ABSTRACT
This paper analyses the complex location decision for multinational firms who situated outside of their parent country. Besides the extensively researched effects of economic factors, this study also explores the effects of some social variables on the number of foreign controlled enterprises (FAT) across 21 European host countries, over the period 2006-2009. As theoretically predicted, the tax rate has a negative effect on the number of FATs per country. The size of the country in GDP, low wages, a skilled labour force, and the importance of university education in the hosting country has a positive effect on the number of FATs. Inconclusive results are obtained on the effects of the number of foreign languages spoken per tertiary educated person, as the data shows a surprising negative effect of this variable on the number of FATs. The results regarding the effect of the percentage of the population with a tertiary education also display an astounding negative relationship.
Contents
1.Introduction
2. Theoretical Background
i) Tax effects
ii) Agglomeration effects
3. Methodology
i)Data
ii)Empirical Specification
4. Estimation Results
i) Description of the data
ii)Binary Logistic Regression Results and Assumption Tests
ii)OLS results and Assumption Tests
5. Conclusion
6. References
7. Appendix
1.Introduction
As more globalization takes place it seems that international borders have become a problem of the past. Governments worldwide are adopting more lenient migration policies, the most astounding of which is that of the European Union (EU). In its Council Directive of 29 April 2004, the EU declared free migration and residence of member citizens in all its territories a right. Since many multinational enterprises, MNEs, operate in many places outside their parent countries, it only seems reasonable as to why these MNEs would choose Europe for their host location, with its rather unrestrictive migration policy.In an article in the Fortune Magazine of 2007, out of the top ten cities with the most headquarters, six of the ten are European including Paris, London and Munich among others.[1] Furthermore, Van Dijk et al. (2000) studied how firms choose to relocate within the Netherlands. About one third of total firm demographics are concerned with firm relocations, whether of MNEs or local businesses. This proves the inevitable importance of the firm’s location decision.
The location choice for multinationals is a complex decision that depends on many different determinants. This study seeks to find which different economic and social factors affect the probability of hosting foreign affiliates, FATs, of multinational enterprises in different countries.The effects of these variables on the number of FATs are also analyzed.
Over the past two decades much relocation of large MNEs has occurred from the US and Europe to other European countries. Especially Western EU countries, with their well-developed and maintained infrastructures, are more than reasonable choices when deciding where to situate. According to Bruinsma and Rietveld (1993), those cities among the top ten regarding accessibility by air, rail and road are mainly located in the UK, France, Germany the Netherlands and Belgium. Although Western Europe could seem more attractive for MNEs due to these findings, some MNEs also situate in Eastern Europe. Evidence on this is provided in Faggio (2001). After MNEs have chosen to locate in Eastern Europe, they calculate the probability of choosing one of three Eastern EU members to locate to.
Many European countries have been trying to compete for foreign direct investments, FDIs, a part of which consists of foreign affiliates, FATs. Many nations have taken different approaches to attract large foreign firms. A major factor on which European states are competing for FDIs is through applying different tax policies.Avast amount of the literature is on the location of multinationals and the effects of taxes, especially on the US case. Slemrod (1990), Hartman (1984) and Swenson (1994), among others, have studied the tax effects on MNEs located in the US. As taxes seem to be of vital importance to the location decision, they have also been included in this study.
Although tax is an important factor to consider, it is not the only factor on which multinationals base their location. Another economic factor which has been extensively researched is that on agglomeration economies and their effects on the location choice for MNEs studied by Crozet et al (2003), and Brülhart et al (2012). As suggested by Davis and Henderson (2003), the assortment of services available as well as the presence of other firm headquarters attracts MNEs to situate at these sites, which weakens the tax effect on this location decision.
Some research has been done by Basile et al (2008) and Hajkova et al.(2006) on social factors affecting the locations of subsidiaries, but less research has been done compared to taxes. In the study by Hajkova et al. (2006), she states that when border and labour policies are considered along with taxes, the size of the tax effect is also tremendously reduced. This is one of the main reasons why social factors are a main focus of this study. A major concern for US investors is also the level of education. In a study on location choices of R&D MNEs locating in Europe by Siedschlag et al (2009), they findthat when education levels are included, the amount of taxes charged becomes a less important factor when deciding where to locate. The educational attainment per country is included to examine whether the tax effects on the probability of MNE location is changed. Anotherrather neglected but potentially important aspect for MNE location is the ease of communication as suggested by Welch et al. (2005). When people speak the same language, it is much easier to get the desired message across to employees. Perhaps the number of languages spoken by the residents of a country makes it easier for different cultures to integrate into the host society. Factors regarding the labour market also seem to suppress the effects of taxes regarding the site choice for subsidiaries, as discussed in Faggio (2001).
Plenty of research exists on the country level, where MNEs need to decide in which region of a country to locate, after already having picked the country. However, less literature exists to explain how the country location is chosen. In this study the focus is on 21 European countries, including countries from the North and South, as well as the East and the West.
The structure of this thesis is as follows: in the following section, some of the related literature on the subject is discussed. Section 3 describes the data and methodology used. The results are presented in Section 4, and finally the conclusions and recommendations for future research can be found in Section 5.
2.Theoretical Background
In this section, an overview of the main literature related to the subject studied onmultinational enterprise locations is provided. Much of the literature has focused on the effects of taxes on different types of foreign direct investment. This is discussed in the first subsection. The second subsection shows that when other factors are included, especially agglomeration effects, the effects of taxes on the location decision of foreign investment are weakened.
i) Tax effects
In recent years, quite some research has been done on MNEs and their location choice. In a study by the OECD (2007), “Taxation and Foreign Direct Investments”, the OECD provides an overview of the existing literature on taxes up until 2007. An analysis is given on the first major studies carried out on time series, cross sectional, panel and discrete choice models. One of the first studies done in the field of FDI´s locations was on taxation by Hartman(1984).A basic time series specification is built to test whether tax rates in the US have any explanatory power for the sites FDIs choose to situate at. He took this data on US MNEs for the period 1965-1979. In his methodology, he regresses the FDI on variables in logarithmic forms, such as the rate of return on inbound FDI and a variable measuring the after tax rate of return on US capital. In this specification, only FDI financed by new equity is used, instead of FDI which is financed by retained earnings. Due to this, there are no mixed effects between completely new FDI entering the country for the first time and already existing FDI. His findings empirically confirmed that a higher after-tax rate of return stimulates higher FDIs. Also, the higher the amount of retained earnings, the more attractive a location becomes for FDI. Another interesting finding is that the personal income tax rate as a percentage of the corporate tax rate has a negative effect on the amount of FDI. The downside to Hartman’s research is that it only uses data on the aggregate level for all FDI in the US. FDI may consist of different investments, i.e. new foreign affiliates, mergers and acquisitions and plant expansions. Special rules may apply to each specific type of investment, so conclusions drawn about the effects of taxes should not be drawn for both physical and capital FDIs.
The findings for the times series studies seem to be in line with those on a cross-sectional basis. The cross sectional studies also find evidence to support the assumption of distortive tax effects on the amount of FDI. Slemrod (1990) uses data from 1960-1987 studying the US tax rate on inbound FDI,which adds to the literature by including investments made by countries from all over the world. Two groups are compared. The first group consists of Canada, the Netherlands, France and the former West-Germany, and the other group includes Italy, the UK and Japan. This study focuses on the repatriation of foreign sourced income, where a distinction is made between countries who exempt FDI from taxes and those who use a credit tax system. Three consecutive effective marginal tax rates, EMTRs, for the host country, the US, are included in this study. The EMTR is defined as the marginal tax rate applied on a single investment, unlike the effective average tax rate, which takes an average value for a number of different projects. Also specified is the difference between home and host country taxes on corporate and personal income. He finds that the EMTR has a negative effect on FDI. Tax exemption in the home country is an insignificant predictor for the amount of FDI in the hosting US. However, a larger host (US) tax rate, results in fewer FDI funds being sent to the host country. In sum, the amount of FDI from each of the countries is negatively affected by the US, host country’s tax rate. Interestingly enough, inbound FDI financed by retained earnings are not affected by the host tax rate. Grubert & Mutti (1991)also specify a model studying the effects of taxes on property, plant and equipment, where physical rather than capital investments are analyzed. The data covers US MNEs from the manufacturing sector investing in 33 host countries. They find that outbound investment is negatively affected by the host country tax rate. However, time series analyses have their drawbacks.
The highest number of significant tax elasticities has been found on the panel basis, amounting to two thirds of all studies on panels in the paper of the OECD (2007). Swenson’s (1994) study shows the effects of US tax reforms on the foreign and native investor. A total of 18 industries over the period 1990-1991 have been examined, with data distinguishing between physical and capital investments. The investors studied have been split into two groups. The first group accounts for investors using international laws for their financial reporting. The other group consists of investors who use national accounting rules. Swenson has used logarithmically transformed variables to model FDI on the average tax rate (ATR), the exchange rate, industry dummies and time fixed effects. Remarkably, a larger ATR has been linked with a larger amount of FDI, which is in line with the prediction by Scholes and Wolfson (1990). This result may be due to the fact that non-tax predictors have been added to the equation. The sign of the effect of tax rates on FDI may change when one or all of these variables is accounted for.
Furthermore, Pain and Young (1996) have studied German and UK outbound FDI between the periods 1977-1992 into other, mainly European countries. Among other logarithmic variables, a first order lag of previous FDI has been regressed to determine the effects of independent tax rates on the FDI. Only for the UK, and not for Germany, it is found that the larger the home country tax the more sensitive an MNE is to the host country tax rate.
Finally, discrete choice models are considered, where the decision to locate is mutually exclusive. As in the current estimations that follow in section 4, the studies in the OECD Tax Policy Studies examinethe effect of host country taxes on the probability of a foreign affiliate choosing a location. Swenson (2001) studies the different types of FDI; new plants, plant growths, mergers and acquisitions, joint ventures and the FDI due to the increase in equity of firms. Instead of studying dollar terms of FDI, count data on the number of firms per location is used. The effects of statutory tax rates, which are the legally allocated tax rates in a country, are examined. A total of 46 countries investing into 50 different US states have been sampled. It is found that the FDI in its physical form is also negatively affected by state tax rates, especially for the new plants and growth of plants. Stöwhase (2003) uses a sample from 1991-1998 to study how outbound German FDI is affect by the taxes in 8 European countries. A number of fixed effects are included, namely the size of the country, labour costs and the amount of government expenditure on investment in the sectors of production and services. The average tax rate has a negative effect on new production plants being set up in a specific country. For the FDI in the service sector, the statutory taxes also have a negative effect on the investments made in new plants opened under the service sector. In conclusion, the OECD study finds that not only the amount of capital invested in a country, but also the location specifically chosen for a subsidiary is negatively impacted by tax rates
Other studies besides those reviewed in the OECD provide diverse results. Devereux et al. (2003) find similar results. They analyze data for the US from 1979-1999 on investments on simple plant and machinery in European countries like France, the UK and Germany. Like Slemrod (1990), in their methodology they use a weighted average of effective average tax rates (EATR), the effective marginal tax rate (EMTR) and the applicable statutory tax rate. They specify a range for the possible taxes due on the different earnings of projects unlike specifying only the EMTR, which is limited to the analysis of a single project only. The UK seems to be the most favoured location, especially for highly profitable investments due to high tax credits offered on repatriation of foreign income.
Hines et al(1996) examined what effect different corporate tax rates among 50 US states had on the FDI location of MNEs. Evidence is found supporting the hypothesis that foreign MNEs are sensitive to the foreign tax rates of the countries hosting them.
Voget (2011) studied the probability of parent companies to relocate given an increase in the tax rate. Data is available on an international level, so a large country-comparison can be made. The data included in this study extends over the period 1997-2007 for a worldwide database of 33 countries. His main finding was that an additional tax on income from abroad, after paying taxes for income produced within the home country, would make countries less attractive to locate headquarters at. In this study, a comparison is made between two groups of MNEs, mainly those that were taxed on worldwide profits and those exempted from paying these taxes. An interesting methodology is used. The probability of relocating or not, a binary outcome variable, is predicted using a logistic regression. A number of explanatory variables are used in different specifications, including a ratio of MNE tax payments on their income, dummies denoting whether foreign tax credits are provided and whether more than 95% of dividend repatriations are exempted from taxation. This method is very similar to the one which will be used in this paper’s empirical specification in section 4. The MNE may decide to relocate either to avoid controlled foreign company legislation (CFC laws) such as taxes on royalties or dividends, or to avoid restrictive rules which do not allow tax deferral until the repatriation of foreign income. A dependent variable reports the value 1 if the headquarter has relocated in the recent past, and 0 otherwise. The independent variables included are the EATR and the presence of CFC legislation in the specific year of interest. The market value of the MNE, leverage ratio, earnings over total assets as well as whether or not a firm is positioned in the high tech sector are the variables included to control for any firm specific effects. The results show that CFCs increase the chances of an MNE’s headquarter to relocate. A foreign country with a lower foreign tax rate would have a higher chance to attract the headquarters of an MNE to relocate to it.