FDI inflows in the Netherlands: what effect do cultural distance, spatial distance and institutional distance have?

(1)

1

FDI inflows in the Netherlands:

what effect do cultural distance, spatial distance and

institutional distance have?

Abstract:

As FDI has a prominent role in economic research and development, this thesis analyses the effect of cultural distance, spatial distance and institutional distance on the FDI inflows in the Netherlands. A dataset consisting of FDI inflows from 64 different countries over a period ranging from 2008 until 2012 is used. Results show a significant negative coefficient for spatial distance and its influence on FDI inflows, whereas cultural distance shows an insignificant negative coefficient for its influence on FDI inflows. Furthermore, institutional distance shows positive effects on FDI inflows, all being economically significant.

University of Groningen, Faculty of Economics and Business MSc International Economics & Business

Author: Jasper Timmerman

E-mail address: j.timmerman.7@student.rug.nl Student number: S2337207

(2)

2 Tabel of Content Introduction ... 3 Structure ... 4 Literature review ... 5 Theoretical Background ... 5

Production cycle theory of Vernon ... 5

The theory of Exchange rates on Imperfect Capital market ... 5

The Internationalisation theory ... 5

The Eclectic Paradigm by Dunning... 6

FDI ... 6

Spatial distance ... 7

Cultural distance... 7

Institutional distance ... 9

Empirical studies on the determinants of FDI ... 11

Spatial distance ... 11 Cultural distance... 11 Institutional distance ... 12 Methodology ... 12 Dependent variable ... 12 Spatial distance ... 13 Institutional distance ... 13 Cultural distance ... 13 Control variables ... 14 Research Design ... 17 Analysis ... 21

Cultural Distance – Regression 1 ... 21

Institutional Distance -Regression 2 ... 22

Spatial Distance – Regression 3 ... 22

Limitations... 32

Conclusion ... 33

References ... 34

Appendix ... 40

Appendix 1 ... 40

FDI dataset changes ... 40

Appendix 2: ... 41

(3)

3

Introduction

The global economy is characterized by international transfers in services, goods and capital. Foreign direct investment (FDI) is one of the most important components of such transfers (Iamsiraroj, 2016). FDI has seen substantial growth since 1992 (UNCTAD, 2017). Growth of FDI levels is associated with an increase in economic growth and vice versa (Hansen & Rand, 2006; Ayanwale, 2007; Liu, Burridge, & Sinclair, 2002) and is therefore important to globalized countries. The Netherlands is highly dependent on FDI inflows, as they are one of the most economically internationalized countries in the global economy (Centraal Bureau Statistiek, 2014), for the period of 2005 – 2015 the Netherlands had on average an absolute increase in inward and outward FDI flows and stocks (OECD, 2017). This is identical to the global trend, however 63% of Dutch outward FDI went to the EU-15 and the United States (Centraal Bureau Statistiek, 2014). The spatial distance between these trading partners differs significantly, raising the question what influences the Dutch FDI flows. Former research analysed the cause of differences in accumulation of FDI for specific countries before, and found evidence for psychic distance (Beckerman, 1956), cultural distance (Hofstede, 1980) and institutional distance (Kostova, 1999) influencing FDI inflows.

The objective of this thesis, is to analyse the effect of spatial, cultural, and institutional distance on the level of Dutch FDI flows. For analysing spatial distance the database on spatial country distances of the Centre d’Etudes Prospectives et d’Informations Internationales (CEPII) will be used. The analysis of cultural distance will follow with the help of previous literature and an adjusted version of the accompanying cultural distance index developed by Kogut & Singh (1988). Whereas institutional distance will be calculated with the help of Global Competitiveness Reports of the World Economic Forum. The corresponding FDI flows will be taken from the Dutch National Bank.

The Netherlands is located in West-Europe, with similar cultures as Germany, Belgium and France . The spatial distance between Netherlands and the rest of Europe is low. Research has shown that developed countries like to invest in culturally similar countries (Kogut & Singh, 1988). Continuing, low institutional distances, the extent of similarity or dissimilarity between the regulatory, cognitive and normative institutions of two countries (Kostova, 1996), will result in full ownership investments such as FDI (Xu & Shenkar, 2002). All these results show the trade-off to take into account for FDI flows, and form the foundation of the following research question:

Do cultural distance, spatial distance and institutional distance influences the level of Dutch FDI inflows?

(4)

4

shows an insignificant negative coefficient for its influence on FDI inflows. Furthermore, institutional distance shows positive effects on FDI inflows, all being economically significant.

Structure

(5)

5

Literature review

FDI is a complicated field of study, researched by many scholars. The largest and most popular theoretical strands about FDI will be discussed in this section. Furthermore, background information on cultural, spatial and institutional distance will be provided. Finally, relevant literature will be reviewed. First we will provide theoretical background on the foundation of FDI flows.

Theoretical Background

Production cycle theory of Vernon

The production cycle theory of Vernon is the first influential piece about FDI and its movements along locations. Vernon (1966) proposed four stages, which he called production cycle theory, to explain US FDI investments in Western Europe. The four stages of this cycle were: innovation, growth, maturity and decline. The stage of innovation was the vocal point of Vernon in explaining the FDI of US transnationals at that time. He argued, that thanks to the demand for the innovative American products after the second World War, the US firms began to export these products to Western Europe. However, the advantage of possessing new technologies fades away as the product develops (Denisia, 2010). And with the awareness of the innovative technologies, Western Europe began to imitate these products. This specific cycle forces the American firms to start production facilities overseas, also known as FDI. This theory held its own in explaining FDI for a decade, however the more the world globalized, the less cases the production cycle theory applied to. In an attempt to provide a more empirical analysis for FDI, the theory of exchange rates on imperfect capital markets was established in 1981 by Itagaki.

The theory of Exchange rates on Imperfect Capital market

In 1981, Itagaki proposed a different view on FDI flows in comparison to the production cycle theory of Vernon. The newly introduced theory had real exchange rate changes as the driving force behind FDI. The concept behind this was that, when real exchange rates for a country appreciated, FDI increased thanks to carrying ‘more‘ value abroad, whereas the opposite was expected at a depreciation. This theory was tested empirically by Cushman in 1985, and he found a significant correlation between the real exchange rate of the USD, and American FDI levels. The results showed an increase in FDI when the USD appreciated, whereas the opposite was seen with a depreciation. The limitation with this theory, is the applicability for simultaneous in- and outflows of FDI between different countries in different currencies. During the same time period, another influential theory regarding FDI was developed, the Internalisation theory.

The Internationalisation theory

The internationalisation theory was founded by Hymer in 1976, and predicts the underlying determinants of FDI. The theory moves beyond that of international capital movements and those movements in response to differences in interest-rates. Hymer (1976) gives two causes for FDI. “(1) Firms control enterprises in many countries in order to remove competition between them. (2) Firms undertake operations in a foreign country in order to appropriate fully the returns to certain abilities which they possess” (p. 3).

(6)

6 The Eclectic Paradigm by Dunning

The eclectic paradigm which was originally presented in 1976 at a Nobel Symposium in Stockholm, had its origins in the 1950s. At that time, Dunning proposed the first of the following three forces in determining the level of FDI, ownership-specific effects, later accompanied by location-specific and the internationalization effect. These three determinants are also known as the OLI-framework (Hagen, 1997). The propositions of the eclectic paradigm were presented as follows by Dunning (2001):

(1) The (net) competitive advantages which firms of one nationality possess over those of another nationality in supplying any particular market or set of markets. These advantages may arise either from the firm’s privileged ownership of, or access to, a set of income-generating assets, or from their ability to co-ordinate these assets with other assets across national boundaries in a way that benefits them relative to their competitors, or potential competitors.

(2) The extent to which firms perceive it to be in their best interests to internalise the markets for the generation and/or the use of these assets; and by so doing add value to them.

(3) The extent to which firms choose to locate these value-adding activities outside their national boundaries. (p. 176)

Ownership advantages includes aspects such as trademark or entrepreneurial skills, whereas location advantages benefits for example from natural resources availability or tariffs. Even though the global economy and the subsequent growth of global and financial capitalism have fundamentally affected the way in which multinational corporations (MNCs) activities are undertaken and organised since the birth of the OLI-framework. The paradigm continues to provide a framework which facilitates how best to synthesise relevant complementary theories, or how to choose between potentially competing theories, and helps to operationalise them (Cantwell & Narula, 2001).

FDI

After seeing different influential theoretical strands on foreign direct investment, we will elaborate upon the subject of FDI more in depth in the following section. FDI is when a firm invests directly in facilities to produce and/or market a product in a foreign country (Hill, 2003). This differentiates FDI from other international capital flows, thanks to its specific focus on market/product facilities, and notion of direct control. Firms also have other options to access foreign markets, such as licensing and exporting, these however have disadvantages such as transportation costs and trade barriers, and the loss of technological know-how control (Hill, 2003).

(7)

7

In further discussing trends and theory about FDI, a differentiation has to be made between FDI flows and FDI stocks. The OECD (2014) provides the following definitions: “FDI flows record the value of cross-border transactions related to direct investment during a given period of time, usually a quarter or a year. Whereas FDI stocks measure the total level of direct investment at a given point in time, usually the end of a quarter or of a year” ( OECD Data, FDI Flows). Data provided by the IMF, shows that both FDI flows and FDI stocks have been increasing for the past decades, with the first decrease seen around the 2000s recession (Patterson, Montanjees, Motala, & Cardillo, 2004). The dominant sector in which FDI flowed changed, as shown by Doytch (2011):

The shares of the FDI net inflows (the difference between purchases and sales of domestic assets by foreigners) by sectors display very similar patterns. During 1990– 2004, the period of study of this paper, the share of the service FDI net inflows in the sample of 60 examined countries rose by 11%, from 44 to 55%, while the share of manufacturing FDI net inflows fell by 12% from 33 to 21%. (p. 412)

The trends in FDI have been shown to change repeatedly, which further emphasises why research on FDI has to be up to date with the latest changes, since it can have major inferences for both location as level of FDI.

Spatial distance

As we have previously seen, theory on FDI started to get popularized in the 1960s, the years in which FDI became a fundamental driver of economic growth discussions (te Velde, 2006). Especially in the 1960s, there was no common consensus on the impact of FDI on economic growth. Some authors argued that FDI leads to economic growth and productivity increases in the economy as a whole and hence contributes to differences in economic growth as argued by a multitude of research (Zhang, 2001; Adams, 2009; Baliamoune-Lutz, 2004). Others stress the risk of FDI destroying local capabilities and extracting natural resources without adequately compensating poor countries (te Velde, 2006).

Thanks to the increasing popularity of FDI, researchers started to focus more on variables affecting these flows. In 1966, H. Linnemann published one of the earliest empirical tests on the effect of geographical distance on FDI flows between countries, which found that the larger the geographical distance, the smaller the level of FDI flows. Considering the timeframe of Linnemann’s research, it was not surprising geographical distance negatively affected FDI flows, since information technologies and transportation quality was still underdeveloped. The research by Linnemann (1966) was pivotal in the continuing strand of research on geographical distance and FDI flows, fuelled by both information and transportation revolutions. This ever changing environment is one of the reasons why geographical distance and its influence on FDI is still an important research subject, since it asks for up to date analyses.

Cultural distance

(8)

8

between 1967 and 1973. Hofstede (2011) gives the following definitions of the four dimensions:

Power Distance

Power Distance has been defined as the extent to which the less powerful members of organizations and institutions (like the family) accept and expect that power is distributed unequally. This represents inequality (more versus less), but defined from below, not from above. It suggests that a society's level of inequality is endorsed by the followers as much as by the leaders. Power and inequality, of course, are extremely fundamental facts of any society. All societies are unequal, but some are more unequal than others.

Uncertainty Avoidance

Uncertainty Avoidance is not the same as risk avoidance; it deals with a society's tolerance for ambiguity. It indicates to what extent a culture programs its members to feel either uncomfortable or comfortable in unstructured situations. Unstructured situations are novel, unknown, surprising, and different from usual. Uncertainty avoiding cultures try to minimize the possibility of such situations by strict behavioural codes, laws and rules, disapproval of deviant opinions and a belief in absolute Truth; 'there can only be one Truth and we have it'.

Individualism

Individualism on the one side versus its opposite, Collectivism, as a societal, not an individual characteristic, is the degree to which people in a society are integrated into groups. On the individualist side we find cultures in which the ties between individuals are loose: everyone is expected to look after him/herself and his/her immediate family. On the collectivist side we find cultures in which people from birth onwards are integrated into strong, cohesive in-groups, often extended families (with uncles, aunts and grandparents) that continue protecting them in exchange for unquestioning loyalty, and oppose other ingroups. Again, the issue addressed by this dimension is an extremely fundamental one, regarding all societies in the world. Table 3 lists a selection of differences between societies that validation research showed to be associated with this dimension.

Masculinity – Femininity

(9)

9

In 1988, Kogut & Singh were the first to propose a mathematical concept for calculating cultural distance, which was based on the deviation between the specific countries and the four cultural dimensions, and then correcting it for variance differences. Kogut & Singh (1988) provided us with the following index:

𝐶𝐷𝑗 = ∑{ (𝐼𝑖𝑗 − 𝐼𝑖𝑢)2_{/ 𝑉𝑖) } /4} 4

𝑖=1

“Where CDj stands for the cultural distance between country j and the United States. Iij stands for the ith cultural dimension and jth country, Vi is the variance of the index of the ith dimension, u indicates the United States “ (Kogut & Singh, 1988, p. 422). This index will be adjusted as to match the purpose of this thesis. Since the country of comparison in this case is the Netherlands, the 𝐼𝑖𝑢 will be changed to include the Netherlands instead of the United States, 𝐼𝑖𝑛. Where 𝐼𝑖𝑛 stands for the ith cultural dimension for country n, indicating the Netherlands.

The cultural dimensions faced some critique in the last decade, on which Berry, Guillen and Zhou (2010) introduced the following four points of critique, based on the supporting literature. The first point of critique is the fact that all cross-country differences are assigned towards a difference in cultural distance, which removes the ability of other arrays of distance to explain differences (Ghemawat, 2001). Second, the cultural dimensions were proposed to be non-affected by time, whereas recent studies have shown political, cultural and economic distance can change swiftly change over time (Inglehart & Baker, 2000; Shenkar, 2001). The final two points relate to the interpretability of the cultural dimensions, since they cannot be used for individualistic inferences, even though Hofstede assumed the sample size of one firm was representative for the overall population (Berry et al., 2010).

Institutional distance

(10)

10

Regulative distance: property rights, intellectual property rights protection, judicial independence, burden of government regulations, efficiency of legal framework, transparency of government policymaking.

The normative distance: ethical behaviour, strength of auditing and reporting standards, efficacy of corporate boards, quality of management schools, local availability of specialized research and training services. (p. 187)

(11)

11

Empirical studies on the determinants of FDI

The following section will review the relevant literature available on the variables. The review will consist of different findings on the specific subjects, providing a broad spectrum of views to take into account, resulting in the research hypotheses.

Spatial distance

Traditional International Business theory suggests a preference for physically closer countries and correspondingly predicts a negative relationship between country distance and FDI flows (Chari & Shaikh, 2017). Thanks to larger transferability costs of competitive advantages (Markusen, 1984) Ragozzino (2009) showed in his research that U.S firms resorts to higher level of acquiring investments if the target is geographically closer. A similar research was done by Chakrabarti and Mitchell (2013) showing that firms prefer geographically proximity targets. Literature shows identical conclusions when looking at FDI and geographical proximity. Kolstad & Wiig (2012) find that larger distance from China deters non-OECD countries from FDI. There is a staggering amount of support for the negative relationship of country distance and FDI flows (Ahern, et al., 2015 ; Banalieva & Dhanaraj, 2013 ; Ghemawat, 2001 ; Rugman & Verbeke, 2004 ; Slangen & Beugelsdijk, 2010 ). The prior discussed literature brings us to the following hypothesis:

H1: spatial distance will have a negative effect on the level of FDI inflows in the Netherlands

Cultural distance

(12)

12

situation in which there is contradicting literature, I will take the situation of the Netherlands into account. The largest extent of the FDI inflows of the Netherlands comes from the Europe and the US, these regions, and the countries in it, show great similarity in terms of the previously discussed Hofstede’s cultural dimension. Therefore I hypothesize the following:

H2: cultural distance negatively affects the level of Dutch FDI inflows Institutional distance

Previous research on institutional distance has covered a large array of international trade and investment aspects. These prior studies found that institutional distance effects the choices of MNE’s organizational knowledge and practices transfers (Kostova, 1999). The choices of MNE’s on entry mode, ownership decisions and country choice (Xu & Shenkar, 2002) and the survivability of foreign subsidiaries (Gaur & Lu, 2007). More specifically, prior literature found a multitude of correlations between FDI and institutional distance. Arslan and Larimo (2010) found a non-significant positive correlation between institutional distance and the level of FDI for Finnish FDI in Central and Eastern Europe. Countries seem to adjust to institutional quality and the accompanied institutional distance, and if related to the previously mentioned OLI paradigm, further inferences can be made. Basically, looking at institutions, they form a foundation for the O and I advantages for a firm. If local institutions interfere with the maximizing of profit of the multinational, this should have a negative impact on FDI inflow in the host country (Choi, Lee, & Shoham, 2016). Furthermore, literature suggests that firms go abroad to attain competitive advantages, which can then be transferred to the home market to surpass domestic competition (Hymer, 1976). Keeping this literature in mind, combined with the level of institutions in the Netherlands, we would expect countries to invest less in countries with higher institutional distance, and thus countries with a higher institutional score will be investing more in the Netherlands than the other way around. We thus formulate the following hypothesis:

H3: Institutional distance will have a positive effect on the level of FDI inflows in the Netherlands

Methodology

In the following section the empirical part of this thesis will be discussed, starting with information on the data collection used for both dependent and independent variables. Followed by the proposed econometric model. The dataset consists of 64 different countries, all reporting FDI inflows in the Netherlands, during the period 2008-2012.

Dependent variable

(13)

13

Netherlands in year T, a year later however, the firm is in heavy weather and decides to abort the project, removing what is left from the initial investment, let’s say 1 million USD. The second year will show as a negative net inflow of -1 million USD. This is technically not a net outflow, since it is not investment from the Netherlands into the foreign country, and it is therefore reported as a negative net inflow.

Spatial distance

The independent variable, spatial distance, is assessed with the help of the CEPII database. This database provides us with multiple observations for spatial distance, such as the celestial distance between the two capitals of the respective countries, and the calculated distance between two countries based on bilateral distances between the biggest cities of those two countries, those inter-city distances being weighted by the share of the city in the overall country’s population (Head & Mayer, 2002). The data will be transformed using a natural log, many economic variables, including prices, incomes, and wages, have skewed distributions, and the use of logarithms in models for such variables is common to make this more symmetric (Carter Hill, Griffiths, & Lim, 2011).

Institutional distance

institutional distance will be compiled with the help of the Global Competitiveness Guidebook, as research by Arslan and Larimo has shown, the following six indicators form the foundation for the institutional distance variable, property rights, intellectual property protection, judicial independence, burden of government regulation, efficiency of legal framework in settling disputes, efficiency of legal framework in challenging reps, transparency of government policymaking. All these indicators are scored on a scale of 1 to 7 in the Global Competitiveness Guidebook, and I will compute the average score of these six indicators to indicate institutional distance. Since this thesis will be comparing the differences in institutional distances to that of the Netherlands to analyse the effect on FDI flows, average scores will be subtracted with the average score of the Netherlands. Negative values implies an institutional score lower than the Netherlands, whereas positive values imply a higher level of institutional score.

Cultural distance

Cultural distance observations will be collected with the help of the database provided by Geert Hofstede on www.geerthofstede.com, which consists of the most up to date dataset of the cultural differences dimensions of the specific countries. These cultural dimensions will be transformed according to the paper of Kogut and Singh (1988), providing us with a single cultural distance index for the relevant countries. As the original index intended to analyse cultural differences between countries and the United States, adjustments are made to adjust this for the Netherlands instead. This provides us with the following index:

𝐶𝐷𝑗 = ∑{ (𝐼𝑖𝑗 − 𝐼𝑖𝑛)2/ 𝑉𝑖) } /4 4

𝑖=1

(14)

14

Control variables

(15)

15

The following table provides the summary statistics of all the relevant variables.

Variable Observations Mean Standard

Deviation

Minimum Maximum

FDI Inflow of the Netherlands ( in million USD) 255 187.8588 3,312.036 -22,926.9 16,922.5 Cultural Distance Index 255 3.204314 1.882216 0.13 10.33 Institutional Distance 255 -1.241597 1.016801 -4.098571 0.33 Spatial Distance 255 7.925049 1.25181 4.95192 9.819122 Population Growth 255 0.4534745 0.060403 0.37 0.514 Infrastructure 255 37.74644 1.598812 35.173 39.811 Education 234 -27.6283 14.2743 -63.9522 9.325 GDP Growth Differential 252 1.638888 3.546991 -11.93242 13.83771

(16)

16

is -27.6%, which implies that most trade partners have a lower level of gross secondary enrolment than the Netherlands, which is not surprising, only being surpassed by the likes of Australia. Furthermore, we have 252 observations for the growth differential between investing and host countries, positive values imply a higher growth in the foreign country that year, and vice versa. The largest difference in growth differential was for Singapore in 2010, in which Singapore reported a 15.2404 % growth, whereas the Netherlands reported 1.4202%, resulting in a staggering 13.837% growth differential. Estonia accounted for the largest negative growth differential, with a decline of 14.7% in 2009, in the year that the Netherlands reported a decline of -3.767%, resulting in a -10.96% growth differential.

As the summary statistics show the minima and maxima for all the relevant variables, the question remains if outliers are apparent, and if so are they restricting the empirical analysis. Visual representation of these outliers will be provided in appendix 2. In estimating if outliers distort the accuracy of the regression, the Cook’s distance will be used. Cook’s distance measures overall influence of outlying observations (Mendenhall & Sincich, 2012). As the visual representation shows in the appendix, all variables have observations outside the fitted model. After investigating these outliers, checking them for data entry errors e.g. typos, and comparing them to the data source they originated from, these observations shown not to be thanks to data entry errors, and thus, the values of the Cook’s distance will be analysed to make sure they will not distort the regression. A value higher than 1 for the Cook’s distance shows high likeliness of the outlier distorting the regression (Mauro, 1998; Cook & Weisberg, 1982). None of the Cook’s distances come remotely close to the threshold of 1, therefore it suggests these outliers do not distort the regression. To further verify this is the case, regression results will be compared before and after dropping these outliers, to detect any relevant changes. The regressions results show similar signs expect for education, however the size of the coefficient is reasonably smaller. As the regression results show that the outliers do distorts the regression somewhat, a look has to be taken as what solutions remain. The two methods most used are trimming and winsorizing, whereas trimming is the exclusion of observations, and winsorizing replaces outliers with the closest relevant data (Dixon & Yuen, 1974). As investigating these outliers, checking them for data entry errors e.g. typos, and comparing them to the data source they originated from, displays that the outliers are not thanks to data entry errors, trimming of these observations is not appropriate, as trimming is most commonly used for data entry errors (Wilcox, 2010). Therefore to make sure the regression is not distorted, the data will be winsorized on the 99th_{percentile. The 99}th percentile is picked as the Cook’s distance in appendix 2 shows a relatively small number of outliers concerning for distortion, and thus a lower percentile choice is unnecessary.

(17)

17

consistent” (p.3). The extremely large consensus on importance of normality in both statistical as econometric literature e.g. Principles of Econometrics (Carter Hill, Griffiths, & Lim, 2011), Data Analysis Using Regression and Multilevel/Hierarchical Models (Gelman & Hill, 2007), Introduction to Econometrics (Stock & Watson, 2011) outweighs this smaller extent of contradicting literature. Therefore as a consequence of not having normally distributed error terms, results have to be interpreted cautiously.

Research Design

(18)

18 Table 2: Correlation matrix

None of the variables seem to be highly correlated with each other, with cultural distance and institutional distance being moderately correlated (r=0.4076). To make sure this is the case a Variance Inflation Factor (VIF) test is also ran, as shown in the appendix, and these results show no signs of multicollinearity problems. Thus the following regressions are presented below:

Eq. (1): 𝐹𝐷𝐼𝑖𝑡 = 𝛽0 + 𝛽1 𝐶𝑢𝑙𝑡𝑢𝑟𝑎𝑙 𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒𝑖𝑡 +

𝛽2 𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝐺𝑟𝑜𝑤𝑡ℎ 𝐻𝑜𝑠𝑡 𝑐𝑜𝑢𝑛𝑡𝑟𝑦_𝑡+ 𝛽3𝐼𝑛𝑓𝑟𝑎𝑠𝑡𝑟𝑢𝑐𝑡𝑢𝑟𝑒 𝐻𝑜𝑠𝑡 𝐶𝑜𝑢𝑛𝑡𝑟𝑦_𝑡+ 𝛽4 𝐸𝑑𝑢𝑐𝑎𝑡𝑖𝑜𝑛𝑖𝑡 + 𝛽5 𝐺𝑟𝑜𝑤𝑡ℎ 𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡𝑖𝑎𝑙𝑖𝑡 + 𝜀𝑖𝑡

FDI is the dependent variable, β is the measured coefficient, with β 1 being the indicator for our independent variable, and with β2 till β5 being the estimated coefficient for the control variables. i is the country indicator, whereas t is the time indicator in years. ε is the error term and accounts for the within country errors.

Eq. (2): 𝐹𝐷𝐼𝑖𝑡 = 𝛽0 + 𝛽1 𝑆𝑝𝑎𝑡𝑖𝑎𝑙 𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒𝑖𝑡+

𝛽2 𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝐺𝑟𝑜𝑤𝑡ℎ 𝐻𝑜𝑠𝑡 𝑐𝑜𝑢𝑛𝑡𝑟𝑦_𝑡+ 𝛽3 𝐼𝑛𝑓𝑟𝑎𝑠𝑡𝑟𝑢𝑐𝑡𝑢𝑟𝑒 𝐻𝑜𝑠𝑡 𝐶𝑜𝑢𝑛𝑡𝑟𝑦_𝑡+ 𝛽4 𝐸𝑑𝑢𝑐𝑎𝑡𝑖𝑜𝑛𝑖𝑡 + 𝛽5 𝐺𝑟𝑜𝑤𝑡ℎ 𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡𝑖𝑎𝑙𝑖𝑡 + 𝜀𝑖𝑡

FDI is the dependent variable, β is the measured coefficient, , with β 1 being the indicator for our independent variable, and with β2 till β5 being the estimated coefficient for the control variables. i is the country indicator, whereas t is the time indicator in years. ε is the error term and accounts for the within country errors.

(19)

19

FDI is the dependent variable, β is the measured coefficient, , with β 1 being the indicator for our independent variable, and with β2 till β5 being the estimated coefficient for the control variables. i is the country indicator, whereas t is the time indicator in years. ε is the error term and accounts for the within country errors.

(20)

20 Table 3:

Regression Table Dutch FDI Inflows

(21)

21

Analysis

As seen in table 3, there are noticeable differences between the pooled OLS and BE models, therefore the discussion will be on the specific samples and adjusted for both approaches. The discussion will start with the analysis related to regression 1.

Cultural Distance – Regression 1

(22)

22

increase in the level of education is associated with a decrease of FDI inflows of -7.86 million USD for the Pooled OLS model, while it reports a -6.27 million USD for the BE model, as education differs highly, as shown by the standard deviation of 14.27, a change of 1 percentage is extremely reasonable. A decrease of -7.86 million USD for a 1 point increase of education is a 4.2% decrease in average FDI inflows, whereas a 6.27 million USD decrease per 1 point is a 3.3% decrease, and thus Education is classified as economically significant. Lastly, the coefficient of Growth Differential is negative for the pooled OLS model, whereas it is positive for the BE model. A 1 point increase in growth differential reports a 48.42 million USD, and a 25.8% decrease in FDI flows and FDI inflows to GDP for the Pooled OLS model, whereas the BE model reports an increase of 24.45 million USD or 13% for FDI per 1 unit increase. since growth differential has a mean of 1.6 with a standard deviation of 3.5 a change of 1 point is reasonable and thus in combination with the percental change of FDI inflows classified as economically significant.

Institutional Distance -Regression 2

The Institutional Distance variable is positive and not significant. An increase of 1 point in the institutional distance indicator is reported to increase FDI inflows between 421.73 million USD to 402.71 million USD, which is an increase of 224.6% of FDI inflows for the Pooled OLS model, and a 214.4% increase of FDI inflows for the BE model. Furthermore Institutional Distance has a standard deviation of 1 and ranges from 4 till 0.33 a 1 point change is reasonable and Institutional Distance is thus economically significant. Infrastructure is not significant for both specifications and an increase of 1 point is reported to decrease FDI inflows with 21.39 million USD, or a 11.38% decrease in FDI inflows for the pooled OLS model, whereas it reports an increase of FDI inflows of 236.13 million USD or a 125.73% increase for FDI. As infrastructure has a mean of 37.7 and a standard deviation of 1.6, a change of 1 point is reasonable to expect, and thus in combination with the percental change of FDI inflows classified as economically significant. Education is negative over both the BE and pooled OLS model, while not being significant. A 1 point increase in the level of education is associated with a decrease of FDI inflows of 14.55 million USD for the Pooled OLS model, while it reports a decrease of 12.69 million USD for the BE model, as education differs highly, as shown by the standard deviation of 14.27, a change of 1 percentage is extremely reasonable. A decrease of -14.55 million USD for a 1 point increase of education is a 7.7% decrease in average FDI inflows, whereas a 12.69 million USD decrease per 1 point is 6.7% decrease, and thus Education is classified as economically significant. Finally, the coefficient of Growth Differential is negative for the pooled OLS model, whereas it is positive for the BE model. A 1 point increase in growth differential reports a 52.05 million USD, and a 27.7% decrease in FDI flows for the Pooled OLS model, whereas the BE model reports an increase of 14.96 million USD or 7.9% for FDI inflows and FDI inflows to GDP per 1 unit increase. Since growth differential has a mean of 1.6 with a standard deviation of 3.5 a change of 1 point is reasonable and thus in combination with the percental change of FDI inflows classified as economically significant.

Spatial Distance – Regression 3

(23)

23

significant. A 1 point increase in the level of spatial distance is associated with a decrease of FDI inflows of 462.60 million USD for the Pooled OLS model, while it reports a decrease of 503.15 million USD for the BE model, Since Spatial Distance has a mean of 7.9 and the standard deviation is 1, a change of 1 is reasonable. A decrease of 462.60 million USD for a 1 point increase of spatial distance is a 246% decrease in average FDI inflows, whereas a 503.15 million USD decrease per 1 point is 267% decrease, and thus Spatial Distance is classified as economically significant. Population Growth is in line with all the former results, showing a positive effect for the pooled OLS, and a negative effect for the BE. As seen in the former regressions. The coefficients for the BE model range from increasing FDI inflows with 2,246 million USD to decreasing 3,194.91 million USD, for an increase of 1% in population growth. At first this seems extremely large, however a population growth of 1% in the Netherlands is highly unlikely, as our samples shows with a minimum of 0.37%, a maximum of 0.51% and a standard deviation of 0.06%. Still a change of 0.1% would bring an increase of 224 million USD or a decrease of 319 million USD, which are a 119% increase and a 169% decrease in FDI inflows and thus Population Growth will be classified as economically significant. As infrastructure has a mean of 37.7 and a standard deviation of 1.6, a change of 1 point is reasonable to expect. A 1 point increase in the Infrastructure indicator is reported to decrease FDI inflows with 60.35 million USD for the Pooled OLS model, whereas it reports an increase of 48.82 million USD for the BE specification. A decrease of 60.35 million USD is a 32% decrease of FDI inflows per 1 point increase, whereas an increase of 48.82 million USD would increase FDI inflows with 26% per 1 point increase. Therefore Infrastructure will be classified as economically significant. Education is negative over both the BE and pooled OLS model, while not being significant. A 1 point increase in the level of education is associated with a decrease of FDI inflows of -11.56 million USD for the Pooled OLS model, while it reports a -5.85 million USD for the BE model, as education differs highly, as shown by the standard deviation of 14.27, a change of 1 percentage is extremely reasonable. A decrease of -11.56 million USD for a 1 point increase of education is a 6.2% decrease in average FDI inflows, whereas a 5.85 million USD decrease per 1 point is a 3.1% decrease, and thus Education is classified as economically significant. Finally, the coefficient of Growth Differential is positive for both models. A 1 point increase in growth differential reports a 4.92 million USD, and a 2.6% increase in FDI flows for the Pooled OLS model, whereas the BE model reports an increase of 98.81 million USD or 52.6% for FDI inflows per 1 unit increase. Since growth differential has a mean of 1.6 with a standard deviation of 3.5 a change of 1 point is reasonable and thus in combination with the percental change of FDI inflows classified as economically significant.

(24)

24

(25)

25 Table 4:

Regression Table Dutch FDI Inflows – Cultural Distance Pooled OLS

t statistics in parentheses