International Investment Treaties & Foreign Direct Investment Is there a connection?

(1)

Page | 1

International Investment Treaties

&

Foreign Direct Investment

Is there a connection?

The Dutch Perspective on outward FDI, SPEs and M&As

Reinoud Adriaan Sluijk

(2)

Page | 2

Author Reinoud Adriaan Sluijk

Address Nijlande 17

9452 VA Nijlande

Telephone number 0031(0)6 2914 7527

E-mail reinoudsluijk@gmail.com

Student registration number s1908898

Masters MSc International Financial Management

Graduation subject Foreign Direct Investments

University Rijksuniversiteit Groningen

Faculty Faculty of Economics and Business

Thesis supervisor Dr. H. Vrolijk

(3)

Page | 3

Acknowledgement

Writing a master thesis is challenging and requires effort, perseverance and the will to step outside one’s comfort zone. The Latin adagium “luctor et emergo” is a very fitting way of describing this feeling. At the same time it increased my will to develop myself and grow. I enjoyed grasping the opportunities to develop and improve my skills. It is something I will cherish and regard as valuable for my future career.

In completing my master thesis I would like to thank those who contributed to my academic

development. My thoughts in this respect principally go to Dr. Vrolijk. Thank you very much for being my sparring partner and guiding me through the last phase of my studies!

I would furthermore like to thank my parents, my two sisters and Sanne for their unconditional support.

(4)

Page | 4

List of abbreviations

BIT Bilateral Investment Treaty

BTT Bilateral Tax Treaty

DTT Double Taxation Treaty

FDI Foreign Direct Investment

IIT International Investment Treaties

M&A Mergers & Acquisitions

MNE Multinational Enterprise

OECD Organisation for Economic Cooperation and Development

SPEs Special Purpose Entities

UNCTAD United Nations Conference on Trade And Development

WESP World Economic Situations and Prospects

(5)

Page | 5

Executive Summary

The Netherlands is positioned in the top 10 countries of Foreign Direct Investment (FDI) activities and also it is ranked high on International Investment Treaties (IITs). This thesis examines if there is connection between these two aspects. It is tested using (lagged) annual growth rates of FDI stocks and cumulative IITs whether IITs and FDI mutually encourage each other, IITs stimulate FDI or if FDI stimulates IITs. Dutch outward FDI is divided into FDI excluding Special Purpose Entities (SPEs) and FDI consisting of SPEs only. As an additional research, Mergers and Acquisitions (M&As), a specific form of FDI, are also studied in relation to IITs. IITs for this study consist of the cumulative Bilateral Investment Treaties (BITs) and cumulative Double Taxation Treaties (DTTs). The results principally comply with current literature on other developed countries in that IITs stimulate FDI although existent literature mostly studied the United States of America or certain groups of countries such as members of the Organisation for Economic Cooperation and Development (OECD).

For all FDI, including SPEs and M&As, IITs create an environment that attracts FDI. There is no support for IITs and FDI mutually benefitting each other nor is there support for FDI encouraging IITs to be signed. This thesis shows that, for The Netherlands, the growth in FDI (excl. SPEs) lags four years behind the entry into force of IITs. For M&As, this lag time is one year, if data on the Internet Crisis is omitted. However, this relationship is only applicable for FDI excluding SPEs.

(6)

Page | 6

1 Introduction

On the 4th of April 2014 Philip Morris (PMI) announced that it will end cigarette production at its factory in Bergen op Zoom. Drago Azinovic, President European Union at PMI said: “While the decline [in sales] is partially driven by societal and economic factors, PMI has consistently expressed its concern over the negative impact of excessive fiscal and regulatory policies […]. Unfortunately, new regulations affecting the industry, such as the recently agreed EU Tobacco Products Directive, do nothing to address this growing problem" (PMI, 2014). This is one recent example illustrating how regulation affects investment climate for foreign direct investment.

More recently, on the 25th of May 2014, the Sunday Times reported that Amazon, an American company, funnelled £11bn through its small office in Luxembourg last year to minimise its tax bill in Britain and across the European Continent. The online retail giant even claimed a £4m tax rebate in Luxembourg, where it has based its European operation for tax purposes. The company paid only £4.2m in UK corporation tax last year, despite generating sales of £4.3bn there. It slashes UK liabilities by channelling profits through the Luxembourg subsidiary, called Amazon EU SARL. This is an example of a special purpose entity (SPE) set up for tax optimisation purposes. Large amounts of money flow through SPEs without adding any economic activity in the country where they are located. SPEs will therefore be treated separately in this paper when discussing Foreign Direct Investments (FDI).

The topic of tax havens as in the former example is very actual one. As recent as 11 June 2014, the Volkskrant (a Dutch newspaper) published a story about The Netherlands’ position in tax evading practices. Two American organisations (Citizens for Tax Justice and Public Interest Research Group) conducted a research on US multinationals from the Fortune 500 list, which lists the 500 largest US companies and where they direct their money for fiscal policy reasons. The result was that

(8)

Page | 8 Countries and trade blocks between countries, such as the EU, seem to be engaged in a sort of

beauty contest to lure foreign direct investments into their territories. Depending on what a company intends to do, it apparently chooses a particular country to originate its foreign direct investment from. This depends on national and supranational legislation, the particular legal entity used, and it does so with a particular end in mind. The stories above also demonstrate that there are trends in the relative attractiveness of countries. It is clear that international treaties play an

important role in defining which countries are attractive for FDI. This paper investigates the relationship between the number of treaties and the development of FDI, specifically for The Netherlands.

When looking at FDI flows out of The Netherlands it turns out that since 1980 the country ranked among the countries with the highest outward FDI stock (excluding SPEs; figure 1). The Dutch

ranking, varying between 4th and 9th place for a period of 32 years, indicates that The Netherlands is a global player when it comes to outward FDI. There is a clear trend break visible 2010, where

The Netherlands’ position among the top outward FDI countries deteriorates yearly to the 9th place in 2013. It must be noted that this FDI rank is a relative measure. A country can move up in the ranks by increasing its outward FDI but it can also move up the ranks if other countries decrease its outward FDI. Nonetheless, the fact that The Netherlands is among the top ten for 32 years indicates that it is an important player in outward FDI. The data used to rank the countries on outward FDI

(9)

Page | 9

Figure 1: Dutch Outward FDI (excl. SPEs) rank 1980-20121 Source: UNCTADstat

Table 1 and figure 1 raise the question as to whether there is a relationship between the high BIT score and high outward FDI score of The Netherlands.

International Investment Treaties play an important role in steering investment flows, so they warrant closer investigation. In this thesis, specific attention is given to the amount of treaties. Table 1 shows the top ten countries that have the most bilateral investment treaties (BITs) in 2013; The Netherlands occupies the 7th position. The numbers are based on a table taken from the UNCTAD investment report 2013, included in Appendix B. A noteworthy remark is that seven of the top ten countries with the most BITs in 2013 are European, with China, Egypt and the Republic of Korea being the exceptions respectively on the 2nd, 6th and 10th spot. Furthermore, it is striking is that the United States of America occupy the 52nd position. This is remarkable because the United States is one of the main economic powers and one of the main players on the international economic play field. Lejour and van ‘t Riet (2013) state a possible explanation can be that US BITs tend to be more uniform, inflexible and pro-investment relative to BITs signed by other developed nations.

1

(10)

Page | 10

Table 1: Top 10 BIT countries in 2013 Based on: UNCTAD World Investment Report

Rank

Country

BITs

1 Germany 136 2 China 128 3 Switzerland 118 4 United Kingdom 104 5 France 102 6 Egypt 100 7 Netherlands 96 8 Belgium/Luxembourg2 93 9 Italy 93 10 Korea, Republic of 90

Definition of the research question

The guiding research question for this thesis is:

"How do the amount of International Investment Treaties and Dutch outward Foreign Direct

Investments influence each other?"

This thesis will attempt to define a relationship between IIT and FDIs. In order to give a deliberate answer to this research question it is positioned in context and several supporting questions are answered:

 How did FDI (excluding SPEs) develop over time?

 How did SPEs develop over time?

 How did IITs develop over time?

 Is there a relationship how IITs’ and FDIs’ growth rates change over time?

 Is there a relationship how IITs’ and SPEs’ growth rates change over time?

A more specific type of FDI is M&As. This thesis will elaborate on this specific type of FDI in order to analyse if M&As act differently than the overall FDI.

 Is there a relationship how IITs’ and M&As’ growth rate change over time?

In order to grasp FDIs and IITs, this thesis commences by explaining the notions FDI and IITs. Subsequently, existing literature is discussed. The next chapter consecutively provides a conceptual model combined with the methodology followed by an analysis. The thesis ends with conclusions.

2_{Belgium’s and Luxembourg’s BITs are counted in this overview as concluded by the Belgo-Luxembourg}

(11)

Page | 11

2 Foreign Direct Investment

2.1 Definition of FDI

Multinational enterprises (MNEs) expand their operations abroad via Foreign Direct Investment (FDI). Theoretically, FDI is “the allocation of resource bundles (combinations of physical, financial, human, knowledge and reputational resources) by an MNE in a host country, with the purpose of performing business activities over which the MNE retains strategic control in that country” (Verbeke, 2009, p.31). However, in this thesis the operational definition of FDIs from the IMF’s balance of payment manual, and the Organisation for Economic Cooperation and Development’s (OECD) is applied: “international investment by an entity resident in one economy in an enterprise resident in another economy that is made with the objective of obtaining a lasting interest” (IMF, 2003, p.23). This definition stresses the long-term relationship between the investor and the enterprise in combination with a substantial degree of influence on management.

2.2 Stages of internationalisation and forms of FDI

The simplest form of internationalisation is traditional export and import for which it is not necessary for an MNE to have a branch or factory in that country. Therefore, exporting has a relatively small cost. Through export a company can obtain experience in foreign markets. And if a product is unsuccessful abroad and the company decides to no longer pursue that market, the sunk costs3 are relatively low. A MNE gets more involved when it establishes a sales office abroad and becomes physically present in the foreign country. This stage entails a larger cost commitment than the exporting stage.

When a company decides to start production abroad, cost commitments become larger. There are three strategies that are commonly used to install a production unit abroad. The first and most widely used mechanism is Mergers and Acquisitions (M&As). M&As are popular because by merging or acquiring a foreign company, the obtaining company takes on board existing market presence and market knowledge. However, mergers and acquisitions (M&As) are not easy, because it can be very hard to effectively integrate companies. This might be caused by e.g. cultural differences or an internal management struggle for power. Nevertheless, M&As are the principal method for MNEs to expand abroad (Sachs and Sauvant, 2009).

3_{Sunk costs are costs incurred that cannot be redeemed in any way. Since these costs are already incurred and}

(12)

Page | 12 The second strategy is to build a production site abroad fully owned by the parent company, a so called greenfield investment. Starting a greenfield is generally viewed as risky if the parent company has limited knowledge of the foreign country. Codes of conduct and how people are perceived to do business can be very different in different countries and people react differently to the same

products.

The third approach entails obtaining a franchise as a form of FDI. When applying this strategy a company’s product is being sold in a specific area using that company’s name. It lets entrepreneurs sell its products and in return asks for a fixed fee and a certain percentage of the gross sales, so called royalty fees.

2.3 Global Foreign Direct Investments over time

Globally the amounts involved in FDIs run into the thousands of billions of dollars per year, and have been growing exponentially. Figure 2 shows global FDI since 1980. Figure 2 divides the world into three economies: developing economies, such as China, Brazil or many countries in Africa, developed countries like the US, The Netherlands or Denmark, and transition economies, such as Russia or Albania.4

4

(13)

Page | 13

Figure 2: The World's Outward FDI stock (excl. SPEs) 1980-2012 and total FDI stock (excl. SPEs)5 Source: UNCTADstat

The blue line indicates the global FDI stock excluding SPEs and has grown exponentially over the years. In fact, developed, developing and transition economies have all increased their outward FDI stock nearly every year. Furthermore, the graph shows that the world’s outward FDI stock is largely dependent on developed countries although developing countries are taking a larger share of total outward FDI for their account. Sachs and Sauvant (2009) name cross border mergers and acquisitions (M&As) as the primary driver for the increase of FDI shown in figure 2 because M&As are the main form of FDI in the developed world. At the same time M&As are becoming also increasingly dominant in developing nations.

Two other observations from this graph are the growth pause during 2000-2001 and the enormous setback in 2008 preceded by excessive growth. During 2008, the entire world economy suffered a severe setback due to the financial crisis resulting in a decline of outward FDI in all three types of economies with developed countries being hit the hardest. UNCTAD (2009) provide two reasons for this significant decrease in FDI. The first emphasises tougher access to finance; both internally due to decreased profits and externally due to a lower availability and higher cost of finance. Secondly, economic prospects deteriorated quickly serving as a disruption on investments, especially in

5

The Internet Crisis in 2000 put a halt to outward FDI growth in developed countries and developing countries whereas transition economies seemed unaffected given they grew by more than 100% during those years although in absolute numbers, although the amount of outward FDI was still very small in 2001 for this subgroup. Furthermore, there were some minor interruptions during other years. In 2011, the absolute world FDI stock increased thanks to the increase of transition- and developing countries’ outward FDI stock despite the decrease in developed countries.

(14)

Page | 14 developed countries. After 2008, the outward FDI stock recovered and when averaging the large increase in 2007 and the loss in 2008 it conforms to the longer term trend.

2.4 Special Purpose Entities

MNEs can also set up a special kind of subsidiary called a Special Purpose Entity (SPE). An SPE is commonly established in a different country than the country of origin of the mother company and engages primarily in international financial transactions. In addition, it adds little to nothing to local economic activities in the host country and do therefore not create many (if any) new jobs.

FDIs measured according to the international standards include transactions that arise from SPEs’ direct investment relationships6. However, when SPEs have financial intermediation as their primary function, the international standards recommend that transactions with affiliated banks and

affiliated financial intermediaries, except transactions in equity capital and permanent debt, are excluded from FDI statistics” (IMF, 2003, p.49). This exclusion was initiated because SPEs can have the effect of inflating FDI statistics7. This can be demonstrated by the example from The Netherlands (figure 3). The Nederlandsche Bank publishes FDI data with and without SPEs, because it is important to distinguish money flows that merely pass through the country from money flows that result in economic activity8.

6

The international standards refer to the standards recommended by the IMF. Not all organisations comply with these standards though. UNCTAD and OECD for instance measures FDI data excluding SPEs.

7_{SPEs increase FDI data because they are shell companies that merely pass through funds. It is correct that}

funds flow into the country and then leaves the country. However, the country does not benefit from these flows because there are no local activities. The money just passes through.

8_{According to IMF (2003), The Netherlands and Luxemburg are the only OECD countries that differentiate}

(15)

Page | 15

Figure 3: Total Dutch Outward FDI (excl. SPEs) and Dutch Outward FDI stock (SPEs) Source: DNB

Figure 3 shows that FDI exiting The Netherlands have increased over time and for a large part consist of investments via SPEs. In 1999, the difference between outward FDI transaction data excluding SPEs and FDI of SPEs only already comprises a factor 2 and increases over the years to a factor 3 in 2007. From 2007 it increases further to reach the factor 4 in 2012 and then drops back to 3 in 2013. Especially investments through SPEs located in The Netherlands have grown. However, there is evidence of a halt to this trend since 2011, when the financial crisis was biting badly.

2.5 Why is FDI important?

There are two core reasons why FDI is a key component of a countries development strategy. The most important one is that FDIs can create jobs and attract capital into the host country. Secondly, FDIs can be characterised as and often are knowledge-intensive (Kerner, 2009), which can be an attractive proposition for the host country. Moving or reproducing this knowledge in other countries creates knowledge spill-overs, which increases welfare. These spill-overs are caused by competition among firms and by the allocation of employees who can bring sophisticated foreign production techniques to the host country. Hence, in order to attract economic activity to (developing) countries the country must attract FDI. Thus, there is a variety of benefits (financial capital, human capital, knowledge and spill-over activity) that make it attractive for a country to look at ways to attract (additional) FDI. € 0 € 500,000 € 1,000,000 € 1,500,000 € 2,000,000 € 2,500,000 € 3,000,000 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11 20 12 20 13 M ill io n € Time (year) Total Dutch FDI

(16)

Page | 16

3 International Investment Treaties

With FDI stocks increasing, countries increasingly start to make agreements with each other in order to regulate FDIs. This chapter will shed some light on the two main types of IITs, the bilateral

investment treaty (BIT) and the bilateral double taxation treaty (DTT). According to Sachs and Sauvant (2009), BITs and DTTs are the primary IITs that significantly address investment issues.

3.1 Bilateral Investment Treaty

A BIT is an agreement between two sovereign states. The capital importing country's main objective is to attract FDI while the capital exporting country's principal concern is protecting the investments made by its nationals in the foreign country. BITs were originally made mostly between developed and developing countries where the former used to be the main or only source of FDI in the latter. The treaties were originally designed to protect against the risks of investing in developing countries where the environments to do business were often riskier than in developed countries. Historically, developing countries have at times expropriated foreign assets without compensating the owners. BITs tackle most of these problems.

3.1.1 Signing date and entering into force

(17)

Page | 17 favourable treatment of foreign investors must be reflected in domestic laws, emphasising the need to modify these laws accordingly. Adjusting domestic laws can cause a reduction in government revenue and reduce capacity to dictate terms of future investments, which is offset by the chance of attracting increased economic activity and local employment opportunities. Violating a mutually ratified treaty can initiate a binding and costly international arbitration process which could result in large fines. Literature research is suggesting that, for a country to reap the expected rewards of a bilateral treaty, it must not only be signed, but it must also be ratified by parliament in both countries (Haftel, 2010).

3.1.2 Functions of Bilateral Investment Treaties

(18)

Page | 18

3.2 Double Taxation Treaty

Along with the rise of FDI and the BIT came the double taxation treaty which is another investment agreement between two sovereign states. The objective is complementary to that of BITs. DTTs serve to agree on how revenues generated from the foreign investments are to be treated. One of the main issues these treaties deal with is how to allocate taxes on revenues. Please find an example of how taxes can be treated under a DTT below:

Say a foreign affiliate earns a profit of €10. In its home country the tax rate is 30% while in the host country the tax rate is 20%. The DTT avoids the affiliate to pay double taxes and provides the affiliate the possibility to claim credit for taxes paid abroad. Thus, the affiliate already pays €2 tax in the host country which can be deducted from the tax owed in its home country. Therefore, instead of €3, it owes on €3 - €2 = €1 in taxes to its home country. Total taxes to be paid are then €3 (€2 abroad and €1 to the home country) instead of €5 (€2 abroad and €3 in the home country). Moreover, it is possible that the clause in the DTT specifies that the affiliate only has to pay fictitious tax in the host country due to financial incentives. This leaves merely the €1 to be paid in taxes to its home country as total tax payable for the multinational.9

Probably the most important factor when analysing tax is the existence of Bilateral Tax Treaties (BTT) between countries. Double taxation on foreign earned income is one of the major stumbling blocks for cross border investments. Egger et al. (2006, p.902) clearly explain the definition of double taxation: “if the same tax base (e.g. income or wealth) of a specific taxpayer (i.e. a person or

enterprise) is taxed in two or more jurisdictions”. Bilateral tax treaties are typically a way of avoiding double taxation and as such are widely implemented by countries to increase their attractiveness for cross-border investments (Lejour and van ‘t Riet, 2013). Consequently, the number of signed tax treaties has risen significantly over time (Neumayer and Spess, 2005).

Sachs and Sauvant (2009, p.15) summarise the effect of DTTs clearly in one sentence: “The

preferential tax and related arrangements in DTTs reduce the administrative complexity of foreign investments and facilitate the flow of goods and services between the treaty partners”. In sum, both BITS and DTTs seek to increase the attractiveness of a country for FDI.

9

(19)

Page | 19

3.3 Time lag between Bilateral Investment Treaties and Double Taxation

Treaties?

In theory, one would expect a time lag between the signing of BITs and DTTs. It would be logical to sign a BIT before a DTT, because a DTT revolves around taxes from revenues generated by the investment and it takes some time before the investment breaks even with its initial investment costs. Hence, until the investment actually starts to produce profits, there would be no need yet for a DTT. Although this holds true for individual greenfield investments, this is not necessarily the case when there are multiple investments at different stages of maturity. It is probably also not the case for M&As because M&As are acquisitions of existing companies that already produce profits.

3.4 Treaty shopping

With the advent of many bilateral investment and tax treaties it has become possible for MNEs to choose optimum combinations of countries to originate FDIs from. And they can also optimise (for tax and non-tax purposes) the route that money flows through. This leads to the occurrence of a practice called treaty shopping. Basically, this means that MNEs take into account the international treaties that countries have signed when considering FDIs. MNEs have become much more aware of the possibilities that this offers to globally optimise tax and protect its investments. Treaty shopping entails that companies “shop around” to see what combination of countries has most favourable treaties for them. Weyzig (2013) calls treaty shopping a diversion of FDI motivated by tax- and non-tax reasons to maximise the economic benefit of the MNE. A non-tax-related reason to invest indirectly can include a reduction in taxes payable or avoidance of host country tax on capital gains. SPEs are often the vehicle used for treaty shopping. For example:

(20)

Page | 20

Figure 4: Example of indirect FDI

In this simple example, The Netherlands is classified as country C. The consequence is that

The Netherlands has a lot of FDI flowing in and out of the country which is merely passing through and thus artificially inflates the real FDI in- and outward FDI-figures of The Netherlands.

Treaty shopping for tax reasons has come more and more into the spotlight. Some of the very largest technological companies such as Amazon (see example in introduction) but also Apple and Google globally pay very little tax, because they have raised treaty shopping to a fine art. This is increasingly frowned upon by society and is currently the subject of public debate in many countries.

Other countries

BIT/DTT Country A&C

FDI from country A to C

Country C

Country B

Country A

FDI from country C to B

BIT/DTT Country B&C

(21)

Page | 21

4 Previous research

The world is becoming increasingly globalised and national economies are increasingly intertwined with each other. This is evident from rapid growths in exports and imports all over the world. For The Netherlands, Lejour and van ‘t Riet (2013, p.4) saw an increase of imports and exports on average of 7.5% per year between 1960 and 2005. Companies joining this trend may opt to break up production processes to be able to produce at locations that suit them best from a low cost and/or high productivity perspective. Many companies have evolved into such multinational organisations. This means that the MNEs have to deal with numerous governments, including regional ones such as the European Union. The exceptional increase in FDI during the 1990s sparked a series of research studies into the factors affecting FDI. This section discusses previous research related to international treaties and their effects on FDI activity.

4.1 Bilateral Investment Treaties (BIT)

Neumayer and Spess (2005) did research on FDI flows from the OECD to developing countries that have signed a BIT with an OECD country. More specifically, they studied whether a developing country has more FDI inflow when it has more bilateral tax treaties. They have used a larger and more representative sample compared to previous studies comprising of the FDI inflow of 119 developing countries from the period 1970-2001. Neumayer and Spess utilised a non-dyadic research approach meaning that they analyse the total OECD FDI flows to developing countries rather than country-to-country flow. They found a significant, robust and positive effect between the amount of tax treaties and FDI inflow of developing countries. Neumayer and Spess (2005) used a non-dyadic approach because they believe that BITs have positive spill-over effects. They base their opinion on the signalling effect, which is explained in paragraph 3.1.2. Especially, signing BITs with large capital exporting countries sends out a signal to potential investors that the country is serious about the protection of foreign investors. However, as Haftel (2010) states, ratifying a treaty is an even better signal than just signing the treaty. Furthermore, Coupé, Orlova and Skiba (2008) claim the signalling effect to be of less value because then countries would only sign treaties with major capital exporting countries. It is Coupé, Orlova and Skiba’s (2008) view that that is not the case in practice and therefore they prefer the dyadic research design.

Kerner (2009) places emphasis on whether BITs have a signalling effect in addition to a commitment effect on FDI. Basically, this is also what the discussion about utilising dyadic or non-dyadic data comes down to. In terms of FDI data, a dyadic approach uses country-to-country FDI flows while a non-dyadic research design measures the aggregate FDI flow. A dyadic design is criticised for

(22)

Page | 22 signatory country however it implicitly signals to other countries it is serious about FDI protection. This could create positive spill-over effects of FDI from non-signatory countries into the host country as well. A dyadic research design, focussing on country-paired FDI, does not account for this spill-over effect. The non-dyadic approach, which uses the entire FDI flows does account for this effect.

However, the signalling effect has been criticised by Coupé, Orlova and Skiba (2008). They argue that countries would then sign BITs with just a few major capital exporting countries and observe that this is not the case in practice. Using dyadic data, Kerner (2009) found a positive effect of BITs on FDI for both the signalling and the commitment effect. Nonetheless, Kerner’s results indicate that the commitment effect is stronger than the signalling effect.

Haftel (2010) studied United States’ BITs in relation to American FDI flows into developing countries. He noted that the difference between signed and ratified BITs is often underappreciated in existing literature. He added emphasis to whether a BIT was signed or actually entered into force and concluded that only the BITs that entered into force had a positive influence on FDI inflows. He argues that signing a BIT is rather costless for a government and does not provide foreign investors adequate information on how the host government will treat them. In contrast, BITs that enter into force are costly yet create a credible commitment to signal favourable treatment of FDI.

Busse, Königer and Nunnenkamp (2010) analysed the effect of BITs on FDI activity towards

83 developing host countries. They used 28 source countries, including various non-OECD countries in an attempt to cover the recent rise of FDI flows from developing countries to other developing countries. Their dyadic approach overcomes the critique of Neumayer and Spess of limited host country data coverage by exploiting UNCTAD’s largely unpublished bilateral FDI database. They conclude that BITs do encourage FDI to developing countries.

4.2 Double Taxation Treaties (DTT)

Blonigen and Davies (2004) are the first researchers to publish studies on FDI and bilateral tax treaties. Their follow up paper investigates the effect of DTTs on inbound and outbound US FDI for the period of 1980 to 1999. They found no systematic evidence that DTTs have an effect on US FDI activity and used the most comprehensive dataset at the time consisting of 88 partner countries. Their research separates old from new treaties, the former being tax treaties established before the sample period and the latter from 1981 onwards. Blonigen and Davies argue there might be an endogeneity problem10 with the old treaties because the US tends to conclude DTTs with highly

10_{With this endogeneity problem is meant that DTTs may be correlated with unobserved variables and}

(23)

Page | 23 active FDI countries such as Canada, Japan, Australia, New Zealand and Western European countries at an early stage.

In 2006, Egger et al. conducted a study on the role of tax treaties for bilateral stocks of outward FDI. Their model weighed the impact of tax treaties on both welfare and outward FDI. They assumed a bilateral tax treaty is only implemented if both countries gain in welfare. Egger’s study had two particular hypotheses: firstly, they studied under which conditions are countries inclined towards implementing a bilateral tax treaty: welfare effect of tax treaties. And secondly, of more importance to this thesis, whether FDI rises or falls after implementing tax treaties, given the assumption that a tax treaty increases welfare for both countries. The result of the second hypotheses suggested a negative correlation between tax treaties and outward FDI stocks. The reason Egger et al. provide for this negative relationship is that the tax treaty reduces tax revenue in the parent country. If a country uses tax revenues to finance public infrastructure, it reduces plant set-up costs. However, reduced tax revenues lead to higher plant set-up costs for an MNE due to having to compensate for lacking infrastructure. Egger et al. state that this effect may outweigh the FDI enhancing effect the tax treaty has due to the abolishment of double taxes. However, combining the first and second hypothesis, they conclude that tax treaties do not have a clear-cut predictive value on FDI. There may be another possible explanation for this negative correlation: Egger’s assumption that FDI is only performed when tax treaties increase welfare of both countries may not be valid. In a pessimistic scenario it could be possible that it is a sufficient condition for a country to sign an IIT into force that anticipated FDI benefits some key members of a ruling coalition. A more positive view could be that a host country is actually prepared to give up some short term income in order to build infrastructure that will pay off in the longer term. In reality there are probably examples of both.

(24)

Page | 24 The analysis done by Azémar and Delios (2008) focussed on the outward FDI from Japanese

investments to developing countries between 1990 and 2000. Their research focuses on corporate tax rates although they additionally examine the effect of a specific clause in bilateral tax treaties, called tax sparing, on corporate tax rates. They concluded that when corporate tax rates increase in host developing countries, Japanese firms invest significantly less in those host countries. However, some countries signed also a DTT with Japan. And for those countries that have signed a DTT with Japan, the corporate tax rate has no longer a significant effect on the location choice of Japanese multinational firms. For countries that do not have a DTT, the strong negative relationship between tax rates and FDI is still observed. Indirectly, this research shows an effect of tax treaties on FDI: Azémar and Delios (2008) conclude that higher tax rates reduce Japanese outward FDI. The other way around, lower tax rates then encourage FDI. Tax treaties have the effect on tax rates that the importance of tax rates diminishes. Consequently, DTTs make for a more stable environment which is favourable for FDI.

Barthel, Busse and Neumayer (2010) focus their research on the impact of double taxation treaties on FDI. They used 30 FDI source countries, of which 10 are developing countries and 105 FDI host countries of which 84 are developing countries. Their major contribution is that they overcome either the absence of information on bilateral FDI or when researchers did use a bilateral FDI data, they suffered from a small and unrepresentative sample. Alternatively, Barthel, Busse and Neumayer (2010) were able to create a dyadic country dataset in which both developed and developing

countries were broadly represented over a long period of time (1978-2004). They found a statistically significant and substantial result, namely that DTTs increase the bilateral FDI stock between 27% and 31%.

(25)

Page | 25 Lejour (2014) demonstrates in his study on the impact of bilateral tax treaties on FDI for all OECD countries from 1985 onwards that, similar to what he found with van ‘t Riet in 2013, new bilateral tax increase FDI by 21% and after ten years the effect fades away. He also compared country pairs that did sign a treaty with pairs that did not. Lejour concluded that countries that do sign treaties on average experience a higher growth in FDI stocks than the pairs that did not sign a DTT.

Consequently, Lejour (2014) and also Lejour and van ‘t Riet (2013) find that bilateral tax treaties increase FDI.

4.3 Both BITs and DTTs

(26)

Page | 26

Table 2: Summary of previous literature

Authors BIT and/or DTT? Effect on FDI

Neumayer and Spess (2005) BIT Positive

Kerner (2009) BIT Positive

Haftel (2010) BIT Positive

Busse, Königer and Nunnenkamp (2010) BIT Positive

Blonigen and Davies (2004) DTT None

Egger et al. (2006) DTT Negative

Neumayer (2007) DTT Positive

Azémar and Delios (2008) DTT Indirectly positive

Barthel, Busse and Neumayer (2010) DTT Positive

Lejour and van ‘t Riet (2013) DTT Positive

Lejour (2014) DTT Positive

Coupé, Orlova and Skiba (2008) Both BIT: Positive

DTT: None

(27)

Page | 27

4.4 Conceptual Model

It is assumed that FDIs and IITs lead to economic growth. More treaties can attract more companies and also facilitates the administrative burden for both companies and the country itself. Hence more economic activity is being stimulated. FDI also encourages economic growth by the creation of e.g. new jobs and a competitive industry. Although these factors foster economic growth and welfare, they lay outside the scope of this thesis.

From the above chapters the preliminary conclusion could be drawn that IITs, such as BITS and DTTs, are a precursor of FDI growth. However, the causal relationship between IITs and FDI remains unclear. In this subchapter a conceptual model will be presented illustrating three alternatives on how IITs and FDI could be related. Since previous research generally found a positive impact of IITs on FDI, there are three alternative ways IITs and FDI can stimulate each other. Figure 5 illustrates the conceptual model comprising these three alternative ways of stimulating the wheel of economic growth.

Figure 5: conceptual model

Interntional Investment Treaties (IITs) Foreign Direct Investment (FDI)

2 3

(28)

Page | 28 Alternative 1 illustrates the relationship where IITs and FDI mutually encourage each other. This is evident if the rise of FDI and IITs occurs at the same time. Alternative 2 indicates that countries that already have IITs in place could become more popular candidates for FDI. However, alternative 3 shows that it can also be the case that FDI is the source of IIT growth. In this scenario, countries that already receive many FDI might decide to create more IITs in order to protect its high inward FDI. More IITs would also facilitate the existing FDI and reduce administrative complexity for both the country and the investors.

In order to study the relationship of Dutch FDI with IITs, it is necessary to map all aspects through time. Figures 6a to 6c depict the theoretical shapes of the graphs following the conceptual model, which derives the three hypotheses this thesis will examine:

1. Alternative 1 (figure 6a) demonstrates that FDI and IIT grow at the same time indicating they affect each other positively (alternative 1 from the conceptual model).

Figure 6a: Alternative 1

Alternative 1: IITs

FDI

(29)

Page | 29 2. Figure 6b illustrates that there is a lag time between IITs and FDI: IITs increase first followed

by FDI. There is a lag time because it takes some time for MNEs to realise that those agreements have made the host country more attractive for foreigners to invest in (alternative 2 from the conceptual model).

Figure 6b: Alternative 2

3. In figure 6c FDI rises before the IITs (alternative 3 from the conceptual model). This could be the case when countries are put under pressure by MNEs to make agreements to protect the rising or high level of FDIs, to encourage further growth of FDIs and to decrease

administrative complexity.

Figure 6c: Alternative 3

FDI

IITs are a precursor to FDI with a certain lag time

FDI

(30)

Page | 30

5. Methodology

FDI is mostly measured in either inflows or stocks. The difference between measuring by inflows or stocks can be found in the fact that FDI inflows refer to the capital of foreign investors that invest into a foreign country during a year, while FDI stocks reflect the total value of foreign owned assets at a given point in time. Treaties are related to existing outstanding investments and that is shown by FDI stocks. Moreover, flows fluctuate a lot which makes it difficult to discover long term trends. Therefore, this thesis uses FDI stocks as a measure for the total value of FDI over time. This is also the preferred measure in existing literature, at least for US DTTs studies (Neumayer, 2007).

FDI data for the world is extracted from the UNCTAD database and converted to euros according to OECD historic exchange rates. This is done to be able to compare it better to the Dutch FDI data which is already in euros. The annual outward Dutch FDI stock with and without SPEs are provided by the Dutch national bank, de Nederlandsche Bank. Unfortunately though, outward FDI of SPEs is available from 1999 onwards. This limits the ability to evaluate the effect of early treaties on growth. Concerning FDI data, one has the option to utilise dyadic or non-dyadic data. In an ideal world, a dyadic research design would be preferred because of the bilateral information making the dataset more extensive and therefore more accurate than non-dyadic data. However, it is hard to obtain such a complete dataset because such data is not freely available. Therefore, non-dyadic data is used in this thesis. The consequence is that this thesis measures the total effect and therefore cannot distinguish between the signalling effect or the commitment effect treaties have.

Data about BITs and DTTS is extracted from the UNCTAD’s Country specific Lists for BITs and DTTs. This study has followed Sachs and Sauvant’s (2009, p.7) advice where they claim that “UNCTAD has the best database of BITs”. Additionally, the UNCTAD website provides lists of DTTs with signing dates per country. Furthermore, given Haftel’s (2010) argument that merely signed BITs do not have an influence on FDI while ratified BITs do, this study uses the date the BIT entered into force for the analysis. The data of BITs and DTTs were in need of a couple of changes to be able to work with them. First, they were sorted by year. Next, the cumulative numbers were created leaving gaps in which years there was no treaty signed (or ratified in the case of BITs). The list of BITs is updated until 1 June 2013, while the list of DTTS is updated until 1 June 2011. To facilitate the analysis, BITs signed during 1 January 2013 and 1 June 2013 are ignored11. The same applies for DTTs, albeit in 201112.

11

During the period of 1 January 2013 to 1 June 2013, which is excluded from the analysis, no BITs were signed.

12_{During the period of 1 January 2011 to 1 June 2011, which is excluded from the analysis, no DTTs were}

(31)

Page | 31 Furthermore, when IITs as a whole are considered, data is updated until 2011 because DTT data limits more recent analysis.

Both FDI and IIT data is graphed through time and the logarithms are taken to be able to better spot changes in growth rate. Taking it a step further, year on year growth rates are determined for both FDI and IITs revealing the annual growth rates.

In order to define the type of relationship the shape of the cumulative IIT curve is matched to the FDI curve. If the two are re-enforcing each other (Relationship 1) than a good correlation would be expected between the two curves without any time lag. However, if IITs trigger FDI (Relationship 2), than a good correlation might be expected when the FDI curve from later years is compared with the cumulative IIT curve from earlier years. The FDI curve is kept in place while the IIT curve is lagged, by moving it to the right, to check if Relationship 2 from the conceptual model holds. On the other hand, when FDI triggers creation of IITs (Relationship 3), then the cumulative IIT curve from later years would correlate with the FDI curve. In order to find out whether such a relationship exists, the IIT growth rate curves are compared to the FDI growth rate curve. Relationship 1 is tested by looking at the IIT growth rates and FDI growth rates of the same years. For Relationship 2 and 3, a certain lag has to be implemented. In both cases, the IIT growth rates curve is one that is being moved while the FDI growth rates curve remains at its original time scale. The IIT growth rates curve is shifted in a window from minus 10 years (IIT curve moved left testing Relationship 3) to plus 10 years (IIT curve moved to the right testing Relationship 2). The minus 10-year to plus 10-year window is chosen arbitrarily. However, for SPEs and M&As, 10-year movement of the IIT growth rates was not feasible due to the lack of data from previous years and the first completed M&A on the list was in 1997 even though the M&A data starts at 1980. Shifting the IITs curve left for SPEs was limited to maximum of 6 years. And for M&As, IITs could be shifted left for 8 years maximum. Correlation coefficients and R-squared are calculated for each of these comparisons.

It must be noted that FDI is an umbrella term for all sorts of investments abroad. As was touched upon in chapter 2 the lion’s share of FDI is a consequence of M&As. Therefore it is of high interest to study whether the results are similar when looking at M&As instead of FDI. The Bureau van Dijk’s Zephyr database is used as a source for data on M&As. Information such as the type of deal (merger or acquisition) and the date the deal was completed of Dutch acquirer firms are extracted from Zephyr. Information provided by Zephyr relating Dutch to Dutch target firms13 were deleted from the extracted file, because they do not contribute to outward FDI. Furthermore, the column called

13_{A Dutch firm that merges or acquires another Dutch firm is not relevant for this research because these}

(32)

Page | 32 deal status was cleaned of ”withdrawn deals” and “assumed completed” deals14. The database distinguishes between the acquirer and the target company. Considering a merger, it is difficult to decide which company is supposed to be the acquirer and which one the target firm. Therefore, mergers are also deleted from the sample. In addition, acquisitions with an unknown stake

percentage are deleted because the unknown stake cannot provide any indication whether there is a lasting interest nor if there is a large enough stake to have a significant influence on management (which is at least 10%).

Then, the year in which the acquisition was completed was used to continue the analysis. For each year, the sum of the deal value was calculated. The summed deal value per year instead of the total number of acquisitions is used because for FDI the outstanding FDI is used as well and not the number of investments that occurred. As an additional analysis, M&As were investigated omitting the Internet Crisis data points from 2000 to 2002 because this particular crisis hit M&As severely.

14_{“Withdrawn deals” are deals that did not happen and thus are excluded from this research. “Assumed}

(33)

Page | 33

6 Analysis

In this chapter, analyses are performed on Dutch outward FDI vs. world outward FDI, Dutch IITs vs. Dutch outward FDI and the causal relation between Dutch IITs and outward FDI.

6.1 Foreign Direct Investment

The Dutch outward FDI stock including and excluding SPEs are compared with the world’s FDI outward stock excluding SPEs as illustrated in figure 7 below. For both the world and

The Netherlands, it is not until the new millennium that FDI becomes of some significance. Until 2007, the trend is that FDI mostly increases. However, in the period of 2000-2001, the total world’s FDI stock declines by 3%. To the contrary, Dutch outward FDI by SPEs and outward FDI excluding SPEs increased with 12% and 15% respectively during that time period. Then, in 2008, global FDI

plummeted by 15% whereas The Netherlands was not affected as hard given that outward FDI without SPEs declined by 0.1% and SPEs’ FDI even increased by 12% in 2008. The world already redeems itself in 2009 by nullifying the 2008 decline and firmly continues its upward trend after that. Dutch outward FDI maintains its increasing nature although SPE growth fell back significantly since 2012. Where on average they grew 12% per year during 2008-2011, they grew just 3% in 2012 and even -1% in 2013.

Figure 7: Dutch Outward FDI stock (excl. SPEs) and Dutch Outward FDI stock (SPEs) vs. world FDI stock (excl. SPEs) Sources: UNCTADstat, IMF and www.dnb.nl

€ 0 € 2,000,000 € 4,000,000 € 6,000,000 € 8,000,000 € 10,000,000 € 12,000,000 € 14,000,000 € 16,000,000 € 18,000,000 € 20,000,000 € 0 € 500,000 € 1,000,000 € 1,500,000 € 2,000,000 € 2,500,000 € 3,000,000 19 80 19 82 19 84 19 86 19 88 19 90 19 92 19 94 19 96 19 98 20 00 20 02 20 04 20 06 20 08 20 10 20 12 W or ld o u twar d FD I excl . S P Es in M ill ion € D u tc h o u tw ar d FD I in M ill io n € Time (year)

(34)

Page | 34

6.2 International Investment Treaties

Figure 8 exhibits the cumulative Dutch IITs over time, namely BITs and DTTs. The Netherlands has experienced a growth of both BITs that entered into force (red bars) and DTTs (blue bars). It also reveals during which years BITs entered into force and/or DTTs were signed. Years in which a bar is drawn are years in which one or more BITs became legally binding instruments and/or a DTT was signed. Years that do not possess a bar did not experience a BIT that entered into force and/or a DTT signed. The added line simplifies noticing discrepancies between BITs and DTTs. In addition, the total IITs (sum of BITs and DTTs) is indicated by the green line.

Figure 8: Cumulative International Investment Treaties over time

The first BIT The Netherlands had was initiated in 1965 and entered into force in 1966. Since then, BITs increased slowly throughout the 1970’s. It can be derived from figure 8 that many Dutch BITs became ratified towards the end of the 1980’s. That is where the upward trend accelerates. During the 1990’s until 2008 the number of BITs increased explosively. Although in 2008 there were no new BITs ratified, 2009 continued the upward trend. However, since 2009, no new BITs have entered into force (the dataset used to produce this graph is updated until and 1 June 2013 for BITs and until 1 June 2011 for DTTs). Both the semi years not taken into account, thus BIT data is extracted until 2013 and for DTTs until 2010. This means that there have not been Dutch BITs entering into force for the last 3.5 years. A possible explanation could be that since the Dutch BIT network expanded rapidly, The Netherlands has BITs in place with most or all of its usual trading partners. In that case there is no need for new BITs.

(35)

Page | 35 The first Dutch DTT was signed in 1951 and the last one in 2010. The growth rate picked up in the 1970’s followed by a short period of no new DTTs during 1982 to 1985. Through 90’s the number of Dutch DTTs signed increased the most. In the early 2000’s, DTTs kept on growing albeit somewhat slower than in the 90’s. As figure 8 shows, this results in a somewhat flattened S-shaped graph. The graphs of DTTs and total IITs also take this shape.

It is apparent that DTTs were signed before BITs were signed and entered into force15. The difference between the first DTT and the first BIT entered into force is 15 years. This indicates that some countries first signed a DTT rather than a BIT. Apparently certainty about double taxation is a prerequisite for creating an environment suitable for investment. Another indisputable fact is that there are more DTTs in total than BITs. There always has been and as of 1 June 2013 there still are. The BIT and DTT growth curve are fundamentally similar apart from their different starting point and a couple of discrepancies. Both experienced a significant boost in the late 1980s all the way until 2010 although DTTs grew slightly faster than BITs.

The timing of such international investment agreements is already touched upon in chapter 3 paragraph 3. It was stated that if BITs preceded DTTs, probably greenfields were the main form of FDI. This is because taxes are not relevant until the investment breaks even its initial investment. When DTTs precede BITs, as is the apparent case for The Netherlands, probably M&As are the principal form of FDI. When M&As are performed, profits are made straight away. The MNE obviously prefers not to pay double taxes as well as protect its investment in the new company.

6.3 Dutch outward FDI and Dutch IITs

Figure 9 demonstrates cumulative IITs and Dutch outward FDI (excluding SPEs and only SPEs) in absolute numbers in one graph. The graph shows that both BITs and DTTs increased in numbers before outward FDI starts to make its rise. Where IITs begin to increase substantially around the 1990s, FDI without SPE started its rise at the end of the 1990s. It is not possible to infer whether SPEs started increasing before IITs or vice versa because the data on SPEs starts in 1999. From 1999 through 2004, outward FDI of Dutch SPEs seems to increase parallel to both BITs and DTTs. As of 2005, outward FDIs of Dutch SPEs increase faster than IITs indicated by the steeper line. FDI excluding SPEs also rises but not as fast as IITs do.

15_{The first Dutch BIT was signed by Côte d’Ivoire on 26-04-1965. This was not the first BIT to enter into force}

(36)

Page | 36

Figure 9: Dutch Outward FDI stock (excl. SPEs), Dutch Outward FDI stock (SPEs) and cumulative BITs and DTTs against time

To have a closer look at the growth of FDIs and IITs, figure 10 shows the log of the FDI and IIT data. When the line rises linearly, growth increases exponentially. The steeper the line, the faster it grows. Alternatively, when the line drops, there is negative growth. This happens occasionally for FDI. Furthermore, when the line is horizontal it means no growth or decline has taken place. There are a couple years for IITs for which this is the case.

Figure 10: Log Dutch Outward FDI stock (excl. SPEs), Dutch FDI (SPEs) and log of both BITs and DTTs

0 20 40 60 80 100 120 140 0 500000 1000000 1500000 2000000 2500000 3000000 19 50 19 52 19 54 19 56 19 58 19 60 19 62 19 64 19 66 19 68 19 70 19 72 19 74 19 76 19 78 19 80 19 82 19 84 19 86 19 88 19 90 19 92 19 94 19 96 19 98 20 00 20 02 20 04 20 06 20 08 20 10 20 12 C u m . n u m b e r o f B ITs & DTTs Ou tw ar d FDI w ith & w ith o u t SPE s Time (year) Total Dutch FDI outward excl. SPEs

Total Dutch outward SPEs Cum. BITs Cum. DTTs 0 0.5 1 1.5 2 2.5 3 3.5 0 1 2 3 4 5 6 7 19 50 19 52 19 54 19 56 19 58 19 60 19 62 19 64 19 66 19 68 19 70 19 72 19 74 19 76 19 78 19 80 19 82 19 84 19 86 19 88 19 90 19 92 19 94 19 96 19 98 20 00 20 02 20 04 20 06 20 08 20 10 20 12 Log c u m . n u m b e r o f B ITs & D TTs Log o u tw ar d FD I wi th & w ith o u t SPE s Time (year) log Dutch FDI outward excl. SPEs

(37)

Page | 37 Figure 10 reveals that BITs consistently grow faster than DTTs from 1985 onwards until 2010. From that time The Netherlands did not sign or ratify new DTTs or BITs and growth patterns levelled out. In general, it can also be seen that IITs grew faster at first but since the start of the millennium FDI has higher growth rates, partly because both BITs and DTTs diminished their growth. It will later be shown that this can possibly be attributed to the lag time between treaties and FDI of around 4 years.

Figure 11 depicts the Year on Year (YoY) growth rates of IITs. From 1987, the growth in IITs

experiences an upward trend until 1995. Since then, the growth in IITs slows down. The black lines indicate trends over several years.

Figure 11: YoY IIT growth rates

6.4 Dutch outward FDI excluding SPEs vs IITs

(38)

Page | 38

Figure 12: YoY Dutch Outward FDI growth rates (excl. SPEs)

Looking at the longer term trend lines in figures 11 and 12, it seems that outward FDI without SPEs and IITs follow a similar pattern of first increasing growth followed by period of decreasing growth. Figure 13 depicts the growth rates per year relative to the previous year for IITs and Dutch outward FDI without SPEs. However, figure 13 shows that their apparent analogous growth path does not occur at the same time. It seems that the growth of Dutch outward FDI excluding SPEs lag a couple of years behind IITs’ growth path.

Figure 13: YoY Dutch Outward FDI growth rates (excl. SPEs) and YoY IIT growth rates

-10% -5% 0% 5% 10% 15% 20% 25% 30% 35% 19 85 19 86 19 87 19 88 19 89 19 90 19 91 19 92 19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 Year o n Year gr o wt h r ate s Time (year)

(39)

Page | 39 In order to find out whether FDI excluding SPEs’ growth lag behind IIT growth, correlations and R-squares were calculated for different lag times, which are shown in figure 14. In addition, the correlations and R-squares are shown to test the whether FDI precedes IITs (Relationship 3). This is shown by negative years lagged.

Figure 14: Correlations and R-square values for YoY Dutch Outward FDI growth rates (excl. SPEs) and YoY IIT growth rates

Figure 14 clearly shows that at 0-years lagged, the correlation and R-square coefficients are very low (Alternative 1). Testing for Alternative 2 (whether IITs precede FDI), one would have to look in the plus 1 to plus 10 year range of Figure 14. At 4-years lagged, both the correlation and R-square coefficient are at their highest level, respectively 0.72 and 0.51. These are fairly high coefficients and therefore there is an indication that it may take four years before the effect of new treaties can be seen on FDI (excluding SPEs). Figure 15 shows how 4-year lagged IITs growth rates relate to FDI excluding SPEs. Moving on to the third alternative (testing if FDI excl. SPEs precede IITs), the range minus 1 to minus 10 is inspected, because the IIT curve is moved left. One can see that there are correlations and R-squares increase while the IIT cure is moved to the left. The highest correlation is at -0.71 and R-square is 0.50. However, this value is reached when IITs are moved by 9 years. It is unlikely for a country to implement IITs based on FDI (excl. SPEs) from 9 years ago with a negative correlation: i.e. the more FDI grew 9 years ago, the fewer IITs are implemented. Therefore, this result is ignored. -0.80 -0.60 -0.40 -0.20 0.00 0.20 0.40 0.60 0.80 -10 -5 0 5 10 Co rr e lation an d R -sq u ar e c o e ff ic ie n t

Time (years lagged)

Correlation excl. SPEs R-square excl. SPEs

Alternative 1

(40)

Page | 40

Figure 15: YoY Dutch Outward FDI growth rates (excl. SPEs) and 4-year lagged YoY IIT growth rates

6.5 Outward FDI of Dutch SPEs vs IITs

Looking at Dutch SPEs’ FDI, the same pattern can be distinguished as with the growth rates of IITs and FDI excluding SPEs (figure 16). Generally, SPEs’ FDI increase its growth rate slightly from 1999 to 2006, albeit with large spikes. Subsequently, FDI growth rates diminish yearly, apart from a minor rebound in 2011, to even being negative in 2013.

Figure 16: YoY Dutch Outward FDI growth rates (SPEs) -10% -5% 0% 5% 10% 15% 20% 25% 30% 35% 19 85 19 86 19 87 19 88 19 89 19 90 19 91 19 92 19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 Year o n Year gr o wt h r ate s Time (year)

(41)

Page | 41 The outward FDI of Dutch SPEs vs. IITs is drawn in figure 17. Although a similar growth rate path is distinguished for FDI of SPEs and IITs’ growth, just as with outward FDI without SPEs and IITs, their timing is imbalanced. It appears that the growth of outward FDI of Dutch SPEs also lag a couple of years behind IIT growth.

Figure 17: YoY growth rates for Dutch Outward FDI (SPEs) and YoY growth rates for IITs

To determine if there is a good fit with a particular lag, correlations and R-squares are computed for different lag times. The result is shown in figure 18.

Figure 18: Correlations and R-square values for YoY Dutch Outward FDI growth rates (SPEs) and YoY IIT growth rates 0.00% 5.00% 10.00% 15.00% 20.00% 25.00% 30.00% 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11 Year o n Year gr o wt h r ate s Time (year)

YoY FDI growth SPEs YoY IIT growth

-0.80 -0.60 -0.40 -0.20 0.00 0.20 0.40 0.60 -8 -6 -4 -2 0 2 4 6 8 10 Co rr e lation an d R -sq u ar e c o e ff ic ie n t

Time (years lagged)

Correlation SPEs R-square SPEs

Alternative 1

(42)

Page | 42 Without a lag for IITs and FDI (SPEs), Alternative 1, there is a pretty strong correlation and R-square coefficient of -0.57 and 0.32 respectively, indicating a negative influence of IITs on FDI (SPEs). For Alternative 2 (IITs precede SPEs’ FDI), the best correlation and R-square coefficient is at three years lagged, although a 2-year lag and 4-year lag is a close second best. However, none of the coefficients are convincing, with the correlation coefficient for 3-years lag being -0.53 and an R-square of 0.29. The two and four year lag coefficients are 0.47 for at two years lag and 0.41 at four years lag. R-squares are merely 0.22 at two years lag and 0.17 for four years lag. All these low and fluctuating coefficients indicate little predictive value. Alternative 3 (SPEs’ FDI precede IITs) has its best correlation and R-square coefficient at a 1-year FDI lag with 0.51 and 0.26 respectively. However, when one takes a bird’s eye view of the entire graph it becomes clear that the correlation coefficient is very unstable. It goes from positive to negative values and back very often. These fluctuating values restrain from drawing conclusions.

6.6 M&As vs Dutch IITs

This paragraph zooms into a specific form of FDI, namely M&As. Figure 19 shows that M&As in value have experienced an extraordinary peak of €70,523 million around the millennium change, the period preceding the Internet Crisis. After that, the yearly amount of Euros invested via M&As has dropped to an average of €7,335 million.

Figure 19: The sum of M&A deal value per year and the number of M&As per year

(43)

Page | 43 There is no apparent relationship between IITs and the sum of deal values per year (Figure 19). DTTs and BITs have both risen steadily over the years, while the deal values are still the same order of magnitude as in 1996. Figure 20, showing the logs of the cumulative BITs and DTTs and the log of the summed deal value per year demonstrates this better. Apparently, there are other factors that influence the way the M&As have developed over time, which are beyond the scope of this paper. During the 1990s there was a belief that a new economic era had begun, driven by the seemingly endless possibilities that the internet was providing. This period was characterised by enormous investments in electronic commerce initiatives (Walden and Browne, 2008), without paying attention to P/E ratios and without a good analysis when companies would actually become profitable, if at all. While in hindsight it was true that some companies were very successful and have grown

exceptionally fast (Microsoft, Apple, Google), it transpired during the Internet Crisis that exceptional growth was not meant to be for all Internet companies. Consequently, investor perceptions changed from basing their expectations on the future during the build-up of the bubble, to tending to rely more on actual performance when the bubble deflated (Aharon, Gavious and Yosef, 2010). Therefore, a lot of the M&A activity from those years turned sour, triggering the Internet Crisis.

Figure 20: Log of M&A deal value per year and log of BITs and DTTs

Looking at the year on year growth rates in figure 21, there appears to be only one downward trend.

0 1 2 3 4 5 6 7 8 9 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11 20 12 lo g M & A d e al v al u e Log B ITs & log D TTs Time (year)

(44)

Page | 44

Figure 21: Year on Year growth rates of the sum of M&A values

Year on Year growth rates for M&As and IITs are combined in figure 22.

Figure 22: YoY growth rates for M&A deal values and YoY growth rates for IITs

-100% -50% 0% 50% 100% 150% 200% 250% 300% 350% 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 Year o n Year gr o wt h r ate s Time (year)

YoY M&A growth Linear (YoY M&A growth)

-100% -50% 0% 50% 100% 150% 200% 250% 300% 350% 0% 5% 10% 15% 20% 25% 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 YoY M & A g ro wt h Year o n Year gr o wt h r ate s Time (year)

International Investment Treaties & Foreign Direct Investment Is there a connection?