Predicting the performance of large
European multinational enterprises based
on their strategic decision
Predicting the performance of large
European multinational enterprises based
on their strategic decision
20.06.2014
Rijksuniversiteit Groningen Faculty of Economics and Business
Nettelbosje 2 9747 AE Groningen
Supervisor: Dr. J. A. Neuijen Wordcount: 13.970
Student:
Sebastiaan van den Bos
Student number: S2401088Address: Oostersingel 154a 9711 XL Groningen The Netherlands Phone number: 0031 6 388 96 332
E-‐mail: s.j.van.den.bos@student.rug.nl Course: Master’s Thesis IB&M
Course number: EBM719A20. 2013-‐2014.2
ABSTRACT
This thesis studies the relation between the strategic decision and performance of Multinational Enterprises (MNEs). A theoretical contribution to the existing literature is provided by making a distinction between firm strategy and firm orientation. Additionally, a distinction is made between regional orientation and global orientation of firms, which are decomposed into four diverse integration types (regional integration, global sales, global production and full global integration). By gathering performance data of 85 large European MNEs and linking this to the four integration types, it is found that there are no statistical significant differences in performance between the selected integration categories. There is a small-medium estimated effect size found of 3.1 %, which indicates that there is a relation between these factors. In general it can be stated that firms that have a global orientation, tend to posses over higher means in performance, compared to regional orientated MNEs.
Key words: Strategic decision, regional strategy, global strategy, MNE orientation, MNE performance
Research theme: Drivers for success in international business Supervisor: Dr. J. A. Neuijen
Co-assessor: Dr. K. van Veen Manuscript type: Empirical
TABLE OF CONTENTS
ABSTRACT ... I LIST OF MODELS AND FIGURES ... IV LIST OF TABLES AND FORMULAE ... IV LIST OF ABBREVIATIONS ... V ACKNOWLEDGEMENTS ... VI CHAPTER 1. INTRODUCTION ... 1 1.1 Problem formulation ... 1 1.2 Research question ... 2 CHAPTER 2. THEORY ... 3 2.1 Background ... 3
2.2 Transnational corporations and global diffusion ... 4
2.3 Performance of MNEs ... 5
2.4 The determinants of strategy ... 8
2.5 Strategic orientation of MNEs ... 8
2.6 Operationalization of MNE strategy ... 10
2.7 Industry level of globalization ... 12
2.8 Motivation of the study ... 18
2.10 Main research question & propositions ... 19
2.11 Conceptual model ... 20
2.11 Hypotheses ... 21
CHAPTER 3. DATA & METHODOLOGY ... 22
3.1 Study population ... 22
3.2 Variables and measures ... 22
3.3 Research method ... 27
3.4 Line of reasoning ... 29
CHAPTER 4. RESULTS ... 30
4.1 Descriptives ... 31
4.2 Robustness checks ... 33
4.3 Post Hoc test ... 35
CHAPTER 5. CONCLUSION AND DISCUSSION ... 39
5.1 Conclusion ... 39
5.2 Discussion ... 40
5.3 Implication ... 44
CHAPTER 6. LIMITATIONS AND FUTURE RESEARCH ... 45
6.1 Limitations ... 45
6.2 Further research ... 45
REFERENCES ... 47 APPENDIX A – STUDY POPULATION ... A-1 APPENDIX B – RETURN ON ASSETS DATA ... B-1
LIST OF MODELS AND FIGURES
FIGURE 1. Operationalization of Concepts ... 15
FIGURE 2. Regional Integration and International Strategies ... 16
FIGURE 3. Conceptual Model ... 25
FIGURE 4. Categorization ... 29
FIGURE 5. Study population ... 29
LIST OF TABLES AND FORMULAE
TABLE 1. One-way Descriptives ... 37TABLE 2. Test of homogeneity of variances ... 38
TABLE 3. Anova ... 38
TABLE 4. Robust tests of Equality of Means ... 39
TABLE 5. Post Hoc – Multiple Comparisons ... 41
TABLE 6. Univariate Analysis of Variance ... 42
TABLE 7. Descriptive Statistics ... 43
TABLE 8. Test of Between-subject Effects ... 44
TABLE 9. Integration ... 44
LIST OF ABBREVIATIONS
ANOVA Analysis of Variance
ASEAN Association of Southeast Asian Nations BP British Petroleum
CEO Chief Executive Officer
EU European Union
GDP Gross Domestic Product GLM General Linear Model
IRATA Intra Regional Assets to Total Assets IRSTS Intra Regional Sales to Total Sales
MERCOSUR Mercado Comun del Cono Sur (Southern Cone Common Market) MNE Multinational Enterprise
NAFTA North American Free Trade Agreement ROA Return on Assets
ROE Return on Equity
TNC Transnational Corporation US United States
ACKNOWLEDGEMENTS
It is a pleasure to thank those who made this thesis possible.This thesis was made possible through the support and guidance of my supervisor, Dr. J. A. Neuijen. I am grateful for his valuable advice and insightful comments so I could finish my Master’s thesis within 5 months.
CHAPTER 1. INTRODUCTION
Globalization is a process of interaction and integration among the governments, businesses and inhabitants of different nations. A tremendous amount of literature is written about the rise of globalization in the past decades. Evidence by scholars has supported the phenomenon of increased globalization by analyzing world trade data (Hindle, 2008). Some scholars argue that the importance of distances has decreased as a result of increased globalization and the world has become “flat” (Friedman, 2007). This would suggest that Multinational Enterprises are not limited by distances any longer and are free to produce and sell in foreign regions without experiencing more difficulties as opposed to operating in their domestic region. Contrasting to globalization, there is the process of regionalization. Regionalization means that firms are mainly active in their region of origin (Ghemawat, 2007). Firms can make use of globalization in their strategic decisions. Hence, globalization enables firms to produce and sell their products and services outside their domestic country. Certain businesses decide to operate around the world, and others have a preference to operate relatively close to home. Hence, there are differences between a regional orientation and a global approach and there are reasons why firms prefer one approach over another.
1.1 Problem formulation
1.2 Research question
As an extension to the existing literature about globalization and regionalization, this thesis studies the difference in performance of various types of global integration. This thesis provides a literature research that shows the differences between various forms of
globalization, their causes and their effects. There is a clear distinction provided between the strategic decision of a firm and the orientation of a firm. The strategic decision is defined as an overall total set and combination of choices, which together form the firm strategy. This thesis also provides a new definition namely: firm orientation. The firm orientation is defined as the geographical location of upstream and downstream of a firm, and is a part of the strategic decision. Furthermore, the four largest European industry sectors are discussed in order to determine if there is relation between integration categories, industry sectors and performance.
This thesis will provide four different global integration categories and position 85 large European MNEs in the appropriate integration categories. Subsequently the performance of each of the European MNEs will be measured. Afterwards, the four integration categories will be compared with each other in order to determine if there are differences in performance between the categories.
To draw the most complete conclusion and discussion for the interests described, the research question of this thesis is formulated as follows:
CHAPTER 2. THEORY
2.1 BackgroundUnder the influence of globalization, different international business strategies require support of unique international practices in order to make a firm successful. Globalization enables businesses nowadays to take benefit of advantages that a foreign region has to offer.
Globalization is a process of interaction and integration among the governments, companies and populations of different countries.
In the present day a great amount of ambiguity exists about the differences in performance between firms that are either “regional” or “global” oriented. Globalization is said to be increasing. This can be supported by measuring the growth of world trade expressed as a percentage of total global GDP. In 1990 the ratio was approximately 15%, these days circa 20% and calculations show an estimated grow up to 30% by 2015 (Hindle, 2008). In 1991, Bartlett and Ghoshal already emphasized the powerful impact of global strategies
implemented by large MNE’s, and explained this as a result of the magnitude and rapidity of change of environmental forces, industry changes and firm behavior (Bartlett & Ghoshal, 1991). Their research is extended in many forms in order to determine variables that influence the chosen strategy of firms. Most people agree that the world has become more “globalized”, however the opinions of the most accurate definition of globalization and whether this
phenomenon is positive or negative, differs. Arguments that portray globalization in a positive perspective are that it leads to the increase of competition, which encourages innovation, and that it spreads prosperity and opportunities (Stiglitz, 2006). Arguments that place
In contrast to the phenomenon of globalization, Ghemawat (2007) determined the concept of regionalization. Regionalization means that firms operate within their region of origin. Examples of regions are the NAFTA, ASEAN and the European Union (Agh, 2010). Ghemawat (2007) published an interesting study about the effect of globalization on firm developments and found that manufacturing processes of large multinationals are often divided in different regions. Due to the importance of variances and differences through regions and nations, he concluded that the world cannot be considered global and stated therefore that the world is semi-global. In fact, Ghemawat emphasizes the importance of considering the significant differences between countries, and warns MNEs not to
underestimate the effects of these differences when implementing a global strategy. Hence, although the world has become more globalized, firms operating in foreign regions still experience difficulties as a result of differences with their domestic country or region of origin. Therefore Ghemawat (2007) determined the concept of regionalization, where firms operate in their region of origin since they experience fewer difficulties in this region as in a foreign region.
It can be concluded that globalization is proven not to be worldwide uniformly because of these differences in regions. Linking economic integration with corporate strategy provided the need to divide the global integration of MNEs into two segments. The first segment contains firms that are pursuing global integration and the second segment is firms that pursue regional integration (Rugman and Verbeke, 2004).
2.2 Transnational corporations and global diffusion
TNCs must adjust to foreign formal and informal regulations in pursuit of their own learning and resource allocative goals (North, 2005). Hence, the TNC is embedded in multiple
different institutions that can differ highly across the globe. It can be stated that institutions influence the determinants and impacts of TNC activity and therefore the regional
development (Yeung, 2009). Therefore using a transaction based view and defining TNCs as outside and autonomous economic agents who pursue uniformly decision making and
corporate behavior around the globe, is incorrect. Hence, TNCs have to make strategic decisions to decide in what region they want to place their upstream and downstream. This strategic decision whether to pursue a global or regional strategy can be based on growth and trade theories (Romer, 1994). Growth and trade theories suggest that both regional
specialization and interregional trade can lead to economies of scale (Yeung, 2009). Based on Growth and trade theories, it can be concluded that firms can spread assets and sales globally. However, Oh and Rugman (2012) found that despite the global possibilities, 60 per cent of the MNEs locate their upstream and downstream in the identical region as they are born. Oh and Rugman also found that 23 per cent locate both up- and downstream in another region as the region of origin. This leaves the resting 17 per cent of the MNEs make a combination of being active in multiple regions at the same time.
Can it be stated that the most MNEs are unable to achieve a fully globally strategy, or are these MNEs operating in their home regions on purpose? Considering that the most actual literature suggest that there is regionalization trend, combined with the findings of Oh and Rugman (2012) that most MNEs implement a regional orientation, Proposition 1 is proposed:
Proposition 1: Large European multinational enterprises prefer a regional orientation
over a global orientation.
2.3 Performance of MNEs
It is widely recognized that the environment, strategy and performance relationship of a firm according to the contingency theory, is essential to the understanding of the strategy and performance of a firm (Jemison, 1981; White and Hamernesh, 1981).
The literature suggests that firms have different strategies in order to achieve a certain level of performance. In this chapter the different factors that influence the strategic decision of firms to either implement a global strategy or a more regional oriented strategy, are explained. To survive the challenge of increased globalization, firms are forced to compete with foreign competitors. Fawcett and Scully (1995) found that MNEs adapt their strategy on globalization effects that influences the firm. Firms that do not adjust their strategy in relation to
globalization, have a lack of global awareness. Fawcett and Scully (1995) therefore suggest that firms who are more globally aware, are not only better able to understand the competitive threats of increased globalization but are also better at recognizing the opportunities that come alongside with globalization (Fawcett and Scully, 1995). Hence, Fawcett and Scully suggest that firms are affected by their external environment and do adapt their strategy in order to improve performance (Fawcett and Scully, 1995). The understanding and adjusting to
globalization will lead to an above-average level of competitiveness, which is important when the amount of competitors increases (Fawcett and Scully, 1995). Petrovka (2013) corroborates that the greatest threat of globalization is that it leads to increased international competition (Petrovka, 2013). The main opportunities that are created by globalization are market integration and reduced trade barriers. Globalization may also lead to an increase in demand of new and more diversified products through cross-cultural exposures and high end products and services as a result of increased prosperity (Petrovka, 2013). Sustar (2004) suggest that firms should search for and create common characteristics of consumption in multiple foreign regions across the globe and by the use of standardization, increase their markets.
The budget for the EU is over 150 billion US dollar and Brussels spent between 2007 and 2013 over $ 1 trillion, and this growth is expected to continue (Newman, 2013).
These factors indicate that MNEs have good reason to prefer integrating their upstream and downstream within the mega-region they are located in. To provide clarity, the concepts of up- and downstream are defined. The upstream is defined as the assets of a firm, and the downstream is defined as sales of the firm. A more elaborative explanation of the process of measuring up- and downstream can be found in chapter 3 “Data” of this thesis.
Despite the profitable tariff reductions and the approximation towards more uniformed rules and regulations within the EU, the benefits of a global integration indicates that it leads to a higher performance of MNEs. A regional integration seems to be the less risky decision whereas the global orientation creates more possibilities for higher financial returns and therefore performance.
Proposition 2: The degree of financial returns is the highest for firms located in the
full global integration category.
Strategic decisions of firms do not necessarily have to be either only focused on a fully globally oriented strategy or a fully regionally oriented strategy. Combinations of these two are a possibility as well. According to the determined threats and benefits is it possible to separate the upstream and downstream of firms. This generates the opportunity for firms to spread the location of assets and sales across the globe.
The ability of MNEs to leverage the innovative and entrepreneurial potential of their
dispersed assets is a fundamental strategic imperative (Bartlett and Ghoshal, 1989). It creates the opportunity for firms to achieve the benefits of one strategy and eliminate the threats that are tied to that specific strategy. An example of the separation of up- and downstream can be; benefiting from high demand outside the mega-region by locating the sales outside this region, but remain the assets within the region in order to benefit from transparent rules and regulations.
It can be stated that it is currently unknown which large European MNEs that implement either a regional or global strategy, or are orientated according these principles, retain over the highest performance. Due to this remarkable research gap this thesis will investigate the differences between performances of firms that have either a regional or global orientation. 2.4 The determinants of strategy
The starting point of this thesis is the strategic decision of European MNEs. According to Porter (1996), strategy is about being different; deliberately choosing a set of activities to create value and above all, it is essential to superior performance. The strategic decision is the total choices made in order to interpret the means to meet the strategy (Porter, 1996).
The definition of strategy used in this thesis will be aligned with Mintzberg’s (1994) definition. According to Mintzberg, strategy consists mainly out of four factors. The first issue is: strategy is a plan; it is a means of getting from A to B. Secondly Mintzberg considers strategy as a pattern in actions over time. This can include the decision for ultimate quality of goods, either high-end or low-end. Thirdly, strategy is position. This decision selects what specific products can be placed in particular markets. The fourth and last factor covers
strategy as perspective, which includes the direction and vision of the firm. Mintzberg (1994) argues that strategic decisions and actions can collide with a changing environment but also accommodate it (Mintzberg, 1994).
2.5 Strategic orientation of MNEs
Initially, Oh and Rugman (2012) use the word “strategy” in order to indicate where an MNE is active. Oh and Rugman’s work lacks an exact definition of strategy. Since they determine the chosen firm strategy on the geographical location of organizations, this thesis will use the concept of “orientation” instead of strategy. The abstract definition of orientation used in this study is: the geographical locations of upstream and downstream of firms determine whether a company is either regional or global oriented. A distinction is made between the definitions of “strategy” and “orientation”. Whereas orientation covers the geographical area that a firm is focusing on with locating their assets and sales, is strategy according to the literature a more complete set of total activities of the firm.
Determining MNE orientation by analyzing upstream and downstream activity is the most reliable and representative method in determining where firms are active (Oh and Rugman, 2012).
2.6 Operationalization of MNE strategy
According to the literature, the strategic orientation of a firm is a segment of their overall strategy, and firm orientation can be divided in global integration and regional integration. The literature also represents that separation of assets and sales is possible. Oh and Rugman (2012) therefore suggest dividing the “global oriented” category into three subcategories in order to determine the degree of integration, namely: global sales, global production and full global integration. The illustrative operationalization of concepts is provided in Figure 1.
FIGURE 1. Operationalization of Concepts
Source: Author’s work
In order to clarify the differences between the selected orientations defined in the thesis, Figure 2 is inserted in order to illustrate the four different integration types.
FIGURE 2. Regional Integration and International Strategies
Source: Oh and Rugman, (2012)
Since the four categories are described, it is interesting to discover in which categories the large European MNEs can be found. Therefore there is need for an overview of firms per category. This will be provided in Chapter 4 Results, after the data collection occurred. Since for each single category there exists multiple contradicting underlying literature what category has the highest probability of retaining the highest performance, the performance of all categories have to be compared with each other. Hence, it should be determined if there are differences in performance between the different integration types and what category possesses over the highest performance. The existing literature provides motivations from different angles but is inconsistent and contradicting in suggesting what the strategy or
orientation leads to, or which one retains a higher performance. It is found that most European MNEs implement a regional integration strategy what could be an indication for high
2.7 Industry level of globalization
Porter (1986) argued that the degree of international competition differs from industry to industry. Underlying reason for this is that each industry possesses over unique combinations of technological, economic and competitive pressures (Porter, 1986). These pressures may increase the need for MNEs to coordinate or distribute their value-added activities across the globe. Each industry’s different mixture of competitive pressures that is experienced,
influences the strategies of firms and leads to different levels of global integration. These competitive pressures consist of tangible and intangible assets. (Prahalad and Doz, 1987; Yip, 1992). Tangible assets consists primarily fixed assets, such as property, plant, equipment, land and machinery, and current assets such as inventory. Prahalad and Doz (1987) argued that value-added activities can be generated on a global integration focus, by access to raw
materials, generating scale economies, costs reorganization, high technological intensity and a highly standardized level of customer preferences. Global industry conditions may differ significantly from each other. This is explained by differences in market structures,
distribution channels, rules and regulations and distinctive local preferences. These industry conditions determine the level of global integration for businesses, which generates the industry level of globalization. (Yip, 1992). In other words; the industry level of globalization can be determined by analyzing the global integration level of firms.
Due to differences in competitive pressures, industries are rarely identical and experience multiple pressures that form and shape an industry and their propensity on globalization. Industries are widely diversified, which triggers a significant variation to the extent that industries show global integration (Yip, 1992).
The second type is technological knowledge, which is defined as the beneficial results of new technological knowledge on the productivity and innovative capability of other firms (Peri, 2009). Differences in managerial knowledge and practices have a significant influence on cross-country productivity. Bartlett et al. (2002) found that MNEs posses more advanced managerial knowledge than local business and this advanced knowledge creates benefits for the MNE in order to obtain spillovers from local companies (Bartlett et al, 2002). Spillovers can occur within a nation (Orlando, 2004) and across nations through international investment (Branstetter, 2006). Hence, in all of the four former defined integration-categories, industry spillover possibilities are feasible and MNEs are relatively good at obtaining these industry-spillovers compared to local and usually smaller businesses.
Although many scholars gave multiple reasons for global production or sales, some specific arguments are interesting for certain industries. An interesting example is labor resources in a specific region. The textile industry is an (in)famous example of locating the upstream in low-wage countries but with good production knowledge (Tokatli, 2008). Also the unemployment rate within a region is found to have significant influence on firm creation and location (Storey, 1991). Additionally are both the availability of human resources and universities, found to have a significant influence on the decision on locating upstream or downstream in a specific region (Evans and Leighton, 1990, Kirchhoff et al. 2007).
The traditional distinction between industries is that firms are divided in either the
manufacturing industry or the service industry. Service industries contain mainly intangible downstream and manufacturing firms provide tangible downstream. However, these days a distinction between only service and manufacturing industries is deficient. Therefore the main European industries will be discussed.
Financial sector
Financial markets become more globalized. Especially banks have increased their markets as a result of growing networks of physical branches and subsidiaries in other regions as their domestic markets (Cetorelli & Goldberg, 2012). Cetorelli & Goldberg (2012) found that as a result of banks being more globally active, banks can respond to domestic liquidity shocks by shifting funds based on relative needs. Hence, banks are able to shift funds from headquarters to their subsidiaries when necessary. At first this might seem a factor that positively
influences the liquidity of banks, it turns out that the most benefits occur for firms lending by this banks, except not for the banks self. Lending by global banks has a decreased risk of domestic liquidity shocks but the banks foreign lending can be affected by this phenomenon. This may leads to an increased international wave of liquidity shocks, supported by global banks (Cetorelli & Goldberg, 2012). Hence, the globalization of banks created benefits for banks, but these benefits are aligned with negative factors as well. Therefore the performance can rise as a result of being globally active, but there is no guarantee in receiving a higher performance.
Primary Sector and Chemicals industry
The primary sector consists out of businesses making direct use of natural resources. This in contrast to manufacturing companies that produce manufactured goods. Therefore
Based on historical industry-energy prices, it is expected that the chemicals industry is mainly active in the global production category and “gas, water & electricity” providing businesses in Europe is having a relative low performance compared to other industries. Underlying
motivation is that this branch is experiencing reduced sales for multiple continues years, due to the replacing of assets of big demanding industries to foreign regions. This decrease in demand is expected to continue in the upcoming decade (Novum, 2014).
Proposition 3: The global production category contains mainly European
multinational enterprises that are active in the primary and chemicals industry.
Manufacturing industries
Service industries
Millar and Choi (2011) studied explicitly the effects of globalization on service industries. They argued that standardization and globalization are aligned, which indicates that
standardization is a condition for a high global performance (Millar and Choi, 2011). Service oriented firms need to have high market- and operational efficiency in order to be
competitive.
Therefore these firms’ implement a long-term strategy that reduces their environmental impact in the regions where they are located (Lee and Pati, 2012). According to Lee and Pati (2012) it is expected that MNE that are active in services have a high performance. This statement is confirmed by a study from Morikawa (2011). This study found that all selected service industries possessed over economies of scale (Morikawa, 2011). This finding may indicate that service industries are good at generating benefits and analyzing a beneficial location for up- and downstream. In all observed service industries, significant economies of population density where found. Also the observed firms showed an increase in productivity from 7% to 15% when firms changed their geographical location of assets or sales
(Morikawa, 2011). The strength of service industries is emphasized when a look is taken at manufacturing industries. Their sales destinations are geographically less restricted. Hence, in service industries the demand density of an area is extremely important in order to have a high performance (Morikawa, 2011). The importance of customer demand for service industries is also confirmed by Hiroshi, Takao and Makoto (2011). Their study found that two main factors are important for service industries, namely; (1) market size, which determines the demand and (2) strategic interaction between firms, which implies the competition in an industry sector. This study argues that firms active in services should find a market with a high demand and the appropriate amount of competition (Hiroshi, Takao and Makoto, 2011). The literature already determined that service providing MNEs are outstanding in finding the appropriate market and are able to improve productivity by shifting markets in order to increase their performance.
This can be based on the fact that the regional integration category is the largest group of firms, but also because it is determined that despite the increased globalization, there are still distances between countries and even more between mega-regions. The literature tends to state that the physical distance is aligned with the geographical distance. So it is expected that in each of the four categories, a high amount of firms are implementing a regional integration and that only a small amount of firms in each category are implementing a global orientation. A small amount of firms are expected to implement a global orientation, this is because only the most efficient firms can overcome the difficulties that accompany a global integration form (Bernard et al. 2007). Hence, the firms that do implement a global integration must have very good reasons to do so and must be extremely efficient, therefore are these firms expected to have higher performances as firms that implement a regional orientation. This theory is in line with a study by Rugman and Oh in 2010. This study considers that firms view the region of origin as a learning pool in order to increase efficiency and gain experience in expanding outside the domestic country. Eventually a regional orientation may lead to a more global integration of firms (Rugman and Oh, 2010).
According to the existing literature Proposition 4 is represented in this thesis:
Proposition 4: Regional integration is the majority orientation in all four selected
2.8 Motivation of the study
There exists a lot of ambiguity over when to use global orientation and when to implement regional orientation. The question: ‘To what extent is this problem worth researching?’ can be answered by looking at the value of a research. This study determines and analyzes if there are differences in performance of multiple MNEs. The importance of performance of MNEs is widely recognized (Conolly, Conlon & Deutsch, 1980). Performance is the time test of any strategy (Schendel & Hofer, 1979). Oh and Rugman (2011) determined already in their research which variables have a significant influence on the probability of choosing a certain strategy. However there still exists ambiguity about the differences in performance between various integration types. The new value that this master thesis will provide, is that it will give an insight in what the differences in performance are between MNEs with different levels of integration. Hence, the problem is that it is unknown if there is a significant relation between the orientation of large European MNEs and their differences in performance. This thesis will provide an answer to this problem.
This thesis contains an up-to-date research that evaluates the performance of multiple currently used integration levels by large MNEs. A database is provided which gives an overview of the performances of firms that pursue a regional or global orientation. Furthermore a research is provided which determines the differences in performance. By fulfilling this research, modest contribution to the existing literature is provided.
The objection of this thesis is two-fold. Firstly it aims at determining the performance of each integration category. Secondly, it analyzes the differences in performance between each orientation, and between different integration types.
2.10 Main research question & propositions
After determining the academic problem, the research question of this master’s thesis is formulated as follows:
Does the choice between implementing either a global- or a regional orientation result in differences in firm performance?
The theory provided numerous propositions. These propositions are discussed in Chapter 5 Conclusion & Discussion of this thesis. The propositions are summarized in chronological order of the theory.
Proposition 1: Large European multinational enterprises prefer a regional orientation
over a global orientation.
Proposition 2: The degree of financial returns is the highest for firms located in the
full global integration category.
Proposition 3: The global production category contains mainly European
multinational enterprises that are active in the primary and chemicals industry.
Proposition 4: Regional integration is the majority orientation in all four selected
2.11 Conceptual model
After defining the problem indication and the research question, an abstraction of the research that will be implemented can be given. In this paragraph the conceptual model is represented in Figure 3. The conceptual model is a graphical form that indicates the variables that will be researched in order to provide an answer the research question. The conceptual model can be seen as an illustration of the research of this thesis.
Research units: Large European multinational enterprises Concepts: Strategic decision and performance
Relations: Between firm strategy and firm performance FIGURE 3. Conceptual Model
Source: Author’s work
The conceptual model represents the construction of the thesis. The most left four
Hence, the arrows used in this conceptual model are representing the relations between the multiple variables. In words, the research consists out of a test in order to determine if there exist a difference between the independent variables (the four strategies) and the dependent variable (performance of the MNE). Thereafter, their performance will be determined and represented in a ranking from high to low.
2.11 Hypotheses
In order to provide a good answer to the research question, two main matters have to be answered in the results and discussion of the thesis. The first problem is to determine the performance of the European multinational enterprises of each single strategy. Second objective is to analyze whether there is a difference in performance between the levels of integration. This will be tested by a null hypothesis (Ho) and an alternative hypothesis (Ha). These hypotheses study the relation of one independent variable on the dependent variable. The null hypothesis is that the means are all equal. The alternative hypothesis is that at least one of the means is different. Hence, two hypotheses will be formulated in the thesis.
Null hypothesis: There is no difference in performance between the four types of
integration (global sales, global production, pure global integration and regional integration)
Alternative hypothesis: The performance between the four integration types (global
CHAPTER 3. DATA & METHODOLOGY
3.1 Study population
In this chapter a more in-depth view of the data used in this thesis is provided. The MNEs that are selected have to be large European MNEs. Therefore the sample will be drawn out of the
fortune global 500. Fortune global 500 is an annual ranking of the top
500 corporations worldwide as measured by revenues. The criterion for an MNE to be
selected for the sampling frame is that these firms have to be originated in Europe. According to Thomas, who states a general rule for the thesis sample size, namely: ‘As few as you must, as many as you can.’ (Thomas, 2004). Therefore, the sample size will consist out of 85 European MNEs listed on the global fortune 500. Underlying motivation for this is that only 249 European firms are listed in the Fortune Global 500 between 2003 and 2007. Thirty of these MNEs have not provided their geographic location of assets and sales. 169 of these selected firms did not provide insight in either or both total assets and net income. By leaving these firms out of the sample, the total used sample size for the thesis will consist out of 85 units. Data will be collected from 4 years in order to create reliable results. The sample consists out of firms originated in: Austria, Belgium, Denmark, Finland, France, Germany, Ireland, Italy, Luxembourg, the Netherlands, Norway, Poland, Portugal, Russia, Spain, Sweden, Switzerland, Turkey and the United Kingdom.
3.2 Variables and measures
The concepts need to be related to the empirical reality. The concepts used in the master thesis are “strategic orientation of MNEs” and “performance”. The orientation needs to be
operationalized into four indicators as presented in the conceptual model. In this paragraph the dependent and independent variables are made measurable in order to apply them for statistical purposes.
3.2.1 Categorization of dependent variables
The categorization of the selected MNEs will be based on the degree of integration. By looking at the geographical location of upstream and downstream of the firm (inside Europe, outside Europe or separated), it can be determined to what category the firms belong. Hence, two indicators, namely; assets and sales will measure the degree of integration. The used scale indicator in order to measure assets is “intra-regional assets-to-total assets” also known as IRATA. The scale indicator for measuring the sales is the IRSTS, which is an abbreviation of “intra-regional sales-to-total sales”. According to Oh and Rugman (2012) a threshold of 60% for the IRATA and the IRSTS can be used best for converting the strategic firm orientation into the four subcategories. Although using a 50% threshold seems more logical, this threshold could be too strong and overdo the regional firms (Osegowitsch and Sammartino, 2008). Besides, the IRATA is generally higher as the IRSTS and therefore a 60% threshold is preferred and is used in this research. Hence, the abstract definitions for each single category are:
(1) Category 1: Global integration: European MNEs that have less than 60% of their
sales and assets in Europe (or more than 40% of their sales and assets outside Europe) is implementing a global integration strategy.
(2) Category 2: Global sales: European MNEs that have less than 60% of their sales,
but more than 60% of their assets in Europe (or more than 40% of their sales, but less than 40% of their assets outside Europe) is implementing a global sales strategy.
(3) Category 3: Global production: European MNEs that have more than 60% of their
sales, but less than 60% of their assets in Europe (or less than 40% of their sales, but more than 40% of their assets outside Europe) is implementing a global production strategy.
(4) Category 4: Regional integration: European MNEs that have more than 60% of
Similar research is implemented earlier by Oh and Rugman (2012). Although the sales and assets information from Oh & Rugman is not available, this data is public and could be gathered from the annual reports of the large European MNEs and were provided by the OSIRIS database of Bureau van Dijk.
Data from 187 firms is gathered for a period of eight years, starting in the year 2000.
However, not all firms provided annually complete data. Finally, 1100 firm-year observation are made. The outcomes of the categorization process are as following:
Of all observed firms by Oh and Rugman (2012) is 23 % placed in category 1, which represents the Global integration category. The second category; Global sales maintains 12 per cent of the sample size. 5 Percent of all observed firms are allocated in the third category, namely: Global production. The major part of the sample size and therefore the most
implemented form of integration is placed into the final category (category 4), which is the Regional integration category. Hence, 60 per cent of all studied European MNEs implement a regional integration strategy. To summarize and clarify the categorization data and their results, an overview is provided in Figure 4.
FIGURE 4. Categorization
3.2.2 Performance measure for independent variables
The importance of performance of MNEs is widely recognized (Conolly, Conlon & Deutsch, 1980). Performance is the time test of any strategy (Schendel & Hofer, 1979). Oh and
Rugman determined in their research already which variables have a significant influence on the probability of choosing a certain strategy. The contingency theory views a firm as an open system that is affected by their external environment. The contingency theory suggests that firms will therefore adapt their strategies to the changing environment. Hence managers will use new value-added strategies in order to enhance performance as the competitive
environment changes (Hofer, 1987).In order to be able to compare differences between orientations of MNEs, which is the goal of this thesis, the actual performance needs to be measured. The performance measures are the dependent variables of the research. Measuring performance is challenging. There are many different measures in order to measure the performance of the firm. There is not a single best measure, consequently it not possible to find the perfect measure. Hence, the most appropriate measure has to be found. Samuel DiPiazza, Jr., CEO of accounting firm PriceWaterhouse- Coopers, and performance measurement expert Robert Eccles have said in 2002, ‘‘Measurement plays a dual role: It focuses attention on what is important, as determined by the company’s strategy, and it monitors the level of performance along those dimensions in the effort to turn strategy into results”. Hence, performance is considered as a suitable indicator for measuring the success of the implemented strategy of the firm.
For providing an answer to the research question it is necessary to compare the performance of firms based on assets and sales, therefore other factors will be excluded. The goal of this thesis is not to provide the best overall and most complete performance measure of
The performance measure that provides the most representable performance of the relation between assets and sales is the “Return On Assets” (ROA). The ROA provides an insight of how efficiently the firm is in converting the invested capital (total assets) into net income (Hagel, Brown & Davison, 2010). The ROA is formulated as a percentage, the higher this percentage is, the higher the abilities of a firm are in order to convert investment into net income. The formula for calculating the ROA is represented in Formula 1:
FORMULA 1. Return On Assets
Net Income ROA=
Total Assets
By way of illustration; Firm A has total assets in worth of € 8 million and a net income of € 2 million, hence the ROA is 25%. Firm B possesses over total assets of € 10 million and a net income of 3 million euros, hence the ROA is 30%. This suggests that firm B is better at converting their investment into profits. By calculating the ROA, the differences in height of investments made by the firms in the sample do not influence the performance. In this study the performance of firms will be aligned to their ROA percentage.
Hence, performance will be measured by return on assets. Reason for this is that the ROA is considered a better performance measure compared to the other often used performance measure: Return on Equity (ROE). Not a single metric is perfect but ROE can be problematic since MNEs are able to influence the performance easily in order to hide a decreasing
performance (Saraç, Ertan, & Yücel, 2014). Analyzing the ROA provides a better insight of fundamentals of the firms, including asset utilization. Allocating the resources of a business wisely is an important job of the management (Hagel, Brown & Davison, 2010). Therefore the ROA is considered as a respectable performance indicator. All financial data will be provided in U.S. dollars.
3.3 Research method
In this paragraph the research method is described. In order to test the hypothesis of this thesis, a One-way ANOVA test will be used.
A one-way ANOVA test can be used in order to determine whether or not there is a
statistically significant difference between the observed categories. If there is a statistically significant difference, then it can be stated there certain categories have a higher performance compared to the other three categories. An ANOVA is a collection of statistical models, which can be implied to analyze the variances between group means and procedure such as variation between and within groups. ANOVA offers a statistical test to determine if the means of multiple groups are identical, so ANOVA generalizes the t-test to multiple
categories. An alternative research method that could have been used in this thesis, would be implementing numerous two-sample t-tests. However, doing this raises the chance of
committing type I errors. This would unnecessary decrease the reliability of the results of this study. Hence, implementing the ANOVA test is a good test to compare the four selected categories for statistical significant differences.
Before the ANOVA test can be implied, the variance of homogeneity needs to be tested first. The Levene Statistic is an inferential statistic used to measure the equality of variances for a variable calculated over two or more groups. Hence, the Levene test is a good test to examine if the population variances are equal. An alternative for using the Levene statistic is the Bartlett test. However, the Levene statistic is less sensitive from departures to normality. Therefore the Levene statistic is considered as the better alternative.
Based on the data collection process, the variances are expected to be homogenous and
Hence, The whole idea behind the analysis of variance is to compare the ratio of between group variance to within group variance.
After the ANOVA is realized, two robustness checks will be implied in order to check if the outcome of the ANOVA still provides the same outcome when assumptions are altered. The first robustness check that will be implemented is the Welch test. The Welch test is similar to what the ANOVA is doing, except for that it accommodates heterogeneous variances. This means that the Welch test provides a different outcome as ANOVA since it includes differences in sample sizes. The second robustness check is the Brown-Forsyth test. The added value of this robustness check is that the medians are deducted away and the only distinction between the four categories is their variability. The Brown-Forsythe test is also included as a robustness check to compare the results with the Levene test since the Levene Statistic tends to overcompensate for type I errors. The Brown-Forsythe test can review this outcome.
After the ANOVA test, a Post hoc test will be implied. When there is homogeneity of variances, which is expected, the Bonferroni method will be implied. The Bonferroni test provides an insight in actual differences of performance between the different categories. There is a second method of implementing a one-way ANOVA, which provides an estimate of effect size. This can be obtained by calculating a Univariate General Linear model. This test is not a statistical test which can determine if there is statistically significant difference in means or to determine if there is a correlation that is statistically significant, but since it provides an estimate of effect size, it can be found if the effect is small, medium or large. The norms for defining the effect size are as following: a small effects is about 1 percent (η2 = 0.01), a medium effect is round 6 % (η2 = 0.06) and the large effect is circa 14 % (η2 = 0.14) (Field, 2009)
The advantage of calculating an estimate of effect size is that in case the results of the
The research question of this thesis is a closed question. The correlation of the research is that the strategic decision varies the performance of the European MNE. For the time sequence applies that the cause comes before the effect. Hence, the strategic decision can only be determined after actual implementation of the strategy. This means that the chosen strategy or orientation can be ascertained by looking at the current geographical location of the assets and sales of the MNE. Although many factors can influence the performance of a firm, this is not the case for the question which strategy over the highest performance possesses. There are no moderators of mediators applied in this research.
Hence, an ANOVA (F-test) analyses will be used in this research. Before this test can be applied, the Levene test will be executed in order to determine the homogeneity of variance. After the ANOVA, which is able to determine if there are significant difference between the diverse integration categories, the Bonferroni will be applied as a Post Hoc test in order to determine the actual differences in performance. Two robustness checks will be applied in order see if the outcomes do not change when the assumptions are altered. Furthermore a general linear model will be generated in order to find the estimate of effect size. For this later test, the Levene statistic will be necessary again. The estimated effect size will be calculated and provided with multiple descriptive and integration models. The estimated effect size will be given in a percentage and classified as small, medium or large.
Based on the research gap earlier defined in this thesis, the independent variables and their relation with the dependent variable are expected to be able to explain the differences in variance in the dependent variables accurate, when the effects of other variables are controlled.
3.4 Line of reasoning
By observing the conceptual model (Figure 3), it stands to reason that the hypotheses will provide an answer to the main research question. This indicates that when the formulated hypotheses are tested, the answer to the main research question can be provided and
CHAPTER 4. RESULTS
In this chapter the empirical results of the analyses are represented. An ANOVA test is applied with the assistance of SPSS in order to determine if there are significant differences between the four selected categories. The output of the SPSS analysis is provided in this chapter. Initially more clarity about the final positioning of the individual European MNEs is provided in Figure 5. The complete list of the study population in alphabetical order can be found in Appendix A – Study population. The outcomes of performance measurement process on firm-level are due to the detailed and large amount of data provided in Appendix B – Return on assets data.
4.1 Descriptives
In Table 1 the descriptives of the data is represented. This table shows the statistics of the four categories. N represents the sample size of each category. Hence, in the global sales category 10 firms are observed over a period of four years. The mean of their performance is 23.711. The standard deviation for this category is 19.767 and the standard error is high with 6.251. With a confidence of 95 % it can be stated that the mean performance in this category is between 9.570 and 37.850. The lowest performing firm in this category had a performance of 1.220 and the highest measured performance in this category is 67.215.
The second category, which is regional integration, retains the highest sample size containing 36 firms, observed over a period of four years. The mean is the lowest of all categories, barely 14.767. The standard deviation is 19.498 and the standard error is 3.250. Based on a
confidence level of 95 per cent the confidence interval for the mean is between 8.170 and 21.364. The lowest measured performance in the regional integration category is negative 4.910 and the highest observed performance is 111.615, which is also the highest overall observed performance.
The third category is concerning firms that pursue a global integration. The sample size for this category is 29 MNEs, observed over a timespan of four years. The measured mean is 20.511. The observed standard deviation is 18.138. A standard error of 3.368 is given. The lower bound is 13.612 and the upper bound is 27.411 based on a confidence level of 95 %. The overall lowest measured performance is found in this category, knowing: -14.359. The highest observed performance in this category is 59.461.
The fourth and last integration category contains the firms that implement global production integration. The sample size of this category is 10 and the mean is 18.384. The observed standard error is 1.989. Based on a confidence level of 95 per cent, it is discovered that the mean performance can be found between 9.710 and 27.057. The lowest observed performance in this final category is 1.102 and the highest perceived performance is 37.487.
TABLE 1. One-way Descriptives ROA N Mean Standard Deviation Standard Error 95% Confidence
Interval for Mean Minimum Maximum
Bound Lower Upper Bound
Global sales 10 23,7101 19,766836 6,250823 9,56976 37,85044 1,22 67,215 Regional integration 36 14,76669 19,497672 3,249612 8,16963 21,36376 -4,909 111,615 Global integration 29 20,51134 18,137733 3,368093 13,61212 27,41057 -14,359 59,461 Global production 10 18,3836 12,124251 3,834025 9,71043 27,05677 1,102 37,487 Total 85 18,20432 18,335398 1,988753 14,24946 22,15917 -14,359 111,615
Source: Author’s calculations
In order to determine if the chosen Post Hoc test is correct and suitable for this research, the homogeneity of variances have to be tested. This is a vital part for the analyses of the results, since the Bonferroni method is selected as a Post Hoc test. The Bonferroni test is only reliable when the homogeneity of variance is not violated. If the homogeneity of variance was
violated, a Post Hoc test that assumes that the variances are heterogeneous should have been selected. The homogeneity of variances is tested by the Levene Statistic, which can be found in Table 2. When the outcome of the Levene Statistic is significant, the variances are
heterogeneous. Hence, if the Sig. is lower then .05 then the assumption of homogeneity of variance is violated. When a look is taken at Table 2 the Sig is .692.
Since this is greater than .05, it can be stated that the assumption of homogeneity of variance is not violated. Hence, the variances are homogeneous in this test. Since the variances are homogenous in this test it is confirmed that the Bonferroni method is a correct Post Hoc test for this thesis.
TABLE 2. Test of homogeneity of variances
ROA
Levene
Statistic Df1 Df2 Sig.
0,487 3 81 0,692
* p < .05
In order to determine which group possesses over the highest performance, it first needs to be confirmed that there is a significant difference between the four selected categories. This can be done analyzing the output of the ANOVA test. The ANOVA output is provided as Table 3.
TABLE 3. Anova ROA Sum of Squares Df Mean Square F Sig. Between Groups 883,228 3 294,409 0,872 0,459 Within Groups 27356,466 81 337,734 Total 28239,694 84 * p < .05
Source: Author’s calculations
ANOVA examined the variability amongst the means and compares it against the variability within each mean in terms of the firms within each group. However, since Table 3 provides a significance level of the ANOVA, the Sig. value can also be used in order to prove whether or not the results are statistically significant. In order to be statistically significant the Sig. value has to be lower as, or equal to .05 per cent.
According to Table 3 the Sig. is .459, which indicates that there is no statistically significant difference among the variances based on the 95 per cent rule. There are no statically
significant differences between group means as determined by the one-way ANOVA (F(3,81) = .872, p= .459). Hence, there is not sufficient evidence to reject the null hypothesis.
4.2 Robustness checks
TABLE 4. Robust tests of Equality of Means ROA Statistica Df 1 Df 2 Sig. Welch 0,751 3 27,237 0,531 Brown-Forsythe 0,97 3 45,722 0,415 a. Asymptotically F distributed. * p < .05
Source: Author’s calculations
The first robustness check that is used is the Welch test. The Welch test is very close to what ANOVA is, except that it is overcoming the Behrens-Fisher problem. This problem occurs for independent samples, when differences between two normally distributed populations of which the variances of these populations are not assumed to be equal (Welch, 1947). In Table 4 it can be seen that the Welch detected a difference in degrees of freedom. The ANOVA detected degrees of freedom within groups of 81 and in the numerator the degrees of freedom is 3. According to the Welch test the numerator is similarly 3 but the denominator is adjusted substantially to 27.237. However, the Welch statistic is not statically significant since the Sig. is .531, which is higher as .05.
4.3 Post Hoc test
Although there was not enough evidence to reject the null hypothesis, the results of the Post Hoc are provided in Table 5. For the Post Hoc test, the Bonferroni method is used with the ROA as the dependent variable.
TABLE 5. Post Hoc – Multiple Comparisons
Source: Author’s calculations
Table 5 represents the mean differences between the groups. (I) Is compared with (J). Hence, each category is individually compared with the other three categories. The results of this comparison are provided in the second left column: “mean difference (I-J)”. Each comparison provides one value that represents the actual difference.
Before analyzing these differences, it should be stated that none of the results provided in Table 5 are significant considering all the Sig. values are 1.000.
Since the output would only have been significant if these values would have been equal or lower as .05, statistical conclusions cannot be based on this output. However, these results can be used for a discussion. The differences in performance between the four categories can be read as following; the mean from Global integration (20.511) is deducted from Global sales (23.710). The value that is found in the second column is 3.198. Hence, the global sales category has a higher performance compared to global integration. Again, it has to be stated that this outcome is not significant. Based on Table 5 it appears that the global sales category possesses over a higher performance as the three other groups. Firms in the regional
integration category possess over the lowest performance.
4.4 Estimate of Effect size
Since the ANOVA turned out not to be statistically significant, the estimate of effect size is calculated. Table 6 verifies that the exact same data as for the ANOVA test is used.
TABLE 6. Univariate Analysis of Variance Between-subject factors Value Label N Integration 1 Global sales 10 2 Regional integration 36 3 Global integration 29 4 Global production 10 Source: Author’s calculations
TABLE 7. Descriptive Statistics
Dependent Variable: ROA
Integration Mean Deviation Std. N
Global sales 23,7101 19,766836 10 Regional integration 14,76669 19,497672 36 Global integration 20,51134 18,137733 29 Global production 18,3836 12,124251 10 Total 18,20432 18,335398 85
Source: Author’s calculations
For the Univariate General Linear model an homogeneity of variance test is necessary as well. The Levene test is used to determine if the homogeneity of variance is violated. Since the values and output of the Levene test does not change, the results of this test can be found in Table 2: Test of homogeneity of variances.
In Table 8 is the test of between subject effects that provides the estimated effect size. The F values and the degrees of freedom are similar as in the ANOVA table (Table 3). The added value of this table is the column with partial eta squared. This column is the measure of effect size and by looking at the category row; it is found that the .031 of the variability in
TABLE 8. Test of Between-subject Effects
Dependent Variable: ROA
Source Type III Sum of Squares Df Mean Square F Sig. Partial Eta Squared Corrected Model 883,228a 3 294,409 0,872 0,459 0,031 Intercept 22826,103 1 22826,103 67,586 0 0,455 Category 883,228 3 294,409 0,872 0,459 0,031 Error 27356,466 81 337,734 Total 56408,455 85 Corrected Total 28239,694 84
a. R Squared = .031 (Adjusted R Squared = -.005) Source: Author’s calculations
By looking at Table 9, this percentage can be explained if the means are analyzed. The mean of the global sales category is 23.71. The regional integration category retains a mean of 14.767. The Global integration quadrant has a mean of 20.511 and the Global production category possesses over a mean of 18.384. The fact that there is diversion between these means, explains the partial eta squared percentage.
TABLE 9. Integration
Dependent Variable: ROA
CHAPTER 5. CONCLUSION AND DISCUSSION
5.1 Conclusion
It is found that large European multinational enterprises implement different application of global integration. The goal of this thesis was to determine if there are differences in
performance of large European Multinationals that implement different forms of integration. The integration categories are based on the geographical location of their upstream and downstream. After implementing an ANOVA analyses, it is determined that there is no statistically significant difference in performance between the four different integration categories (F(3,81) = .872, p= .459). The ANOVA results have shown that there is not sufficient evidence to reject the null hypothesis. Hence, based on the ANOVA analyses, the orientation, and therefore the attendance in a category of an MNE, has no statistically significant influence on retaining either a higher or lower performance. This outcome is supported by two robustness checks, which are the Welch test and the Brown-Forsyth test. The Welch test is F(3,27.53) = .872, p= .531 and the Brown-Forsyth test turned out to be
F(3,45.72) = .872, p= .415. It is found that there is homogeneity of variances over the units in
the sample size by applying the Levene statistic.
Based on the Post Hoc test, although not statistically significant (p= 1 for all comparisons), it is found that firms located in the global sales category retain over the higher performance compared to firms located in the three other integration categories. From high to low, the category-performance ranking is as follows: (1) Global sales (2) Global integration (3) Global production (4) Regional integration.
Since there is not sufficient evidence to reject the null hypothesis, there was need for an estimate effect size. Therefore the GLM method is used. This method found that the variability of the performance of firms is (F(3,81) = .872, p= .459, ηp2 = .031). Hence, the
categorization of firms influences their performance with 3.1%. Since the GLM is ηp2 = 0.031,
5.2 Discussion
Large European multinationals are influenced by globalization; they can take advantage from globalization by locating upstream and downstream outside their country of origin. This leads to MNEs implementing diverse integration practices. The goal of this research was to
determine if there are difference is in performance between these diverse forms of integration practices.
As the first bid to determine the differences in performance of MNEs this thesis determined which different forms of integration are used at the present time. Initially, there is a distinction made between firm strategy and the orientation of a firm. Strategy is defined in the literature as a widespread set of undertakings in order to achieve the strategic goals of the firm. The orientation of a firm is defined as a part of the firm strategy and is therefore viewed as one single strategic decision of the complete firm strategy.
The theory found that various forms of orientation are used these days. Firm orientation can be either regional or global. Based on the geographical location of the firm’s assets and sales, the two different forms of orientation can be split into integration categories. The global orientation can be divided into global sales, global production and a pure global integration practice. Firms that implement global sales have their assets mainly located inside the EU but sell most of their products or services outside the boundaries of the EU. In this category are 12 % of the large European multinational enterprises located. Firms that are positioned in the global production category produce outside the EU and sales occur mainly inside the EU. With holding only 5 % of all observed firms, this is the smallest category. The final category of firms that have a global orientation, produce and sell outside the EU and are therefore defined as the pure global integration category. This category contains 23 % of all observed firms. The regional orientation covers only one single form of integration, which is the pure regional integration and contains firms that have located both assets and sales within the boundaries of the European Union. This fourth category is also the largest of all four, possessing over 60 % of all observed MNEs.
