging countries using a static and dynamic panel gravity model approach.
Wiley. https://onlinelibrary.wiley.com/doi/abs/10.1111/ecot.12224
Bellos, S., & Subasat, T. (2012a). Corruption and foreign direct investment: A panel gravity model approach. Bulletin of Economic Research. https://doi.org/10.1111/j.1467- 8586.2010.00370.x
Bellos, S., & Subasat, T. (2012b). Governance and foreign direct investment: A panel gravity model approach. International Review of Applied Economics.
https://doi.org/10.1080/02692171.2011.587110
Bénassy‐Quéré, A., Mayer, T., & Coupet, M. (2007). Institutional Determinants of Foreign Direct Investment. Wiley. https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1467- 9701.2007.01022.x
Blundell, R., & Bond, S. (1998). Initial conditions and moment restrictions in dynamic panel data models. Elsevier.
https://www.sciencedirect.com/science/article/pii/S0304407698000098
Buchanan BG, Le QV, Rishi M (2012) Foreign direct investment and institutional quality:
some empirical evidence. Int Rev Financ Anal 21:81–89
Busse, M., & Hefeker, C. (2007). Political risk, institutions and foreign direct investment.
Elsevier. https://www.sciencedirect.com/science/article/pii/S0176268006000267 Butkiewicz, J. L., & Yanikkaya, H. (2006). Institutional quality and economic growth:
Maintenance of the rule of law or democratic institutions, or both? Elsevier.
https://www.sciencedirect.com/science/article/pii/S0264999306000368 Carbonell, J. B., & Werner, R. A. (2018). Does Foreign Direct Investment Generate
Economic Growth? A New Empirical Approach Applied to Spain. Taylor & Francis.
https://www.tandfonline.com/doi/full/10.1080/00130095.2017.1393312
21
Clark, T., & Linzer, D. (2015). Should I Use Fixed or Random Effects? Political Science Research and Methods, 3(2), 399-408. doi:10.1017/psrm.2014.32
Coase, R. (1960). The Problem of Social Cost. Journal of Law and Economics,.
Dunning, J. H. (1993). Multinational Enterprises and the Global Economy.
https://www.scirp.org/(S(i43dyn45teexjx455qlt3d2q))/reference/ReferencesPapers.asp x?ReferenceID=1558384
Globerman, S., & Shapiro, D. (2002). Global Foreign Direct Investment Flows: The Role of Governance Infrastructure. Elsevier.
https://www.sciencedirect.com/science/article/pii/S0305750X02001109 Habib, M., & Zurawicki, L. (2002). Corruption and Foreign Direct Investment.
https://link.springer.com/article/10.1057/palgrave.jibs.8491017
Hines, J. R. (1995). Forbidden Payment: Foreign Bribery and American Business After 1977.
NBER. https://www.nber.org/papers/w5266
Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Elsevier.
https://www.sciencedirect.com/science/article/pii/0304405X7690026X Johnson, A. (2006). The Effects of FDI Inflows on Host Country Economic Growth.
https://static.sys.kth.se/itm/wp/cesis/cesiswp58.pdf
Neuhaus, M. (2006). The Impact of FDI on Economic Growth: An Analysis for the Transition Countries of Central and Eastern Europe (Contributions to Economics) (2006th ed.).
Physica.
North, D. C. (1990). Institutions, Institutional Change and Economic Performance (Political Economy of Institutions and Decisions) (59262nd ed.). Cambridge University Press.
OECD. (2002). Foreign Direct Investment for Development.
Https://Www.Oecd.Org/Investment/Investmentfordevelopment/1959815.Pdf.
https://www.oecd.org/investment/investmentfordevelopment/1959815.pdf
Pesaran, M. H. (2004). General Diagnostic Tests for Cross Section Dependence in Panels.
https://ideas.repec.org/p/cam/camdae/0435.html
[(Portfolio Selection: Efficient Diversification of Investments )] [Author: Harry M.
Markowitz] [Apr-1971]. (1959). Yale University Press.
Ricardo, D. (2004). The Principles of Political Economy and Taxation (5.12.2004 ed.). Dover Publications.
Roodman, D. (2009). How to do xtabond2: An introduction to difference and system GMM in Stata. https://journals.sagepub.com/doi/pdf/10.1177/1536867X0900900106
22
Ross, A. G. (2019). Governance infrastructure and FDI flows in developing countries.
Routledge. https://www.tandfonline.com/doi/full/10.1080/19186444.2019.1640572 Sachs, J. (2005). The End of Poverty: Economic Possibilities for Our Time (Annotated ed.).
Penguin Books.
Sheednile, A. (2019). Property rights and transaction costs in developing countries: A political settlement perspective. Developing Economics.
https://developingeconomics.org/author/abdulrasheednile/
Stock, J. H., & Watson, M. W. (2014). Introduction to Econometrics, Global Edition (3rd ed.). Pearson Education.
Torres-Reyna, O. (2007). Panel Data Analysis Fixed and Random Effects using Stata [Slides].
Princeton University. http://www.princeton.edu/~otorres/Panel101.pdf
Wheeler, D., & Mody, A. (1992). International investment location decisions: The case of U.S. firms. Elsevier.
https://www.sciencedirect.com/science/article/pii/002219969290050T
Williamson, O. E. (1975). The Economics of Organization: The Transaction Cost Approach.
The University of Chicago Press. https://www.jstor.org/stable/2778934?seq=1 Worldbank. (2020). Worldwide Governance Indicators. World Bank.
https://info.worldbank.org/governance/wgi/
23
Appendix
Appendix 1
Figure 1
Heterogeneity across countries
Figure 2
Heterogeneity across years
24
Appendix 2
Table 3
Description of the variables
Variable Description Source
FDI FDI net inflows UNCTAD
GDPi Real GDP of the host country (constant 2015 US dollars) UNCTAD GDPj Real GDP of the source country (constant 2015 US dollars) UNCTAD GDPDIF Difference in GDP per capita in thousands of US dollars (PPP) IMF
DIST Shortest distance between major cities in kilometers World Bank VOA Voice and accountability which measures the extent a country’s
citizens are able to express themselves freely, freedom of joining associations and to what extent they can participate in selecting their government
World Bank
PSTAB Political stability and absence of violence measures the likelihood of political instability and political violence
World Bank GOVE Government effectiveness measures the quality of civil and
public services, how independent they are from political
pressures, the quality of policy formulation and how committed the government is to follow such policies
World Bank
REQ Regulatory quality captures the ability of a government to implement and follow sound policies that promote the private sector
World Bank
COC Control of corruption measures the extent to which public power is exercised for private gain
World Bank ROL Rule of law measures to what extent agents follow and abide by
the rules of society. In particular, the quality of contract enforcement, property rights, police and the courts
World Bank
INFL The macroeconomic stability of the host country IMF
OPEN Exports plus imports as percent of GDP World Bank LFPR The proportion of the population aged 15-64 that supply their
labor to the production of goods and services
World Bank INFRA Infrastructure which is measured by the amount of internet
users in the country
World Bank RMAT Total natural resources rents are the sum of oil rents, natural gas
rents, coal rents (hard and soft), mineral rents, and forest rents.
World Bank
