Institutional Influences on Cross-National
Parent-Subsidiary Technology Transfer
Faculty of Economics and Business
Rijksuniversiteit Groningen
Master Thesis in Internat. Business & Management
June 2014
Luisa Anna Kitiratschky
Student Number: s2538695
2 ABSTRACT
I study institutional influences on technology transfer, an aspect that is typically neglected in research on this topic. Thus, I combine institutional research with research on technology transfer. As internal knowledge- and technology sharing is increasingly recognized as a key factor for multinational enterprises (MNEs) to gain comparative advantage and increase competitiveness, it is important to understand the considerations behind these processes. Furthermore, MNEs play a major role in the diffusion of technology through transfers from the HQ to subsidiaries located in foreign countries. Especially for developing countries such technology transfer can facilitate growth and development. While other researchers focus on firm specific characteristics that favor technology transfer, I examine which cultural and institutional characteristics of the technology receiving country encourage technology transfer. I expect that HQs are more likely to transfer technology to their subsidiaries in countries with favorable cognitive, regulatory and normative institutions. My results show that especially control of corruption, as part of the regulatory framework, as well as the normative dimension of institutions influence the likelihood of receiving new technology. The suggestions in this work shed light on a new aspect of technology transfer and thus help to understand the motivation of such transfers more thoroughly. Furthermore, I encourage researchers, as well as managers and governments, to consider the influence of national institutions on technology transfer in their studies, firm strategies and policies.
3 TABLE OF CONTENTS LIST OF TABLES 5 INDEX OF APPENDICES 5 LIST OF ABBREVIATIONS 6 1. INTRODUCTION 7 2. THEORETICAL BACKGROUND 9 2.1.THE TRANSFER OF TECHNOLOGY 9 2.2.THE TRANSFER OF TECHNOLOGY WITHIN THE MNE 10 2.3.INFLUENCES ON TECHNOLOGY TRANSFER 11 2.4.INSTITUTIONAL INFLUENCES ON CROSS-NATIONAL PARENT-SUBSIDIARY TECHNOLOGY TRANSFER 13
3. DEVELOPMENT OF HYPOTHESES 15
3.1.THE INFLUENCE OF COGNITIVE INSTITUTIONS ON TECHNOLOGY TRANSFER 15 3.2THE INFLUENCE OF REGULATORY INSTITUTIONS ON TECHNOLOGY TRANSFER 17 3.3.THE INFLUENCE OF NORMATIVE INSTITUTIONS ON TECHNOLOGY TRANSFER 20 3.4THE INFLUENCE OF THE INSTITUTIONAL PROFILE ON TECHNOLOGY TRANSFER 23
4. DATA AND METHOD 25
4.1.SAMPLE 25
4.2.VARIABLES AND MEASURES 26 4.3.METHOD AND GENERIC EMPIRICAL MODEL 31
4.3.1.EMPIRICAL MODEL 31 4.3.2.STATISTICAL TECHNIQUE 32 5. RESULTS 36 5.1.BASELINE RESULTS 36 5.2.ROBUSTNESS CHECKS 38 6. DISCUSSION 43
5 LIST OF TABLES
Table 1. Summary of Hypotheses... 24
Table 2. Country Institutional Profiles ... 32
Table 3. Mean Values & Standard Deviations ... 34
Table 4. Institutional Influences on Parent-Subsidiary Technology Transfer ... 37
Table 5. Robustness Checks decomposing the regulatory variable, using other data sources, sample sizes and control variables ... 40
Table 6. Summary of Results ... 42
INDEX OF APPENDICES Appendix A – Descriptive Statistics ... 57
Table A1. New Technology Distribution Among Firm Size ... 57
Table A2. New Technology Distribution Among Industries ... 57
Table A3. New Technology Distribution Among Countries ... 58
Appendix B – Multicollinearity ... 59
Table A4. Correlation Matrix ... 60
Table A5. VIF Test for Cognitive Dimension ... 61
Table A6. VIF Test for Cognitive Dimension with Control Variables ... 61
Table A7. VIF Test for Regulatory Dimension ... 61
Table A8. VIF Test for Regulatory Dimension with Control Variables ... 61
Table A9. VIF Test for Normative Dimension ...62
Table A10. VIF Test for Normative Dimension with Control Variables ...62
Appendix C – Control of Corruption Index ... 63
6 LIST OF ABBREVIATIONS
BEEPS Business Environment and Enterprise Performance Survey e.g. Exempli gratia (for example)
et al. Et alii
FDI Foreign direct investment
GLOBE Global Leadership and Organization Behavior Effectiveness
HQ Headquarter
INSEAD The Business School for the World
IPCC Intergovernmental Panel on Climate Change IPR Intellectual property right
Isl. Island
MNE Multinational enterprise
N Number of observations
NAFTA North American Free Trade Agreement
Nr. Number
OECD Organization for Economic Cooperation and Development
p. page
R&D Research & Development
Rep. Republic
S.E. Standard Error
SME Small & medium enterprises
St. Saint
St. Dev. Standard Deviation
U.K. United Kingdom
UN United Nations
UNCTAD United Nations Conference on Trade and Development USA / U.S.A. United States of America
VIF Variation Inflation Factors WDI World Development Index WGI World Government Index
7 1. INTRODUCTION
New technology is vital for a country’s economic growth and development. Anyhow, most new technology gets invented by industrialized, developed countries, not by developing countries who would need high grow rates in order to compete with more developed countries. This is revealed by the Global Innovation Index (Dutta & Lanvin, 2013). Although innovation and new technology are important for the development of industrializing countries, these countries fall behind in technology development (Munir, 2002). Comin & Ferrer (2013) even found out that the adoption patterns for technology account for the majority of divergence of income growth between countries. This implies that not only the mere innovation is of importance, but also importing contemporary technology can be of substantial advantage for a country. Multinational Enterprises (MNEs) play an important role in the diffusion of technology. These international players not only innovate a lot, but also share new technology with their entities in other countries. Technology transfer from Headquarters (HQs) to their subsidiaries into such countries is an important way to receive new technology (Manolopoulos, Papanastassiou & Pearce, 2005; Steensma & Corley, 2000; Heidenreich, 2012).
Not only does knowledge and technology provide advantages to countries, but also to its firms. Innovation and knowledge creation are two major drivers of firm’s comparative advantage. Technology is crucial for the competitiveness, economic growth and even survival of firms. MNE’s internal knowledge- and innovation flow enables them to gain firm-specific advantages. These transfers bear some problems however. One of them stems from the embeddedness of firms. Firms try to fit to their local environment in order to gain legitimacy. In the case of MNEs, subsidiaries and HQ act in different locations and are thus embedded in different environments that shape their firm culture and action. The different national institutions that compose this environment have thus an influence on firms and on the way HQs and subsidiaries interact (Heidenreich, 2012).
8
manner to attract technology. However, scholars have only a limited understanding of why technology transfer rates and success vary cross-nationally.
Extant research predominantly considers firm characteristics to explain success in technology transfer either between HQs and foreign subsidiaries or within international joint ventures (Simonin, 2004; Munir, 2002). However, the institutional environment in which the acquiring firm is embedded in is also very important, and likely to provide additional power to explain differences in technology transfer across countries. In this work I develop the theory that national institutions influence cross-national parent-subsidiary technology transfer. Splitting institutions into three pillars according to Scott (1994) and Kostova (1997), I claim that more favorable cognitive, regulatory and normative institutions in the receiving country enhance technology transfer from the HQ to a subsidiary in that country. I draw on theories and insights from innovation and investment literature, as well as knowledge transfer to combine the two strands of institutional theory and technology transfer theory. My aim is to explain cross-national differences in the transfer of technology, thereby suggesting a new channel through which culture and institutions can affect economic performance of countries and firms.
Logistic regression analysis is used to test these claims, taking firm and country level data for over 150 cases in 10 countries. Data is taken from several sources of the World Bank (BEEPS, 2005; WGI, 2005; WDI, 2005), Barro & Lee (2013; data for the year 2005) and Hofstede (1980). Statistics are significant in the expected direction for only one part of the regulatory institutions. Control of corruption statistically significantly influences parent-subsidiary technology transfer. The normative dimension also effects technology transfer, but this effect is not unambiguous throughout the empirical tests. Thus, future research is needed to legitimate the theory about institutional influences on cross-national inter-firm technology transfer.
9 2. THEORETICAL BACKGROUND
This chapter summarizes the positions taken by other researchers concerning the topic of technology transfer. I start by defining technology and broadly explaining the issue of technology transfer, then I narrow down the topic reaching in the end my specific field of research, the influence of institutions on cross-country parent-subsidiary technology transfer. Thereby, I point out that inter-firm technology transfer is an important technology source for countries but that the adaption of the technology to the host country environment causes problems due to cultural differences. These differences are generally neglected by researchers of the field, who hold firm-, technology- and relationship specific factors responsible for the success or failure of technology transfers. These theories are discussed in the following as well. Thereafter, I explain why it is important to consider the institutional context of the transfer.
2.1. The Transfer of Technology
Technology is an important aspect to take a look at, because technology co-determines the transaction costs, as well as the transformation costs of a firm. Technology is employed in order to make production of goods and services more efficient. Unique technology or technology that is embedded in a company in a way that it enables unique capabilities can lead to firm-specific competitive advantage (North, 1994; Haug, 1992; Dunning & Lundan, 2008). Technology is a broad term, defined by researchers in different ways. The United Nations Conference on Trade and Development (UNCTAD) draft International Code on the Transfer of Technology defines technology as “the systematic knowledge for the manufacture of a product, for the application of a process or for the rendering of a service, including any integrally associated managerial and marketing techniques” (Roffe, 1985, chapter 1, para.1.2.). This definition only includes knowledge, but not its embodiment, meaning “machinery and devices developed from scientific knowledge” (Oxford University Press, 2014). This work however considers both aspects. It also follows Haug’s (1992) specification that technology only includes the tangible and intangible assets that “contribute to the economic, industrial, or cultural development of a country” (p. 211). Technology is very complex to transfer, because most components are intangible and tacit, thus not clearly outspoken. Simply buying technology is not sufficient, but technology needs to be adopted and absorbed properly (Alexander, 2012).
10
private sector entities, financial institutions, NGOs and research/education institutions (Metz & Turkson, 2000; Blakeney, 1989). Here, the focus is on transfer between private sector entities, specifically between HQs and subsidiaries. This focus restricts the influence of governments on the transfer, but does not erase it totally, as will be explained later.
2.2. The Transfer of Technology within the MNE
In a MNE, the HQ and its subsidiaries should follow the same organizational goal. The subsidiaries hereby mostly represent the firm in other countries and are often responsible for the operations in the emerging markets. Their HQs support them to reach the organizational goal by providing resources to the subsidiaries. One of these resources committed by HQs is technology (Kaufmann & Roessing, 2005). Firm-specific technology may facilitate competitive advantage and the penetration of foreign markets, because it can lead to the development of unique capabilities within the MNE. Thus, MNEs try to accumulate knowledge internally to develop unique organizational competencies and exploit technological asset interdependencies. A study conducted by Manolopoulos et al. in 2005 shows that MNE internal sources are the main technology providers to foreign subsidiaries. In other words, cross-national parent-subsidiary technology transfer is the main channel through which subsidiaries receive state-of-the-art technology. They also highlight that subsidiaries are not pure technology receivers anymore, but also play an important role in technology development nowadays.
Technology transfer to subsidiaries in less developed countries enhances the productivity growth of the receiving country itself if a certain human capital threshold is reached that is needed to absorb the new technology. Countries below this threshold cannot exploit technology spillover effects, meaning the positive effects of the transferred technological knowhow on productivity and innovative ability of the subsidiary (Xu, 2000). Other benefits of technology transfer are the generation of exports and foreign exchange, tax revenues and employment, as well as accumulated capital and entrepreneurship skills at the subsidiary (Reddy & Zhao, 1990). These benefits can only be realized if the transferred technology is adapted properly.
11 have trouble in dealing with the technology received or it might simply not fit the local market or resources (Reddy & Zhao, 1990).
2.3. Influences on Technology Transfer
Researchers found different influences on technology transfer, mostly concerning firm-, technology- or relationship specific characteristics. These lines of research found convincing evidence on the proposed relationships, but still don’t picture the whole process when they neglect the (institutional) environment in which the transfer is embedded. As these microeconomic-focused theories have extensive explanation power anyhow, they are examined in the following paragraphs in order to understand the context of extant research in the technology transfer field.
One of the firm characteristics widely studied is the ability of the acquiring subsidiary to absorb the offered technology successfully in their own organization. There are several theories trying to explain differences here. One of them is the learning theory that argues that firms learn faster when they already have existent knowledge related to the field. It assumes, that the acquisition of knowledge is a cumulative process. New knowledge can be used effectively, when prior knowledge exists on which the new knowledge can be built on. For technology this means that prior knowledge about related technology enables successful valuation and application of the received technology (Hitt, Dacin, Levitas, Arregle & Borza, 2000; Cohen & Levinthal, 1990; Simonin, 1999; Inkpen, 2000). Furthermore, learning capacity and the intention to learn also play a role in absorptive capacity. Learning intent describes the desire or willingness of an organization to learn from another organization; in this case, the willingness of the subsidiary to learn technological knowhow from its HQ. Some technology might be imposed on subsidiaries from the HQ without mutual agreement, so subsidiaries do not have real intentions to absorb it. Learning capacity is restricted by bounded rationality, in this case meaning the limited capacity of the firm to obtain, store, process and share information properly (Simonin, 2004; Szulanski, 1996).
12
Other lines of research focus on the relationship between the two entities involved. The more intensive the relationship and especially the more trusted the relationship, the better technology transfer works. When relational ties between the two entities are strong, technology is transferred more willingly and the transfer occurs frictionless, because of frequent and effective communication (Wahab, Rose & Osman, 2011; Minbaeva, 2007; Szulanski, 1996).
Minbaeva (2007) already criticized that most researchers focus on only one of the above described dimensions. Researchers either examine characteristics of the knowledge, characteristics of the sender and receiver respectively, or characteristics of the relationship between the two entities. Minbaeva (2007) combines all four dimensions and empirically tests their joint impact. But still, even Minbaeva (2007) does not cover all aspects of a transfer. Only examining differences on firm and relationship level is a too narrow viewpoint and neglects the context in which such transfers occur. These researchers mostly focus on the quality of the transfer and do not ask what hinders HQs to transfer technology to a distinct subsidiary in the first place. The willingness to transfer technology at all depends also on the environment of the receiver. This environment consists of country specificities such as physical and legal infrastructure, as well as national culture and institutions.
Few researchers focus on institutional differences as a possible explanatory variable for the ability of firms to absorb technology. Most research in this area in turn focuses on the view of the entity providing technology and thus deals with problems such as property right protection (Branstetter, Fisman & Foley, 2006; Kaufmann & Roessing, 2005). Papers focusing on the recipient side mostly try to find out which type of technology should be acquired (Minbaeva, 2007; Zander & Kogut, 1995). Some researchers claim that local research and development (R&D) activity has an influence on the quality of transferred technology, meaning that state-of-the-art technology is only transferred to countries with own R&D facilities (Glass & Saggi, 1998; Maskus, 2004; Reddy & Zhao, 1990).
13 2.4. Institutional Influences on Cross-National Parent-Subsidiary Technology Transfer
The institutions of a country affect technology transfer in a number of ways. First, because even large MNEs conduct their R&D mainly in their home country. As a consequence, the resulting innovations and technology are strongly influenced by the home country environment, its culture and institutions. The local quality of basic research, workforce skills and private as well as public investment play a large role. Government policies, the educational system, technical and scientific institutions and cultural traditions in turn affect these characteristics. Technology is thus invented with this background and is constructed to work well under the particular national - or sometimes even local - conditions. This means that the technology might not be effective anymore under other conditions (Carlsson, 2006; Pavitt & Patel, 1999). Second, the most important innovative contributions are “person-embodied and institution-embodied tacit knowledge” (Pavitt & Patel, 1999, p. 103). This makes the transfer even more difficult, as persons have to transfer this knowledge face-to-face and because institutions differ between countries and cannot be transferred along with the technology. Carlsson (2006) also points out that national institutions can create a country-specific technological advantage.
Institutions can be defined as social structures that enable meaningful social interaction – or the rules of the game in a society (North, 1994). Scott (1994) structured institutions into three pillars: A regulatory, normative and cognitive pillar. Regulatory institutions are the rules and laws of a national environment. The cognitive dimension is the social knowledge shared by the people in a country; structures that shape how people select and interpret information. The normative pillar consists of social norms, values, beliefs and assumptions about how to behave that are shared and carried by the individuals of a society. Together, these three pillars construct the “country institutional profile” (Kostova, 1997). This profile “reflects the institutional environment in that country, defined as “the set of all relevant institutions that have been established over time, operate in that country, and get transmitted into organizations through individuals“ (Kostova, 1997, p. 180).
14
15 3. DEVELOPMENT OF HYPOTHESES
Now that I pointed out why it is important to examine the institutional context of cross-national technology transfers within MNEs, I build hypotheses about these institutions. This part of my work explains in detail which types of institutions are favorable for parent-subsidiary technology transfer. As already explained before, the country institutional profile established by Kostova (1997) cannot be used in a holistic manner, but has to be established for each individual situation. Thus, the following sections explain which aspects of the institutions in a country influence technology transfer and suggest reasons for these influences.
3.1. The Influence of Cognitive Institutions on Technology Transfer
The cognitive dimension of the institutional profile consists of the widely shared social knowledge in a country (Kostova, 1997). For technology transfer, it is the knowledge and skills about technology possessed by the people in a country.
Cognitive models are mental models about the structure of the environment. They contribute to goal setting, planning and the attainment of goals. Basically, cognitive models are intrapersonal rules, which estimate conditions and actions. These internal rules influence human reasoning, learning and decision-making, thereby guiding individual behavior. These cognitive models assimilate between people within the same societal group through communication and interaction. Thus, people within a society tend to interpret reality in the same way, what gives rise to collective decision-making. In new and unknown situations, the cognitive models might not fit. In this case the individual has to create a new cognitive theory or may act according to another existing cognitive rule. If he chooses to create a new cognitive theory, learning occurs. Learning on the other hand enables change in society, policy, economy and organizations. Knowledge grows through collective learning and problem solving within societies (Budzinski, 2003; Mantzavinos, North & Shariq, 2004; Rosenbaum, 2001).
16
As explained in the previous section, technology is not adjusted to the receiving country’s needs when transferred from HQ to subsidiary. HQs provide information on product and technology, but do not invest in adjustments to the local environment. Thus, these countries have to develop their own technological capabilities, meaning they have to develop a new cognitive model that tells them how to deal with the new situation. In order to do so, people need a certain level of human capital to be able to successfully adapt new technology (Reddy & Zhao, 1990; Xu, 2000; Tung, 1994).
One success story in this regard is Japan in the period between 1950 and 1970. In that time, Japan imported technology from abroad in the value of $3 billion, but invested five times more in its adaption, precisely in adaptive engineering and R&D activities. Investment in training was an essential part of its spending. People were educated to assimilate technological advances made elsewhere to local demands and on the other hand learned to develop, produce and sell goods and services with worldwide appeal. In consequence, the country not only recovered from the 2nd world war, but also became a leading power in industry and economics (Tung, 1994).
Lasserre (1982) and Manimala & Thomas (2013) among others also highlight the importance of training for the transfer of technology. Mutual training between transferring HQ and receiving subsidiary is a crucial success factor for the transfer. Technology transfer consists to a major part of knowledge transfer – and for understanding this knowledge education is fundamental. Technology transfers include a huge amount of tacit knowledge and technology per se is complex. Therefore communication and training are important. Manimala & Thomas (2013) found out in a case study analysis that training and education ranked number one of the critical factors affecting the success of international technology transfer. Training is meant to help employees in the subsidiary creating a cognitive model for the interaction with the new technology. Cognitive rules how to behave in new situations are normally created over time through experimenting and experience, but this process might be alleviated through training. Higher education in turn eases training efforts to create these models. According to Xu (2000), technology transfer only enhances productivity growth in a country when this country’s average male secondary school attainment surpasses around 1.9 years. Xu (2000) calls this number “human capital threshold”. Accordingly, he reasons, that countries with less education do not benefit from transferred technology.
17 technology and the ability to cope with new situations quickly. Otherwise a transfer would not pay off, as it bears costs but would not bring benefits. Thus, I have the following hypothesis:
Hypothesis 1: HQs are more likely to transfer technology to subsidiaries in countries with favorable cognitive institutions.
3.2 The Influence of Regulatory Institutions on Technology Transfer
The regulatory dimension covers the laws, regulations and government policies concerning a topic. In case of technology transfer, they include regulations that support the implementation of new technology financially or give such firms privileges, as well as government-sponsored programs that give advice and assistance in the acquisition process and later with the use of the new technology (Kostova, 1997; Scott, 1994).
Even though technology transfer within a MNE does not involve governments directly, governments set the framework within which the transfer occurs. In comparison, transfer from a research institution to a company would involve a public institution directly, but a transfer within a MNE occurs between two entities of one firm (Lynn, Mohan Reddy & Aram, 1996). Nevertheless, governments can create favorable conditions for technology transfer. One of these conditions is to attract foreign direct investment (FDI). Several researchers argue that technology follows investment. Accordingly, FDI is a prerequisite for technology transfer. Despite direct incentives, such as tax cuts and financial subsidies, countries can attract FDI and technology by appearing stable and reliable (Haug, 1992; Archibugi & Pietrobelli, 2003; Bitzer & Kerekes, 2008). Maskus (2004) emphasizes transparency and stability of the government as a mean to attract foreign technology.
Governments also influence technology transfer directly, by imposing constraints on technology imports through tariffs and trade barriers. The trend however goes more in direction of liberalizing technology flow. Governments can enhance technology transfer by improving public and private institutions, connecting them to the institutions of the providing countries and by providing incentives to transferring firms (Maskus, 2004; Reddy & Zhao, 1990).
18
established a Code on the Transfer of Technology already in 1977, the Paris Convention of 1833 elaborated an international law on patents and the OECD proposed a “Multinational Agreement on Investments” (Haug, 1992; Maskus, 2004; Metz & Turkson, 2000). Two specific examples of how supranational organizations powerfully influence technology transfer are Joint Implementation and Clean Development Mechanism of the Kyoto protocol. Industrialized countries with emissions exceeding their upper limit can invest in environmentally friendly technology in developing countries in order to generate more emission credits for their own use. Thus, industrialized countries can reduce their obligations through technology transfer to other countries. This is a huge incentive for technology transfer that might even dominate other influences or doubts with regard to the receiving countries, such as political instability or missing property right protection (Grubb, Vrolijk & Brack, 1999; Schneider, Holzer & Hoffmann, 2008).
Nevertheless, national governments are still important, because international regulations are difficult to enforce, do not cover all subjects and are not implemented in all countries. On top of that, they only provide incentives, but still the national regulative environment can hinder or facilitate decisions. Thus, national regulations need to be taken into account here. Governments make up one of the three institutional pillars examined in this paper, because of their impact on economic activity (Rodriguez, Uhlenbruck & Eden, 2005). Through their laws and regulations, governments set up the reward structures of an economy.
19 occurs at times. Expropriation involves economic loss and results in diminishing competitive advantage if new technology is copied that could have been sold at high prices otherwise (Kaufmann & Roessing, 2005). Hoekman, Maskus & Saggi (2005) explain why there are problems in enforcing IPRs. Developing countries do not fund IPR enforcement, because merely firms from developed countries, where technology comes from, benefit from these rights, but not local firm. Therefore, it is even more important that governments appear trustworthy to protect firms from imitation; otherwise it would be too risky for MNEs to share their technology with the subsidiary in that particular country (Schneider et al., 2008). Confidence in the quality of the legal system is a critical factor for technology transfer, especially the confidence in contract enforcement, property rights, police and courts. Firms feel safe when they feel that they can bring IPR violation to court and get a fair trial (World Bank, 2005b).
Besides convincing firms of the reliability of laws and government services, governments should reduce investment-, as well as political, risks. Access to capital - which is a vital component of technology transfer - is restricted if investors worry about political risks and enforcement of rules and regulations. Thus, political stability and legal security are vital for the HQ’s decision to transfer technology to a country (Schneider et al., 2008; Oleschak & Springer, 2007). Government processes have to function effectively, law processes have to be transparent and public regulation has to be reviewed by independent authorities. If these criteria are guaranteed, firms worry less about contract enforcement and IPR violation and feel therefore more secure (Metz & Turkson, 2000). When firms have more trust in government and national institutions, they are more willing to make high-value investments in e.g. technology in that country. Firms are especially careful with these kinds of commitments, as they are expensive and long-term, thus involve higher risks (Anokhin & Schulze, 2009).
20
Corruption is especially a problem in international business, because (1) levels of corruption vary across countries and thus also tolerance and the way of dealing with corruption varies and (2) corruption favors local players with connections to authorities that international firms usually lack (Rodriguez et al., 2005). If personal connections are needed e.g. for the delivery of services or to gain bureaucratic requirements, MNEs have a disadvantage. MNEs with HQs in countries with lower corruption might not be willing to accept this disadvantage and avoid making huge commitments in that country, because these can offset the potential gains of a technology transfer and hence make it more costly (Anokhin & Schulze, 2009). Countries with less corruption tend to invest in countries that also have a low level of corruption, whereas countries with higher levels of corruption tend to invest in highly corrupt countries as well. Countries possessing high technology are predominantly little corrupt and thus countries with a high level of corruption are less likely to receive technology from them. Anokhin & Schulze (2009) found a positive relationship between control of corruption and level of innovation in a country. This phenomenon can be transmitted to technology transfer as well, as explained above.
This section explained how regulations and governments influence technology transfer. The legal regulations in a country shape the environment in which technology transfer takes place. MNEs do care about the risks and opportunities offered by this regulatory environment. Dominant issues are especially the stability of politics, enforcement of IPRs and the level of corruption in a country. Therefore, I conclude that:
Hypothesis 2: HQs are more likely to transfer technology to subsidiaries in countries with favorable regulatory institutions.
3.3. The Influence of Normative Institutions on Technology Transfer
21 component reflects the “values, beliefs, norms and assumptions about human nature and human behavior held by the individuals in a given country” (Kostova & Roth, 2002, p. 217).
The normative dimension of institutions is important in the case of technology transfer, because it influences the way the new technology is adopted and implemented. If the employees of the subsidiary accept the technology, it is more likely to be fully internalized and will function more efficient. Even though HQ efforts, such as training, are supportive, ultimately social norms enforce the use of new technology. Technology that is consistent with the norms and values held by the receiving country’s society and thus the subsidiary’s employees, will be better internalized and integrated (Kostova & Roth, 2002).
As already explained before, innovations are created in the supplier’s country context and are therefore difficult to transfer to another context smoothly (Carlsson, 2006; Pavitt and Patel, 1999). Much of the underlying culture that is incorporated in new technology stems from normative influences. All kinds of social relations, interests, ideologies and structures that are prevailing in the inventor’s society are put into the technology, which are not visible in the end product. The shape of institutions – such as reporting hierarchies, organizational knowledge and standard operating procedures – influences the final innovation in a way that is implicit and may not even be realized by the innovator or the recipient (Munir, 2002).
22
This was not a problem of training and education, but a problem of normative understanding of the technology. Even though employees were taught how to react, they would behave intuitively, following the mental model they internalized since childhood. People act according to experience and routine (Cummings & Teng, 2003). This model would normally work to interpret the environment correctly, but does not fit to foreign technology. Mental models – or normative learning structures – evolve with feedback from new experiences and environments, but this is a gradual process, taking some time (North, 1994). Accordingly, technology transfer is a challenge, because “technologies and institutions co-evolve in organizations” (Munir, 2002, p. 1415), but as the technology was not co-evolved in the organization – but simply put in – no normative institutions could simultaneously evolve. Hofstede (1980) describes this learning model as culture, “the collective programming of the mind which distinguishes the members of one human group from another … the interactive aggregate of common characteristics that influence a human group’s response to its environment” (Hofstede, 1980, p. 25). Culture is put into practice in the institutions of a society, which strengthen the mental programs. Hofstede’s cultural dimensions (1980) power distance, individualism/collectivism and masculinity/femininity all capture aspects of expected social behavior and peoples attitudes.
23 Concluding, it can be said that norms, values and culture of a country influence technology transfer, because they are embedded in the technology itself and on the other hand guide the handling of the new technology at the subsidiary. If the subsidiary is able to absorb the given technology properly, the MNE gains, because the technology can be used efficiently. On the other hand, if the subsidiary is not able to incorporate and deal with the new technology, this can be disastrous and cause tremendous losses to the company, as explained in the Bhopal example (Munir, 2002). Thus, I expect that:
Hypothesis 3: HQs are more likely to transfer technology to subsidiaries in countries with favorable normative institutions.
3.4 The Influence of the Institutional Profile on Technology Transfer
Together, the three dimensions cognitive, regulatory and normative, described in the previous sections make up an institutional profile for a country. According to the hypotheses about the single dimensions, it can also be expected that a country performing better in all three of them is better suited to receive technology. Thus:
Hypothesis 4: HQs are more likely to transfer technology to subsidiaries in countries with a favorable institutional profile.
24 Table 1. Summary of Hypotheses.
Hypothesis Hypothesized Effect
H1 HQs are more likely to transfer technology to subsidiaries in countries with favorable cognitive institutions.
H2 HQs are more likely to transfer technology to subsidiaries in countries with favorable regulatory institutions.
H3 HQs are more likely to transfer technology to subsidiaries in countries with favorable normative institutions.
25 4. DATA AND METHOD
The statistical analysis was conducted to answer the research question of whether institutions influence cross-national technology transfer, in the special context of parent-subsidiary transfer. Logistic regression was used to test whether national institutions in the receiving country have an influence on the likelihood that HQs transfer technology to subsidiaries in that country. This chapter describes the sample and data used, the models established to answer the research question and the statistical technique applied to produce the results explained in the subsequent section.
4.1. Sample
This analysis was designed to examine country-level differences in the amount of technology transfers from HQs to their foreign subsidiaries. The unit of analysis in this study is thus the country. Subsidiaries are seen as representatives for the institutional framework of their countries. Data on firms that transfer technology, firm size and industry is on firm level, while data on institutional dimensions, infrastructure and IPRs is on county level.
The sample for this analysis was drawn from the Business Environment and Enterprise Performance Survey (“BEEPS”) in 2005, a joint initiative of the European Bank for Reconstruction and Development (EBRD) and the World Bank Group. This survey covers approximately 9,500 enterprises in 28 countries. Questions are answered in face-to-face interviews with interviewer and top managers or business owners. The survey contains a wide variety of firm characteristics and examines a wide range of interactions between firms and the state. After filtering the survey for foreign owned subsidiaries that acquired new production technology in the last 36 months, 384 cases remained.
Foreign owned subsidiary is defined here as a firm owned 50% or more by “private foreign individual(s)/company(s)/organization(s)”. This is a quite broad definition of ownership, whereas stricter ownership indications, up to 100% owned by the aforementioned group would also have been possible, but would have shrank the sample size essentially. A narrower definition of ownership is considered as robustness check.
26 4.2. Variables and Measures
Technology transfer. The dependent variable is taken from the BEEPS survey of 2005
(EBRD & World Bank). Firms were asked (1) if they acquired new technology in the last 36 months and (2) if they did acquire new technology, what the most important ways their firm acquired this new technology were. One possible answer was “Transferred from parent company”. This answer was taken as dependent variable and was coded as 1; all other answers were recoded as 0 (EBRD & World Bank, 2005).
The BEEPS survey results in 384 cases of foreign owned subsidiaries that acquired new technology in the last 36 months. 39 of these cases transferred their technology from the parent company, 345 cases used other ways to acquire new technology. The study includes 29 countries. This makes a mean of 1,34 transfers from parent per country. Poland, Ireland, Romania, Ukraine, Turkey and Azerbaijan are the countries with most subsidiaries that acquired new technology in the sample. Nonetheless, these countries are not the main receivers of technology from the HQ. Most cases where new technology is received from the parent come from Estonia (4), Ireland (4), Hungary (3), Kazakhstan (3) and Poland (3). Table A.3 in the Appendix shows how subsidiaries with new technology are distributed among countries. It shows that most countries with subsidiaries that received new technology are located in Eastern Europe.
Cognitive dimension. The independent variable cognitive dimension is operationalized by
27
Regulatory dimension. The independent variable regulatory dimension consists of three
distinct measures that are aggregated. The measures are taken from the Worldwide Governance Indicators (WGI) by the World Bank for the year 2005, which measures the quality of governance in over 200 countries. It is based on almost 40 data sources by over 30 organizations. From the six measures for government provided in the WGI, I selected three for the purpose of my study. These dimensions in turn consist of over 100 individual variables. Data was collected in surveys from national respondents, as well as experts from private firms, public organizations and non-governmental organizations. Despite the variety of sources, one possible downturn of the data is its subjectivity. The surveys reflect the views on governance in a country, not the actual governance (World Bank, 2005b; Kaufmann, Kraay & Mastruzzi, 2010). For the 10 countries analyzed in this work, values for the aggregated measure, consisting of the three beneath described dimensions, range from 0.03 (Bulgaria) to 1.63 (Ireland), see Table 3.
Rule of Law expresses the extent of confidence agents have in the rules of society. Key
issues are the quality of contract management, as well as the effectiveness of police and the courts and accordingly the likelihood of crime and violence (World Bank, 2005b). Indicators for the year 2005 rank from a minimum of -2.2091 to a maximum of 1.9749. Countries from the BEEPS study rank from -0.9046 (Russia) to 1.5801 (Ireland), see Table 3.
Governance Effectiveness includes “perceptions of the quality of public services, the
quality of the civil service and the degree of its independence from political pressures, the quality of policy formulation and implementation, and the credibility of the government's commitment to such policies“ (World Bank, 2005b). Indicators rank from -2.1728 to 2.1578. For countries included in the BEEPS study, indicators rank from -0.4594 (Russia) to 1,7335 (Ireland), see Table 3.
Control of Corruption is amongst others measured by the frequency of bribery and
corruption and the perception thereof (World Bank, 2005b). The lowest indicator has a value of -1.6784, the highest indicator a value of 2.3518. Russia has the lowest indicators in the BEEPS study (-0.7812), Ireland the highest (1.5752). The descriptive statistics for the BEEPS study are displayed in Table 3.
28
= + +
3
A higher score in this aggregated value indicates a better regulatory environment for technology transfer.
Normative dimension. The level of uncertainty avoidance in a country indicates this
institutional pillar. Data is taken from Geert Hofstede’s cultural dimension study (1980) where he surveyed 88,000 employees in more than 40 overseas subsidiaries of IBM. This data is available on Hofstede’s official website (www.geerthofstede.nl). As Hofstede surveyed people in comparable positions from only one company, he could ascribe his results to county-specific differences in values and norms of their people. This study has been checked for internal and external validity and reliability thoroughly. Hofstede created ordinal scales for countries for this dimension based on standardized factor analysis of questionnaires (Shane, 1995).
The level of uncertainty avoidance in a society indicates whether people are comfortable with uncertainty and ambiguity or not. Societies with high indicators for uncertainty avoidance stick closely to principles about beliefs and behavior and are intolerant for differing action. Unusual situations make people feel uncomfortable and thus foreign ideas and strange behavior are avoided in high uncertainty avoidance countries. On the other hand, societies with low uncertainty avoidance are more open for innovations and foreign ideas, hence more receptive for technology (Hofstede, 1980; Tung, 1994). A higher score in this measure is expected to have a negative influence on technology transfer, as these countries are not open for the unknown. Thus, this indicator has to be interpreted the other way round as the other indicators, with a low score being favorable for technology transfer and a high score inhibiting transfer. It is also the only measure that is not specifically tailored to 2005. Culture is supposed to change only incrementally and very slowly, thus no data for a specific year is needed (Hofstede, 1980). In Hofstede’s data set, scores range from a minimum of 8 to a maximum of 112. The cases in the sample used here take values between 35 (Ireland) and 95 (Russia) with a mean of 74.6. This is also shown in Table 3.
29 dimension is the GLOBE study, reported by House et al. (2004). This study tried to improve Hofstede’s cultural dimensions. The project involved 200 researchers and thus offers more diverse data than Hofstede. Furthermore, data from the GLOBE study is more recent, compared to Hofstede’s study that is based on surveys from the 1960s and 1970s. Even though culture is supposed to be stable over decades, it is better to have more recent data for the purpose of comparison. The correlation between House et al.’s (2004) measure of uncertainty avoidance and Hofstede’s (1980) original measure equals 0.393 (N=42; p = .010). This alternative source is used in the robustness checks.
Institutional profile. To establish country institutional profiles, an enlarged sample that
includes all countries for which data on the three institutional dimensions is given, neglecting the country’s representation in the BEEPS 2005 (EBRD & World Bank), was used. Table 2 shows the institutional profile established for 63 countries and ranked according to their performance on each dimension and overall. The United States rank highest in the cognitive dimension of the profile, Singapore in the normative dimension due to low uncertainty avoidance and Finland ranks first in the regulatory dimension. The best overall institutional profiles for technology transfer do Sweden, New Zealand and Denmark have. From the countries included in the sample used for my statistical analysis, Ireland ranks highest in the institutional profile and Turkey ranks lowest. The mean values and standard deviations for the institutional dimensions of the included countries are displayed in Table 3.
Intellectual property rights. National IPRs seem to have a considerable influence on the
willingness to share knowledge with a local firm, as already indicated in the section about regulatory dimension. The regulatory dimension however focuses more on the actual application of IPRs, not on their mere existence. Thus, the existence of IPRs is included separately as control variable. IPR effectiveness is measured by “charges for the use of intellectual property, payments (BoP, current US$)” from the World Development Indicators (WDI) in 2005 established by the World Bank.
30
Infrastructure. Infrastructure is an important characteristic for technology transfer, because
most technology needs a certain level of infrastructure to function efficiently. Infrastructure has many aspects, but the most important ones for technology transfer are (1) information infrastructure (Metz & Turkson, 2000, p.35) and (2) energy/electricity supply (Metz & Turkson, 2000, p.35; Tung, 1994). Thinking of machinery, as well as information technology, electricity is essential. One example is a transfer to Chinese Wuhan Steel Mill from Germany. The machinery could only be used for some limited time, even though the machinery would have been able to run much longer, but the city had no reliable supply of electric power, thus electronic outrages restricted the productivity of the technology (Tung, 1994). Infrastructure is measured through “electric power consumption (kWh)” in the WDI (World Bank, 2005a), because this is a more narrow measure than commonly used infrastructure indicators like e.g. percentage of paved streets and is more related to the phenomenon under investigation. Electric power is elementary for almost every new technology, particularly in manufacturing and information technology. Values in the sample range from 2018.66 (Turkey) to 6357.42 (Czechia).
Firm size. Firm size is taken as control variable here, as one would expect that large firms
play a more important role in innovation and technology transfer than small firms, thus that there are differences (Schneider et al., 2008). Small firms are less attractive for international technology, because they lack information and access to capital (Lukacs, 2005). On the other hand, researchers also claim that small enterprises are more innovative, and thus more favorable for new technology (Schneider et al., 2008; Shane, 1992). It is however questionable if firm size is of importance in the particular context of this work, because it takes a look at the HQ-subsidiary relationship. One might assume that subsidiaries are rather small, because they are split-offs from their parent.
31
Industry. Industry might also have an influence on technology transfer, because innovation
might be easier in some industries than in others. Such industries are characterized as producing physical products, being able to exploit scale economies and being supported by government. Industry differences are important when studying country-level differences, as some countries might have more innovation-friendly industries than other countries and have thus an advantage in technology transfer (Scott, 1994). On top of that, other industry specific effects can favor or impede technology transfer, such as the Clean Development Mechanism favors technology transfer in the energy industry, as explained earlier. Because of these special mechanisms in the Kyoto protocol, industrialized countries transfer “green technology” to developing countries in order to improve their carbon dioxide emissions (Grubb et al., 1999; Schneider et al., 2008). In order to take such effects into account, industry is controlled for in this study.
The BEEPS 2005 (EBRD & World Bank) questionnaire includes a question about the industry of the particular firm. Firms can indicate which percentage of their sales comes from a specific sector. This variable is recoded into a binary variable for the purpose of this analysis. The industry where the majority of sales can be dedicated to is chosen as the primary industry. The respective industry is coded as 1, all other industries as 0. Most new technology occurs in the manufacturing industry (57,6%), followed by wholesale, retail, repairs industry (18,5%) and transport storage and communication industry (6,8%). The distribution of parent-transferred new technology is accordingly. Numbers are shown in Table A.2 in the Appendix. 4.3. Method and Generic Empirical Model
4.3.1. Empirical Model
Several models are established to test the hypotheses. Model 4 turned out to be the most accurate model and is therefore presented beneath this paragraph. In the models, ! is the number of firms in country i which got technology transferred from their parent in year t. " is the constant, ! is the independent variable cognitive dimension (measured in years of
education) for country i in year t, ! represents the regulatory dimension (measured as rule of law, government effectiveness, control of corruption) in country i at time t, # represents the normative dimension (measured as level of uncertainty avoidance) in country i. The control variables are included as $% ! for the intellectual property rights present in country i
32
final control variable with industry dummies, where value 1 is given to firms whose main sales are in that industry at time t, 0 otherwise, and is the error term.
Model 4:
!= " + () !+ (* !+ (+# + (,$% !+ (- !+ (. & !+ (/ '!+
4.3.2. Statistical Technique
Data from the different sources was combined using SPSS Version 20. The regressions to calculate the models were also conducted with SPSS. Binary logistic regressions were run, because the dependent variable is a dummy variable, taking the value of 1 if firms answered that they got their new technology from their parent, 0 if technology was received through other ways. By using multiple regression, values for the dependent variable can be predicted from the values of several independent variables (Hair et al., 2006, p. 186).
Table 2. Country Institutional Profiles.
34 Table 2. (continued) Countries Cognitive Dimension Regulatory Dimension Normative Dimension Institutional Profile Mexico 43 44 42 47 Indonesia 58 57 15 48 Costa Rica 46 37 49 49 Portugal 50 23 62 50 Brazil 52 47 38 51 Uruguay 47 30 60 52 Argentina 38 51 49 53 Turkey 55 39 44 54 Ecuador 48 61 30 56 Morocco 62 46 31 56 Colombia 53 48 39 57 Peru 41 55 52 58 Bangladesh 59 63 27 59 Pakistan 60 59 33 60 Venezuela 56 62 38 61 El Salvador 51 53 57 62 Guatemala 63 60 61 63
Cognitive Dimension: rank 1 = most years of school; Regulatory Dimension: rank 1 = highest score in aggregated measure of rule of law, government effectiveness and control of corruption; Normative Dimension: rank 1 = lowest score of uncertainty avoidance;
Institutional Profile: mean of all rankings
Table 3. Mean Values & Standard Deviations.
36 5. RESULTS
This chapter presents the outcomes of the logistic regression. First, a regression was run testing the baseline model and the influence of the control variables. In a second step, the results on the hypotheses were proved on their consistency by conducting several robustness checks.
5.1. Baseline Results
Model 0 in Table 4 is my most basic model, including only the control variables. Results are not in line with prior research, as all variables lack statistical significance at usual levels (p > .05). Whether industry dummies are taken into account or not does not make a considerable difference concerning statistical significance. Nonetheless, when they are left out in Model 0a, model fit decreases (the test-statistic for the Chi-square test equals 4.334 at three degrees of freedom). Thus, industry dummies are kept in the following models. When the control variables are taken out in Model 5, the explanatory power of the model drops from 24.2% (Nagelkerke R squared = 0.242 in Model 4) to 0.7% (Nagelkerke R squared = 0.007 in Model 5). Thus, control variables are kept in the model.
Hypotheses 1 through 3, which claim that HQs are more likely to transfer technology to subsidiaries in countries with favorable cognitive (hypothesis 1), regulatory (hypothesis 2) and normative (hypothesis 3) institutions cannot be confirmed. When their measures are included in the model separately (Models 1 through 3), they all lack statistical significance at usual levels (p > .05). Results are shown in Table 4. The measures for the cognitive and regulatory dimensions show the expected positive sign, whereas uncertainty avoidance, the measure for normative institutions, has a slightly negative influence on technology transfer from the parent. This direction is also as expected. As the variables lack statistical significance however, the impact of the individual institutional dimensions is not significantly different from zero.
Table 4. Institutional Influences on Parent-Subsidiary Technology-Transfer.
Model 0 Model 0a Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
IPR -.011 (.009) -.011 (.008) -.008 (.010) -.009 (.009) -.012 (.009) .009 (.013) Infrastructure .000 (.000) .000 (.000) .000 (.000) .000 (.000) .000 (.000) .000 (.000) Firm Size -.815 (.550) -.771 (.512) -.798 (.551) -.989 (.568) -.822 (.553) -1.199 (.602) -1.210 (.560) Industry dummies?
yes no yes yes yes yes no no
Cognitive dimension .223 (.276) .334 (.271) .071 (.166) Regulatory dimension .925 (.573) 2.837** (1.267) 1.472** (.619) 2.006** (.698) Normative dimension -.003 (.015) .056** (.026) .034** (.017) .040 (.018) Constant -1.206 (1.436) -1.454* (.790) -2.683 (2.342) -.190 (1.552) -.870 (2.444) -7.873** (3.578) -5.858** (2.455) -5.575*** (1.745) N 154 154 154 154 154 154 154 154 -2 log likelihood 119.553 128.942 118.901 116.185 119.523 110.062 126.225 121.093 Nagelkerke R squared .147 .048 .154 .181 .148 .242 .077 .131
38
and normative dimensions are statistically significant at the 5% level. The regulatory dimension goes in the expected direction (B = 2.837). A one unit increase in the regulatory dimension increases the probability of a technology transfer by 17.063 times (EXP(B) = 17.063). The variable for the normative dimension on the other hand is contrary to expectation positive (B = 0.056). This suggests that HQs are more likely to transfer technology to subsidiaries in countries with more uncertainty avoidance. This effect has an odds ratio of 1.057. Consequently, only one of the three independent variables in this model shows the hypothesized effect. Therefore, hypothesis 4 cannot be confirmed.
The results of Model 4 have to be interpreted with caution, because the odds ratio for the only hypothesized relationship, regulatory dimension, is extremely high. The standard error exceeds 1 (S.E. for regulatory dimension in Model 4 = 1.267), which means that the sample is likely to give us estimates not representative for all firms conducting technology transfer. Reasons for distortions in the regression might be due to the small sample size or multicollinearity1. The model has a statistically significant Chi-square of 23.214, but at 13 degrees of freedom. Therefore, model fit is doubtful as well. Nevertheless, Model 4 is taken as my baseline model in the following.
The most accurate model is created by backward stepwise analysis (Model 6). This model only includes firm size, regulatory dimension and normative dimension, of which only regulatory dimension and the constant are statistically significant at any usual level.
Regulatory dimension influences technology transfer from parent positively, with an odds ratio of 7.430. Model 6 only explains 13.1% of the change in the independent variable (Nagelkerke’s R squared = 0.131).
5.2. Robustness Checks
Robustness tests shall check whether there are biases in the findings. Different models are developed to test the robustness of the above results. The first check takes a closer look at the three different components of the regulatory dimension measure by disaggregating them. In the second and forth check I increase and narrow down sample size. Third, I replace Hofstede’s (1980) measure of uncertainty avoidance with House et al.’s (2004) uncertainty avoidance measure. Finally, I include local R&D activity as control variable, to check if a country’s affection towards R&D influences technology transfer as well. Table 5 reports results for the robustness checks. The three independent variables were checked for outliers by applying the outlier labeling rule (Tukey, 1977). No outliers were detected.
1
39 Regulatory dimension is the only variable that shows the expected influence and is statistically significant throughout the basic models. Its effect varies greatly, ranging between odds ratios of around 17 (Model 4) and 4 (Model 5). The value for the regulatory dimension was calculated out of three different variables: rule of law, government effectiveness and control of corruption. Their single influence on technology transfer cannot be determined from the above models. It is however expected that all three components have a positive influence on technology transfer to that country, thus that HQs are more likely to transfer technology to entities whose country has higher values in the three measures. Model 7 reveals that only control of corruption is statistically significant at usual levels. Rule of law and control of corruption both have tremendously high odds ratios (EXP(B) = 57.267 for rule of law and EXP(B) = 438.079 for control of corruption), whereas government effectiveness does not explain the dependent variable at all (EXP(B) = 0.000). Uncertainty avoidance wears a positive sign, contrary to expectations but in accordance with Model 4.
This implies, that government effectiveness could also be dropped from the equation, whereas most of the effect of the regulatory dimension seems to stem from its control of corruption part. This is confirmed in Model 7a, where taking out control of corruption leaves the other two regulatory measures statistically non-significant. Consequently, hypothesis 2 can be restated, now claiming that HQs are more likely to transfer technology to subsidiaries in countries with a higher level of control of corruption. Model 7 has a good and statistically significant model fit but at high degrees of freedom (the test-statistic for the Chi-square equals 27.772 at 15 degrees of freedom).
40
Table 5. Robustness Checks decomposing the regulatory variable, using other data sources, sample sizes and control variables.
Model 7 Model 7a Model 8 Model 9 Model 10 Model 11
IPR .016 (.015) .009 (.013) -.006 (.008) .003 (.016) .012 (.026) .010 (.013) Infrastructure .000 (.001) .000 (.001) .000 (.000) .000 (.000) -.001 (.001) .000 (.000) Firm Size -1.295 (.610) -1.151** (.588) -.533 (.410) -1.135* (.594) -1.356 (.858) -1.215** (.605) High-Tech Exports .016 (.050) Industry dummies?
yes yes yes yes yes yes
Cognitive dimension .303 (.341) .164 (.300) .133 (.184) .240 (.260) .752 (.469) .281 (.315) Regulatory dimension .526 (.348) .590 (.596) 3.740 (2.460) 2.692** (1.263) Rule of Law 4.048 (3.774) 3.897 (3.151) Government Effectiveness -7.921 (5.210) -1.760 (3.569) Control of Corruption 6.082** (2.860) Normative dimension .069** (.029) .052** (.025) .337 (.975) .073* (.044) .061** (.031) Constant -9.562** (3.894) -7.273** (3.304) -1.762 (1.930) -24.419 (200086.792) -12.745** (6.281) -8.130** (3.626) N 154 154 249 145 82 154 -2 log likelihood 105.504 110.445 203.824 115.393 50.229 109.963 Nagelkerke R squared .285 .238 .083 .130 .350 .243
41 the HQ has an effect on the amount of technology transfer between these entities. Probably HQs are more willing to transfer technology to subsidiaries that are “closer” to them and which they control more. Results show that the normative dimension is statistically significant at the 1% level, whereas all other variables, except the constant, lack statistical significance at usual levels. The measure for normative dimension wears a positive sign. Compared to the other models, the only difference is the lack of statistical significance for the regulatory dimension. This is not sufficient evidence to propose the suggested relationship between ownership and technology transfer.
Replacing Hofstede’s (1980) measure of uncertainty avoidance with House et al’s (2004) measure for the same variable in the GLOBE study reduces model fit compared to Model 4 (the test-statistic for the Chi-square test equals 11.446 at 13 degrees of freedom). The results for Model 9 show that when taking GLOBE data, uncertainty avoidance lacks statistical significance. It wears a positive sign, contrary to hypothesis 3. The only statistically significant variable in this model is firm size at the 10% level. The negative coefficient indicates that MNEs are more likely to transfer technology to smaller firms, which is in line with the sample that showed that more small firms received technology from the parent than large firms (see Table A2 in the Appendix). The explanatory power in this model drops to 13%, which is quite low, compared to 24.2% in Model 4 where Hofstede’s (1980) data was used. Thus, the GLOBE data reveals weaker results than Hofstede’s cultural dimension data. Due to these distinct results for uncertainty avoidance, no clear direction for the influence of the normative dimension on technology transfer can be detected.
