The impact of
investment and tax treaties on FDI:
evidence from The Netherlands
Master Thesis International Financial Management
Georgia Dimitriou
2
Author
Georgia Dimitriou
Address
Merwedestraat 22, 9725KD, Groningen
Telephone
0031 6 17917662
georgiadimitriou.cy@gmail.com
University
Rijksuniversiteit Groningen
Program
International Financial Management
Student Number
S2619261
Supervisor
Dr. Hein Vrolijk
Co assesor
Dr. Wim Westerman
3 Acknowledgements
The completion of this thesis would not be accomplished without the grateful help of my
supervisor, Dr. Hein Vrolijk, who helped me with patience to achieve the goals and objectives of this thesis. I would like to express my gratitude to him for the inspiration he gave me, his useful remarks, and his guidance to this learning journey. Furthermore, I would like to thank Francis Weyzig and Klaes Wester for their insights and for sharing information about my topic. Finally, very warm thanks to my family, friends and George for their support and encouragements.
4 Abstract
This thesis examines the impact of Bilateral Investment Treaties (BITs) and Double Taxation Treaties (DTTs) on outward Foreign Direct Investment (FDI) of The Netherlands for the period of 2009 to 2013. The Netherlands is chosen because on average 78.8% percent of the Dutch outward FDI comes from Special Purpose Entities (SPEs).
Two identical cross sectional models are used to find out if the determinants of outward FDI are different for SPEs than for productive firms, non-SPEs. The findings show that the DTTs are a significant determinant for Dutch SPEs but not for productive firms. The BITs do not have any impact on the outward FDI of the Dutch SPEs. However, BITs have a negative influence on the FDI outward of the productive firms. This means that SPEs located in the Netherlands increase their outward FDI to host countries that they have DTT with the Netherlands. On the other hand, FDI of corporations that are productive investors is reduced when a BIT exists between the Netherlands and the host economy.
Keywords: FDI, Bilateral Investments Treaties, Double Taxation Treaties, Special Purpose Entities
5 Contents Table of Figures ... 6 Table of Tables ... 6 Abbreviation List ... 7 1. Introduction ... 8 2. Literature Review ... 10
2.1. Foreign Direct Investment ... 10
2.2. International Investment Treaties ... 11
2.3. Bilateral Investment Treaties (BITs) ... 11
2.4. Double Taxation Treaties (DTTs) ... 12
2.5. Previous studies on Treaties effects on FDI ... 12
2.6. Special Purpose Entities (SPEs) ... 15
2.7. Closing Remarks ... 16
3. Descriptive Analysis ... 18
3.1. The trend of the inward and outward of FDI for the years 2008-2013 ... 18
3.2. Geographical Breakdown of FDI ... 20
3.3. Treaties ... 22
4. Data and Methodology ... 23
4.1. Data collection... 23
4.2. Methodology ... 23
4.3. Selection of the variables and estimation model ... 26
5. Regression Analysis Results and Discussion ... 28
5.1. Results of multivariate regression models ... 29
5.1.1. Non-SPEs cross sectional regression results... 29
5.1.2. SPEs cross sectional regression results ... 32
5.2 Introduction of the Inward FDI ... 33
5.3 Robustness Test ... 39
6
7. References ... 44
I. Appendix A: General ... 46
II. Appendix B: Figures of Descriptive Analysis ... 51
III. Appendix C: Statistic results ... 54
Table of Figures Figure 2.1: International Investment Treaties over time, Source: UNCTAD ... 11
Figure 2.2: The increasing trend of Dutch SPEs, source: Claasen and v.d.Dool, 2013 ... 17
Figure 3.1: Graph of the FDI position in the Netherlands for the year 2008-2013 ... 18
Figure 3.2: FDI stocks of non SPEs in the Netherlands for the year 2008-2013 ... 19
Figure 3.3: Geographical Breakdown of average SPEs FDI for the period 2008 - 2013 ... 21
Figure 3.4: Geographical Breakdown of average productiveFDI for the period 2008 - 2013 ... 21
Figure 5.1: Indirect round-tripping example ... 34
Table of Tables Table 2.1: Summary of the research design of the previous studies ... 15
Table 3.1: Number of BITs and DTTs the Netherlands on 2012 ... 22
Table 4.1: Summary of the model used in the previous studies ... 24
Table 4.2: Summary of the control variables used in the previous studies ... 25
Table 5.1: Descriptive Statistics of parameters... 28
Table 5.2: Regression Results of non SPEs outward and treaties... 30
Table 5.3: Regression Results of SPEs outward and treaties... 32
Table 5.4: US SPEs FDI in million euro... 35
Table 5.5: Pair correlation Matrix of Non SPEs ... 36
Table 5.6: Pair correlation Matrix of SPEs ... 36
7 Abbreviation List
Foreign Direct Investment FDI
Bilateral Tax Treaties BIT
Double Taxation Treaties DTT
Multi-National Corporations MNC
Special Purpose Entities SPE
United Nation Conference on Trade and Development UNCTAD
Mergers and Acquisitions M&A
International Monetary Fund IMF
International Investment Agreements IIA
Institute of Chartered Accountants in England and Wales ICAEW Organization for Economic Co-operation and Development OECD
8 1. Introduction
Globalization gives the opportunity to corporations to cross the border and invest in other economies. The Foreign Direct Investment (FDI) activity has significantly increased in the past few decades (World Bank, 2013). Multinational Corporations (MNCs) tend to invest in
countries where they can gain benefits, especially in terms of taxation. There is also an increasing interest on the investment policies used by MNCs. The term "offshore", "Special Purpose Entities", "shell companies" are well discussed through the last years since many MNCs use these types of activities for their tax planning.
Many MNCs establish Special Purpose Entities (SPEs) in countries famous as tax havens and deduct their tax liability significantly; among them are Apple, Nike, Dell, and Oracle.1
Consequently, in some cases the effective tax rates of corporations with high profits are lower compared to entities with low profit margins. Some companies achieved the minimization of their taxes by establishing SPEs in countries with favorable tax treaties network. In this way, they redirect their profits through them to tax haven countries.
One example is the case of "Double Irish Dutch Sandwich". Google in order to minimize its taxes hold subsidiary in Ireland. Google establishes one more subsidiary, the Google Netherlands Holdings, which is a shell company with no operations, held in the Netherlands. Through this company, Google shifts its profits to Ireland and then from Ireland to Bermuda Island, which is a tax haven in order to optimize its tax payable amount.2
Offshore in practical sense means the country-territory that offers benefits for tax concessions. One of the benefits can be the extensive tax agreements network of the offshore center.3 The use of a country’s tax treaties network by the parent company can lead to the reduction of the tax liability of the income received from the host country. A distinction can be made between
Bilateral Investment Treaties (BIT), which tend to protect and promote the investments, and Double Taxation Treaties (DTT) which are allocating the taxable income between the parent and
the host country. Many studies are trying to find out the impact of the treaties on FDI. There is
1 Apple is not alone, Citizens of Tax Justice, 2013 2 Dutch Sandwich saves Google million in taxes 3
9 not a conclusion as to whether the treaties increase or decrease the FDI since the findings of the previous literature found both positive and negative results.
This thesis examines the relationship between FDI and treaties in the Netherlands. The
Netherlands, as mentioned above, is one of the countries that has a high level of offshore FDI. The FDI of the country is consisted of a high percentage of the SPEs. According to UNCTAD World investment report (2013), the SPEs percentage of Dutch inward and outward FDI was more than 75% in 2011. Only about a quarter, 25%, comes from enterprises located in the Netherlands having real production, the productive firms.
I want to answer the following research question:
What are the main differences between the productive firms4 and SPEs in The Netherlands , especially in terms of the effects of treaties on outward FDI?
This question is answered by using identical cross sectional OLS models for both, productive outward FDI (OFDI) and SPEs outward. The control variables used are the same in both models. The results show that the DTT are significant and positive in the SPEs outward activity but
insignificant in the productive OFDI activity. The BIT does not affect the SPEs outward Investment. On the other hand, BIT influences negatively the productive FDI outward of the Netherlands.
The chapter 2 of the thesis presents the literature of this thesis. The descriptive analysis of the data is discussed in chapter 3. The section 4 shows the data and methodology used. Finally, the results, discussion and conclusion are presented in chapters 5 and 6 respectively.
10 2. Literature Review
This part of the thesis, firstly presents the terminology of FDI. Then, I introduce the types of treaties as well as their benefits and costs. A presentation of the previous findings in regards to the impact of the treaties on the FDI follows next. Finally, an explanation of the SPEs terms and the policies used by MNCs in regards with the FDI and tax planning are described.
2.1. Foreign Direct Investment
According to the United Nation Conference on Trade and Development (UNCTAD), "Foreign Direct Investment (FDI) refers to an investment made to acquire lasting interest in enterprises operating outside of the economy of the investor". International Monetary Fund (IMF) defines FDI as "A category of international investment that reflects the objective of the resident in one economy (known as direct investor) obtaining a lasting interest in an enterprise resident in another economy (known as the direct investment) where the investment is treated as direct investment when the direct investor has obtained 10 percent or more of the ordinary shares of that entity".
Examples of FDI are the cross border Merger and acquisitions, the joint ventures agreements, the development of a new entity in a foreign economy and the reinvestment of earnings. FDI can be measure as stock or flow. Flows are the financial transaction for example in one year period between affiliated companies. Positions are the stock of the investment at a particular point in time for example at the end of the year 2013.
During the last decades, the FDI has significantly increased (World Bank, 2014). The globalization gives the incentives to corporations to cross the border and invest in other
economies. According to Sachs and Sauvant (2009) the FDI increasing trend is mostly due to the increasing trend of the cross border Mergers and Acquisitions (M&A) for the last years.
According to Ernst &Young 2012 Survey, the political stability and transparency along with the legal and regulatory environment are the most important drivers that define the decision
11 between the home and the host countries makes the regulatory environment of the foreign entity more transparent and stable.
2.2. International Investment Treaties
An international Investment agreement is a treaty between countries ron cross border
investments. The general aim of the treaties is the promotion and protection of the FDI. The IIAs consist of Bilateral Investment Treaties (BITs) and Double Taxation Treaties (DTTs).
According to UNCTAD (2011) in 2010 there were in total 6092 IIAs from which 2807 were BITs and 2976 DTTs. The Figure 2.1 below shows the increasing trend of the sign of
investment and tax treaties. After 1985 the number of treaties signed, has risen considerably.
Figure 2.1: International Investment Treaties over time, Source: UNCTAD
2.3. Bilateral Investment Treaties (BITs)
According to UNCTAD, the BIT is an agreement, signed by two countries, which reciprocally encourages, promotes and protects the investments in each other's territory. Many developed countries prefer to sign a BIT with developing countries in order to gain the maximum safeguard for their investments in these economies.
12 One benefit of the BITs is the protection of the investors by ensuring a stable legal environment in the host country. A treaty between two countries reduces the risk of foreign investment because it first explains all the procedures and secondly, it ensures the participate countries that the particular investment will be safeguarded. In addition, in BITs, countries can negotiate and adjust the terms of treaties to their own preferences and needs in contrast to multinational treaties (Gazzini and Brabandere, 2012). The signing of a BIT also has some costs for both countries; the negotiation and the documentations involved in the treaties demand time and lots of work from professionals.
2.4. Double Taxation Treaties (DTTs)
According to the Institute of Chartered Accountants in England and Wales (ICAEW), the DTTs are the agreements between two countries, which allocate the taxable income to each country in order to eliminate the double taxation issues.
The DTTs prevent the double taxation of the investor in both the host and the home country. Most countries use, as subject to tax matters, the worldwide income and therefore the foreign income is taxable in the resident country and the source country as well. DTTs provide
provisions and rules for the withholding taxes on dividends, interest and royalties payments. In addition, the definition and calculation instructions included in the treaties help the parent company to form a better understanding regarding the foreign tax system (Barthel et al, 2010). The DTTs protect the integrity of the tax systems in both countries by exchanging information and ensuring that tax residences will not hide their revenues (Pickering, 2013).
Nevertheless, the elimination of DTT reduces the tax revenues in the parent country. The taxable income will be subjected to tax liability in the source country and through the tax credit relief or exception method will be offset in the tax liability of the parent country. Therefore, the parent country receives lower taxes rather than in the case of non-DTT.
2.5. Previous studies on Treaties effects on FDI
13 either the FDI inbound or FDI outbound. Some studies use dyadic data while others use non-dyadic data but the aggregate amount of FDI.
In theory both treaties, investment and tax treaties, have a positive effect on the FDI. Treaties strengthen the possibility that the corporations will cross the borders and invest abroad. The elimination of the double taxation obtained by DTTs minimizes the obstacle of the double taxation that firms are facing when they invest in foreign countries. In addition, the BITs offer enhances the trust between the home and host economies. Therefore, one can expect that the benefits of treaties will increase the FDI of the signatory countries.
One study that confirms the findings of several studies about the BIT causal effects on FDI is the meta-analysis study of Bellak (2013). He includes in the meta-analysis 33 studies published between 1998 and 2012. The results indicate a positive relationship of BIT and FDI. The limitation of these studies concerns the measurement of the treaty. The most effects of BITs, he suggests that should be measure by the provisions in the treaty and not with the number of treaties signed by countries.
In addition, papers that focus on particular countries conclude that the DTTs increase the FDI activities. Murthy and Bhasin (2013) find that the introduction on DTTs treaties has a positive effect on the FDI inward stocks of India. They examine the effect of DTTs on inward FDI of India from the top ten investing countries (among them is the Netherlands) for the period 1991 to 2008. They use fixed effect regression analysis and principle component analysis, including variables based on the knowledge capital model. The study of Braun and Fuentes (2014) examines the effects of DTTs on FDI between developing countries and Austria. They use logistic and count data regression model with lag variables while they examine the implication of the DTTs of Austria on its FDI in developing countries. They find a significant positive
relationship between the DTTs and FDI activity of Austria in developing countries.
Studies, which use panel data with larger samples, find positive evidence in relation to treaties and the FDI. Lejour (2014) concludes that DTT treaties increase the FDI significantly. The sample consists of Organization for Economic Co-operation and Development (OECD)
14 Barthel et al. (2010), the positive outcome of DTT treaties is stated. The authors examine the effect of DTTs on FDI stocks for the period of 1978 until 2004. They use a dyadic dataset of a large sample in both developed and developing countries for FDI stocks. Their outcome is that DTTs lead to higher FDI stocks and that the effects are substantively important as well.
On the other hand, some studies find evidence that treaties reduce FDI. The study of Egger et al. (2006) finds a negative impact of the DTT treaties on outward FDI stocks of OECD countries for the period 1985-2000. They solve an equilibrium model of FDI, which is similar to Knowledge Capital model. They compare the period two years before and after the conclusion of the treaty.5 Bloningen and Davies (2002) examine the effect of the BITs signed by OECD countries for the period 1982-1992 on FDI using the Knowledge Capital model. They split the treaties into old and new treaties. They find evidence that old treaties are indeed increasing FDI but new treaties do not affect the FDI positively and reduce tax evasion rather than promote the FDI.
The table 2.1 presents the above literature findings, the countries' sample, the treaties effects considered and the type of FDI measured. The study of Weyzig (2013) is one study that is not mentioned above since it is the only one taking into account the SPEs amount of the FDI. This study is described in the next section.
5
15
Table 2.1: Summary of the research design of the previous studies
Author and year
Countries Period FDI Flows/ Stocks
BITs/DTTs Findings
Blongingen &Davies (2002)
OECD 1982-1992 Out Both DTTs -
Lejour (2014) OECD 1985-2011 Inward Stocks DTT, BIT + Egger et al.
(2006)
OECD 1985-2000 Out Stocks DTTs -
Murthy & Bhasin (2013) India 1991-2008 In flows DTTs + Barthel et al. (2010) 30 countries
1978-2004 In Stocks DTT & BITs +
Bellak (2013) 1998-2012 Both Both BITs +
Braun & Fuentes (2014)
Austria 1990-2011 Out None, (no. of
Projects)
DTTs +
2.6. Special Purpose Entities (SPEs)
According to OECD, the SPEs are legal entities of multinational enterprises that are established in a different location of their parent companies and are not used for the contribution to the production but in order to hold assets or liabilities. In the Appendix A, there is a list showing the criteria of a company to be called SPEs according to OECD terminology.
16 The conduit entities are the holding companies that set up in the third country and serve as a pipeline for the income from the host country B to the parent company in country A. The establishment of base companies is another way through which the multinational entities avoid the tax at the home country. (Desai et al., 2003) One other type of companies, the mixing ones, is used for mixing the low and high taxes of countries in order to offset the taxes paid in the parent company. (Dolan and Weil, 1995)
Many multinationals enterprises use SPEs investing in a corporation in one country and then reinvest the investment in another corporation and in another country. This behavior refers to the tax treaty shopping which is a particular form of tax avoidance by MNCs. It involves the
diversion of FDI through a third country in order to achieve reduction of withholding taxes under favorable tax treaties (Kingson, 1981). The intermediate country is more likely to have a
favorable tax treaty network, which benefits the parent company to avoid the taxes that have to be paid in the host country.
Tax treaties reduce or eliminate these withholding taxes on a bilateral basis, providing an advantage to foreign investors from the partner country. When multinationals engage in treaty shopping, they may obtain benefits, which a host country would otherwise not provide them. Weyzig (2013) made an important analysis of the tax treaty shopping and the FDI routed through the Netherlands. He analyzes the determinants of FDI diversion with the usage of micro data from Dutch SPEs. The findings confirm that tax treaties are a key determinant of FDI routed through the Netherlands.
2.7. Closing Remarks
Most of the previous studies found a positive relationship between the FDI and the treaties, either BIT or DTT. Based on their results, it is expected that the treaties will have a positive causal effects on the OFDI of the Netherlands.
Hypothesis 1: There is a positive relationship between treaties and OFDI in the Netherlands.
17 Netherlands. This treatment of the FDI data for the Netherlands, implies that the country report major percentage of SPEs. The Figure 2.2 shows the increase of the Dutch SPEs as well. The amount of FDI outward including SPEs is increasing rapidly after 2004. Therefore, the investigation of the causal effects of treaties on FDI of the Netherlands, should take into
consideration that the FDI arise from two different categories of corporations; the SPEs and the productive entities. .
Figure 2.2: The increasing trend of Dutch SPEs, source: Claasen and v.d.Dool, 2013
18 3. Descriptive Analysis
In this chapter, I will describe the FDI activity of the Netherlands in the period 2008-2013. The aim of this chapter is to strength the argument of splitting the FDI into SPEs activity and
productive FDI. Firstly, a trend of the FDI is presented and then a geographical breakdown of the FDI follows. Finally, a description of number of treaties that the Netherlands signed closes this chapter.
3.1. The trend of the inward and outward of FDI for the years 2008-2013
The Figure 3.1 presents the FDI stocks in the Netherlands for the period between 2008 and 2013. The most important observations of the graph presented below are (1) that the total IFDI and OFDI follow the same trend and (2) the SPEs activity moves with the same trend as the total FDI and (3) the high percentage of SPEs in the total FDI of the Netherlands.
Figure 3.1: Graph of the FDI position in the Netherlands for the year 2008-2013
The inward and outward FDI position rose in value from 2008 to 2013, both by around 35%. Both inward and outward FDIs, showed their first big increase on 2010 with 6% and 8% higher values than 2009. The next year, 2011, the increase picked the 12% for the inward and 10% for the OFDI. 0 500.000 1.000.000 1.500.000 2.000.000 2.500.000 3.000.000 3.500.000 4.000.000 4.500.000 2008 2009 2010 2011 2012 2013 m ill io n e u ro 's
FDI of SPEs and total FDI of the Netherlands
19 In general, the FDI outward is higher by 18% of the total inward in 2013. The highest difference between inward and outward was on 2010 where the outward recorded as 23% higher than the inward.
Table II.3 on Appendix B presents the average percentage of SPEs inbound and outbound for all the countries reported in the DNB report. On average each country, reports 83% IFDI from SPEs and 79% outward for the years 2008-2013. The FDI of SPEs indicate a huge increase the last years with reporting 39% higher inward of SPEs in 2013 compared to 2008 and 38% higher outward for the same period. Overall, the percentage of the SPEs is very high which proved that in the Netherlands, MNCs establish SPEs hold the major amount of FDI activities.
In Figure 3.2 below shows the FDI position of the productive FDI in the Netherlands for the period between 2008 and 2013.
Figure 3.2: FDI stocks of non SPEs in the Netherlands for the year 2008-2013
0 500.000 1.000.000 1.500.000 2.000.000 2.500.000 3.000.000 3.500.000 4.000.000 4.500.000 2008 2009 2010 2011 2012 2013
Productive and Total FDI of the Netherlands
20 The amounts of FDI excluding SPEs, which are the investments of the productive companies, consist on average of the 17% of the total IFDI and 21% of total OFDI in the years 2008 to 2013. However, the outward FDI by these type of entities are higher than the inward which is
consistent with the higher outward observed in total FDI and in FDI from SPEs data analysis. The data for productive firms' FDI do not have any increasing or decreasing trend during the last six years. The amounts are fluctuated and the increase of inward FDI for the period is only 8%. However, the increase of outward FDIs of the productive corporations established in the
Netherlands reports 21% higher in 2013 compared to 2008. Compared to the previous graph, the trend of the FDI of productive firms does not follow the same trend of the overall FDI. This is not surprising since these types of investments are consisted by productive firms invest in foreign affiliates and do not use policies of profit shifting.
3.2. Geographical Breakdown of FDI
On average European Union (EU) hold around 55% of the inward and outward FDI reported in and out of the Netherlands. The 40% of the EU investments consists of EMU countries. This thesis focus on the differences between SPEs and Non-SPEs FDI and therefore the geographical breakdown of the SPEs and the productive FDI is following.
The Figure 3.3 shows the geographical breakdown of the average FDIs origin and destinations in regards with SPEs. The EU corporations are tha major origin of the FDI in and out of the
Netherlands. The US ranks as the second investor. The 9% and 14% of inward and outward respectively, hold by the all other countries not included in the sample while in Non SPEs chart the percentage holding by these countries is relatively lower, 3% inward and 9% outward. The countries included in the sample can be found in the Appendix A.
This chart shown in Figure 3.4, presents the investments of FDI to Dutch corporations and the FDIs by productive corporations to foreign countries. The mainly origin for the FDI of
21 Netherlands via SPEs and productive firms shows more interest to invest in Switzerland
compared to SPEs.
Figure 3.3: Geographical Breakdown of average SPEs FDI for the period 2008 - 2013
22
3.3. Treaties
The Netherlands has concluded in total 91 BITS and 95 DTTs. The countries reported in the data of DNB, mostly have a DTT with the Netherlands. For this reason, an analysis of the countries that are not reported in the DNB report of FDI has been made and the result is that the most treaties signed by the Netherlands are DTT and not BITs. BITs are concluded with mostly developing countries and not developed economies. Netherland has both DTT and BIT with 56 economies, and there are 105 countries that the Netherland does not have neither BIT nor DTT.
Table 3.1: Number of BITs and DTTs the Netherlands on 2012
DTT
BIT
YES NO
YES 56 35
NO 39 105
The major question is whether other countries can benefit from the treaties network of the Netherlands. This could be the case if the inward FDI and the outward FDI from/to a particular country have significant difference. If a country report significant different inward and outward, is an indication that the country redirect the investment in other economy and therefore use the treaty network of the country. One example is the case of Denmark, which reports on average 30% higher SPEs 'outward than inward. Denmark has a DTT treaty with the Netherlands but not a BIT. This might be a sign that MNCs invest in the Netherlands by establishing SPEs and then redirect their investment to Denmark in order to benefit from the DTT of the Netherlands and Denmark. In the Appendix B, there is a table presenting the countries that report high differences between inward and outward. In chapter 5, I elaborate on this topic by using inward FDI as one of the determinants of outward FDI.
23 4. Data and Methodology
4.1. Data collection
The Nederlandsche Bank (DNB) provides the data for this thesis. The data are available for both inward and outward stocks of FDI and the amount of SPEs, for 33 countries for the period 2008 until 2013.
The data for the explanatory variables of the model are obtained from different sources. The BITs are collected from the UNCTAD organization and the DTTs from the tax authority of the Netherlands. The World Bank database is used for the collection of the data for the GDP per country and GDP growth. The data in regards with the distance between the Netherlands and the host country are collected by nl.distance.to. I use the KPMG global website for the data for the corporation tax rates.
4.2. Methodology
The first interpretation of the data is performed by a descriptive analysis. For the descriptive analysis, I calculate the average of the SPEs share for each country. I conduct several identical cross sectional regression models to find out the difference on effects of the treaties on the FDI of SPEs and Non SPEs. The table 4.1 presents the estimation techniques and components of the methodology and the table 4.2 shows the control variables used by the previous studies. Most of the studies used the fixed effect OLS method for estimating the relationship between inward or outward FDI activity and the BITs or DTTs. Some studies (Bloningen and Davies 2002, Egger et al. 2006, Lejour 2014) parameterized the knowledge-Capital Model, which developed by
Markusen and Maskus (2001), and used most of its variables for testing the relationship between the treaties and FDI. The model includes variables as the sum of the GDP in both home and host country, the difference in GDP per capita of the two countries, the difference on the two
24
Table 4.1: Summary of the model used in the previous studies
Author(s) and year
Estimation Technique
Dependent variable Explanatory variables
Bloningen &Davies(2002)
Fixed Effect Model OLS
(1) FDI stocks outward (stocks) and (2) FDI outflows6
BITs(dummy)
Lejour (2014) Fixed Effect OLS model
FDI outward stock (log) DTTs, BITs and Multilateral treaties, (dummies)
Egger et al (2006) Propensity score matching approach
Changes on FDI outward stocks (log)
Bilateral tax treaties(dummy) Murthy and
Bhasin (2013)
Fixed Effect model FDI inflows(log) Treaty dummy, Treaty Age
Barthel et al. (2010)
Fixed Effect OLS model
Inward stocks(log) DTTs(dummy)
Weyzig (2013) Tobit Share of diverted FDI through the Netherlands
DTTs and BITs(dummies)
Braun and Fuentes (2014)
Fixed Effect OLS model
Number of FDI projects in host country
DTT (dummy)
6
25
Table 4.2: Summary of the control variables used in the previous studies
Author(s) and year Control variables Bloningen &Davies
(2002)
Log of the Sum of GDP of both countries; home and host, log of the difference of the GDP of the two countries, log of the difference in capital labor market, distance, trade openness, FDI openness
Lejour (2014) Log of the Sum of GDP of both countries; home and host, log of the difference of the GDP of the two countries, log of the difference in capital labor market, distance, trade openness, FDI openness, dividend tax rate
Egger (2006) Log of the GDP sum, difference in the GDP, log of the difference in capital labour market, differences in corporation tax, expenditure to GDP ratio
Murthy and Bhasin (2013)
Ratio of GDP of home to host country, Ratio of GDP per capita of the two countries, Distance, Trade openness, FDI openness Barthel et al. (2010) Log of the GDP of host country, log of the GDP per capita of host
country , log of the GDP deflator, Population of host country , Trade openness, dummy for BITs
Weyzig (2013) Home and host gravity variables (Ratio of FDI stocks)1/2 , Investment
treaties variables, Developing host, Home and host corruption7 Braun and Fuentes
(2014)
GDP per capita(similarity index), Trade openness (total exports plus imports divided by GDP), Corporation tax rate of the host economy, corruption index of the host country, quality of infrastructure of host country
7
26
4.3. Selection of the variables and estimation model
In order to find the differences between the SPEs and Non-SPEs, I am using a simple version of the models reported in the table above. The dependent variables are the natural log of the average outward FDI stocks for the period 2009 to 2013. The natural log increases the fitness of the model while it reduces the asymmetry of the distribution of the dependent variable.
The explanatory variables included in the model are both, BIT and DTT, treaties, measured by a dummy variable, which will take the value of 1 if there is a treaty between the destination economy and the Netherlands.
In order to control for other effects that might explain the increase on the FDI outward in the examined period, I add in the model a few control variables. I select two of the control variables in accordance with the gravity model, which is the natural logarithm of the sum of the GDP of the Netherlands and the destination economy and the log of the distance between Amsterdam and the capital of host country. In accordance with this model, larger economies are expected to invest more. Therefore, it is expected that the variable of GDP will have positive relationship with the FDI activity since is reflecting the potential FDI. The distance is expected to have a negative relationship since the farther are the countries, the higher information asymmetry and risk therefore the lower FDI.
Furthermore, I introduce the GDP growth of the destination economy and the log of the
corporation tax rate of the host country. The GDP growth of the destination country is expected to have positive impact on FDI outward. The higher the growth in the host economy, the more Dutch firms are willing to invest in this economy. The corporation tax rate is expected to have negative effects on FDI outward. The higher the corporation tax in the host country, the lower wiliness of firms to invest in this particular country.
27 The two estimated fixed effect OLS models are the following:
where in equation, (1) the dependent variable is the log of the average outward FDI of SPEs and in equation (2) is the average outward of productive FDI to destination economy i, Sum of GDP is the log of the average sum of the GDP for the Netherlands and the destination economy i ,distance is the log of the distance between Amsterdam and the capital of host country, growth of GDP is the average growth of GDP for the destination economy i, tax rate is the log of the average corporation tax rate of the destination economy i, BITs dummy and DTT are dummies variables which takes the value 1 if the Netherlands has a tax treaty with county i and 0 if there is not treaty between them. The period which is examined is between 2009 and 2013.
28 5. Regression Analysis Results and Discussion
Several models are estimated in order to find out the causal effects of BITs and DTTs on FDI activity of SPEs and Non SPEs. The table 5.1 presents the descriptive statistics for the regression variables of the productive firms' and SPEs models. The sample consist of thirty-three
observations which are the countries reported in the DNB report. From the report of the DNB, the Other European countries, other countries of the world and ABC-SS islands are not included in the sample due to the difficulty of allocating the amounts of GDP and the treaty implications.
5.1 Descriptive Statistics
The table 5.1 below shows the descriptive statistics of the parameters which of the models.
Table 5.1: Descriptive Statistics of parameters
The Outward of productive FDI ranges between 39.6 million and 99,881.4 million euro's. The average outward of SPEs to host country is around 90,000 million with a minimum of 227.8 and maximum 375,945.4 million euro's. For both outwards ' FDI, the minimum destination economy is Latvia and the highest FDI outward is hosted in UK.
In regards with the control variables, the GDP sum of the Netherlands and host economy varies between 861,786.1which is the sum of the Dutch and Malta GDP million to 16,419,432 which is the US GDP plus the Dutch one. The distance of the countries included in the sample is around
Variable N Mean Median STD Minimum Maximum
29 2,588 kilometers with the closest country the Belgium, only 173Killometers far and the most far away country is Japan. The corporation tax rate of host countries varies between 10 % and 40%. The GDP growth is on average 0.586 and there are cases, which the growth is negative. The most negative GDP growth is observed in Greece while Chine increase on average its GDP by 8.857% the period between 2009 to 2013.
5.1. Results of multivariate regression models
The several cross sectional models are presented in the tables below; the productive FDI regression results in table 5.2 and SPEs regression results in the table 5.3. The models are identical in order to find the differences between the causal effects of treaties on productive outward FDI and SPEs. Firstly, the models include only the gravity variables used in most of the previous studies, GDP and distance. In the second model, I introduce the GDP growth and corporation tax of the host economy. Finally, the BIT and DTT, are added in the models.
5.1.1. Non-SPEs cross sectional regression results
The results of the model with dependent variable the non SPEs outward are presented in the table 5.2.
The first model, which includes the gravity variables, support that the GDP and the distance variables are significant and their signs indicate the right relationship predicted by the gravity model. The higher the sum of GDP of home and host economies is the FDI outward increases. The distance between home and host country is one important driver of the FDI outbound. The results of the model 1, confirm the argument of the negative relationship between the distance and the FDI.
The next model, 2, includes the variables of corporation tax and GDP growth of the destination economy. The variable of corporation tax rate of the destination country does not sign a
30
Table 5.2: Regression Results of non SPEs outward and treaties
Model 1 Model 2 Model 3 Model 4 Model 5
DTT_Dummy -1.053267 (0.765315) 0.016166 (0.675091) BIT_Dummy -1.167487*** (0.267511) -1.170363*** (0.297892) GDP 2.146727*** (0.396688) 1.988866*** (0.526086) 1.221945** (0.446291) 1.938966*** (0.519151) 1.220821** (0.457201) Distance -0.416481*** (0.124791) -0.416759*** (0.136159) -0.222802* (0.115102) -0.447755*** (0.135915) -0.221848* (0.123869) Corporation tax rate 0.165317 (0.426082) -0.166200 (0.340829) 0.391502 (0.450488) -0.170488 (0.390767) GDP Growth 0.016241 (0.054486) 0.137208** (0.050729) 0.027974 (0.054309) 0.137326** (0.051929) Constant -13.51953*** (4.578680) -12.11752** (5.501154) -2.721885 (4.799672) -10.97916 (5.478178) -2.716199 (4.896788) Observations 33 33 33 33 33 Adjusted R2 0.476139 0.443606 0.661669 0.460822 0.648664
Note: * indicates statistically significance on 10%, ** on 5%, *** on 1% confidence level. The number in brackets presents the standard deviation of each variable.
31 fitness and it is not indicate any significant relationship with the FDI outward. The DTT along has a negative causal effects on FDI.
The final model, which includes both dummies, indicates again the negative relationship of BIT and FDI. The gravity variables are significant and the GDP growth as well. The corporation tax is not significant but the sign is the one that it was expected; the higher the corporation tax rate in the destination economy the lower FDI outward is reported to that economy. The DTT has a positive coefficient in the model 5, but it is not a significant driver of FDI productive outward. The pair correlation of the independent variables is included in the Appendix C. There is no indication for any multicollinearity of the variables. Furthermore, for all the models, I test for heteroscedasticity in residuals and the null hypothesis of homogeneity of the residuals cannot be rejected.
Overall, the results of Non SPEs model shows that the outward FDI of productive activity is affected negatively by BITs and not affected at all by DTTs. Furthermore, the outward FDI of entities other than SPEs, influence by the GDP and the GDP growth of the host economy and the distance.
The question arise from the above multivariate analysis is the inconsistency of the results with the previous studies. This thesis, finds evidence for a negative relationship between BIT treaties and outward FDI arise from productive entities. How do other studies find positive causal effects of BIT on FDI? One more question arise from this analysis, is how do other studies find
significant impact of DTT on FDI? Especially this question is even more strong for studies which include The Netherlands in their sample, like the study of Lejour(2014), which finds evidence of positive relationship with tax treaties and FDI.
32
5.1.2. SPEs cross sectional regression results
The results of the model with dependent variable the SPEs outward are presented in the table 5.3.
Table 5.3: Regression Results of SPEs outward and treaties
Model 1 Model 2 Model 3 Model 4 Model 5
DTT_Dummy -0.472395 (0.793577) 0.352437 (0.787400) BIT_Dummy -0.839971** (0.313210) -0.902677** (0.347450) GDP 1.696581*** (0.400452) 1.666754*** (0.530781) 1.114977** (0.522532) 1.644133*** (0.538323) 1.090483* (0.533262) Distance -0.336252** (0.125975) -0.312594** (0.137374) -0.173048 (0.134765) -0.326496** (0.140934) 0.152258 (0.144475) Corporation tax rate 0.121110 (0.429874) -0.117407 (0.399053) 0.222555 (0.467124) -0.210897 (0.455776) GDP Growth -0.024881 (0.054972) 0.062151 (0.059396) -0.019619 (0.056315) 0.064422 (0.060568) Constant -7.956277* (4.622127) -8.133487 (5.550242) -1.373605 (5.619612) -7.622928 (5.680486) -1.249875 (5.711419) Observations 33 33 33 33 33 Adjusted R2 0.349386 0.309755 0.434757 0.293859 0.417505
Note: * indicates statistically significance on 10%, ** on 5%, *** on 1% confidence level. The numbers in brackets present the standard deviation of each variable.
33 and the distance is negative related to FDI. In the second model, the introduction of the GDP growth and corporation tax does not affect the results of model 1 and the fitness of the model deceases.
In models 3 and 4 the main explanatory variables, BIT and DTT, are introduced respectively. Both variables, sign a negative coefficient when they added in the model separately. However, the last model includes both dummies and the fitness of the model rises. The GDP is the only control variable that is significant. The main explanatory variables, treaties, indicate the same causal effects on SPEs outward like in the previous model of Non SPEs outward.
Comparison of the two models
Both models, find evidence for a negative and significant relationship between BITs and FDI. Furthermore, the results support an insignificant effect of tax treaties on outward FDI. The main difference between the models is the explanation of the OFDI by the control variables. The productive FDI is explained by the GDP growth and gravity variables, GDP of both home and host economies and distance. However, the FDI arise from SPEs is explained only by the GDP variable. Therefore, the SPEs outward is driven by other factors. It can be observe that the R square of the models has highly difference. Even though, both models are identical, the productive FDI model has higher R square compare to SPEs outward model. This is an indication that the SPEs outward is explained by other factors rather than the parameters included in the model. Finally, the hypothesis of positive relationship between treaties and FDI cannot be accepted neither in the case of BIT nor in the case of DTT.
5.2 Introduction of the Inward FDI
The adding value of this thesis is the consideration of the tax planning policies by MNCs. Previous studies, do not consider the round tripping model and the correlation between inward and outward FDI. The round tripping policy motives the introduction of the Inward FDI variable in the model. This section of the thesis, presents the results of adding to the model the variable of inward FDI.
What is round tripping?
34 redirect the funds via outward FDI to home country A. The corporations, which are using this policy, report similar inward and outward of FDI in the "channeling" economy. One method to achieve the minimization of their home country's taxes is by establishing an SPE in the host country, channel their investment via the SPE to this country and then redirect it back to the home country in the form of outward FDI. Furthermore, an indirect way of round tripping policy is presented in figure 5.1. The firm in country A report inward in country B. Then country B shift part or all of this inward to a tax haven country in order to minimize the taxes. Later, the
investment is invested back to the "channeling" economy, country B. Finally, the FDI is redirected as outflow from country B in the home country A. In this policy, there is not only precise for round tripping but for profit shifting as well. The shifting of the profits from the host economy to another country, is refer as profit shifting model.
Figure 5.1: Indirect round-tripping example
How does round tripping is related to the Dutch FDI?
The descriptive analysis of the SPEs in the Netherlands shows high correlation between inward and outward FDI in aggregate level and in some cases in country level as well. The aggregate level comparison between inward and outward is presented in figure 3.1. In some cases, the reported inward FDI from particular countries is similar to the reported outward to these
35 economies; therefore, MNCs are expected that they use the round tripping policy for their FDI in the Netherlands. In the Appendix B, the table II.2 shows that countries on average have 4.4% difference between their inward and outward SPEs. The percentage is not big enough to support that there is no existence of round tripping policy.
I explain all the above using the example of the US. US is on the top of the list of the FDI inward from SPEs and second in the FDI outward, after UK.
The following table presents the FDI inward and outward of US for the examined period.
Table 5.4: US SPEs FDI in million euro
SPEs US 2009 2010 2011 2012 2013
Inward 377,337 454,413 619,120 587,834 671,068
Outward 290,444 300,986 352,679 390,444 473,041
US investments via SPEs in and from the Netherlands are increasing during the period. The 2009 the inward from US was 377,337 million euro. The amount of SPEs sending back in US is 290,444 million euro. The missing FDI of 86,893 millions, redirect to other economies via profit shifting policies. One way of tax planning by corporations is the direct round tripping. I this case the 290,444 million invested via SPEs in the Netherlands to benefit from the tax benefits of the country and then redirect back to US via SPEs. One other way of tax planning by US firms is the diversion of their FDI. The 377,337 millions might be redirected from the Netherlands to
countries that offer tax benefits. Then the 290,444 reinvested back to the Netherlands from these economies and finally from the Netherlands in the form of outward invested back to headquarters in US.
This example presents how MNCs use SPEs for their tax benefits and the usage of the round tripping policy. Considering all the above, the log of the inward FDI of the SPEs and non-SPEs variable is added to the model. The round tripping is a phenomenon where the inward and outward FDI of one country to another, are similar. Therefore, the sign of inward SPEs variable in case of round tripping is expected to be positive.
36 fact that the dummy of BIT is one of the main explanatory variable, the model has to exclude the inward of Non SPEs. Therefore, in order to solve the multicollinearity issues, the inward of FDI is not included in the model. The results of pair correlation for Non SPEs and SPEs are presented in the table 5.5 and the table 5.6 respectively.
Table 5.5: Pair correlation Matrix of Non SPEs
Variable Outward of Non SPEs Inward of Non SPEs GDP Growth Distance Corporation Tax rate BIT Dummy DTT Dummy GDP Outward of Non SPEs 1 Inward of Non SPEs 0.853336 1 GDP Growth 0.161083 -0.049181 1 Distance -0.171621 -0.357816 0.379197 1 Corporation Tax rate 0.444876 0.356366 0.213022 0.075985 1 BIT Dummy -0.623899 -0.816798 0.438677 0.358499 -0.313199 1 DTT Dummy 0.041678 -0.043645 0.153541 -0.111502 0.408732 0.142522 1 GDP 0.571434 0.408229 0.400373 0.388302 0.583647 -0.200567 0.166323 1
Table 5.6: Pair correlation Matrix of SPEs
Variable GDP Growth Distance Corporation tax rate BIT Dummy DTT Dummy GDP Outward SPEs Inward SPES GDP Growth 1 Distance 0.379197 1 Corporation tax rate 0.213022 0.075985 1 BIT Dummy 0.438677 0.358499 -0.313199 1 DTT Dummy 0.153541 -0.111502 0.408732 0.142522 1 GDP 0.400373 0.388302 0.583647 -0.200567 0.166323 1 Outward SPEs 0.048083 -0.158458 0.382712 -0.568240 0.080947 0.495171 1 Inward SPEs 0.007657 -0.162478 0.455704 -0.694806 -0.020276 0.478439 0.940274 1
In contrast, the inward of SPEs does not cause high multicollinearity in the model. The only correlation that is above 0,750 is the correlation between Outward SPEs and Inward SPEs but is not consider as multicollinearity since the outward is the dependent variable.
37
Table 5.7: Introduction of Inward FDI
Previous Model Model 1 Model 2 Inward FDI 0.766653*** (0.067557) 0.753276*** (0.064187) DTT_Dummy 0.352437 (0.787400) 0.587493* (0.324424) 0.604103 * (0.320399) BIT_Dummy -0.902677** (0.347450) 0.287848 (0.177245) 0.215428 (0.142819) GDP 1.090483* (0.533262) 0.362960 (0.228446) 0.317255 (0.216857) Distance 0.152258 (0.144475) -0.044948 (0.060153) 0.152258 (0.144475)
Corporation tax rate -0.210897
(0.455776) -0.380295* (0.187999) -0.398926** (0.184302) GDP Growth 0.064422 (0.060568) -0.018245 (0.025955) Constant -1.249875 (5.711419) -0.896755 (2.348621) -0.896755 (2.067187) Observations 33 33 33 Adjusted R2 0.417505 0.901519 0.903435
Note: * indicates statistically significance on 10%, ** on 5%, *** on 1% confidence level. The numbers in brackets present the standard deviation of each variable.
38 negative coefficient of BIT in the comparable model, the dummy of the BIT shows not
significant causal effects on the SPEs outward activity. However, the coefficient is positive even though it is not significant. The corporation tax of the destination economy becomes significant and negative in this model. Finally, in model 2 the GDP growth is excluded which indicated an increase in the fitness of the model. The significant of the rest variables remain the same with no changes on their signs.
What is the relationship between inward and outward FDI from SPEs? The inward of the SPEs in the country explains a lot the outward since firstly is significant and secondly the adjusted R square of the model increases significantly. The inward of SPEs stimulate the outward of SPEs and the two variables are highly correlated. The correlation between inward and outward is an indicator of the round tripping policy used by MNCs.
The DTT have significant causal positive effect on DTT on SPEs activity but an insignificant positive relationship between the BIT and the SPEs. The corporation tax of the destination economy influence negatively the SPEs outward which means that the higher the corporation tax on the destination economy, the lower SPEs activity is expected in this economy. The GDP growth is insignificant and negative therefore does not included in the final model while the sum of the GDP is insignificant and positive. The distance is negative but not significant as well. In other words, only the variables that have an implication with the tax issues, is significant in the model of SPEs. The 92.15% of the dependent variable are explained by the independent variables.
39 The BIT is the treaty that protects the environment of the investment and increase the confident of the entities located in the home country. Mostly, the BITs are signed with developing
countries in order to increase the confident of the stable environment in the host economy of the investment. This thesis finds positive relationship between the FDI and BIT but, contradictory to the most of studies that are included in meta analysis of Bellak (2013), does not find evidence that the BIT is significant in the FDI activity of SPEs located in the Netherlands.
In regards with the productive corporations, explanations of the negative relationship between the BIT and the FDI outward could be the fact that productive firms drive their investments for other reasons rather than the treaties. They do not concern the corporation tax or DTT but the gravity variables, which corporations consider on their cross border investments. This is
contradictory with the SPEs outward determinants. SPEs affect their decisions for their outward FDI by the tax benefits that can obtain from the host economy. One more explanation of the negative impact of BITs might be that most of BITs are signed between developed and developing countries and this might imply that the productive firms in the sample invest in corporations located in developed world and not developing.
5.3 Robustness Test
Most of the countries included in the sample have DTT with the Netherlands. Therefore, I check one other model with the age of the treaty as the explanatory variable instead of the dummies. I prepare two sets of panel data, one for SPEs and the other for Non SPEs for the period 2008 to 2013. A fixed effect OLS of 198 observations, is used to measure how the number of years the BIT and DTT are in force impact the FDI activity. For the age of the treaty, I subtract the year of the enforcement of the treaty from the examination year; the examination period is 2008 to 2013. This measurement, confirms the results of the multivariate cross sectional regression. The DTT is positive and significant and the BIT is positive but not significant in the model of SPEs.
40 6. Conclusion
The tax policies used by MNCs are a phenomenon that is well discussed and criticized over the last years. Many MNCs establish conduit companies, known as SPEs, and through round tripping or diversion of their FDI, avoid their tax liability. The last years there is an increasing attention to researches to examine whether investment and tax treaties are associated with the observing increase of the FDI.
In this thesis, I investigate the impact of treaties on FDI outward the Netherlands during the period 2009 to 2013. Describing the FDI activity of the Netherlands for this particular period, the phenomenon of SPEs in the country is also observed. Around 80% of the FDI outward activity of the country is invested by SPEs. The mainly focus of this thesis is on the different causality effects of tax treaties on the FDI activity by productive firms and SPEs.
This thesis also introduces the inward SPEs as one explanatory variable of the outward FDI that arises from SPEs. In the descriptive analysis, it is observed that for some countries the inward FDI is almost as high as the outward FDI, especially for the case of SPEs. I strength this
argument furthermore since the trend of the outward SPEs trigger the trend of the inward. This is evident in the round tripping policy that is used by MNCs to avoid their home taxes. Therefore, I add in the model the variable of the inward FDI, which is positive and significant.
The results indicate that the tax treaty network of the country affects the SPEs activity positively. The tax treaties do not influence the productive outward FDI, since it shows insignificant
positive results. In accordance with the BIT, this does not influence the destination of the
41 The issue that arises from this thesis is the limitation of samples. The developing economies included in the sample are a minority compared to the developed countries. Most of the countries listed in the sample are countries of the EU and countries that signed a DTT with the
Netherlands. Even though the robustness of the results used another measurement of the treaties, which is the age of the treaty, the sample consists of countries where the treaties are old. This can explain the negative relationship of BIT and FDI since countries that signed a BIT are mostly developing.
I suggest that, for further research, a panel data with fixed effect model would be a more appropriate technique. In order to achieve that, a bigger sample of destination economies and a wider estimation period are needed.10 A descriptive comparison with other countries that used the same pattern, like the Netherlands, would give an evidence of the examination of more countries that use the same pattern of FDI. These countries could be Austria, Luxembourg and Hungary, which as mentioned in a previous stage, report separately the FDI activity of SPEs and Non SPEs. The graph in Appendix A shows the trend of the FDI inward and outward of these countries. A bigger sample of the origin and the destination of the FDI, for a longer period, would provide a more precise evidence of the differences between SPEs and other FDI in terms of BITs and DTTs.
Finally, the provisions that treaties include would be a better measurement of the effects of treaties on FDI. The quality of the treaties that countries sign will represent accurate results instead of the number included in the treaty network. Particularly with the DTT, the withholding taxes of interest, dividends and royalties can influence the importance of the tax treaties for corporations.
Relevance for policy and management
At the end of 2014 there was a lot of publicity about the tax facilities of Luxembourg, Lux Leaks. If there would be a ranking of similar countries, the Netherlands would be probably second on that list. Around 80% of the Dutch outward FDI is from SPEs, set up in the
42 impact on their outward FDI. In contrast: BITs and DTTs seem to have a negative impact on the outward FDI by productive firms that are located in the Netherlands.
This raises the question; which firms are actually benefiting from the Dutch treaties? The firms that have production and employment in the Netherlands, does contribute to the Dutch economy? Or foreign MNCs using the Netherlands for their tax planning? These are relevant questions because making and sustaining treaties bring costs, mainly for the Dutch taxpayers. These costs might be higher than the benefits for the Dutch productive firms, and for the Dutch economy as a whole.
The role of the Netherlands in the tax planning activities of foreign MNCs may also have
negative consequents in the longer term. The last few years the tax planning activities are seen as tax evasion end meet a growing opposition, also internationally. Countries like South Africa and India want to end existing treaties that no longer meet their demands and expectations. Countries like Luxembourg and the Netherlands, giving easy entry to notorious tax havens. One OECD project started in 2014 is against Base Erosion and Profit Shifting, which will hurt countries like the Netherlands. The whole system of international taxation is going to change.8 This does not only mean the shrinkage of the Dutch industry of lawyers and tax experts working for SPEs, but might also give problems for Dutch companies that want to invest abroad, if they can no longer benefit from tax and investment treaties. In other words, Dutch firms will benefit most if the Netherlands gets a better reputation in the field of international investment and taxation.
Views from Experts
In order to have an insight of how companies view the treaties, I contacted two people, Francis Weyzig who is a policy advisor and Klaes Wester who is working in a trust company.
Firstly, Weyzig wrote a PhD dissertation on Taxation and development and publishes articles
about the tax avoidance by MNCs. He is now working as a policy advisor on tax justice at Oxfam Novib, a worldwide development organization against poverty and inequality. Particularly, I asked him if the companies consider BIT and DTT in their decision making process of their foreign investment. In his opinion, companies are interested in treaties but treaties are not the only drivers of their investments. He thinks that companies do not consider
43
treaties for their expansion of operations in foreign countries. However, if the focus of a company is to invest in foreign country for the first time, treaties are important and affect the decision making process of the investment. The high percentage of SPEs established in the Netherlands, indicates that companies take into account the tax treaties when invest in the country since the Netherlands have many favorable tax treaties. SPEs are type of companies that mostly consider the DTT to get the maximum benefit in regards with their tax liability.
Furthermore, I had a conversation with Klaes Wester who is working in FirstNames Group in the department of corporate and institutional services; MNCs. FirstNames Group is a trust global company operates who operates in the Netherlands as well as in other countries. One of the services of the company is the maintenance and registration of foreign entities, which are part of (mostly American) multinationals, who seek to conduct business in the European Union. These entities are part of the tax structure of this multinational and are mostly registered in the
Netherlands and Luxembourg for their favorable tax climate.
I asked Wester to tell me about how treaties influence the MNCs investment decisions in other countries. He thinks that companies are more interest in countries that offer a favorable tax treaty network instead of the existence of the treaty between home and host country. One example, he mentioned, is about a multinational firm, which has its regional headquarter in Germany and hold entities in the Netherlands as well. He said that these entities are hold for purely tax reasons. The tax authority requires a company to have real substance in the country to prove that it is indeed a real entity (e.g. not just tax deduction). Therefore, it rents an office space, directors and sometimes-even employees from the FirstNames Group. The cost of the service for this
particular client is around 100 thousand euro's per year. The benefits that the MNC gain from the tax treaties are far more than the costs of the transaction.
44 7. References
Barthel, F., Busse, M., Neumayer, E., 2010. The impact of double taxation treaties on foreign direct investment: evidence from large dyadic panel data. Contemporary Economic Policy, 28(3), 366-377.
Bellak, C., 2013. How Bilateral Investment Treaties Impact on Foreign Direct Investment: A Meta-analysis of Public Policy. Draft version August 2013 prepared for 2013 MAER Network Colloquium, University of Greenwich (London, UK)
Blonigen, B., A., Davies, R., B., 2002. Do bilateral tax treaties promote foreign direct investment? Working Paper No. w8834. National Bureau of Economic Research. Available
online at <http://www.nber.org/papers/w8834.pdf> [Accessed at 10th of August 2014].
Braun, J., Fuentes, D., 2014. A legal and economic analysis of double taxation treaties between Austria and developing countries. Vienna Institute for International Dialogue and Cooperation. Claassen, P., Dool, G., 2013. The Effects of Including SPEs on BOP and FDI Statistics. De Nederlandsche Bank.
Desai, M. A., Foley, C. F., Hines Jr, J. R., 2003. Chains of ownership, regional tax competition, and foreign direct investment. Springer Berlin Heidelberg, 61-98.
Dolan, D.K., Walsh Weil, M.F., 1995. Use of Holding Companies in International Tax Planning. Taxes, 73,873-888.
Egger, P., Larch, M., Neymar, M., Winner, H., 2006. The impact of endogenous tax treaties on foreign direct investment: theory and evidence. Canadian Journal of Economics/Revue
canadienne d'économique, 39(3), 901-931.
Kingson, C.I., 1981. The Coherence of International Taxation. Columbia Law Review, 81(6), 1151-1289.
45
Markusen, J. R., Maskus, K. E., 2001. Multinational firms: reconciling theory and evidence. Topics in empirical international economics: A festschrift in honor of Robert E. Lipsey. University of Chicago Press, 71-98.
Murthy, K. V., Bhasin, N., 2013. The Impact of Bilateral Tax Treaties on FDI Inflows: The Case of India. Available at SSRN:
<http://ssrn.com/abstract=2234966 orhttp://dx.doi.org/10.2139/ssrn.2234966> [Accessed 2 September 2014].
Weyzig, F., 2013. Tax treaty shopping: structural determinants of Foreign Direct Investment routed through the Netherlands. International Tax and Public Finance, 20(6), 910-937. Pickering, A., 2013. Why Negotiate tax treaties. United Nations. Available at:
<http://www.un.org/esa/ffd/wp-content/uploads/2013/05/20130530_Paper1N_Pickering.pdf> [Accessed 14 October 2014].
Gazzini, T., Brabandere, E., 2012. Bilateral Investment Treaties, International Investment Law, Available at SSRN: http://ssrn.com/abstract=2030872.
World Investment Report 2011: Non-Equity Modes of International Production and Development. United Nations Publication.
World Investment Report 2013: Global Value Chains: Investment and Trade for Development. United Nations Publication.
World Bank investment report (2014): Investing in the SDGs: An action plan. United Nations Publication.
46 I. Appendix A: General
A. The Variables used measurement, sources and expectations For all the variables, I calculated the average of the year 2009-2013.
Table I.1: Data Sources and measurement of the variables used in the Regression Analysis
Variable Measurement Source Expected
Sign OFDI_SPEs Log of the outward FDI of SPEs Dutch Netherlands
bank (DNB) OFDI_NonSPEs Log of the OFDI of total minus the
OFDI of SPEs
Own calculation
IFDI_SPEs Log of the inward of FDI from SPEs Dutch Netherlands bank (DNB)
+
IFDI_Non SPEs Log of the inward of FDI from Non SPEs
Own calculation +/-
DTT Dummy equal to 1 in the year of the effectiveness of DDT between Netherlands and the host economy
Netherlands Tax Authority
+
BIT UNCTAD +
GDP_sum Log of the sum of the GDP of the Netherlands and the destination country
World Bank +
Distance Log of the distance between the capital of home and host country
Nl.distance.to -
GDP growth GDP growth of the destination economy
World Bank +
Corporate tax rate
Log of the Corporate tax rate of the host country
47 B. BITs and DTTs countries
The following Table I.2 and Table I.3, compare the BITs and DTTs concluded by the Netherlands. The collection of the treaties was obtained by UNCTAD and belastingdienst.nl
Table I.2: Countries which The Netherlands have a BIT and/or a DDT in 2012
Countries included in the FDI report for the Netherlands of DNB
Double Taxation Treaty
YES NO
Bila
teral Inv
estm
ent
Trea
ty
YESAlbania, Belarus, Bosnia and Herzegovina, Bulgaria, China (Mainland), Croatia, Czech Republic, China (Hong Kong), Hungary, Estonia, India, Latvia, Lithuania, Malta, Macedonia, Moldova, Montenegro, Poland, Romania, Russia, Slovakia, Slovenia, Turkey, Ukraine
NO
Aruba, Austria, Australia, Belgium, BES island, Brazil, Canada, Denmark, Finland, France1,
Germany, Greece, Iceland, Ireland, Italy, Japan, Luxembourg,
Norway4, Portugal, Spain, Sweden,
Switzerland, United Kingdom, United States5
Andorra, Cyprus, Gibraltar, Guernsey, Faroe Islands , Jersey, Liechtenstein , San Marino, Vatican City
