1
The effect of the UNASUR on FDI inflow: A panel
data study
Written by: Ruben van der Ploeg (10646795)
Course: Bachelor thesis
Coördinator: Dhr. A. Rilovic Date: January 2017
2
Table of Contents
Introduction... 3
Literature review... 5
Data en methodology... 9
Analysis... 14
Conclusion... 19
Bibliography... 21
Appendix... 24
Statement of OriginalityThis document is written by Ruben van der Ploeg, who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3 Abstract
This paper investigates the effect of the UNASUR on FDI inflow, using two approaches. Using panel data for 12 countries in the years 2000-2013, we first investigated the direct effect of the UNASUR on FDI inflow. Secondly, we investigated whether institutional quality was affected by the UNASUR. The reason behind the second investigation was the possible problem of omitted variable bias caused by the global financial crisis. The regression output for the first investigation showed no significant results concerning UNASUR but did indicate that indeed the model could be affected by omitted variable bias. The results for the second investigation showed that the UNASUR had no significant effect on institutional quality, therefore making it impossible to observe an effect of the UNASUR on FDI inflow through institutional quality.
1
Introduction
International unions or agreements are, now more than ever, under heavy debate. Recently, the United Kingdom voted for an exit from the European Union. The Dutch Financial Newspaper wrote that US president Donald Trump hates the free trade agreement NAFTA. (Gersdorf, 2017) The events indicate that nations are becoming more nationalistic and some trade agreements or trade unions are perceived to be outdated. However, even today, nations are committing to unions in order to stimulate economic development and international cooperation. Most of Latin American countries, for example, have formed a new intergovernmental union in 2007 called the UNASUR. The foundation of the UNASUR gives a good opportunity to observe the effect of a union formation. This paper, therefore, aims to investigate whether any effect on foreign direct investment is apparent as a result of the establishment of the UNASUR.
In this investigation, foreign direct investment (FDI) is taken as a dependent variable. As Buchanan (2012) states, FDI is a major stimulus for economic development. Surprisingly, is therefore, that none of the 21 goals for the UNASUR that are listed on their website is about attracting more FDI. Rather, the goals concern primarily the development of international relations and the improvement of institutional quality. Still, these goals could be set with the aim of improving the FDI inflow for the Latin American countries. Previous studies have already observed a correlation between institutional quality and FDI inflow (Buchanan, 2012),
4
(Masron and Abdullah 2010). Therefore, if the UNASUR has already succeeded in improving institutional quality, we should be able to observe an improvement in FDI inflow.
To observe any change in FDI inflow caused by the UNASUR, we have to control for any change in FDI caused by other determinants. In this respect, Athukorala (2009) describes three motives for multinational companies (MNC’s) to engage in FDI namely, market-seeking, resource-seeking or efficiency-seeking. Still, if all motives are controlled for we face one last problem concerning the time period of the establishment of the UNASUR. The foundation of the UNASUR, namely, took place at exactly the same time period as the start of the global financial crisis. To control for every effect that may have been caused by the global financial crisis is next to impossible and therefore, a model estimating the direct effect of the UNASUR on FDI inflow may be vulnerable to omitted variable bias causing the estimates to become inconsistent. Therefore, this paper considers two approaches to estimate the effect of the UNASUR. Firstly, we analyse the direct effect of the UNASUR on FDI inflow for the UNASUR countries. The second analysis concerns the effect of the UNASUR on institutional quality. This approach is based on the assumption that institutional quality positively influences FDI inflow (Buchanan 2012), Masron and Abdullah (2010). This would mean that if, as we assume, institutional quality is unaffected by the global financial crisis, it would allow us to investigate the relationship between the UNASUR and FDI inflow through institutional quality. As outlined by Kaufmann (2010), this investigation will consider six different measurements of institutional quality: Control of Corruption, Government
effectiveness, Political Stability, Regulatory Quality, Rule of Law, and Voice and accountability. Since the six variables are highly correlated, we will take averages of the six
variables in order to get one dependent variable. This variable is named Governance.
The countries that are used in this investigation will be all members of the UNASUR, except for Suriname. Due to negative data values we had to leave out Suriname. This still leaves us the following eleven countries: Argentina, Bolivia, Brazil, Chile, Colombia, Ecuador, Guyana, Paraguay, Peru, Uruguay, and Venezuela. The investigation uses a panel data for the period 2000-2013. For the second investigation, the year 2001 is excluded since no data on institutional quality is available for this year.
The data shows that the UNASUR has no significant impact on FDI inflow when a direct panel data linear regression is performed. The coefficient measuring the effect of the UNASUR did show a positive value however, the data was likely suffering from omitted variable bias that led to insignificant results concerning the UNASUR. The second investigation shows that the UNASUR has a small and insignificant impact on institutional
5
quality. The findings, therefore, provide no evidence that it is possible to observe any effect of the UNASUR on FDI inflow through institutional quality.
This paper is set out as follows. Section 2 will discuss relevant literature and give the hypotheses for this paper. Section 3 will outline the empirical design of this study. Section 4 will discuss the results and in section 5, we make a conclusion about this investigation and mention some possibilities for a follow up study.
2
Literature review
2.1 UNASUR
To examine the impact of the UNASUR on FDI inflow, we first have to describe what the UNASUR is and why it would affect the FDI inflow in the participating countries. The UNASUR is an intergovernmental organization that ultimately tries to improve cooperation between South American countries in order to boost the development of each member. In December 2004, the CSN (South American Community of Nations) was formed with the purpose of integrating regional processes developed by already existing custom-unions but it was only until the end of 2006 that a common agenda was created that would increase the cooperation between South American Nations on political, economical and social levels. A few months later, in April 2007, the community changed its name to UNASUR, an abbreviation for Union of South American Nations. From this point onwards, the intergovernmental organization created committees in order to develop an own program and policies. Politically, the UNASUR was created as a platform to discuss regional policies that promote democracy and human rights. Economically, the UNASUR intends to transform the economy into one where member states are less dependent on their natural resources and promote the advancement of science and technology. Socially, the UNASUR intends to balance the redistribution of income, decrease the socioeconomic gaps between rural and urban areas, and eliminate regional inequality.
The reason for this investigation is that many unions between nations exist over the world, and all unions aim for a better cooperation and openness between participating countries. As Liargovas and Scandalis (2012) have already argued, a better openness to trade will lead to a higher FDI inflow. Added to that, many of the 21 goals established by the UNASUR concern an improvement in institutional quality among participating nations. A previous study by Buchanan (2012) showed that institutional quality has a positive effect on FDI inflow. This means that if, to some extent, the goals of the UNASUR have been reached,
6
we should observe a positive effect in FDI inflow. Following finding in previous studies by e.g. Andrew (2004) or Alfaro, Chanda, Kalemli-Ozcan and Sayek (2004), this increase in FDI inflow will accelerate the economic growth in the South American nations.
2.2 Foreign Direct Investment
To control for other reasons for multinational companies (MNC’s) to invest in UNASUR countries, we use proxies that should capture the changes in FDI not caused by the UNASUR. The use of proxies is mandatory since the motives for MNC’s to engage in FDI are immeasurable. The proxies will act as variables that represent these motives. To understand what proxies should be implemented in the analysis, we should know the motives for MNC’s to engage in FDI. In this respect, Athukorala (2003) describes three main motives for companies to engage in FDI.
The reason that developed countries attract more FDI than developing countries is based on the market-seeking argument. The argument explains that a higher developed market with higher demand will give more opportunities to serve demand. MNC’s that seek to satisfy domestic demand are market seeking.
The second reason for foreign direct investment is described by the resource-seeking argument. MNC’s can find a country or region attractive to invest in purely for the resources that the company can obtain (Athukorala, 2009). Deng (2007) described in his paper that China, for example, engages in outward FDI mainly for resource seeking arguments. Over 70% of their outward going FDI is explained by resource seeking arguments.
The last reason Athukorala (2009) describes consists of three subcategories. The efficiency seeking argument is divided between resource-based manufacturing, labour-intensive final consumer goods, and assembly processes within vertically integrated global production systems. Efficiency seeking means that MNC’s engage in FDI to improve their production costs. The first category has a parallel with the resource-seeking argument, but the difference between the two is that resource-based manufacturing is described by Athukorala (2009) as a focus on using the resources to improve a company’s efficiency rather than exporting the product. The second argument explains why MNC invest in countries with a low wage rate. Companies that try to maximize their efficiency will also look to places where labour cost is the cheapest. Therefore, countries with a low relative wage rate tend to attract FDI. The last argument by Athukorala (2009) describes the outsourcing of certain processes within the production chain to enhance efficiency. MNC’s could outsource their processes to regions with high education, good infrastructure or other reasons to improve efficiency.
7 2.3 Institutional variables
As previously mentioned, the paper consists of two methods of investigating the relation between the UNASUR and FDI inflow. In the second method, the effect of the UNASUR on institutional quality is calculated. The concept of institutional quality is divided into six different variables, namely control of corruption, rule of law, regulatory quality, government effectiveness, political stability, and voice and accountability (Kaufmann, Kraay, & Mastruzzi, 2010). All data for the variables is retrieved from surveys from four different main sources: households and firms, commercial business information providers, non-governmental organizations, and public sector organizations.
To understand these numbers, we must also understand the concepts of all six institutional variables. Firstly, control of corruption (CC) is the extent to which public power is used for private gains (Kaufmann, Kraay, & Mastruzzi, 2010). A corruptive environment also means that the state is more likely to use the public power to engage in corrupt practices that would harm MNC’s. However, there is also a study by Cuervo-Cazurra (2008) that questions the impact of corruption on FDI inflow and claims that corruptive environments enables the bypassing of inadequate regulation, which could compensate for the risk that a company faces in a corruptive environment. They distinguish two kinds of corruption: known corruption, and unknown corruption, and they argue that FDI inflow would benefit from unknown corruption, yet it would have detrimental effects on FDI inflow when the corruption is commonly known. If we assume that all corruption that is transformed to data is about the known corruption, we can safely assume that corruption harms the amount of FDI inflow.
Secondly, rule of law (RL) is described by Kaufmann, Kraay, & Mastruzzi (2010) as the extent to which agents have confidence in and abide by the rules of society. To obtain information about the rule of law, surveys are taken to establish the confidence that the public has in institutions that manage the order and regulation within a country, like the police or the government. Added to that, hard numbers about crime rate and property rights accompany this.
Thirdly, Kaufmann, Kraay, & Mastruzzi (2010) describe regulatory quality (RQ) as the degree to which the government is able to implement sound policies and regulations that benefit the private sector. This means that a well regulated environment will ease investments in terms of time and costs. The ease of starting up businesses and investment contributes to FDI inflow (Bénassy-Quéré, Coupet, and Mayer (2005). Regulatory quality is calculated
8
mainly by public surveys concerning government intervention, investment freedom or other regulatory matters.
Government effectiveness (GE) captures the perceptions of the quality of public- and civil services, and bureaucratic effectiveness (Kaufmann, Kraay, & Mastruzzi, 2010). Therefore, this variable is obtained with surveys about the perceptions of government services. Added to that, the level of infrastructure, which, in itself is a public service, is an indicator for government effectiveness because it can ease the way in which a government can implement their services.
Political stability and the absence of violence (PV), as the fifth institutional variable, is concerned with the likelihood of politically motivated violence including for example terrorism. A country with a low likelihood for political disruptions is more likely to attract FDI (Kearney, 2004). It is calculated by surveys on for example social unrest, political terror scale, or ethnic tensions but also by index numbers from the international country risk guide about government stability and internal of external conflict.
Finally, voice and accountability (VA) is included as the last institutional variable for institutional quality. Voice and accountability is about the perception to what extent the citizen is able to make their decisions within the society. This includes the vote for government, freedom of expression, and free media. Institutions that measure and promote freedom around the world, like the freedom house, calculate voice and accountability by a freedom of press index, democracy index, or human rights.
2.4 Hypotheses
To finish the literature review, we will now give you the hypotheses for both investigations. Due to the fact that the first investigation might be influenced by the financial crisis, we expect the UNASUR to be negatively influenced by unobserved effects from the crisis, causing omitted variable bias. The only way omitted variable bias can happen is by leaving out determinants that correlate with both the FDI inflow and the UNASUR. Concerning the UNASUR, we therefore expect the impact on FDI inflow to be insignificant. Thus the hypothesis for the first regression is as follows:
H0: 𝛽1 ≤ 0 H1: 𝛽1 > 0
The above stated hypothesis concerns the coefficient of the UNASUR. We will test this hypothesis at a 5% significance level.
9
The second investigation is aimed to measure the impact of the UNASUR on FDI inflow indirectly via institutional quality therefore solving the problem we expect to face in the first investigation about the inability to control for all effects that come with the financial crisis. We expect this investigation to show significant results concerning the impact of the UNASUR on institutional quality, since several goals listed by the UNASUR concern improving the institutional quality. Thus, the hypothesis for the second regression is a follows:
H0: Governance is not affected by the UNASUR H1: Governance is positively affected by the UNASUR
If the results are according to our expectations, the null-hypothesis will be rejected at a 5% level.
3
Data and Methodology
3.1 Direct effect of UNASUR on FDI
3.1.1 Variables and Data
To analyse the effect of the UNASUR on FDI inflow, we use data from the years 2000-2013 to capture both the time period with and without the South American trade union. We take the beginning of the year 2007 as the start of the UNASUR, even though the official establishment date is April 2008. Since the common agenda was already created in December 2006, we expect the possible influences of the UNASUR to start from this point onwards. In this way, both the time period before and after the union will be 7 years. All data is retrieved from the World Bank. Next, the data should be used in the regression model that takes FDI inflow as dependent variable (Stock and Watson 2003). The main independent variable will be the UNASUR, which will act as a dummy variable, taking on the value zero to represent the time period before the UNASUR, and taking on the value one to represent the time period after the UNASUR. To control for other reasons that affect FDI inflow, the model will include control variables. Finally, the estimation error will be included in the regression model, which will result in the following model:
10
Where FDI represents the total FDI inflow for each country, UNASUR represents the dummy variable explained previously, Zi are the control variables used in the model and εi is the error
term.
3.1.2 Control Variables
This paper considers different proxies that will control for factors that influence FDI inflow, other than the UNASUR. The control variables are established by considering the study by Athukorala (2009), who argued that there are three main reasons for MNC’s to engage in FDI, market-seeking, resource-seeking, or efficiency-seeking. The proxies discussed below will serve in place of these immeasurable variables.
To cover the market-seeking arguments for FDI inflow, this paper will include “GDP
Growth”, and “GDP per capita at purchasing power parity”. MNC’s that engage in FDI for
market-seeking arguments are interested in serving the domestic market. Thus, they will look at opportunities in terms of demand or expansion. A growing GDP also means growing markets and a growing demand, therefore a higher GDP growth will result in a higher FDI inflow. Added to that, GDP growth has proven to be a significant explanatory variable for FDI inflow by a previous study (Alfaro et al, 2004), thus the fact that it represents the state of the domestic market and it significantly explains FDI inflow makes GDP growth a viable proxy for market-seeking FDI. GDPpcPPP is important to determine the state of the current market. An already big market with a high demand will give more opportunities for MNC’s to invest in. This makes GDPpcPPP a good proxy for domestic market demand. We chose
GDPpcPPP over the current GDP, because total GDP does not cover for difference in
population size. A country can have a significantly higher total GDP, and still have an unattractive domestic market because the population size of the country is just higher. To adjust for differences between prices and currency between countries, the GDPpc is PPP adjusted.
The resource-seeking argument is fairly straight forward to cover. The proxy should capture the effect that a country’s natural resources has on the FDI inflow, thus to what extend the natural resources are attractive for investment decisions. This means that those natural resources extracted for local consumption should not be approximated. This paper will take the “Total natural resource rents, as % GDP” variable to capture the effect of resource-seeking arguments. This variable calculates the difference between market price and extraction costs as a percentage of the total gross domestic product and represents therefore
11
the contribution of resources to economic output, omitting the extraction on resources used for local consumption.
Most control variables, in this paper, are used to cover the efficiency-seeking argument. To improve on the efficiency, MNC’s will consider those countries that can ultimately provide improvements in their production chain. Makki and Somwaru (2004) have argued in their paper that trade openness facilitates an improvement in the efficiency of the production chain by shifting certain production processes to other countries. Thus, an open economy does not necessarily improve the production chain, yet it does give a good opportunity to easily build up certain production processes. The amount of trade relative to the domestic GDP is used as a proxy for trade openness because a high trade ratio indicates an open economy.
Since the proxy Trade covers the argument for MNC’s to facilitate an improvement in the production chain rather than improve the production chain itself, we still need to control for this. Studies by i.e. Bloningen (2005), Loree & Guisingen (1995), Globerman & Shapiro (2002) have confirmed that infrastructure is an important determinant for FDI inflow. When infrastructure is high, costs for production, transportation and communication are lower and thus will benefit the production chain. To find a good proxy for infrastructure, Loree & Guisinger (1995) have distinguished two kinds of infrastructure, namely telecommunications infrastructure and transportation infrastructure. Telecommunications infrastructure is important to investors because it shows how fast a company can reach its people (Loree & Guisingen, 1995). In this paper, telecommunications is approximated by the amount of internet users for a high amount of internet users means a company can easily reach its audience or contacts via the internet. Transportation infrastructure is important for a company in terms of the transport of its products or activities. Donaldson (2010) finds evidence that a good transportation infrastructure leads to a decrease in trade costs, an increase in international trade, and an increase in the level of real income. To approximate transportation infrastructure, we use the amount of air freight measured in metric tons times kilometres travelled. Like railway transport or car transport, air transport indicates whether a country has a highly developed transport system and is therefore a good approximation for transportation infrastructure. We chose air transport as the proxy used in this paper due to insufficient data for other proxy variables.
Lastly, what we have to remember is that the impact of the UNASUR may also be captured through the above stated control variables. In other words, some proxies may correlate with the UNASUR because the UNASUR may also stimulate FDI inflow through
12
the control variables. To ensure that we capture these effects, we will exclude those variables from the regression that highly correlate with the dummy variable UNASUR.
3.1.3 Empirical design
This paper shows an analysis of panel data of the UNASUR countries. Panel data gives a large number of data points, increasing the degrees of freedom and reducing the collinearity among explanatory variables (Hsiao, 2014). The regression will be a linear panel data regression using fixed effects. In a panel data regression you can either choose random effects or fixed effects. Fixed effects is used when we assume that unobserved country characteristics effect FDI inflow (cultural characteristics for example). Random effects does not make this assumption and therefore it does not control for it as well (Greene, 2000). Hausman (1978) already found significantly different results for both effects. Hausman created a test that would clarify whether to use fixed effects or random effects in an investigation. The outcome for this test gave a chi-squared of 21.32, which means the null hypothesis of random effects model is rejected at a 1% confidence thus, we therefore use the fixed effects model.
Furthermore, we need to decide whether to treat the standard errors as homoskedastic or heteroskedastic. In this respect, we perform a Wald test to test for heteroskedasticity. The outcome gave a chi-squared of 731.77 which is rejected at a 1% confidence interval. This means we have to treat the standard errors as heteroskedastic and therefore, we use a heteroskedastic-robust standard error.
Lastly, the variables are transformed to logarithms. The advantage for logarithms in regression analysis is that the change in the dependent variable due to changes from the independent variables can be interpreted as percentages. Therefore, is facilitates us to interpret the output of the regression model in section 4. The downside was that we had to exclude Suriname from the regression since it had too many negative data points.
3.2 Effect of UNASUR on Institutional quality
3.2.1 Variables and data
The effect of the UNASUR, in this paper, will be captured by the change in the institutional quality before and after the formation. As previously mentioned, institutional quality is divided into six different variables. However, as Table A3 in the appendix shows, the six variables have a high correlation between each other. The high correlation gives us the opportunity to take an average of all six variables and run a regression for 1 dependent
13
variable (Knack and Keefer, 1995). This variable is called Governance, and will act as a proxy for institutional quality. The model for the first investigation will thus be as shown below.
𝐺𝑜𝑣 = 𝛽1 𝑈𝑁𝐴𝑆𝑈𝑅 + 𝛿𝑖𝑍𝑖+ 𝜀𝑖
3.2.2 Control variables
Since the dependent variable has changed, we have to establish new control variables that will approximate the influences on Governance, other than the UNASUR. To establish control variables, we rely on previous studies that investigated determinants of governance variables. Much of the studies took corruption as a main dependent variable however, since the correlation between the institutional variables are high, we assume that any relationship between a control variable and an institutional variable will also hold for the other institutional variables and are thus good determinants for Governance.
To start off, an obvious variable for institutional quality would be the extent to which a country rules as a democracy. A democratic environment means that the government should act according to what the people want, therefore governance will be affected by the extent of democracy (Roberto & Rigobon, 2005). The Freedom House indicates two different variables that can approximate if and to what extent a country is democratic. These variables are
Political rights and Civil liberties. Political rights means the freedom to take part in
governmental practices. Civil liberties means the liberty that one person is granted by the government. Both variables are measured on a scale of one to seven.
Furthermore, we include the variable Free press which should approximate to what extend one person is free to say or write what they want. A study by Brunetti (2003) has showed that free press has a significant negative influence on corruption. Therefore, the proxy is assumed to have a positive relationship with control of corruption. Moreover, we expect that freedom of speech has a positive relationship with voice and accountability, since both measure freedom of expression. The data for free press is retrieved from the Freedom House.
Following papers by i.e. Akça, Ata, and Karaca (2012), Feng (1997), Braun & di Tella (2004) another variable that was found to have a significant negative influence on control of corruption was inflation. Inflation is the rise of prices causing a drop in purchasing power. Several reasons for the relationship between corruption and inflation have been suggested. A high inflation rate causes the purchasing power to drop, also for those who work in a public sector. The lower real income may cause civil servants working in the public sector to engage
14
in corruptive activities that would benefit themselves (Al-Mahrubi, 2000). Furthermore, increasing inflation also raises the government expenditures since more money is needed for activities. However, since tax cannot be raised or otherwise the government would lose the support of their people, the government debt raises which results in corruptive activities (Samimi, Ghaderi, Hosseinzadeh, and Madeni, 2012).
We already explained that many of the papers used to establish the control variables in our investigation, used corruption as a dependent variable. However, a paper by Al-Mahrubi (2004) investigated the determinants of governance as a whole, using the same data provided by Kaufman. This paper found that both economic growth and trade openness were significantly influencing governance. To begin with the first one, an explanation for this is that a prosperous economy gives politicians and people working in the public sector less incentives to engage in corrupt activities. Furthermore, a government is better able to execute and regulate activities simply because they have more money to spend. These factors all contribute to a better institutional quality. Just as in the first investigation, to approximate for economic growth, we include the variable GDP growth.
Lastly, trade openness is included in the regression, following Al-Mahrubi (2004). However, other studies have also verified the positive effect of trade openness on institutional variables. Ades & di Tella (1999), Brunetti & Weder (2003), and Persson, Tabellini, & Trebbi (2003) have estimated a significant negative relation between trade openness and corruption. The argument was that more open economies result in a more competitive market. The greater the competition within markets, the lower the margins. The low margins on products leave little opportunity to engage in corruptive activities (Treisman, 2000). Added to that, Rigobon & Rodrik (2005) found that a higher trade openness also positively influences Rule of Law. Al-Mahrubi (2004) explains that the reason for this is that the only way a country can maintain an open economy is by requiring high-quality domestic institutions. Therefore, a country needs a good Rule of Law, to maintain a stable open economy, and both will positively influence each other. As with the first investigation, we will use Trade as %GDP as a proxy for trade openness.
Just as we did in the first investigation, the variables will be transformed to logarithms to interpret the results as percentage changes. The institutional variables showed data ranging from -2.5 – 2.5 and therefore much of the data showed negative values. Since we cannot take logarithms of negative values, we solved this problem by adding 2.5 to every data point. No data showed values of -2.5 therefore we were able to take logarithms for every data point.
15
4
Analysis
4.1 Correlations
We begin the analysis by discussing the collinearity between the control variables used in this paper. As we stated earlier in this paper, some control variables may correlate with our main independent dummy variable UNASUR. The problem that arises with this issue is that we might find unreliable results for the effect of the UNASUR on FDI inflow, simply because we did not measure the effect of the UNASUR through other channels that we controlled for.
Table 1 shows us that the control variable Internet is the only variable showing a high correlation with the UNASUR. This means that the effect of the UNASUR cannot be well established if we include Internet in our regression model. The problem we face when we include Internet in the panel data regression is that we can tell little about the partial effect of a change in the dummy variable UNASUR. Therefore, we decided to exclude Internet from the regression model.
Table 2 shows us the correlation matrix for the second investigation. Concerning the UNASUR, we do not find any variables that show high correlation. We do, however, see a high correlation between the two variables that we use to approximate the level of democracy, which could cause multicollinearity problems. This means that both variables will capture much of the effect of the other and we should therefore exclude one of the proxies. We chose to exclude Civil Liberties from the regression.
Airfreight 0.0684 -0.0020 0.3213 -0.2093 -0.3841 0.3289 1.0000 Internet 0.7609 0.1729 0.7018 0.0328 -0.0337 1.0000 Trade 0.1523 0.0647 -0.2623 0.2431 1.0000 Naturalres~s 0.1908 0.2337 0.1064 1.0000 GDPpcPPP 0.4068 0.1058 1.0000 GDPgrowth 0.1706 1.0000 UNASUR 1.0000 UNASUR GDPgro~h GDPpcPPP Natura~s Trade Internet Airfre~t
16 4.2 Direct effect of the UNASUR on FDI inflow
Table 3 shows the results for the first regression. The first thing that becomes apparent when analysing the regression output is that the UNASUR does not show to be significant in explaining FDI inflow. The output estimates that the establishment of the UNASUR has a positive effect of 15.56% of the total FDI inflow for all UNASUR countries. The standard error, however is bigger than the coefficient, namely 0.1997, which means that the t-statistic gives an insignificant result for the effect of the UNASUR on FDI inflow. This means that regression output gives insufficient evidence that the null hypothesis can be rejected. As argued earlier, the fact that it is likely that the UNASUR is correlated with unobserved effects from the financial crisis, the insignificant results are likely caused by omitted variable bias.
Furthermore, all control variables show to be insignificant expect for GDP growth, which is estimated to influence FDI inflow with 2.38% for every 1% increase, with a confidence level of 10%. For the rest of the variables, we have to assume that they do not effect FDI inflow, according to our regression output. This does raise some questions about the model. The results contradict previous studies and common consensus that the arguments described by Athukorala (2009), are beneficial for FDI inflow. The model seems to be misspecified. This assumption is enforced by the low squared of the regression. An R-squared of 0.0408 means that only 4.08% of the variance of FDI inflow is explained by the regressors. Added to that, the F-statistic shows that we cannot reject the hypothesis that at least one of the included regressors is not equal to zero, because it is not significant. The contradiction to previous studies added to the low values for the R squared en F statistic makes us conclude that there is a misspecification in this model.
trade -0.0147 -0.0363 -0.0740 0.0171 -0.2780 -0.0966 1.0000 inflation -0.1002 0.3519 0.2744 -0.1749 0.3015 1.0000 freedomofp~s 0.2245 0.7944 0.7024 0.0589 1.0000 GDPgrowth 0.1171 -0.0265 -0.0111 1.0000 CL -0.0330 0.8409 1.0000 PR -0.0095 1.0000 UNASUR 1.0000 UNASUR PR CL GDPgro~h freedo~s inflat~n trade
17
We now turn to the second investigation to see if the effect of UNASUR on institutional quality does give us significant results that would imply a positive effect of the UNASUR on FDI inflow.
4.3 Results for the effect of UNASUR on institutional quality
The first investigation did not give us any evidence for a positive significant relation between the UNASUR and FDI inflow. The second investigation however, should indicate whether
18
UNASUR can positively influence FDI inflow, by working through institutional quality. The investigation is based on the assumption that institutional quality is not affected by the financial crisis. The results for the regression are presented in table 4.
The output shows different results as compared to the first investigation. The F-statistic is 32.82 and now rejects the null hypothesis that none of the variables are significantly different to 0 at a 1% significance level. Added to that, the R-squared is now much higher, namely at 0.7489, thus 74.89% of the variance of Governance is now explained by the regressors. This is a reasonable statistical fit.
Concerning the UNASUR, we do not find any evidence that it significantly effects the institutional quality. The coefficient is estimated at 0.0199 which means that the establishment of the UNASUR positively influenced the institutional quality of the UNASUR countries by 1.99%. The t statistic, however, gives insufficient evidence to reject the null hypothesis. This means that the establishment of the UNASUR is estimated to be unrelated to the institutional quality.
Furthermore, the control variables are insignificant except for the coefficients Political
rights and GDP growth. GDP growth shows to have a significantly positive influence on
Governance. Every one unit change in GDP growth results in an increase in institutional quality of 0.3%. The relation between the two variables is therefore very small in magnitude. Concerning political rights we find a negative relationship. Every 1% increase in the level of democracy, the variable that political rights tried to approximate, results in a 12.59% decrease in institutional quality, with a confidence level of 5%. This result contradicts previous findings about the relation between institutional quality and democracy. Other variables are estimated to have no effect on institutional quality.
To end this section, we have found no significant results concerning a positive relation between the establishment of the UNASUR and the FDI inflow. Although the first model seems to indicate some misspecification, as we already predicted, the second investigation gives us no new information about any effect that the UNASUR might indirectly have on FDI inflow.
19
5
Conclusion
In this paper, we tried to observe any effect that the establishment of the UNASUR could have on foreign direct investment inflow. We anticipated that, since the establishment of the UNASUR coincided with the global financial crisis, we might get biased results. Therefore, we also investigated the effect of the UNASUR on institutional quality. We chose institutional quality for two reasons. Firstly, previous papers have estimated that institutional quality has a significant influence on FDI inflow. Secondly, the goals described by the UNASUR often concern the improvement of institutional quality of participating countries.
The results of the panel data regression showed no significant impact of the UNASUR on either FDI inflow or institutional quality. The first regression did show that the global financial crisis is likely to influence the possible impact of the UNASUR.
20
This study has aimed to understand the effect of a formation of a union, since the effects of unions are questioned by national leaders and politicians like President Trump or PVV-leader Geert Wilders. The fact that the investigation showed no significant results does not mean that unions overall do not contribute to economic development. Further research should be aimed at understanding the effects of the financial crisis to make sure that we can control for those effects. Furthermore, different unions and their effect on economic development should be investigated in order to make a sound estimation of the overall effects of union formation on economic development.
21
Bibliography
Ades, A., & Di Tella, R. (1999). Rents, competition, and corruption. The american economic review, 89(4), 982-993.
Akça, H., Ata, A.Y., & Karaca, C. (2012). Inflation and corruption relationship: Evidence from panel data in developed and developing countries. International Journal of Economics and Financial Issues, 2(3), 281.
Alfaro, L., Chanda, A., Kalemli-Ozcan, S., & Sayek, S. (2004). FDI and economic growth: the role of local financial markets. Journal of international economics, 64(1), 89-112. Al-Marhubi, F. A. (2000). Corruption and inflation. Economics Letters, 66(2), 199-202. Al‐ Marhubi, F. A. (2004). The Determinants of Governance: A Cross‐ Country
Analysis. Contemporary Economic Policy, 22(3), 394-406.
Andrew, M. O. L. D. (2004). FDI and poverty reduction: a critical reappraisal of the arguments. Les investissements directs étrangers dans les pays en développement: la: diversité des impacts, ed. Claire Mainguy, Région et Développement.
Asongu, S. A. (2013). Fighting corruption in Africa: do existing corruption-control levels matter?. International Journal of Development Issues, 12(1), 36-52.
Athukorala, P. C. (2009). Trends and Patterns of Foreign Direct Investments in Asia: A Comparative Perspective. Margin—The Journal of Applied Economic Research , 3 (4), 365-408.
Bénassy‐ Quéré, A., Coupet, M., & Mayer, T. (2007). Institutional determinants of foreign direct investment. The World Economy, 30(5), 764-782.
Blonigen, B. A. (2005). A review of the empirical literature on FDI determinants. Atlantic Economic Journal, 33(4), 383-403.
Braun, M. (2004). Inflation, inflation variability, and corruption. Economics & Politics, 16(1), 77-100.
Brunetti, A., & Weder, B. (2003). A free press is bad news for corruption. Journal of Public economics, 87(7), 1801-1824.
22
Buchanan, B. G., Le, Q. V., & Rishi, M. (2012). Foreign direct investment and institutional quality: Some empirical evidence. International Review of Financial Analysis, 21, 81-89.
Cuervo-Cazurra, A. (2008). Better the devil you don't know: Types of corruption and FDI in transition economies. Journal of International Management, 14(1), 12-27.
Deng, P. (2007). Investing for strategic resources and its rationale: The case of outward FDI from Chinese companies. Business Horizons, 50(1), 71-81.
Donaldson, D. (2010). Railroads of the Raj: Estimating the impact of transportation infrastructure (No. w16487). National Bureau of Economic Research.
Feng, Y. (1997). Democracy, political stability and economic growth. British Journal of Political Science, 27(03), 391-418.
Gersdorf, F. (2017, January 12). Mexico en Canada slijpen messen voor strijd met Trump over Nafta. Het financieel dagblad, pp. 6
Gelman, A. (2005). Analysis of variance—why it is more important than ever. The Annals of Statistics, 33(1), 1-53.
Globerman, S., & Shapiro, D. (2002). Global foreign direct investment flows: The role of governance infrastructure. World development, 30(11), 1899-1919.
Greene, W. H. (2000). Econometric analysis 4th edition. International edition, New Jersey: Prentice Hall.
Hausman, J. (1978). Specification tests in econometrics. Econometrica: Journal of the Econometric Society, 46 (6), 1251-1271.
Hsiao, C. (2014). Analysis of panel data (No. 54). Cambridge university press.
Kaufmann, D., Kraay, A., & Mastruzzi, M. (2010). The Worldwide Governance Indicators: A Summary of Methodology. Data and Analytical Issues, World Bank Policy Research Working Paper 5430.
Kearney, A. (2004). AT Kearney’s 2004 Offshore Location Attractiveness Index. Chicago, IL, USA
Knack, S., & Keefer, P. (1995). Institutions and economic performance: cross‐ country tests using alternative institutional measures. Economics & Politics, 7(3), 207-227.
23
Liargovas, P. G., & Skandalis, K. S. (2012). Foreign direct investment and trade openness: The case of developing economies. Social indicators research, 106(2), 323-331. Loree, D. W., & Guisinger, S. E. (1995). Policy and non-policy determinants of US equity
foreign direct investment. Journal of International Business Studies, 26(2), 281-299. Makki, S. S., & Somwaru, A. (2004). Impact of foreign direct investment and trade on
economic growth: Evidence from developing countries. American Journal of Agricultural Economics, 86(3), 795-801.
Masron, A., & Abdullah, H. (2010). Institutional quality as a determinant for FDI inflows: evidence from ASEAN. World Journal of Management, 2(3), 115-128.
Noorbakhsh, F., Paloni, A., & Youssef, A. (2001). Human capital and FDI inflows to developing countries: New empirical evidence. World development, 29(9), 1593-1610. Persson, T., Tabellini, G., & Trebbi, F. (2003). Electoral rules and corruption. journal of the
European Economic Association, 1(4), 958-989.
Rigobon, R., & Rodrik, D. (2005). Rule of law, democracy, openness, and income. Economics of transition, 13(3), 533-564.
Samimi, A. J., Ghaderi, S., Hosseinzadeh, R., & Nademi, Y. (2012). Openness and inflation: New empirical panel data evidence. Economics Letters, 117(3), 573-577.
Stock, J. H., & Watson, M. W. (2003). Introduction to econometrics (Vol. 104). Boston: Addison Wesley.
Serra, D. (2006). Empirical determinants of corruption: A sensitivity analysis. Public Choice, 126(1-2), 225-256.
Treisman, D. (2000). The causes of corruption: a cross-national study. Journal of public economics, 76(3), 399-457.
24
Appendix
Airfreight 137 427.5415 597.8471 0 2419.933 Internet 137 22.64473 15.39045 .7476307 59.9 Trade 137 58.09967 31.06169 21.85242 206.7686 Naturalres~s 137 12.49198 9.598112 .4754512 43.54277 GDPpcPPP 137 10272.78 4399.586 3497.337 21443.75 GDPgrowth 137 3.967333 4.088638 -10.89448 18.28661 UNASUR 137 .4963504 .5018215 0 1 FDI 137 3.484189 2.476394 .0835751 10.72913 Variable Obs Mean Std. Dev. Min Maxtrade 154 63.77529 37.13338 22.10598 206.7686 inflation 154 10.00786 10.24115 -7.714067 68.44368 freedomofp~s 154 43.62338 14.61179 20 78 GDPgrowth 154 3.885648 3.986725 -10.89448 18.28661 CL 154 2.688312 1.025935 1 5 PR 154 2.428571 .9892041 1 5 UNASUR 154 .5714286 .4964863 0 1 gov25 154 2.257421 .6506505 1.118575 3.748145 Variable Obs Mean Std. Dev. Min Max
VA 0.2282 0.8876 0.9072 0.7170 0.8832 0.8411 1.0000 PS 0.1323 0.7001 0.7319 0.4154 0.6745 1.0000 GE 0.2074 0.9235 0.9473 0.8423 1.0000 RQ 0.2795 0.8170 0.8604 1.0000 RL 0.2230 0.9373 1.0000 CC 0.1962 1.0000 FDI 1.0000 FDI CC RL RQ GE PS VA
Table A3: Correlation table of the institutional quality variables Table A1: Summary statistics
