Bachelor Thesis 2016
The relation between FDI and
growth in less developed
countries
June 2016
Lychelle de Lannoy
Student number: 1052227
Supervisor: Gabriele Ciminelli
Statement of Originality
This document is written by student Lychelle de Lannoy 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 the completion of the work, not for the content.
Abstract
There has been done several researches about the relationship between FDI and growth of a country. In this paper we are going to analyze the relationship, specifically for less developed countries, with the data of 149 countries that are listed at the IMF. The analysis is done for the years 2004, 2008 and 2012. The period is chosen in order to conclude if the financial crisis in 2008 had an effect on the relationship of FDI and the growth of a country. The analysis is done with an OLS-regression, where the main variable is inward FDI. Besides the main variable, we included 4 control variables to correct for omitted variable bias. The 4 control variables are, political stability, government revenue consisting only of taxes, portfolio investment and openness of a country. For the second analysis we also divided the 149 countries into 3 groups depending on their geographical location. The main conclusion of this paper is that FDI has a significant effect on growth, after the financial crisis, in 2012. For the years, 2004 and 2008, it may be insignificant, but it does have a positive relation with the growth of the country. Besides that, this paper concludes that each continent has it’s own aspect regarding the determining factors of the intensity of spillovers, such as regional spillovers or the technological gap between the countries.
Table of Content
Section 1:Introduction………5
Section 2: Literature review………..8
2.1 The channels………...8
2.2 The degree of spillovers……….11
Section 3: Data and Methodology………...14
3.1 Data………...14
3.2 Methodology and hypothesis……….16
Section 4: Results………..20
Section 5: Conclusion………29
Summary……….29
Limitations and improvements……….30
References………..31
Section 1: Introduction
Over the last two decades foreign direct investment (FDI) has increased. This started in the developed countries, but later on also spilled to the less-developed countries. According to the data of United Nation Conference on Trade and Development (UNCTAD) the net FDI inflow for developed countries increased from 170 billion dollars in 1990 to 499 billion dollars in 2014, this is equal to an increase from 64% of the GDP to 87% of the GDP. But for less-developed countries it increased from 566 million in 1990 to 23 billion in 2014, which is from 35% of the GDP to 72% of the GDP. This was mainly due to the policy changes in these less-developed countries, such as easing of the restrictions of capital inflow and the subsidies offered. The policy changes were introduced to attract foreign capital to the less-developed countries in the 1980’s (Kobrin, 2005). The multinational enterprises (MNE) have observed how investing in less-developed countries can benefit the firm (Makino et al, 2004a). Because of the recent activities regarding the attraction of FDI in less-developed countries, it is of great interest to analyze, if the FDI affects the growth of the less-developed countries.
There are several channels through which the FDI activities can benefit growth in the host countries. Most obvious would be to think as Gastanaga et al (1998) wrote in his article; Host country reforms and FDI inflows: How much difference do they make, that FDI would benefit the host country by the formation of capital. Besides the formation of capital, whenever a MNE settles in a country, it also attracts and buys new shares. Through the capital inflow, there will be a increase the balance of payment in the host country and in return also increase the growth of the country.
Nevertheless in existing research the relation between FDI and the growth of a country is tested and has find different results. Borentszein et al (1998) agree that FDI has a positive affect on the growth of a country, provided that there is a minimum stock of human capital. Most macroeconomic studies find positive relationship between FDI and growth. But in contrary, Carkovic et al (2002) state that FDI does not have a significant independent affect on the growth of a
this paper, we try to combine the micro, - and macroeconomic aspects of the host country and analyze what the influence of FDI is on the growth of the less-developed countries, through the domestic firms.
The less-developed countries are mostly situated in South America, Africa and Asia (see list 2,3 and 6). At first these countries did not attract the MNEs to invest in their country. That was mainly due to the unstable political situation in those countries because of the high degree of corruption (Dupasquier, 2006; Zang, 2001). The high degree of corruption allowed politicians to put limitations on the inflow and outflow of capital, which in return affected the financial situation of the countries. For example in South America these restrictions led to the Latin America debt Crisis in the 1980’s and in Asia the Asian Financial Crisis took placein 1997 (Pilbeam, 2013). As can be concluded there have been several crisis’ that have taken place over the years, but the most recent crisis that affected most countries is the financial crisis that started in 2008. Therefore in this paper we want to analyze the relation between FDI and the growth of less-developed countries in the period before and after the financial crisis. Whereas we expect that during the crisis, the FDI activities of the country would decrease compared to the years before the crisis. But that the years after the financial crisis, the FDI activities would have a bigger impact on the growth of a country compared to the years before the crisis.
In this paper we will do an analysis on only the less developed countries that are listed at the IMF. The main analysis will be done to investigate whether FDI has a positive relation with the growth of these countries. Besides the relation, we will also be investigating if the FDI has a significant effect on the growth after and during the financial crisis in 2008. We will be doing that by using 3 different years for the analysis, namely 2004, 2008 and 2012. In 2008 the financial crisis started, therefore we thought it would be a good point to take that year into our analysis. We took a range of 4 years before and after the crisis to analyze the effects. At last, we will analyze if the geographical location of the country has an influence on the effect of FDI on growth. The analysis will be done on the basis of the OLS-estimator, whereas growth is the dependent
variable and FDI is the independent variable. Besides the FDI-variable we include 4 control variables to correct for omitted variable bias.
In this paper we concluded that in 2012 the relation between FDI and the growth of a country is the highest. Besides that there was a strong relation, the FDI is also significant in that year. In 2008 the relation between FDI and the growth is also positive but not significant. This is mainly due to the crisis in that year. After dividing the countries, we conclude that each continent has it’s own aspects which could benefit/disbenefit certain spillover channels in that country.
This paper is structured as follow. In section 2 the different types of channels through which FDI spillovers can takes place in the host country will be discussed, followed by the factors that determine the degree of the spillovers. Section 3 will clarify the econometric model that is used for the analysis and explained were that data is found. The data description will be discussed as well. In section 4 the results of the analysis will be presented and in section 5 we will form a conclusion and mention the limitations of the research and improvements for further research.
Section 2: Literature review
In order to analyze the relationship between FDI and the growth of a country, first we need to look into how FDI of the MNEs can affect the domestic firms in order for the economy to grow. In this chapter we will start by discussing what FDI is. Followed by previous research about the different types of channels that influences the domestic firms. Lastly, we will explain the factors that determine the degree of spillovers.
The definition of FDI stated in the UNCTAD has two key elements. “First, FDI is defined as an investment made by a resident of one economy in another economy, and it is of a long-term nature or of “lasting interest”. Second, the investor has “significant degree of influence” on the management of the enterprise” (UNCTAD, 2016). Before actually investing, the MNEs should take some important aspects into account that will determine how the MNEs will settle in the host country. The aspects are the level of control over the domestic firm and the way the MNEs would like to enter into the host country (Muller, 2007). These aspects will decide how the MNEs will be investing in the domestic country. Therefore there are 2 strategies, namely greenfield investment and brownfield investment. The greenfield investment strategy is when a foreign company starts a new venture in the host country from the ground up. On the contrary, the brownfield investment strategy is when a foreign company merges (or acquires) with a domestic company (Wojnicka, 2001).
2.1 The channels
The impact of FDI on domestic firms of a country is caused by spillovers, which in return has an effect on growth.
Crespo et al (2009) clarifies that there are 2 types of spillovers. Namely, horizontal spillovers, this is when the domestic firms and the MNEs operate in the same sector. Vertical spillovers take place if the domestic firm and the MNE operate in different sectors but are linked to each other. When talking about the domestic firm, we are referring to the firm that is situated in the host country. In the case of brownfield investment strategy the MNE will merge with the domestic
firm. The conclusion is that spillovers are the transfer of knowledge from the MNEs to the domestic firm, which will benefit/disbenefit the firm and therefore also the country. Over the years technological spillovers has become very interesting to the domestic firms over the years (Hale & Long, 2006). Markussen (1995) also denote that FDI is the main cause of the international technological diffusion. The reason for that is the high technology knowledge of the MNE. Bouoiyour (2003b) states that the spillovers are like a black box. It cannot be measured directly and it is difficult to recognize.
The technological spillovers can be transferred to the domestic firms in many different ways. The manners through which the knowledge enters the domestic firms are called channels. Most spillovers, such as technological spillovers, will provide an increase of productivity for the domestic firms, which will benefit the firm and therefore also benefit the economy of the country (Crespo, 2007). In addition, Bramilla et al (2009) states that if there is spillover through these channels, there is a high probability that there will develope new products in the domestic market. This can influence the growth of the country. The main channels for the spillovers are demonstration/imitation channel, labor mobility channel, competition channel, backward/forward channel and the export channel (Crespo, 2007).
The idea of the demonstration channel refers to a situation where the MNE is able to introduce the domestic firm to new technology and teach them how to work with it. The imitation channel has the same idea as demonstration channel, but in this case the domestic firms copies the procedures of the MNE (Crespo, 2007). This channel can also be called the training channel. The entrance of foreign firms may give the domestic firms an incentive to train their staff for them to be able to work with new technology (Lensink et al, 2001). Without this channel the domestic firms would have to invest in a risky projects, whereas through this channel the riskiness for these project will decrease. That makes it more beneficial for the domestic firms to invest in the projects (Lensink, et al 2001).
Second we have the labor mobility channel. This channel count upon that the domestic firms are able to hire people from the more advanced firms (MNE), assuming that the company will benefit from the knowledge and the experience that the workers have gained working at the advanced company (Crespo, 2007). Hale et al (2006) also argues that not only through labor mobility a spillover can take place, but also through the so-called network externality channel. This is the indirect mechanism, due to conferences and other events where workers meet up, where managers and workers learn new techniques due to interacting with workers from the MNEs. However, assuming that the workers of the foreign country are more advanced, to be able to benefit from this channel the domestic firms has to hire more skilled workers. This concludes that the labor market is an important aspect for the technological spillovers.
The domestic firms may benefit from the knowledge gained through the labor market as a consequence of FDI of MNE, but the competition they face has increased as well. Therefore we introduce the competition channel. If a new MNE establishes in a country, the domestic firms will have an incentive to work more efficient to be able to keep up with their competition (Crespo, 2007). In that case the competition channel will benefit the domestic firm due to working more efficiently. On the other hand, Bramilla et al (2009) argues that it can also be a loss for the domestic firm. This occurs when the MNE takes away a large market share of the firm.
The competition channel argues that if domestic firms want to benefit from FDI they have to work more efficiently, that includes lowering costs. The backward/forward linkage channel argues how domestic firms can benefit by having a strong relationship with the supplier. Backward linkage is when a company in an industry buys from suppliers of another industry (inputs). Forward linkages is when a company in an industry sells to company in another industry (outputs). Lensink et al (2001) state that this channel includes the strong buyer-seller relationship between the domestic firm and the MNE. Where due to the transactions between these firms, the different types of spillovers will take place. Such as the spillover of technological knowledge or productivity
knowledge. In the case of technological spillover, the MNE can introduce new technology to his supplier. In addition, Crespo et al (2009) argues that due to stronger linkages the MNEs will have a greater demand for the products of the domestic firms in the case of backward linkage. Also the MNE expect a certain level of quality, which they can get by educating the domestic firms. This can only take place if the relationship between buyer and seller is strong. On the contrary, the chance that a negative effect can occur is also possible if the linkage is broken due to bad quality (Cannon, 2001). The domestic firm will then lose a part of their market share. Clearly, this channel is related with the previous demonstration/imitation channel.
The export channel is also closely linked to the other channels. If all the other channels provide the domestic firms to work more efficiently and produce more products, their export will also increase which is beneficial for the economy (Crespo, 2007). Because this will increase the balance of payment and therefore also the growth of the country.
Besides the technological spillover, the domestic firm can also benefit from knowledge spillovers, managerial skills spillovers, capital spillovers (Temiz et al, 2014). All these types of spillovers have the same affect as technological spillovers, but more or less in different degrees.
2.2 The degree of the spillovers
There are many factors that determine the degree of spillovers through the channels. In this section we will discuss the most common determinant factors that influences the degree of the spillovers.
In every country there are different types of firms. Some are more advanced than others; therefore some firms are able to take in more knowledge than the others. The absorption capacity has a significant role in determining the degree of the spillovers. Absorption capacity is the ability to absorb new information from other firms. Makino et al (2004) argues that less-developed countries have high potential for growth because of the weaker institutional support. That means that the domestic firms in less-developed countries have less regulation and are
willing to learn new things, which means they have a high absorption capacity. On the other hand, firms in developed countries have strong regulations and already build up their own system. This means that it is not that easy to adjust from their system, which implicates that their absorption capacity is limited. Hale et al (2006) found that if the labor mobility and network channel are optimal, absorption capacity does not have an influence on the process of the spillovers. The labor mobility and network channel are optimal when all the aspects ensure that there is a maximal spillover capacity through the channel.
As less-developed countries benefit from the openness to learn from the MNE, it is of crucial role that they have some kind of knowledge of the specific spillover. In the case of technological spillover the technological gap is a determinant for deciding the degree of the spillover. Bouoiyour (2003) states that technology gap can be defined as the distance between domestic and foreign firm in terms of productivity. In order for the spillover to take place, there is a range where the spillover would benefit the recipient country. Additionally, Fagerberg (1987) states that the countries can manage the gap themselves. Innovation is the key to manage this gap. If a country innovates, their knowledge will expand and therefore the gap can become either smaller of larger. Depending on which of the country, recipient country or foreign country, is innovating.
Besides the technological gap, regional effects are also a determinant for spillovers. Crespo et al (2009) argues that the FDI spillover effects can be limited, taking into account the geographical distance between the foreign country and the recipient country. Four reasons are summarized by Crespo et al (2009). First, demonstration effects will be local, since the benefits are likely to be spread at least initially to neighboring firms. Second, a skilled-worker who is looking for a job is more likely to accept a job in the region of his home country. Third, MNE prefer to have a local linkage with domestic firms in order to cut on transaction cost and to improve the communication. Fourth, knowledge spillovers will spread more efficiently at a smaller distance. Thus the closer the foreign
country is situated from the host country, the higher the degree of spillover will be.
The degree of spillovers is also determined by the FDI characteristics and domestic firm characteristics. They are 2 separate factors but very closely linked. The firm specific factors are mainly the way the firm operates and what their believes are (Egelhoff et al, 2000). When a MNE is considering to invest in a country, it has to go through different aspects, namely entry modes, forms of ownership, motives and the location of sights in the country they are wiling to invest (Egelhoff et al, 2000). The answers to these questions will be the characteristics for the FDI. But the believes and the way the MNE operates are also characteristics of the FDI (Crespo, 2007). The closer the characteristics of the FDI and the domestic firm are, the higher the degree of spillovers. For example, if the MNE considers doing a brownfield investment, the degree of spillovers will be higher than when the MNE is considering of doing a greenfield investment. That is due to the high similarities of the domestic firm and the MNE that merges with the domestic firm.
At last, the circumstances of the host country are also determinant factors for the degree of spillovers. Each country has it’s own circumstances. Depending on the country-specific circumstances the spillovers may have an affect. As Holger et al (2002c) argues in his paper there is no effect in Venezuela from FDI. On the contrary, there is an effect in Lithuania, but only due to the inter-industry (vertical) spillovers rather than the intra-industry (horizontal) spillovers. Arestis and Demetriades (1997) explain that the country circumstances rely mainly on the institutional structure of financial system, the policy regimes of the countries and the effectiveness of the government.
Section 3: Data and Methodology
In this section we will discuss how the empirical analysis will be done. First we will start by explaining where the dataset is from and how the variables are constructed. After that the method we used will be explained, followed by the hypothesis.
3.1 Data
For this research, the dataset for the years, 2004, 2008 and 2012 is obtained mostly form the database of the World Bank. Other than the database Word Development Indicators (WDI) of the World Bank, information is also obtained from the data of the Center for systemic Peace.
In order to collect the data for this paper, first we needed to obtain a list of countries that fall into the category of less developed countries. The IMF distinguishes 2 categories, namely the developed countries and the less developed countries (see table 1). The criteria that the IMF used to distinguish the countries are the GDP of a country, the export of goods/services and the population of a country. According to the World Economic Outlook (WEO) of IMF the less developed countries can also be distinguished into 2 separate groups, namely the transition countries and the developing countries. The distinction is done according to the GDP per capita of a country. This limits the research to countries that are listed at the IMF.
The dependent variable is the growth (g) of the countries. Data of the growth (g) of the countries can be found in the database of the World Bank. The annual growth rate is calculated on the basis of the percentage change of the GDP per capita of a country using the market prices expressed in US dollars. In the database of the World Bank, there was no distinction between real and nominal GDP growth. This is also the case for the database of the IMF. The GDP per capita is defined as the gross domestic product divided by the midyear population of the country.
The main independent variable is the inward FDI. The data for FDI can also be found in the database of the World Bank. Borensteinz (1998) argued, that
the IMF provides two types of data regarding the FDI, namely the gross FDI and the net FDI. Gross FDI refers to only the inflow in the host country and net FDI refers to the inflow minus the outflows of the host country.
The data in the database of the World Bank also makes a similar distinction, namely the net inflow of FDI and the net outflow of FDI in each country. In this research we are interested in what the relation is between FDI from MNE and the growth of the host country. Therefore for this research the net inflow of FDI is more appropriate. The World Bank states that the net inflow FDI is defined as the new investment inflows minus the disinvestment. The data is constructed as a fraction of the net inflow FDI divided by the GDP of the country of that year.
The political stability (POL) of a country is the first control variable in the regression.
This variable is included because of the relation it has with the growth of a country. If there is a stable political environment, the country will be able to increase economically (Feng, 1997). The data obtained for this measurement comes from the Center for Systemic Peace. We are going to use the political effectiveness as a proxy for the instability. If the political situation in a country is stable, than we assume that the politicians are effective. The proxy will be as follow: 1 = no political effectiveness, 2 = minimal political effectiveness, 3 = political effectiveness, 4 = maximal political effectiveness. This is also one of the variables that determine a state fragility.
The second control variable is the openness (TRADE) of the country. This variable also has an effect on growth, namely due to the more open a country is, which implicates the more export/import will take place (Harrison, 1996). Which in return can lead to a higher GDP for the country. For this variable we use the proxy of trade. Trade is defined as the total amount of import and export of goods and services of a country. Hereby we can indicate how open a country is. In the database of the World Bank this proxy is given as a fraction of the total amount of trade in US dollars divided by the GDP of that country.
The third control variable is the portfolio investment (PORT) of a country. Also called the Foreign Portfolio Investment (FPI). This is included because of
the effect it has on the growth of a country. The more there is invested in stock and bonds of the country, the more capital inflow will take place and the GDP will increase. This variable, which is also found in the database of the World Bank, will be measured as net portfolio investment in US dollars over the real GDP of a country. Net portfolio investment is defined as portfolio investment that covers transactions in equity securities and debt securities.
The last control variable included is taxation (TAX) of a country. This is included because of the indirect relation it has with the growth of a country. The higher the tax percentages of the country, which means that the government will receive more capital to use for other activities (Engen et al, 1996). The taxes of a country also influence the FDI. This depends on the tax system in the host country, which will influence the amount of taxes the MNEs have to pay in the host country (Edmiston et al, 2003). This variable will be determined as a fraction of the government revenue that consists only from taxation over the GDP of the country. This information can be found in the database of the World Bank.
In table 7, 10 and 13 the data descriptive are given for these variables.
3.2 Method and Hypothesis
We will analyze the effect of FDI on growth with the ordinary least square (OLS) method. To be able to use this method, we need to meet some conditions. First, the independent variables have to have some kind of linear relation with the dependent variable. Second, no perfect multicollinearity between explanatory variables can be present, which we can test by using the variance inflation factor (VIF) method. As can be seen in table 9, 12 and 15, we are not dealing with perfect multicollearity. Third, there should be no heteroskedasticity. With the Cook-Weisberg test, we can conclude that there is no heteroskedasticity (see table 16).
With this analysis we can determine if there is a relation between growth and FDI. But we cannot determine if there is causality between the 2 variables. This is mainly due to the concept of reverse causality. Which means that it can
also be that FDI takes place in the host country, because of the high growth rate in that country.
In this research we are determining the effect of FDI on growth of a country. But according to the paper of Borensztein et al (1998) we have to include some control variables in the regression to prevent omitted variable bias.
The first variable included as a control variable, is the political stability of a country. The degree of political stability is a very important aspect to know when talking about the growth of a country. If a country has a very high degree of political instability, such as a high level of corruption or a high degree of rotation of the government, there will be a conflict of interest. The main focus of the government will be to achieve short run goals instead of to increase the growth of a country in the long run (Wholey, 1997). As for the FDI in that country, the higher the degree of political instability in the host country, the less the foreign country will want to invest in that country due to the high uncertainty ( Fatehi-Sedeh et al, 1989). The political stability of a country also determines which spillover channel is important for the country. We expect the sign for this variable to be positive given that the higher the proxy the more stable the political environment.
The second control variable included in our model is country openness. Country openness is defined as the degree of trading from 1 country to the rest of the world. Hereby export plays a role, but also investing in other countries or cooperating with other countries. This factor is crucial for the growth of the host country, because the more open a country is, the more the MNEs are willing to invest in that country (Harrison, 1996). Which means more capital inflow and knowledge spillovers. One of the investment possibilities is the FDI. We expect the estimator sign to be positive. This is because the more a country exports; the more capital will enter the host country.
The third control variable included in the regression is portfolio investment. The more other countries invest in the government stocks, the more capital inflow. Which will increase the balance of payment of a country and in return
affect the growth of the country (Durham, 2004). FDI is another way of investing in a country. We expected this estimator to be positive as well.
The fourth variable is level of taxation. Engen et al (1996) state that the level of taxation has an effect on the investment rate. Namely, the higher the taxes, the less the investment. Besides the discouragement of investment, the level of taxation also has an effect on the labor supply. Which states that if the h taxes are higher, the company has to pay their employees more. Therefore there is a negative effect between taxation and labor supply. In return it also has a effect on the labor mobility channel and. The effects will affect the overall growth of a country. Because the FDI is a type of investment, the taxation will also affect FDI.
The following equation consists of the FDI and the control variables we discussed.
g = + 1 FDI + 2 POL + 3 TRADE + 4 PORT + 5 TAX + (1)
Where g is the percentage growth of the country, FDI is the total amount of net inflow FDI of a country; POL is the political stability denoted as the political effectiveness of the country, TRADE is the openness of the country expressed as the total amount of trade. PORT is the portfolio investment denoted as the total equity and debt securities and TAX is the total government revenues consisting of taxes.
The first analysis will be done for all the countries for 3 separate years, namely 2004, 2008, 2012. These 3 years are chosen because we want to evaluate if there was a difference due to the financial crisis in 2008. The financial crisis had an effect on the world economy. Ever since the Lehman Brother came down in 2008, the financial world went down as well. All investing activities were affected by the fall of the Lehman brothers, including the FDI activities (Taylor, 2009). Therefore we think that the financial crisis had an affect on the FDI activities in the countries.
For the second analysis we will divide the countries in 3 groups according to their geographical location. This will be done on the basis of the 6 continents, respectively, Africa, Asia, Europe, North America, Oceania and South America. Each group consists of 2 continents. With this method we can evaluate if the location where the country is situated has an effect on the relationship between FDI and growth. We think this may differ between locations, because of the upcoming interest for investing in certain parts of the world, such in Asia or in South America.
Section 4: Results
In this section we will discuss the findings of our analysis for each individual year and conclude if our expectations were correct. Besides that we will also make a comparison between the years. And finally the findings of our second analysis, where we divide the countries up according to their geographical location, will be discussed.
The analysis is done on the basis of 149 countries, whereas 2 countries on list 1 are excluded, namely Eritrea and Syria. The lack of availability of the data is the cause of that. According to the IMF, there is a lack of data because the conditions in these countries are unfavorable. We also plotted the data to see if we have any outlier that could influence the data. For each year there were some outliers that we had to take out of our analysis in order to get reliable results. In the following table the analysis is done for the year 2012.
Tabel 1: Determinants of growth for the year 2012
Table 1 reports the estimators of the main variable, FDI, and the control variables on growth. The first regression, where only the FDI-variable is included,
p-values in parentheses adj. R-sq 0.034 0.042 0.076 0.027 0.074 0.129 R-sq 0.041 0.055 0.088 0.041 0.098 0.187 N 147 147 147 147 77 77 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) _cons 2.035*** 1.622*** 3.259*** 2.032*** 3.825*** 4.272*** (0.036) (0.190) tax -0.134** -0.0853 (0.904) (0.062) port 0.00213 -0.180* (0.007) (0.031) trade -0.0170*** -0.0210** (0.143) (0.517) pol 0.382 0.270 (0.014) (0.007) (0.002) (0.015) (0.018) (0.004) fdi 0.0865** 0.0962*** 0.115*** 0.0866** 0.103** 0.130*** growth growth growth growth growth growth (1) (2) (3) (4) (5) (6)
is called the baseline regression. The analysis for 2012 is done with 147 observations. This is due to the elimination of 2 countries, which caused for unreliable results. In the analysis the estimator for FDI is positive, which implicated that if the FDI inflow increases by 1% per year, the annual growth rate will increase by 0,0865%. This implies that there is a positive relation between the 2 variables. That states that our expectations are correct. But according to the R-squared, FDI only explains a minimal portion of the growth. Therefore in model 2, 3, 4, and 5 we expand the first model by simple adding one control variable in each model. And according to the adjusted R-squared in regression 6, the variables indeed add some value to the growth of a country. All the estimators are determined on the basis of 147 observations, except for model 5. The data for the government revenue that only consist of taxes was limited, therefore the observations were less. In model 6, the variable FDI and all the control variables are included. According to this analysis, the FDI is has a significant relation with the growth of the country. This can be seen in all of the regressions. On the basis of the model, we can conclude that in 2012, the FDI activities were important for the growth of the country according to their significant levels. As Forster et al (2009) stated, after the financial crisis, the interest of investors was to aim towards the less developed countries instead of in developed countries. That is because the financial crisis in 2008 had a bigger consequence for the developed countries, than for the less-developed countries. In model 6 we can conclude that by expanding the model with the control variables, according to the adjusted R-squared, the growth is better explained than in model 1.
Tabel 2: Determinants of growth for the year 2008
In table 2, the same method is applied as in table 1, except for a different year, namely 2008. This analysis is done with 148 observations, because in this year there was one country that caused for an outlier to appear. Which would have made our results unreliable. Also in this year, we can find that estimator of FDI in all the regressions are positive, which supports our expectations. In most of the regression, the FDI is insignificant for the growth of the country. We can also see that in model 6, the trade estimator and the tax coefficient sign are negative. Which is in contrary to our expectations. What also took our attention is that the adjusted R-squared of regression 6 is relatively low compared to the other regressions.
The next table will illustrate the finding for the year before the crisis. * p<0.1, ** p<0.05, *** p<0.01 p-values in parentheses adj. R-sq 0.009 0.023 0.008 0.051 -0.021 0.001 R-sq 0.016 0.036 0.022 0.064 0.002 0.058 N 148 148 148 148 89 89 (0.000) (0.000) (0.000) (0.000) (0.000) (0.003) _cons 2.667*** 2.123*** 3.145*** 2.582*** 3.846*** 3.314*** (0.660) (0.733) tax -0.0182 -0.0152 (0.007) (0.292) port 0.0326*** 0.0791 (0.341) (0.759) trade -0.00644 -0.00274 (0.081) (0.082) pol 0.529* 0.670* (0.128) (0.080) (0.085) (0.112) (0.846) (0.472) fdi 0.0620 0.0714* 0.0733* 0.0634 0.00846 0.0321 growth growth growth growth growth growth (1) (2) (3) (4) (5) (6)
Tabel 3: Determinants of growth for the year 2004
The third table is similar to the previous tables, but this time for the year 2004. This analysis is done with 147 observations. We had to exclude 2 countries for this year as well to be able to get reliable results. By looking at the first 4 regressions, we can conclude that the FDI has a significant relation with the growth of the country. Regressions 5 and 6 have fewer observations, which could be a reason that the FDI has an insignificant affect on growth, in those regressions.
If we compare the 3 tables with each other, we can conclude that the FDI in 2012 is significant and has the highest estimator. Which implies that in 2012 the relationship between FDI and the growth of a country was stronger than the other years. Due to the high correlation between the 2 variables, we can conclude that the spillover effects in 2012 were bigger than the years before. But taking into account that reverse causality could also have taken place. Besides that in 2012 the relation between FDI and the growth of a country is the highest,
* p<0.1, ** p<0.05, *** p<0.01 p-values in parentheses adj. R-sq 0.016 0.020 0.047 0.013 0.025 0.007 R-sq 0.023 0.033 0.060 0.026 0.050 0.070 N 147 147 147 147 80 80 (0.000) (0.000) (0.054) (0.000) (0.000) (0.002) _cons 3.439*** 2.856*** 1.795* 3.382*** 7.063*** 5.623*** (0.048) (0.052) tax -0.153** -0.168* (0.488) (0.991) port 0.0849 0.00155 (0.019) (0.418) trade 0.0236** 0.0120 (0.213) (0.312) pol 0.513 0.533 (0.067) (0.067) (0.242) (0.061) (0.689) (0.616) fdi 0.219* 0.219* 0.143 0.226* 0.0631 0.0819 growth growth growth growth growth growth (1) (2) (3) (4) (5) (6)
in contrary to that, the relation between these 2 variables is the lowest in 2008. This supports our expectation.
The second part of our analysis, we separated the countries according to their geographical location. The countries are divided into 3 groups. The first group consists of the countries that are situated in Africa and in Europe. Second group consist of the countries that are situated in Asia and Oceania. And for the third group, which consists of the countries that are situated in North America and South America. For each group, we took 2 continents. This is because of the minimal observations we want to have for each analysis. Some continents do not have a lot of less-developed countries; therefore we had to combine them with another continent in their region. In list 2, 3, 4, 5 and 6 you can find the continents with the countries that belong to that continent.
We also excluded the TAX control variable from this analysis. The reason for that is the lack of data for this variable per group. If we included the TAX-variable, we would not meet our benchmark of 25 countries per group. In the first table, we start by doing our analysis for the countries that are situated in Africa and in Europe.
Table 4: Determinants for growth for Africa & Europe
* p<0.1, ** p<0.05, *** p<0.01 p-values in parentheses adj. R-sq -0.000 -0.032 0.032 R-sq 0.061 0.032 0.092 N 66 66 66 (0.359) (0.133) (0.562) _cons -4.455 1.785 1.159 (0.968) (0.635) (0.699) port 0.0231 -0.0991 0.157 (0.389) (0.693) (0.072) trade 0.0385 0.00432 0.0339* (0.065) (0.332) (0.911) pol 3.266* 0.523 0.0849 (0.737) (0.518) (0.537) fdi 0.0681 0.0398 0.151 growth2012 growth2008 growth2004 (1) (2) (3)
In table 4 the determinants of the growth for Africa and Europe are presented for the 3 years respectively 2012, 2008, 2004. We can also conclude that in 2012 the variables included explain the growth of Africa and Europe better than the previous years. The conclusion is based on the adjusted R-squaredof the models. Now looking at the coefficient of FDI. For all the years it is positive, as we expected. Which means that for all the years there was a positive relation between FDI and the growth of a country. Where in 2004 the coefficient is the highest.
This group of countries is very diverse. According to the IMF Europe is considered a relatively rich continent, which is in contrast to Africa. Africa is considered a relatively poor continent. Europe is considered relatively rich because most of the countries are considered to be developed countries. Whereas we can conclude that some of the determinants of the degree of spillovers, such as regional effects and technological gap are optimal due to the relative small distance between the countries.
Africa, on the other hand, is relatively poor and less advanced than Europe. Whereas we can assume that the MNEs who want to invest in Africa come from a further distance, which according to regional effect determinants, the degree of spillover will be much less than if a country in the region wanted to invest. Other than that, because Africa is not that advanced it may be a disadvantage for the continent. For example, to be able to benefit from the technological spillovers, the country has to have some knowledge to be able to have a sustainable technological gap between the host country and the foreign country. But on the other hand, it does not have their own system, which means they are open to learn and absorb knew knowledge.
The next table illustrates the findings for the countries situated in Asia and in Oceania.
* p<0.1, ** p<0.05, *** p<0.01 p-values in parentheses adj. R-sq 0.063 0.286 0.187 R-sq 0.138 0.343 0.252 N 51 51 51 (0.002) (0.347) (0.724) _cons 3.967*** 1.113 -0.840 (0.783) (0.008) (0.768) port 0.00593 0.0361*** -0.0544 (0.040) (0.292) (0.124) trade -0.0253** -0.0132 0.0367 (0.504) (0.011) (0.001) pol 0.401 1.541** 4.061*** (0.040) (0.010) (0.374) fdi 0.208** 0.358*** -0.147 growth2012 growth2008 growth2004 (1) (2) (3)
Table 5: Determinants for growth for Asia & Oceania
In table 5, the same analysis is done but for the countries that are situated in Asia and Oceania. Compared to the first group, in this analysis we find a negative estimator for the regression of 2004, but in 2008 and 2012 the estimator has a positive and significant relation to the growth of the country. Where the FDI-coefficient is the highest in 2008.
In this group, most of the countries are situated in Asia, as can be seen in list 2. Because Oceania has limited countries in this analysis, it makes it difficult to make a conclusion for these countries.
Asia has become of great interest for the MNEs to invest, because of the relatively low cost (Holden et al, 2009). Because of the low cost, the MNEs are interested in investing is Asia, but due to the backward/forward channel, the MNEs will improve the quality of the firms in Asia itself. This all can lead to an increase of the growth rate in the countries. The export channel will also play a role in this subgroup. Because of the higher quality, the firm will work more efficiently and therefore can produce more to export.
In the next table the outcomes are listed of the analysis of North America and South America.
Table 6: Determinants for growth for North America & South America
In table 6, the analysis is done for the countries that are located in North America and South America. The FDI-estimators are all positive, but none of them have a significant relation with the growth of the country. In 2004 the relation between growth and FDI was the highest.
Also this group is a special group. South America is considered to be relatively poor continent over the years. In those countries there is a lot of political corruption, which influences the FDI. Before the Latin America debt crisis, the government had a lot of rules regarding the inflow and outflow of capital (Pilbeam, 2013). Therefore there was nearly any activity of FDI. After the crisis, those rules were softened and the FDI activities began to increase. The advantage of South America is that it is situated relatively close to the U.S, whereas the regional effect determinant will make sure that a maximal spillover effect will take place. This can go through the demonstration/imitation channel or even through the labor mobility channel. Because over the years, the
* p<0.1, ** p<0.05, *** p<0.01 p-values in parentheses adj. R-sq -0.049 -0.125 0.062 R-sq 0.087 0.020 0.183 N 32 32 32 (0.960) (0.115) (0.661) _cons 0.0635 2.636 0.889 (0.553) (0.514) (0.043) port -0.119 0.286 0.576** (0.275) (0.812) (0.214) trade 0.0150 -0.00382 0.0210 (0.575) (0.921) (0.571) pol 0.371 0.0506 0.458 (0.709) (0.956) (0.464) fdi 0.0527 0.00463 0.159 growth2012 growth2008 growth2004 (1) (2) (3)
collaboration of North America and South America has increased, which benefits the labor mobility channel of FDI. That is because the people can easily move from one country to another (Schiff, 1992).
North America, which consists mostly of the U.S and Canada, is considered to be a relatively rich continent, in the sense that it is active in the investment world (Wheeler, 1992).
In list 5, we can see that most countries that are situated in North America and are considered to be less developed, are islands in the Caribbean. For an island there is only a limited capacity for FDI. That is because an island is relatively small and cannot expand. The maximum capacity of spillovers has to go through the channels to be able to stimulate the growth. Next to that, the population on the islands is relatively poor compared to the continent (Chai et al, 2008). Which means that the labor mobility channel cannot provide on these islands.
As can be concluded, every group has it’s own characteristics towards FDI. Some started off with a higher coefficient for FDI in 2004 and some ended it with a higher coefficient. For all these continents there are some arguments, some had a war going on recently, other just have more developing countries than transition countries. But overall the financial crisis in 2008 affected the whole world including these countries (Chari et al, 2008).
For further research it would be interesting to divide the countries up into 2 groups, namely developing and transition countries. The purpose of the research would be to analyze if there is a difference in growth due to FDI in developing countries and in transition countries.
It would also be very interesting to divide the countries op according to their continent, but that can only be done with all the countries in the world, instead of only less developed countries. This is due to the limited observations per continent.
Section 5: Conclusion
5.1 Summary
This paper focuses on the empirical relation between FDI and the growth of less developed countries. The analysis is done with the data of 149 countries that are listed at the IMF. The main motivation for this paper was to see if the relation between FDI and the growth of a country has changed in the recent years. In this period the financial crisis also took place, therefore it is easier to conclude if that had an effect on the FDI. Besides the overall relationship between FDI and the growth, we also divided the countries up into 3 groups according to their geographical location, to see if there is a significant difference between the continents on the basis of FDI.
The overall result was that the coefficient of the FDI was higher in the latest year, namely 2012. And the coefficient was the lowest in 2008, which we conclude was due to the financial crisis. Most of the FDI-coefficients were positive. In 2012, we can say that the FDI had a significant relation with the growth of the country, because in all of the regressions, the FDI-variable was significant. After dividing the countries up into 3 groups, based on their geographical location, we can conclude that for each group over the years there was a positive FDI-coefficient. Which means that there was a positive relation between the FDI and the growth of a country. Nevertheless each group has it’s own characteristics. These characteristics determine in what degree of which channel will have a better impact on the growth of the country. For example, in Europe their regional effect is one of the main determinants for the spillovers. As for South America the labor mobility channel became popular after the financial debt crisis in that region. And for Asia the low cost in the region was important for the backward/forward channel.
Concluding that each region has it’s own aspects towards FDI, but that overall in 2012 the relation between FDI and growth was the highest.
5.2 Limitations and improvements
The outcome of our analysis is in line with the research of Borensztein (1998), Kobrin (2005) and Zang (2001). Which states that the FDI has a positive relation with the growth of a country. But the significant level of the FDI– coefficient in the analysis is not in line with any of the researches.
In our analysis we used the simple OLS method with 4 control variables. In reality the growth of a country is determined by much more variables than only the one mentioned in this research. For example, for Barro’s (1996) research the variables human capital, years of schooling and much more are included in the regression. This could be a reason that most variables are not significant to growth. Therefore for future research it may be an idea to use a more complex model with more variables. In that case we would get a more realistic view of the economy in less developed countries.
Besides the over simplistic model, we also had the issue of reverse causality. This is also a reason to use another model for future research. To determine whether the FDI has an effect on the growth of a country, we need to use a model were we can clarify that only the FDI has an effect on growth and not the other way around.
A reason the results could be significant could be the way we entered the variables in the regression. In this research most of the variables are insert as a percentage of the GDP and the growth is determined as the percentage change of the GDP. But it could be that instead of a percentage of GDP, the change in amount of a variable gives better results. In that case we would have to find that data in a different database, because the World Bank has only limited data in that form.
For later research it would be an idea to take a time period instead of just 1 year. Due to a greater time period it can cause for more significant results. Because once a MNE invest in a country, it take time before the spillover effects affects the countries growth (Crespo et al, 2007).
References
Arestis, P., & Demetriades, P. (1997). Financial development and economic growth: Assessing the evidence. Economic Journal 107, 783–799.
Barro, R. J. (1996). Determinants of economic growth: a cross-country empirical
study. National Bureau of Economic Research. Working paper 5698
Bouoiyour, J., & Akhawayn, A. (2003). Labour productivity, technological gap and spillovers: Evidence from Moroccan manufacturing industries. WPCATT,
University of Pau.
Borensztein, E., De Gregorio, J., & Lee, J.-W. (1998). How does foreign direct investment affect economic growth. Journal of International Economics 45, 115–135.
Brambilla, I., Hale, G., & Long, C. (2009). Foreign direct investment and the incentives to innovate and imitate. Scandinavian Journal of Economics, 111, 835–861.
Cannon, J.P., &Homburg, C. (2001). Buyer-Supplier relationship and customer firm cost. Journal of marketing, 65, 29-43
Carkovic, M. V., & Levine, R. (2002). Does foreign direct investment accelerate economic growth?. University of Minnesota Department of Finance
Working Paper.
Chai, J., & Goyal, R. (2008). Tax concessions and foreign direct investment in the Eastern Caribbean Currency Union. International Monetary Fund. 257(8).
Chari, V. V., Christiano, L., & Kehoe, P. J. (2008). Facts and Myths about the Financial Crisis of 2008. Federal Reserve Bank of Minneapolis Working
Paper, 666.
Collier, M. W. (2002, June). The effects of political corruption on Caribbean development. In Paper pre sented at the Caribbean Studies Association
Annual Conference, Nassau, Bahamas, May.
Crespo, N., & Fontoura, M. (2007). Determinant factors of FDI spillovers—what do we really know?, World Development, 35, 410–425.
Crespo, N., Fontoura, M.P., Proença, I. (2009). FDI spillovers at regional level: Evidence from Portugal. Papers in Regional Science, 88(3), 591–607. De Gregorio, J. (1992). Economic growth in Latin America. Journal of
Dupasquier, C., & Osakwe, N. (2006). Foreign Direct Investment in Africa: Performance challenges and responsibilities. Journal of Asian Economics, 17, 241-260
Durham, J. B. (2004). Absorptive capacity and the effects of foreign direct investment and equity foreign portfolio investment on economic growth.
European economic review, 48(2), 285-306.
Edmiston, K., Mudd, S., & Valev, N. (2003). Tax structures and FDI: The determent effects of complexity and uncertainty. Fiscal studies, 24(3), 341-359.
Egelhoff, W.G., Gorman, L., &McCormick, S. (2000). How FDI Characteristics influence subsidiary trade patterns: The case of Ireland. Management International Review, 40 (3),203 – 230.
Engen, E. M., & Skinner, J. (1996).Taxation and economic growth (No. w5826). National Bureau of Economic Research.
Fagerberg, J. (1987). A technology gap approach to why growth rates differ. Research Policy, 16, 87-99.
Fatehi-Sedeh, K., & Safizadeh, M. H. (1989). The association between political instability and flow of foreign direct investment. Management international
review, 4-13.
Feng, Y. (1997). Democracy, political instability and economic growth. British Journal of political science, 27, 391 – 418
Foster, J. B., & Magdoff, F. (2009). The great financial crisis: Causes and
consequences. NYU Press.
Gastanaga, V., Nugent, J. B., & Pashamova, B. (1998). Host country reforms and FDI inflows: How much difference do they make? World Development, 26(7), 1299-1314.
Görg, H., & Strobl, E. (2005). Spillovers from foreign firms through worker mobility: An empirical investigation. The Scandinavian journal of
economics, 107(4), 693-709.
Hale, G., & Long, C. (2006). What determines technological spillovers of foreign direct investment: evidence from China. Economic Growth Center
Harrison, A. (1996). Openness and growth: A time-series, cross-country analysis for developing countries. Journal of development Economics, 48(2), 419-447.
Holden, S. T., Deininger, K., & Ghebru, H. (2009). Impacts of low-cost land certification on investment and productivity. American Journal of
Agricultural Economics, 91(2), 359-373.
Kobrin, S. (2005). The determinants of liberalization of FDI policy in developing countries: 1991–2001. Transnational Corporations, 14(1), 67–103.
Kowalski, E. (2000). Determinants of Economic Growth in East Asia: A Linear Regression Model. Research Honors Project, Illinois Wesleyan University
Lensink, R., & Morrisey, O. (2001). Foreign direct investment: Flows, volatility and growth in developing countries. Globalisation and Poverty DESG 2001. 32, Nottingham.
Makino, S., Beamish P. W., & Zhao B.N. (2004a). The characteristics and
performance of Japanese FDI in less developed and developed countries. Journal of World Business, 39(4), 377–392.
Markusen, J.R.(1995). The Boundaries of Multinational Enterprises and the Theory of International Trade. Journal of Economic Perspectives, 9, 169-189.
Muller, T. (2007). Analyzing Modes of Foreign Entry: Greenfield Investment versus Acquisition, Review of International Economics, 15, 93–111. Pilbeam, K. (2013) International Finance (4th edition), Palgrave Macmillan
Schiff, M. (1992). Social capital, labor mobility, and welfare the impact of Uniting States. Rationality and Society, 4(2), 157-175.
Shin, J., & Chang, H.J. (2005). Economic Reform after the Financial Crisis: A Critical Assessment of Institutional Transition and Transition Costs in South Korea. Review of International Political Economy, 12 (3), 409–433. Taylor, J. B. (2009). The financial crisis and the policy responses: An empirical
analysis of what went wrong. National Bureau of Economic Research. Working Paper No. 14631
Temiz, D., & Gokmen, A. (2014). FDI inflow as an international business
operation by MNCs and economic growth: An empirical study on Turkey. International Business Review, 23, 145–154.
case of US firms. Journal of international economics, 33(1), 57-76.
Wholey, J. S. (1997). Clarifying goals, reporting results. New Directions for
Evaluation, 1997(76), 95-105.
Wojnicka, E. (2001). Foreign direct investment in the privatization of the Polish economy, Intereconomics, 36(6), 305-314.
Zang, K. H. (2001). Does Foreign Direct Investment promote economic growth Evidence from East Asia and Latin America. Contemporary Economic Policy, 19(2), 175-185
Appendix
List 2: Less-developed countries situated in Africa
List 3: Less-developed countries situated in Asia
List 4: Less-developed countries situated in Europe
List 6: Less-developed countries situated in Oceania
List 7: Less-developed countries situated in South America
Table 7: Summary data for the year 2012
Table 8: Correlation between the explanatory variables for 2012
Table 9: VIF test for 2012
growth 147 2.492334 3.229169 -7.801452 12.74801 tax 77 16.06549 5.632957 .0195583 28.40163 port 147 .8131954 15.0065 -18.91694 175.6246 trade 147 80.84688 43.31576 0 202.2168 pol 147 .9455782 1.025579 0 3 fdi 147 5.295159 7.527678 -5.496736 54.06343 Variable Obs Mean Std. Dev. Min Max
pol -0.2114 0.0542 -0.1986 0.0645 1.0000 port -0.0777 -0.0922 0.0472 1.0000 tax 0.2939 0.3256 1.0000 trade 0.3795 1.0000 fdi 1.0000 fdi trade tax port pol
Table 10: Summary data for the year 2008
Table 11: Correlation between explanatory variables 2008
Table 12: VIF test for 2008
Table 13: Summary data for the year 2004 Mean VIF 1.19 port 1.03 0.974598 pol 1.12 0.894868 tax 1.22 0.819776 fdi 1.28 0.783404 trade 1.30 0.768079 Variable VIF 1/VIF
growth 148 3.067948 3.566822 -10.11877 13.31262 tax 89 17.18591 8.133343 1.235817 58.6914 port 148 2.341154 24.06221 -36.08154 288.6346 trade 148 85.48559 45.31212 0 321.6317 pol 148 .9121622 .9754054 0 3 fdi 148 6.464455 7.234281 -6.549732 47.75013 Variable Obs Mean Std. Dev. Min Max
pol -0.2623 -0.1387 -0.1419 0.0650 1.0000 port -0.0252 -0.0720 0.1913 1.0000 tax 0.1638 0.3409 1.0000 trade 0.1817 1.0000 fdi 1.0000 fdi trade tax port pol
Mean VIF 1.14 port 1.07 0.935412 pol 1.10 0.911946 fdi 1.11 0.902131 trade 1.18 0.847325 tax 1.22 0.818515 Variable VIF 1/VIF
Table 14: Correlation between explanatory variables 2004
Table 15: VIF test for 2004
Table 16: Cook- Weisberg test Years chi2 p>chi2
2012 1.04 0.3068 2008 0.21 0.6432 2004 5.83 0.9215 growth 147 4.260985 5.347692 -6.493173 33.57581 tax 80 15.48015 6.919174 .9054617 46.71758 port 147 .3777449 3.614681 -7.244831 26.65294 trade 147 81.65822 45.02225 0 290.4993 pol 147 1.136054 1.070395 0 3 fdi 147 3.751372 3.692378 -2.51332 21.3587 Variable Obs Mean Std. Dev. Min Max
pol -0.2376 -0.1713 -0.1935 -0.0237 1.0000 port 0.0136 0.0345 -0.0999 1.0000 tax 0.1876 0.4304 1.0000 trade 0.1872 1.0000 fdi 1.0000 fdi trade tax port pol
Mean VIF 1.15 port 1.02 0.981145 pol 1.09 0.914230 fdi 1.10 0.912875 trade 1.26 0.793195 tax 1.28 0.779022 Variable VIF 1/VIF
