Supervisor: dr. R.K.J. Maseland Co-assessor: dr. M.S.S. Krammer Student: C. Scheffer Student ID: 2010429 c.scheffer@student.rug.nl
RIJKSUNIVERSEIT GRONINGEN
FACULTY OF ECONOMICS AND BUSINESS
MSC INTERNATIONAL ECONOMICS & BUSINESS
Technology Seeking Foreign Direct
Investment:
The non-negative effects of combing cultures in a
quest for knowledge and competences
1
ABSTRACT
Mainstream cultural distance theory predicts that differences in culture between MNEs home and host countries have negative effects on the entry mode, internationalization and performance of FDI. This paper investigates this assumption in relation to the performance of technology seeking FDI, and finds evidence to reject the negativity hypothesis. This outcome fits within the small but growing group of researches that argues there are non-negative effects of cultural distance. Furthermore a comparison is made between the cultural distance measures of Hofstede and Globe. The outcome slightly favours the former over the latter as a measure of cultural distance in international business studies.
2
TABLE OF CONTENTS
SECTION 1: INTRODUCTION 3
SECTION 2: LITERATURE REVIEW 5
2.1INTRODUCTION 5
2.2FOREIGN DIRECT INVESTMENT 5
2.2.1MAIN CHARACTERISTICS OF FDI 5
2.2.2TECHNOLOGY SEEKING FDI 6
2.2.3LOCATION CHARACTERISTICS OF TECHNOLOGY SEEKING FDI 8
2.3CULTURAL DIFFERENCES AND FDI 8
2.3.1CULTURE IN INTERNATIONAL ECONOMICS AND BUSINESS 8
2.3.2THE NEGATIVE ROLE OF CULTURAL DISTANCE IN TECHNOLOGY SEEKING FDI 9
2.3.3THE POSITIVE ROLE OF CULTURAL DISTANCE IN TECHNOLOGY SEEKING FDI 10
2.4HYPOTHESES 10
2.5ECONOMIC MODEL 12
SECTION 3: RESEARCH METHODOLOGY 13
3.1INTRODUCTION 13
3.2THE ROLE OF CULTURAL DISTANCE IN FOREIGN DIRECT INVESTMENT STUDIES 13
3.3MEASURES OF TECHNOLOGY SEEKING FOREIGN DIRECT INVESTMENT 14
3.4MEASURES OF CULTURAL DISTANCE 15
3.5MEASURES OF TECHNOLOGY CREATION 16
3.6SAMPLE 17 3.7DATA PREPARATION 19 SECTION 4: RESULTS 22 4.1INTRODUCTION 22 4.2STATISTICAL RESULTS 22 4.3DISCUSSION 24
4.4ROBUSTNESS OF THE MODEL 25
SECTION 5: CONCLUSIONS 29
5.1INTRODUCTION 29
5.2THE EFFECT OF CD ON TECHNOLOGY CREATION 29
5.3LIMITATIONS AND FUTURE EXPANSIONS 30
REFERENCES 31
APPENDIX SECTION 3 35
3
SECTION 1: INTRODUCTION
FDI is increasingly seen as the major driver of international trade and shaper of the field of international economics (Navaretti & Venables 2004). Due to the size of these capital flows they are able to affect the world economy and to distribute welfare. Countries like Ireland and South Korea successfully increased their overall standard of living during the 90’s and 00’s. FDI is seen as a major catalyser for this development by increasing investments, knowledge and skills. The drivers behind these investments, multinational enterprises (hence MNEs) therefore increasingly become subject of investigation in international economics and business studies. International economics focuses, by nature, on the bigger picture and consequently study the macro effects and causes of FDI. Therefore they have been considering the MNE as a black box until recently (Navaretti & Venables 2004). However in international business studies the motives for MNE to internationalize and those factors that influence this, have been under investigation for decades and are still at the top of the agenda. A recent motive for the MNE to internationalize is technology seeking according to Makino et al. (2002), which, if successful, should provide the MNE with a comparative advantage. One factor that influences the success of technology seeking, according to international business theory, is the existence of cultural differences, which is the focus of this research.
In mainstream international business theory, cultural differences are argued to have a negative effect on entry mode, international diversification and performance of FDI (Tihanyi et al. 2005). Recently a different way of thinking has been introduced; Shenkar (2001) argued that there is not only negativity to be assumed from cultural differences. Specific in relation to technology seeking there are reasons to expect positive effects from differences. Barkema and Vermeulen (1998) argue that exposure to a different culture enables companies to increase the knowledge and skills, due to the learning by diversity argument. This holds that the MNE that is affected by a workforce and market from a different culture is able to learn from this diversity. The effect of the combination of different cultures on technology seeking FDI is the main topic of this research. This leads to the following research question:
4
The aim of this research is to establish more understanding of the role of differences between national cultures in technology seeking investments. It investigates not only the potentially negative effects of cultural distance but also looks at how multinational enterprises create advantages because of these differences. The outcome of the research is relevant for both policy makers and international business scholars. The latter because it will enrich their field of studies, in fact this contribution fills a gap in the current international business literature as it always assumes negativity to derive from cultural differences. For policy makers the relevance lies there where policy is designed in order to act as a host country to MNEs. Furthermore managerial implications of this research are that it creates more insights into the cultural differences and its role and importance in building comparative advantage.
5
SECTION 2: LITERATURE REVIEW 2.1 Introduction
This section provides a summary of the relevant research that focuses on technology seeking FDI and the role of culture within this topic in international business theory. First the main characteristics and theories of FDI are discussed and how technology seeking fits within this topic. Furthermore an elaboration is provided about the technology seeking type of FDI and the motivations that are provided for this behaviour in the literature. This can all be found in paragraph 2.2. Secondly an overview of the role of culture and cultural differences in FDI research is provided in paragraph 2.3. Based on these three statements are hypothesized in paragraph 2.4, the economic model is derived in paragraph 2.5.
2.2 Foreign Direct Investment
2.2.1 Main characteristics of FDI
FDI is seen as an extraordinary activity, something that is not common for all companies. In fact the MNEs driving FDI are argued to be of a special breed: they are “the happy few” (Mayor and Ottaviano 2007:1). This indicates that there existence is rare and that being a MNE has positive effects. Therefore it is remarkable that international economics has “forgotten” to discuss their existence. Hence traditional neoclassical theory sees the company as a “black box” and it is even argued by some that therefore FDI is impossible (Dunning 2000). The effect of which is that international economic theory does not proceed into in-depth explanations on why MNEs exists and on what their motive for internationalization is. This is even more remarkable when the importance of MNEs and FDI is considered for the world economy. To the contrary: within international business theory the existence of MNEs and their motives for internationalization trough FDI are discussed for over four decades. Therefore the explanations and motivations of FDI and the existence of MNEs, discussed in this subparagraph should be looked at as a backbone for the next sub-paragraphs.
6
motivation behind the foreign activity. Furthermore it is argued that the larger the relative size of the competitive advantages (to other companies, preferably in the host country), the larger the foreign activity is. Regarding the second variable it argues that the immobility of the natural and/or created endowments is a driver of foreign activity. The larger the immobility is the more chance that a company will favour foreign over domestic locations for augmentation or exploitation of the competitive advantages. The third, internationalization variable provides a framework to evaluate the different internationalization methods that and company might use to create or exploit its core competences. Furthermore an important factor of the eclectic paradigm is that is it “strongly contextual” (Dunning 2000:164). This implies that it represents a combination of country or region, industry or activity and in-house factors of an individual company and by that the reason for that company to use FDI. Because the eclectic paradigm varies from company to company there are different motives for international activity. The literature has identified four drivers behind international activity based on the eclectic paradigm (Dunning 2000).
First there is the market-seeking motive aimed at satisfying a foreign market or markets.
Secondly companies internationalize because of resource seeking motives, which is aimed to gain access to natural resources.
Thirdly, in relation to motive one and two there is the efficiency-seeking motive, aimed at a more efficient spread of labor.
The fourth motive is “to protect or augment the existing ownership specific advantages and/or to reduce those of their competitor” (Dunning 2000:165). This is also known as strategic asset seeking.
Makino et al (2002) are one of the first researchers to find evidence for the fourth motive, which they argue to be relative recent. They study the behaviour of Taiwanese companies that engage in FDI and find that FDI investments are not only undertaken to exploit assets (seek markets) but also to developed or acquire specific assets to gain comparative advantage and thereby reduce production cost, hence asset seeking or technology seeking FDI. The augmentation of the existing ownership advantages by MNEs is the focus of this research; therefore the next sub-paragraph elaborates on this technology seeking FDI.
2.2.2 Technology seeking FDI
7
motives and industry motives. The argument behind both motives is that the MNEs lack capabilities in-house to develop specific knowledge and competences. These are a requirement to developed technology and therefore create a comparative advantage. Therefore they internationalize to develop their technologies. Several studies (Dunning 1993, 1995, Chang 1995, Almeida 1996 and Makino et al. 2002) support this conclusion.
With respect to the home country motive the lack is explained by nature of the home country. Makino et al. (2002) discuss this argument in relation to the outward FDI of Taiwanese MNEs. They argue that the motive behind international technology seeking behaviour is to explore industry specific knowledge and competences that the MNE is not capable of developing in the home country. An explanation for this is that the home countries technology is underdeveloped due to their overall level of development (Caselli and Coleman 2006). They argue there is “a skill bias in cross-country technology differences. Higher-income countries use skilled labor more efficiently than lower-Higher-income countries” (2006: 499). This is explained by the fact that developed countries are skilled-labor abundant and therefore choose technologies that are best suited. Hence the same principle applies for companies from developing or low-income countries. They are unskilled-labor abundant and therefore choose technologies that are more suited for unskilled workers. Besides the level of development of the labor force there are other factors in the home country that could hinder the ability to develop technology. E.g. market demand; in less developed markets it could not be profitable to developed technologies because the demand is not sufficient to overcome the cost.
8
and companies that create so-called spillovers. More specific in order to create knowledge and profit from the spillovers a MNE has to locate in the industry specific location, also referred to as “main centres of R&D excellence” by Dunning (1994), Cantwell and Janne (1999), and Pearce (1999). Furthermore Cantwell et al. (2004) suggest that the innovating capabilities of these main centres of R&D excellence are increasingly attractive reasons for FDI.
2.2.3 Location characteristics of technology seeking FDI
The spread of technology is not equally distributed around the world. There are regions with higher technological capabilities and those with lower. The majority of studies focus on technology seeking FDI from developed to developed countries (so called North-North FDI flows). Recently scholars started to investigate this type of FDI from developing countries to developed countries (so called South-North flows). Dunning (1997) argues that since the 1990s the developing countries are counting for an increased share of outbound FDI. Makino et al. (2002) analyse these relatively new flows. With regards to the technology seeking MNEs they found that they were more likely to invest in developed countries. This is in line with Patel and Pavitt (1991) who find evidence that technology creation is to a large extent done by developed countries. This argument is also made by Caselli and Coleman (2006) who argue that companies in skilled-labor abundant countries possess technologies specialized for skilled-labor. Schmidt-Ehmcke and Zloczysti (2011) support this conclusion. They identify country-industry combinations that are leading in the manufacturing industry. They come to this conclusion by focussing on the world technology frontier, which determines the level of technological development of each country based on the efficiency in research and design (R&D). Based on these arguments I assume that technology development is a developed world’s thing and therefore the FDI flows aimed at technology seeking are most likely observed as inward flows in developed countries.
2.3 Cultural Differences and FDI
2.3.1 Culture in international economics and business
9
structures that are deemed essential to the constructed identity of a community”. According to Hofstede (2001) another characteristic of culture is that it enables us to distinguish between groups of people. In addition to this, studies find different properties of cultures, e.g. Au (2000) argues that there is no such thing as one national culture, within a nation there are many different cultures or “intra cultural variation” (Au 2000: 1). Furthermore some sub-cultures from different nations might be closer than sub-sub-cultures within a nation. Even though all these arguments are plausible in themselves, in the present study the focus will be on national culture in line with definitions of Hofstede (2001) and Beugelsdijk and Maseland (2011).
Kirkman et al. (2006) investigate into the use of culture in international business studies. They focus on 180 studies that appeared in top-tier journals between 1980 and 2002. A large amount of these studies (148) had culture as main effect, other as moderator (32). In specific these studies focus on cultural distance (hence: CD), which is defined as the “differences between national cultures” (Tihanyi et al. 2005: 270). Furthermore Hofstede (1980) and Kogut and Singh (1998) link CD to “significant differences between countries norms, routines and repertoires” (Morosini et al. 1998: 153). The CD concept quantifies the role of culture, which increases its usefulness in economic and business studies. The current research will elaborate from this point onwards on CD, by which the cultural differences between two nations is implied.
There is a wide variety of factors that are studied in relation to CD, however he majority of studies focuses on its role in FDI according to Shenkar (2001) and Tihanyi et al. (2005). Shenkar (2001) argues that there are three factors under investigation in studies that focus on CD as a main factor that influences FDI, namely: entry mode decisions, international diversification and performance. Shenkar (2001) furthermore states that there are inconsistent results obtained in the research, which is in line with the argument of Tihanyi et al. (2005). The literature shows opposing findings and/or unproven theories in all three fields of the FDI researches. The next two sub-paragraphs will therefore elaborate on the mainstream negative theory and the recent positive role of CD in FDI research.
2.3.2 The negative role of cultural distance in technology seeking FDI
10
programming of the mind, which distinguishes one culture from another (Hofstede 2001). The suggestion made by, among others, Li and Guisinger (1992) is that overcoming the differences increases costs and failure, and according to Chang (1995) hinders the ability of MNEs to generate rent. Furthermore CD increases the complexity of doing business while MNEs internationalize to increase performance according to Tihanyi et al. (2005). Therefore the goals and forecasts that the MNEs have are not reached.
2.3.3 The positive role of cultural distance in technology seeking FDI
While the benchmark idea leads to the conclusion that CD is negative there are researchers that argue the opposite. The idea that there is a positive effect from CD between the home country of the MNE and the affiliates host country stems from the learning from diversity argument, summarized in Barkema and Vermeulen (1998). This states that the influence of different environments increase the knowledge and skills of a company. Especially “the infusion of new ideas and new practices sparks innovations and boosts technological capabilities” (1998: 8). Similar arguments are found in studies from Levitt and March (1988), Abrahamson and Fombrun (1994), Miller and Chen (1994, 1996). Furthermore Argyres (1996a) argues that a new market provides new needs and possibilities to test the developed technology. They enhance innovation and result in more technological skills. This is also found by Morosini et al. (1998) whom argue that the integration of skills can lead to combinations of resources that are unique when MNE and its affiliate acquired their skills in cultural distance countries. These studies show that a new and different culture increases the knowledge and skills of a MNE, which are a requirement to develop new technology.
2.4 Hypotheses
11
FDI. However there are, in contradiction to this argument, researchers who suggest the opposite. In relation to performance of FDI they follow the learning from diversity argument. This holds that there are positive effects on the knowledge and competences of a MNE due to the exposure to a host countries workforce and market. The cultural differences in home and host countries workforce and market lead to increased technology creation and thus increase the performance of this type of FDI and the MNE. Based on these two opposing theories the following is hypothesized:
H1a: Cultural distance is negatively related to technology creation. H1b: Cultural distance is positively related to technology creation.
The CD literature suggests that familiarity with the culture should decrease the distance between the cultures for the MNE (Reus and Lamont 2009). Therefore it is expected that the effect of CD is lower if the MNE already has experience with the host country’s culture, more formally:
H2: The effect of CD on technology creation is moderated by the MNEs previous experience
with the host countries culture.
In line with hypotheses 1a and 2 there is reason to expect a colonial history between two countries to lower the effect of CD on technology creation. Acemoglu et al. (2001) argue that a previous colonial relationship has an influence on the host country today. Therefore a (previous) colonial relationship will reduce the diversity of the host countries culture for the MNE, more formally:
H3: The effect of CD on technology creation is moderated by a colonial relationship between
the MNEs home and host country/
12
Fourthly Regarding the year of the M&A the argument is that “the economic and financial conditions vary year by year” (Morosini et al. 1998:145). Finally regarding the industry the argument holds that industries have different preferences and therefore performances; these should not be compared without controlling for this. Therefore these five factors are introduced to control the relationship between CD and Technology Creation.
2.5 Economic Model
The following figure 1 is derived from the hypotheses. A short explanation per variable is provided. Section 3 discusses the variables in-depth and also their methods of determination.
Figure 1: Economic Model
Technologyijt+2 = β1 + β2 CDklt + β3 previous country experienceilt + β4 colonial
relationshipklt + β5 (CD * previous country experience)iklt + β6(CD * colonial
relationship)iklt + β7 GDP growthkt + β8 country restrictionslt + β9 MNE Sizeit + β10
Year of M&Aij + β11 Industryijt
In the model I represent the Ith MNE and J stands for the Jth company that is either merged with or acquired by the MNE, K represents the MNEs home country, L represents the MNEs host country, T stands for time of M&A. Because of data limitations only entrance trough M&A are investigated in this research, even though other studies did also focus on Greenfield investments (e.g. Kogut and Singh 1988). The dependent variable Technologyijt+2
measures the difference in technology creation by the affiliate in second year after M&A and compares this to the year leading up to the M&A. This lag originates from the argument that these years are first of all critical to the survival of the M&A and secondly that it takes time to integrate the MNE and affiliate (Morosini et al. 1998). CDijt measures the CD between the
home countries of the MNE and the affiliate at time t. Previous country experienceilt and
colonial relationshipklt measure the experience of the MNE in the country of the affiliate and
if there was a colonial relationship between the two countries. Both variables are included as interaction effects because they moderate the effect, as hypothesized. Finally GDP growthkt,
Country restrictionit,Sizeit, Year of the M&Ait and Industryit are control variables that
13
SECTION 3: RESEARCH METHODOLOGY 3.1 Introduction
This section introduces the methodology for the current research. Furthermore data related issues and the sample are discussed. In specific this overview starts with a discussion of the methodology previously used in research that focuses on CD and FDI. Secondly the literature that identifies different motives behind FDI flows is discussed. Thirdly the methodologies used to measure CD are discussed. Fourthly the methodologies to measure technology creation are discussed. Each of the above paragraphs furthermore specifies which methodology is chosen based on the review of the relevant methodologies. Based on this a sample is selected, which is discussed in paragraph 3.6. Finally the tests and preparations in order to analyse the data are discussed in paragraph 3.7.
3.2 The Role of Cultural Distance in Foreign Direct Investment Studies
The economic model specified above fits the stream of researches that focus on performance of FDI. I want to establish understanding about how performance is affected when CD differs (positive or negative) and how strong this influence is. The large majority of studies focus on regression models to answer similar performance related research questions. At the same time the general assumption in these studies is that cultures do not change over time and therefore their differences do not either (Reus and Lamont 2009). This leads to a tendency to analyse CD at a certain point in time and not over time. Likewise the purpose is to discover the effect of CD on performance and not over time. Therefore time-series models and panel data models have not been used to answer similar research questions.
Regression models are constructed in several forms by previous researchers, e.g. Li and Guisinger (1991) who investigate business failure rate by single linear regression with larger CD as explanation for the business failure rate increases. Other studies focus on more variables that influence the performance. Reus and Lamont (2009) for instance study the performance of a sample of 118 international acquisitions by multiple linear regression. Besides CD other variables influence the relationship. Both these studies are based on a linear relationship between CD and performance. Morosini et al. (1998) focus on a derivation of this model, the log-linear regression, where the dependent variable that measured performance was the percentage growth in sales. Based on a sample of 52 M&A they found it to increase with larger CD.
14
others in a sense that the dependent variable is binary: it takes the form of 0 or 1. E.g. Kogut and Singh (1988: 423) study a sample of 180 entry modes into the USA, their “dependent variable Pij is the probability that the ith MNE enters trough the jth entry mode”. Among others Luo and Peng (1999) use Manova analysis, next to multiple regression analyse, with more dependent variables because they argue that performance is more than only sales. However the research question under investigation focuses on a specific type of performance and therefore Manova analysis is not applicable.
The current research investigates if differences in CD have an influence on technology creation and what this influence is. Furthermore there is a moderating effect from familiarity with the culture trough previous experience with it and if there was a colonial relationship between the MNEs and affiliates home countries. Cross-sectional research is applicable for this type of research. The dependent variable compares technology creation before and after the M&A. Binary dependent variable models would therefore also be applicable; however it would imply that all specific information about the size of the increase or decrease in technology creation is forgone. To include this information a regression model will be used to answer the research question. In specific the economic model as stated above represents a multiple regression model, in line with among others to Reus and Lamont (2009).
3.3 Measures of Technology Seeking Foreign Direct Investment
The FDI related literature specifies that FDI flows could be subject to bias when used in studies. One factor is that the motive behind the flows cannot be derived with certainty when focussing on secondary data. Therefore when dealing with FDI flows some distinctions between motives are derived by previous research. Makino et al. (2002) specify that there are two motives behind the South-North FDI flows under investigation. First there is the asset exploiting, also referred to as market seeking motive, secondly the asset seeking or technology seeking is argued to be a motive. It is not possible to directly distinguish between these motives in data; nevertheless there are factors on which previous research has made this distinction.
15
frontier of the manufacturing industry. This indicates that these countries are likely targets of technology seeking. Furthermore, next to the more obvious low technological developed countries that are not in the OECD they find that especially South Korea, Spain and the United Kingdom have a low technological development in comparison to other OECD countries. Thirdly Chung and Alcácer (2002) argue that in specific FDI from the high-knowledge manufacturing industries such as pharmaceuticals tents to be technology seeking. This argument is also found in Schmidt-Ehmcke and Zloczysti (2011), whom rank industries based on their R&D efficiency. They find that electrical and optical equipment; machinery; non-metallic mineral products; chemicals; and rubber and plastic products are industries with high R&D efficiency. Fourthly Beugelsdijk et al. (2010) discuss the tax haven motive. MNEs seek to reduce their taxes and therefore invest large sums in countries without any significant industry such as Bermuda. A final factor of distinction is the specification of the type M&A, hence specific types as an initial public offering are not likely to expose the MNE to the new culture and to gain technology. Based on these five factors is a differentiation made between the market seeking, technology seeking FDI and other motives for FDI. However this cannot be guaranteed, therefore conclusions from the research have to be drawn with care.
3.4 Measures of Cultural Distance
The current research focuses on secondary data of the CD between the MNEs and affiliate home countries. Hofstede (2001[1980]) was one of the first researchers to establish this. House et al.i (2004) made further investigations with the Globe initiative. Each of these methods defines dimensions based on an analysis of the cultures. A score is assigned to each of these dimensions and the overall score provides a quantitative summary of the culture under investigation. These scores can mathematically be compared and by that indicate the distance between the cultures, hence CD. Beugelsdijk and Maseland (2011) and Harzing (2003) argue that each of the methods (From this point onwards referred to as: Hofstede & Globe) is applicable and furthermore that there is little research that makes empirical conclusions with regards to the differences. Therefore this research will determine the CD with these methods separately. These results will be used to establish more understanding regarding any inconsistencies between the two methods. Therefore each methods dimensions are defined; a more detailed explanation is given in their respective publications.
16
vs. restraint (2010). More recent Globe (2004) defined the following nine dimensions: “power distance; uncertainty avoidance; humane orientation; institutional collectivism; assertiveness; gender egalitarianism; future orientation; in-group collectivism and performance orientation”.
There are several ways of calculating the CD based on the dimensions specified above. Tihanyi et al. (2005) explain that a limited amount of research focuses on the absolute differences in dimensions. However most research focuses on the construct that is designed by Kogut and Singh (1988: 422) ii. This construct determines the CD between two countries on the squared difference in each dimension for two countries. This is then divided by the variance of the dimension (preferably in the sample, but other studies did focus on the original variance of Kogut and Singh (1988). Secondly it is dived by the number of dimensions that is under investigation. Drogendijk and Slangen (2006) and Reus and Lamont (2009) show that each of the (Hofstede & Globe) can be used in this CD construct. Therefore adjustments, based on the sample, with regards to the variance of each dimension have to be made. Furthermore Reus and Lamont (2009) argue that the practises of each cultural dimension of the method of Globe have to be used and not the values. The current research will therefore follow previous research and determine the CD with the constructs described above.
3.5 Measures of Technology Creation
17
3.6 Sample
The data required for the variables is derived from different sources and therefore discussed per variable. The number of patents applied for originates from a patenting database, Espace, managed by the ministry of economic affairs of the Netherlands. This provides an overview of all patents that where applied for by companies in more than 80 countries. In total this database consists of 70 million patents publications. The cultural dimensions that are required to measure the CD are derived from Anne-Will Harzing, who published both on her websiteiii. The data on the MNEs and the affiliates derives from the database Zephyr, which is a subsection of the Orbit database. This database provides information about M&A and is able to specify on all requirements such as industry, time, geographic dispersion etc. Furthermore the Orbit database provides information about the experience and size of these MNEs and affiliates. GDP growth data has been obtained from the World Bank. Data regarding the host country restrictions has been obtained from the Institutional Profiles Database, designed by the French government in cooperation with the University of Maastricht. All sources mentioned above that provide data that has been used in previous research in comparable variables and similar research questions and therefore are argued to be valid and reliable.
The database lists over 600.000 worldwide cross-border M&A’s between 2000 and 2009. A sample of 321 M&A has been selected based on the following criteria. These criteria derive from section 3.3.
M&A: From all deal types specified by the Zephyr database the following where not selected: IPO; planned IPO; institutional buy-out; MBI/ MBO; management buy-out; demerger; share buyback. These are not applicable because these are most likely not aimed to create technology. Therefore only acquisitions, joint ventures, mergers and minority stake are applicable deal types.
Industry: Only M&A in the following manufacturing industries, as specified by the NACE Rev.2iv, are selected: (19) coke and refined petroleum products; (20) chemicals and chemical products; (21) pharmaceutical products and preparations; (26) computer electronic and optical products; (27) Electrical equipment; (28) machinery and equipment; (29) motor vehicles, trailers and semi-trailers; (30) other transport equipment; (72) scientific research and development. These industries are all medium to high in technological development according to Schmidt-Ehmcke and Zloczysti (2011).
18
investigate the technological development of the countries that are OECD member and Caselli and Coleman (2006) who lists all countries according to their technological development. Based on these two the following applies:
From the MNEs from non-OECD countries all FDI towards OECD members are applicable.
From the MNEs from OECD member countries FDI that originates in low technology development towards high technology developed countries are applicable and FDI towards highly technological developed industries in medium technological developed countries.
All M&As in countries that are characterized as tax haven are excluded.
Because of limited data availability this results in the following home and host countries: From non-OECD members are MNEs originating in Brazil, China, India, Malaysia,
Russia, Taiwan, South Africa selected whom M&A with companies in all OECD member countries.
From the OECD members are MNEs from Mexico, Spain, United Kingdom selected who M&A with companies in Belgium, Denmark, Germany, the Netherlands and the United States. Furthermore M&A with companies in the machinery industry in Italy are also applicable for these MNEs because it is argued to be a high technological developed industry (Schmidt-Ehmcke and Zloczysti (2011).
19
3.7 Data Preparation
First I conducted diagnostic checks to control the applicability of the sample. These checks controlled for normality of the dependent variable, heterogeneity, multicollinearity and covariance. This resulted in multiple issues related to the data, especially because there was a large zero bias in the dependent variable. Transformation of the dependent variable to a dummy representing growth or no growth in technology creation and analysis with a binary dependent variable model did not improve the results. Therefore transformations of the data had to be made. An option is to transform the dependent variable with its natural logarithm and thereby creating a growth model. Morosini et al. (1998) argue that this is a common method; especially in strategic management research and research that focuses on post acquisitions performance of M&A. In specific their argument holds that “a cross-border acquisitions might be interpreted as a mechanism for the acquiring (or the target) firm to access different routines and repertoires that are missing in its own national culture” (Morosini et al. 1998: 141). If this interpretation is followed Morosini et al. (1998) argue it should increase the competitive advantage of the MNE and thereby its performance, thereby validating the analysis of only the growth. This is in line with Jemison and Sitkin (1986). Furthermore Morosini et al. (1998) argue that to measure growth is appropriate for “such process-based phenomena as M&A” (1998: 143). Based on these arguments I continue the research with a growth model. This implies that only those M&A that had a positive effect on technology creation are under investigation, which reduced the sample size (N) to 116 M&A. The descriptive of the adjusted sample are presented in figure 2. Furthermore figure 7, (appendix section 3, page 37) summarizes the home and host countries of the new sample. Because I now focus solely on those M&A that had a positive effect on technology creation certainty with which I can conclude that the FDI had a technology seeking motive increases. This increases the reliability and validity of the research.
Before proceeding to the inference, in section 4, the results of similar diagnostic checks, as described above, are discussed. Figure 6 (appendix section 3, page 35-36) presents the correlations coefficients for the variables. This shows that the correlations are well below the 0.80 threshold level of multicollinearity (among others de Jong et al. 2010). Therefore I conclude that this overall level of correlation does not indicate that multicollinearity is an issue in this research.
20
statistic, which confirmed the suspicions (JB statistic = 11.951; P < .05). However the skewness and kurtosis of the transformed dependent variable are close to the desired level (Skewness = .78; kurtosis = 2.78). Furthermore the dependent variable has already been transformed with its natural logarithm and the sample size is large enough (N - number of variables > 50) according to Hill et al. (2008) Therefore I conclude that the normality assumption of the transformed dependent variable is not violated.
Thirdly the linearity of the relationship was tested. A plot of de independent variable CD of both Hofstede & Globe against the transformed dependent variable (figure 9 in appendix, page 38, indicated that this was the case. Fourthly the dependent variables are controlled for heteroskedasticity. This test was done for both the Hofstede and Globe analysis. The Whites test failed to reject the null hypothesis (Hofstede: chi2 = 31.12; P > .118) (Globe chi2 = 29.25; P > .172) of homoskedastic distribution. Figure 10 in appendix section 3, page 38, shows a plot of the distribution of the main CD variables.
Figure 2: Descriptive Statistics
Variable Descriptive Observations Mean SD Min Max
1. Difference in technology creation LN of differences in patents 116 1.642 1.556 0 6.430
2. CD measure of Hofstede Kogut and Singh measure of Hofstede 6 dimensions 116 2.163 1.345 0.094 6.742
3. CD measure of Globe Kogut and Singh measure of Globe 9 dimensions 116 1.429 .706 0.048 4.457
4. Colonial history Binary variable, 1 = colonial history 89 = 0 / 27 = 1 .233 .424 0 1
5. Previous country experience Binary variable 1 = previous country experience 89 =0 / 27 = 1 .233 .424 0 1
6. Home country GDP growth Growth of GDP in MNEs home country in year of M&A, in % 116 5.707 3.590 -8.000 14
7. Host country restrictions Continuous variable 1 = low restrictions 4 = large restrictions 116 2.907 .513 2.480 4
MNE size binary variables: Four binary variables for MNE size: 116
8. MNE size small 1 = small 7 0.060 0.239 0 1
9. MNE size medium 1 = medium 5 0.043 0.204 0 1
10. MNE size large 1 = large 17 0.147 0.355 0 1
11. MNE size very Large 1 = very large 87 0.750 0.435 0 1
M&A year binary variables: Ten binary variables for year of M&A: 116
12. 2000 1 = 2000 3 0.026 0.159 0 1 13. 2001 1 = 2001 5 0.043 0.204 0 1 14. 2002 1 = 2002 7 0.060 0.239 0 1 15. 2003 1 = 2003 18 0.155 0.364 0 1 16. 2004 1 = 2004 13 0.112 0.317 0 1 17. 2005 1 = 2005 12 0.103 0.306 0 1 18. 2006 1 = 2006 13 0.112 0.317 0 1 19. 2007 1 = 2007 18 0.155 0.364 0 1 20. 2008 1 = 2008 22 0.190 0.394 0 1 21. 2009 1 = 2009 5 0.043 0.204 0 1
MNE industry binary variables: Nine binary variables for industry: 116
22. chemicals and chemical products 1 = chemicals and chemical products 18 0.155 0.364 0 1
23. coke and refined petroleum products 1 = coke and refined petroleum products 5 0.043 0.204 0 1
24. computer electronic and optical product 1 = computer electronic and optical product 23 0.198 0.400 0 1
25. electrical equipment 1 = electrical equipment 7 0.060 0.239 0 1
26. machinery and equipment 1 = machinery and equipment 21 0.181 0.387 0 1
27. motor vehicles, trailers and semi-trailers 1 = motor vehicles, trailers and semi-trailers 11 0.095 0.294 0 1
28. other transport 1 = other transport 2 0.017 0.131 0 1
29. pharmaceutical products and preparation 1 = pharmaceutical products and preparation 23 0.198 0.400 0 1
22
SECTION 4: RESULTS 4.1 Introduction
This section introduces the statistical analysis that has been applied as specified in section three. First paragraph 4.2 describes the results of the multiple regression model. Secondly paragraph 4.3 discusses the results and paragraph 4.4 finishes this section with the applied robustness checks. Based on this conclusions are drawn in the following section 5. 4.2 Statistical Results
The results of the multiple regression analysis are presented in figure 3. The figure presents five different columns, were model 1 describes only the control variables and country dummies. The model 2 includes the CD, colonial and previous country experience of the MNE variables. Model 3 also includes the interaction of the CD and colonial and CD and previous country experience. For the two CD methods a separate regression is conducted, which is specified above the model. Each model column presents the regression coefficients, next to it the coefficients standard errors are presented. To prevent for the dummy variable trap the following dummy variables are excluded from the regression: Size: medium, Year: 2003, Industry: other transport. Their coefficients are the base category, if all others are 0, which is captured by the constant.
23
Figure 3: Regression results: The Impact on MNE Technology Creation
Hofstede Globe
Model 1 SE Model 2 SE Model 3 SE Model 2 SE Model 3 SE
CD .107 .145 .153 .1510 .251 -.020 .097 .292
Colonial -.369 .413 .583 .815 -.413 .412 .280 .884
Previous country experience -.284 .447 -.973 1.279 -.283 .456 -2.377 3.876
CD*Colonial Relation .406 .762 1.399 2.518
CD*Previous country experience -.483 .357 -.475 .529
Growth GDP -.063 .058 -.089 .061 -.085 .062 -.084 .061 -.080 .062
Restrictions -.122 .341 .007 .395 -.0591 .425 -.046 .390 -.046 .393
Size of MNE: small 1.997 .994 1.952 1.008 2.038 1.009 2.056 1.032 2.085 1.039
Size of MNE: large .909 .870 .904 .888 .928 .893 1.012 .900 1.009 .905
Size of MNE: very large 1.033 .798 1.139 .829 1.181 .832 1.247 .824 1.208 .829
Year of M&A: 2000 -.200 1.069 .181 1.113 .1789 1.121 .000 1.089 -.048 1.100 Year of M&A: 2001 1.204 .850 1.373 .872 1.464 .920 1.242 .856 1.002 .906 Year of M&A: 2002 -.772 .726 -.580 .749 -.703 .763 -.693 .738 -.824 .757 Year of M&A: 2004 .016 .624 .309 .676 .253 .699 .144 .642 -.050 .676 Year of M&A: 2005 -.693 .631 -.450 .671 -.432 .677 -.606 .640 -.733 .657 Year of M&A: 2006 -.106 .618 .172 .662 .166 .672 .015 .629 -.143 .652 Year of M&A: 2007 .306 .633 .546 .671 .648 .681 .447 .671 .463 .680 Year of M&A: 2008 -.412 .526 -.305 .563 -.264 .566 -.439 .537 -.521 .552 Year of M&A: 2009 .249 .893 .044 .920 .123 .923 .133 .950 -.018 .965
Industry: chemicals and chemical products -.311 1.227 -.498 1.238 -.472 1.239 -.490 1.242 -.583 1.252
Industry: coke and refined petroleum products -.286 1.368 -.632 1.396 -.621 1.397 -.512 1.401 -.299 1.426
Industry: computer electronic and optical product .018 1.238 -.167 1.255 -.1569 1.258 -.069 1.253 -.118 1.261
Industry: electrical equipment -.722 1.346 -.828 1.353 -.737 1.355 -.809 1.365 -.871 1.373
Industry: machinery and equipment -.523 1.217 -.612 1.229 -.608 1.232 -.577 1.234 -.703 1.248
Industry: motor vehicles, trailers and semi-trailers .072 1.276 -.130 1.290 -.239 1.292 -.078 1.292 -.175 1.306
Industry: pharmaceutical products and preparation .012 1.207 -.072 1.213 -.0630 1.213 -.065 1.216 -.160 1.226
Industry: scientific research and development -1.404 1.375 -1.466 1.385 -1.403 1.386 -1.521 1.392 -1.616 1.402
24
4.3 Discussion
This study investigates the effect of CD between home and host country on technology creation of MNEs. The proposed hypotheses derived from the literature are not supported, because all results are insignificant. This, however, is a finding in itself.
The first hypothesis states that CD has a negative effect on technology creation. This research fails to find significant effects and therefore H1a is rejected. In fact the sings of both CD measures are positive. However this is not significant at the 90% level (P > .10), therefore H1b is not accepted. This leads me to the conclusion that the effect of CD on technology creation cannot be determined in this research; nevertheless this is evidence that the CD does not have a negative effect on the performance of technology creation. Secondly the Hypothesis 2 and 3 stated that there is a moderating effect of familiarity. Familiarity of the MNE with the host country was measured in two ways familiarity trough previous experience and familiarity trough colonial history. Model 3 includes the interaction of CD and the two indicators of familiarity. These variables however also do not yield any significant result, which leads me to reject hypothesis 2 and 3. Therefore I can conclude that there is not an effect of familiarity with the host countries culture and technology creation.
25
argument. Thereby explaining the contradictory results found in this research and in the literature of among others Morosini et al. (1998) and Reus and Lamont (2009).
Finally Regarding the measures of CD designed by Hofstede and Globe a limited number of conclusions can be drawn due to the insignificant results of both. When focussing at their correlation it is observed that there is a limited overlap between the two measures (r = .22). This leads me to conclude that these measures, although both designed to measure natural culture, do not measure the same. Furthermore looking at the correlation of each measure with the dependent variable it is observed that, although both are low, especially the Globe measure is very small (r = .04), whereas Hofstede had a correlation of r = 15. Furthermore the R-square, which measures the influence of the model on the technology creation in this case, is lower with the Globe measure than with the Hofstede measure, all other factors hold constant. Finally the adjusted R-square, capable of detecting both positive and negative effects of additional variables, is lower with the Globe measure than with Hofstede measure. It also decreases with the inclusion of the interaction effect whereas the adjusted R-square of the Hofstede model 2 and 3 is constant. Based on these results I can conclude that the Hofstede measure is slightly favoured in this investigation over the Globe measure. However due to the lack of significant results this conclusion is taken with care and strong statements are not given.
4.4 Robustness of the Model
A number of additional test where conducted to determine the robustness of the model and to determine if significance of the model could be discovered. Appendix section 4 (figure 11, page 39) presents the descriptive of the additional variables that were used in the robustness checks.
The first method of controlling for the robustness of the model is by introducing two measures to determine the CD, one based on the Hofstede dimensions and one on those of Globe. Figure 3 shows the results. The model is and remains insignificant and furthermore the sings of the CD measures are positive. Next to this robustness check a transformation of the main variable CD is tested. In line with Drogendijk and Slangen (2006) the Euclidian distance is tested, the results are to be found in appendix section 4 (figure 13, page 41). The transformation does not increase the significance level of the model, nor does it have an influence on the sign of the main cultural and familiarity with the host culture variables.
26
effects dummies are tested. The multiple regression model that controls for the effect of the region is to be found in appendix section 4 (figure 12, page 40). In this model Russia is incorporated in the Europe dummy because its capital Moscow lies in Europe, even though most of the land mass lies in Asia. The introduction of regional dummy variables does not change the regression results: the model remains insignificant. Furthermore the sings of the variables do not massively change: only the CD measure of Globe becomes negative in the model 2. When country fixed effects are included in the model the regression results change. In figure 4 the results are presented of an analysis that excluded the year and industry control dummy variables and included the country fixed effects. In figure 5 the results are presented of the analysis that included industry, year variables and country fixed effects. Especially the first regression in figure 4 shows different results. The F-statistic is significant at the 90% level with the Hofstede CD measure. In the second regression in figure 5 is the effect of some single country variables significant, however the model remains insignificant. It is likely that these slightly more significant results are the result of over controlling. This is the case when two variables measure similar things. Because the host countries culture are close in distant including fixed home country effects would lead to a similar measure of CD and thereby reduce the effect of CD in the model. Which is also observed, especially in the first regression the coefficients of the host countries are much larger than that of the CD measures. Therefore the home country effects were not included in the main research.
27
Figure 4: Regression Results Country Fixed Effects Model
Fixed Effect Model Hofstede Globe
Model 1 SE Model 2 SE Model 3 SE Model 2 SE Model 3 SE
CD .099 .146 .135 .147 .014 .283 .155 .312
Colonial -.196 .477 -.890 1.671 -.138 .489 -.398 5.751
Previous country experience -.337 .400 .924 .793 -.327 .401 .566 .885
CD*Colonial Relation .509 1.016 .269 3.711
CD*Previous country experience -.639* .349 -.613 .538
Growth GDP -.009 .063 .001 .064 .013 .064 -.001 .064 .027 .069
Restrictions .422 .315 .528 .384 .329 .427 .446 .366 .361 .374
Size of MNE: small 1.189 .931 1.255 .941 1.389 .938 1.215 .966 1.315 .973
Size of MNE: large .624 .789 .737 .812 .701 .807 .728 .842 .780 .848
Size of MNE: very large .933 .721 1.092 .758 1.107 .753 1.102 .773 1.118 .777
MNE home country: Brazil 2.307* 1.173 2.511** 1.241 2.928** 1.405 2.278* 1.195 2.320 1.609
MNE home country: China 1.881* 1.055 1.781 1.080 2.072 1.294 1.764 1.106 1.499 1.542
MNE home country: Spain 1.671 1.071 1.640 1.095 1.975 1.292 1.608 1.097 1.661 1.542
MNE home country: United Kingdom 2.016** .991 2.526** 1.115 2.609* 1.362 2.252** 1.043 2.043 1.489
MNE home country: India 1.546* .922 1.720* .963 2.046* 1.148 1.552 .958 1.390 1.439
MNE home country: Korea 2.905*** 1.024 2.934*** 1.045 3.401*** 1.248 2.893*** 1.099 2.778 1.505
MNE home country: Russia 2.290* 1.175 2.461 1.215 3.249** 1.422 2.398 1.387 2.790 1.758
MNE home country: Taiwan 3.425*** 1.070 3.443*** 1.087 3.830*** 1.261 3.416*** 1.090 3.412** 1.525
MNE home country: South-Africa .103 1.223 .479 1.304 1.056 2.019 .204 1.247 .188 2.152
28
Figure 5: Regression Results Country Fixed Effects, Industry and Year Model
Fixed Effect Model Hofstede Globe
Model 1 SE Model 2 SE Model 3 SE Model 2 SE Model 3 SE
CD -0.063 0.175 -0.015 0.175 0.310 -0.380 -0.006 0.342
Colonial -0.102 0.524 -0.894 1.928 -0.098 0.524 -2.899 6.860
Previous country experience -0.337 0.430 1.034 0.841 -0.332 0.430 0.628 0.985
CD*Colonial Relation 0.548 1.180 1.909 4.413
CD*Previous country experience -0.693 0.367 -0.655 0.596
Growth GDP 0.121 0.150 0.032 0.124 0.051 0.123 0.025 0.124 0.061 0.129
Restrictions 0.413 0.361 0.423 0.447 0.239 0.499 0.466 0.427 0.386 0.436
Size of MNE: small 1.174 1.016 1.266 1.036 1.416 1.033 1.362 1.063 1.500 1.076
Size of MNE: large 0.520 0.872 0.650 0.901 0.665 0.892 0.740 0.929 0.748 0.932
Size of MNE: very large 0.950 0.793 1.164 0.840 1.249 0.833 1.204 0.849 1.213 0.856
Year of M&A: 2000 -0.418 1.080 -0.528 1.149 -0.574 1.154 -0.413 1.098 -0.434 1.104 Year of M&A: 2001 1.692 0.855 1.645* 0.870 1.780 0.908 1.673* 0.869 1.533* 0.901 Year of M&A: 2002 -0.378 0.721 -0.428 0.743 -0.532 0.778 -0.429 0.743 -0.542 0.809 Year of M&A: 2004 0.082 0.635 0.039 0.668 -0.040 0.697 0.075 0.653 -0.059 0.678 Year of M&A: 2005 -0.710 0.657 -0.795 0.698 -0.789 0.718 -0.709 0.667 -0.783 0.677 Year of M&A: 2006 0.081 0.631 0.028 0.667 -0.029 0.690 0.100 0.640 -0.063 0.661 Year of M&A: 2007 -0.354 0.698 -0.374 0.743 -0.396 0.755 -0.270 0.718 -0.302 0.722 Year of M&A: 2008 -0.343 0.624 -0.377 0.659 -0.386 0.665 -0.325 0.635 -0.282 0.651 Year of M&A: 2009 -0.325 1.123 -0.213 1.151 -0.231 1.140 -0.202 1.152 -0.265 1.157
Industry: chemicals and chemical products 0.138 1.187 0.087 1.224 0.149 1.214 0.057 1.215 -0.047 1.223
Industry: coke and refined petroleum products -0.083 1.351 -0.066 1.389 -0.104 1.379 -0.054 1.392 0.117 1.404
Industry: computer electronic and optical product 0.310 1.193 0.397 1.230 0.415 1.223 0.343 1.212 0.263 1.222
Industry: electrical equipment 0.162 1.340 0.208 1.381 0.520 1.378 0.141 1.361 -0.011 1.382
Industry: machinery and equipment -0.230 1.179 -0.205 1.214 -0.159 1.202 -0.262 1.205 -0.458 1.221
Industry: motor vehicles, trailers and semi-trailers 0.625 1.221 0.599 1.263 0.562 1.252 0.556 1.249 0.442 1.260
Industry: pharmaceutical products and preparation 0.905 1.204 0.901 1.234 0.999 1.223 0.867 1.224 0.803 1.230
Industry: scientific research and development -0.650 1.338 -0.733 1.359 -0.571 1.348 -0.722 1.358 -0.861 1.368
MNE home country: Brazil 2.586* 1.332 2.503* 1.371 2.991* 1.545 2.523* 1.362 2.066 1.877
MNE home country: China 2.213* 1.304 2.100 1.337 2.435 1.554 2.155 1.353 1.406 1.862
MNE home country: Spain 2.072 1.264 2.115 1.321 2.496 1.505 2.023 1.299 1.576 1.884
MNE home country: United Kingdom 2.230* 1.178 2.428* 1.264 2.588* 1.500 2.516* 1.242 1.908 1.734
MNE home country: India 1.412 1.079 1.387 1.103 1.765 1.307 1.499 1.113 0.830 1.698
MNE home country: Korea 3.152*** 1.141 3.191*** 1.175 3.782*** 1.419 3.266*** 1.211 2.693 1.773
MNE home country: Russia 2.990** 1.435 3.157** 1.475 4.079** 1.688 3.328** 1.582 3.324 2.049
MNE home country: Taiwan 3.546*** 1.153 3.539*** 1.175 3.936*** 1.383 3.525*** 1.173 3.064* 1.773
MNE home country: South-Africa -0.126 1.363 -0.131 1.433 0.511 2.278 0.032 1.391 -0.634 2.487
29
SECTION 5: CONCLUSIONS 5.1 Introduction
This section provides the conclusions that can be drawn with respect to this research. The aim of which is to determine if a MNE that internationalized in search of technology is able to derive non-negative effects of the cultural differences it is faced with. Next to this limitations and future expansions are discussed.
5.2 The Effect of CD on Technology Creation
This research provides empirical support to the small but growing stream of researchers that claims there are non-negative effects of CD for the MNE. Were CD implies the differences in cultures (Tihanyi et al. 2005) also referred to as the differences between countries norms, routines and repertoires (Morosini et al. 1998). This result is derived from a sample of MNEs that originate in four continents, from home countries with different levels of development. By focussing on those MNEs that were able to create technology after their M&A I am able to reject the mainstream hypotheses that CD has a negative effect on performance. This leads the conclusion that larger differences in norms routines and repertoires, thus between cultures, do not have to hinder the ability of the MNE to increase its knowledge and competences and with that increase its technology. Furthermore the insignificant results are robust over two CD measures, which add to this conclusion.
The idea that differences in cultures do not hinder the MNE when internationalizing is in line with Shenkar (2001). Moreover this research adds to the argument of Shenkar et al. (2008) who argue that there are not only positive or negative effects of CD. They argue that there are factors such as the amount of interaction, asymmetry between the countries and power of both parties that create frictions between the MNE and affiliate (Shenkar et al. 2008). However there is an equal assumption in the current CD measure, all MNEs face the same effects of CD and thereby hinder the MNE in a similar way. The rejection of the negative hypothesis in this research is proof that this is not the case. Because the MNE and M&A are different the friction differs and therefore the effect on the MNEs ability to benefit FDI and foreign activity differs. Hence there is no one size fits all solutions to the cultural differences theory in FDI.
30
cultural similar countries. By limiting the distance between the home and host countries culture the negative effect of CD would be lower. However, especially in the case of technology seeking FDI, this is not always the best solution. Technology seeking FDI literature states that there are geographic boundaries of technology and furthermore centres of excellence were an agglomeration of knowledge and skills is found. These are solely found in the developed world. Therefore the MNE, especially those from developing countries, are not able to seek for technology in cultural close countries. Moreover the learning by diversity argument argues that distant cultures are attractive for a MNE to increase its knowledge and competences and thereby create technology. The rejection of the negative CD literature assumption in this research leads to the managerial implication that there is empirical prove to favour the arguments in the technology seeking FDI literature that opt for cultural distant locations over those in the CD literature that opt for cultural close locations.
5.3 Limitations and Future Expansions
A first limitation of this research is that all conclusions derive from insignificant results. This leads to a strong conclusion with regards to the rejection of the negativity hypothesis. However with the respect to all other conclusions this limits the ability to draw conclusions. The insignificance could be solved in future research with more diverse and larger sample. If more data becomes available it is likely that significant results of each variable are found and thereby enabling the research to draw more conclusions. Future expansions of this theory could therefore also focus on primary data, and thereby have a larger guarantee of the reliability of the technology seeking motive conclusions. Furthermore in time it could become possible to assign the motive of FDI more reliable, either through sales data of the MNE or trough government regulations.
31
REFERENCES
Abrahamson, E., & Fombrun, C.J. (1994) Macro cultures: determinants and consequences. Academy of Management
Review, 19: 728-755.
Acemoglu, D., Johnson, S., & Robinson, J. 2001. The colonial origins of comparative development: an empirical
investigation. American Economic
Review, 91: 1369–1401.
Almeida, P. 1996. Knowledge sourcing by foreign multinationals: patent citation analysis in the U.S. semiconductor industry. Strategic Management Journal, 17(winter special issue): 155-165.
Argyres, N. 1996a. Capabilities,
technological diversification and
divisionalization. Strategic Management
Journal, 17: 395-410.
Au, K. Y. 2000. Inter-cultural variation as another construct of international management: a study based on secondary data of 42 countries. Journal
of International Management, 6: 217-238.
Barkema, H.G., Shenkar, 0.,
Vermeulen, F., & Bell, J.H.J. 1997. Working abroad, working with others: how firms learn to operate international joint
ventures. Academy of Management
Journal, 40: 426-442.
Barkema, H. G., & Vermeulen, F. 1998. International expansion through start-up or acquisition: a learning perspective. Academy of Management
Journal, 41(1): 7–26.
Beugelsdijk, S., Hennart, J-F., Slangen, A., & Smeets, R. 2010. Why and how FDI stocks are a biased measure of MNE affiliate activity. Journal of International Business Studies, 41(9):
1444-1459.
Beugelsdijk, S., & Maseland, R. 2011. Culture in Economics, History
Methodological Reflections and Contemporary Applications. New York:
Cambridge University press.
Beugelsdijk, S., & Slangen, A. 2010. The impact of national cultural distance on the number of foreign website
visits by U.S. households.
Cyberpsychology, Behavior, and Social Networking, 13(2): 201-205.
Cantwell, J.A., & Janne, O.E.M. 1999. Technological globalization and innovation centers: The role of corporate technological leadership and locational hierarchy. Research Policy, 28: 119–144.
Cantwell, J.A., Dunning, J.H., &
Janne, O.E.M. 2004. Towards a
technology-seeking explanation of U.S. direct investment in the United Kingdom.
Journal of International Management,
10: 5-20.
Caselli, F., & Coleman II, J.W. 2000. The World Technology Frontier. NBER Working Paper 7904.
Chang, S.J. 1995. International expansion strategy of Japanese firms: capability building through sequential entry. Academy of Management, 38(2): 383-407.
Chung, W., & Alcácer, J. 2002. Knowledge seeking and location choice of foreign direct investments in the United States. Management Science, 48(2): 1534-1554.
Drogendijk, R., & Slangen, A. 2006. Hofstede, Schwartz, or managerial perceptions? The effects of different CD measures on establishment mode choices by multinational enterprises. International
Business Review, 15 (4): 361-380.
Dunning, J. H. 1993. Multinational
enterprises and the global economy.
Wokingham, Berkshire: Addison Wesley. Dunning, J.H. 1994. Multinational enterprises and the globalization of innovatory capacity. Research Policy, 23: 67–88.
Dunning, J. H. 1995. Reappraising the eclectic paradigm in the age of alliance capitalism. Journal of International
