1
Total Costs of Doing Business Abroad
Heterogeneity Across Emerging Country
Cities: a Case Study of China and Indonesia
Master Thesis
MSc. Business Studies – International Management Supervisor: Dr Johan Lindeque
Second reader: Dhr Daniel van den Buuse MSc
Student: Selvano Usmany
Student ID: 6152015
Date: 25/3/2016
Statement of Originality
This document is written by student Selvano Usmany who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion
3 Abstract
This research studied the total Costs of Doing Business Abroad (CDBA) by combining the concepts of Liability of Foreignness (LOF), Liability of Newness (LON), Liability of Origin (LOR) and Liability of Multinationality (LOM). Furthermore, it studied how these concepts could be conceptualized at a city level and not at the traditional country level. By using a qualitative multiple case study design, three multinational enterprises (MNEs) in China and two MNEs in Indonesia have been interviewed. The analysis of this data has revealed two insights. Firstly, conceptualizing disadvantages at a city level has showed that LON is more likely to be a subset of LOF while LOR and LOM can be considered as distinct concepts. This showed that the total CDBA consists of LOF, its subset LON, and LOR and LOM. Secondly, analysing at this level has contributed since it revealed that the concepts LOF and LON, and to some degree LOR, are heterogeneous across emerging country cities, thereby indicating that these concepts can be better conceptualized on city level. In conclusion, this study has showed that adding a city as a level of analysis provides better understanding of the total CDBA MNEs face when internationalizing to emerging countries, which can help MNEs in their choices of location.
Keywords
Total Costs of Doing Business Abroad, Liability of Foreignness, Liability of Newness, Liability of Origin, Liability of Multinationality, location choice, emerging market cities
Acknowledgements
The research included in this report could not have been performed if not for the assistance, patience, and support of many individuals.
First and foremost, I would like to extend my gratitude to my supervisor Dr. Johan Lindeque for supervising me over the course of this project. His guidance, patience, flexibility and constructed feedback have helped me to finalize this thesis. For these reasons, I would sincerely like to thank him for his confidence in me.
Secondly, I would like to extend my deepest gratitude to my parents, without whose support and understanding I could never have completed this thesis.
Finally, I would like thank my brother in law Joris for helping me get in touch with the right people to conduct interviews with, and my friends Tristan and Vincent who have support me throughout the complete process.
5 Table of Contents ABSTRACT 3 KEYWORDS 3 ACKNOWLEDGEMENTS 4 INDEX OF FIGURES 7 INDEX OF TABLES 7 1. INTRODUCTION 8 2. THEORETICAL FOUNDATION 10
2.1MNES AND THE TOTAL COSTS OF DOING BUSINESS ABROAD 10
2.1.1 LIABILITY OF FOREIGNNESS (LOF) 11
2.1.2 LIABILITY OF NEWNESS (LON) 13
2.1.3 LIABILITY OF ORIGIN (LOR) 14
2.1.4 LIABILITY OF MULTINATIONALITY (LOM) 15
2.1.5 TOTAL COSTS OF DOING BUSINESS ABROAD FROM NATIONAL TO SUBNATIONAL LEVEL 16
2.2LOCATION CHOICE 16
2.2.1 WITHIN-COUNTRY SPATIAL VARIATION 17
2.2.2 CITIES 18
3. CONCEPTUAL FRAMEWORK 21
4. METHODOLOGY 23
4.1RESEARCH PHILOSOPHY 23
4.2QUALITATIVE MULTIPLE CASE STUDY RESEARCH DESIGN 23
4.3QUALITY CRITERIA 24
4.4CASE SELECTION –MNES IN EMERGING COUNTRY MARKETS 26
4.5DATA COLLECTION &DATA ANALYSIS 28
5. WITHIN-CASE ANALYSIS 33
5.1WITHIN-CASE ANALYSIS COMPANY A 33
5.1.1 COMPANY A’S TOTAL CDBA 33
5.1.2 COMPANY A’S TOTAL CDBA ACROSS CITIES 38
5.2WITHIN-CASE ANALYSIS COMPANY B 39
5.2.1 COMPANY B’S TOTAL CDBA 39
5.2.2 COMPANY B’S TOTAL CDBA ACROSS CITIES 43
5.3WITHIN-CASE ANALYSIS COMPANY C 44
5.3.1 COMPANY C’S TOTAL CDBA 44
5.3.2 COMPANY C’S TOTAL CDBA ACROSS CITIES 48
5.4WITHIN-CASE ANALYSIS COMPANY D 49
5.4.1 COMPANY D’S TOTAL CDBA 50
5.4.2 COMPANY D’S TOTAL CDBA ACROSS CITIES 53
5.5WITHIN-CASE ANALYSIS COMPANY E 55
5.5.1 COMPANY E’S TOTAL CDBA 55
6. DISCUSSION OF FINDINGS 61 6.1LOCATION CHOICE 61 6.2LIABILITY OF FOREIGNNESS 63 6.3LIABILITY OF NEWNESS 64 6.4LIABILITY OF ORIGIN 65 6.5LIABILITY OF MULTINATIONALITY 67 7. CONCLUSION 69
7.1LIMITATIONS AND SUGGESTIONS FOR FUTURE RESEARCH 70
7.2MANAGERIAL IMPLICATIONS AND SCIENTIFIC RELEVANCE 71
8. REFERENCES 72
9. APPENDIX 77
7
Index of Figures
Figure 1: Total Costs of Doing Business abroad (CDBA) 11
Index of Tables
Table 1: Detailed description of selected cases 27 Table 2: Data collection sources 28 Table 3: Interview questions tied to working propositions 30
Table 4: Coding scheme 31
Table 5: Results Case 1 – Company A 34
Table 6: Results Case 2 – Company B 40
Table 7: Results Case 3 – Company C 45
Table 8: Results Case 4 – Company D 51
Table 9: Results Case 5 – Company E 56
Table 10: Cross-case analysis 62
1. Introduction
A large volume of literature in international business (IB) examines multinational enterprises (MNEs) crossing national borders and for decades scholars have been interested in the location choice of MNEs (Dunning, 1998, Rugman, Verbeke & Nguyen, 2011). Nowadays it is well known that locations are chosen most optimally in order to profit from country specific advantages (CSAs), while at the same time keeping the liability of foreignness (LOF) at a minimum (Hymer, 1960; Goerzen, Asmussen & Nielsen, 2013; Lamin & Livanis, 2013).
This study argues that the concept of LOF is most likely to be incomplete and examines these concepts on the basis of two arguments. Firstly, LOF is inadequate in showing the complete disadvantages an MNE is facing in a host-environment (Ramachandran & Pant, 2010; Sethi & Judge, 2009), showing that the total disadvantages might be richer than LOF alone. Secondly, the analysis in IB literature has been shifting from country level analysis to city/subnational region level analysis (Beugelsdijk & Mudambi, 2013) indicating that disadvantages might be conceptualized at city level and not country level.
The Total Costs of Doing Business Abroad
Several scholars discuss a variety of liabilities or disadvantages. The most famous work on the disadvantages of multinationals in a host-environment is the concept of Costs of Doing Business Abroad (CDBA) (Hymer, 1960) and LOF (Zaheer, 1995). However, multiple authors argue that these liabilities do not add up to the total costs and/or disadvantages an MNE is likely to face (Ramchandran & Pant, 2010). Concepts such as the Liability of Multinationality (LOM) (Sethi & Judge, 2009) Liability of Newness (LON) (Stinchcombe & March, 1965) and the Liability of Origin (LOR) (Ramachandran & Pant, 2010) provide different insights in the costs a MNE faces when internationalizing. As such, this study seeks to understand how these different concepts are related to one another to illustrate the complete disadvantages, or in other words, the total Costs of Doing Business Abroad.
Moving from national to sub-national level
Furthermore, MNEs various authors argue that the context of ‘location’, or spatial dimensions, is far from completely understood and further research needs to be conducted to overcome current weaknesses in the conceptualization of location (Beugelsdijk & Mudambi, 2013). Some authors have found that location advantages reside in countries (Dunning, 1998),
9 while others have argued that these might arise from locating in industrial clusters (Porter, 2000), subnational regions (Chan, Makino & Isobe, 2010), or even global cities (Goerzen et al., 2013). Over time, it can be seen that scholars have demonstrated that the level of analysis has been shifting from national to subnational and city level and that advantages might reside in more specific regions or cities within a country. This paper therefore argues that disadvantages might also be conceptualized at a subnational or city level.
Using cities as the level of analysis and focusing on the total CDBA of MNEs, this paper aims to combine both concepts to show how MNEs deal with location choice and disadvantages that come at city level. The most recent work in the IB literature on cities and LOF has been conducted by Goerzen et al., (2013). They found that MNEs locate near global cities in order to reduce LOF. However, as these scholars have showed how mitigating LOF can arise from locating in global cities, there is little theoretical or empirical work how LOF together with LOR, LON and LOM differs across global and non-global cities. Moreover, little is known about how these concepts interact with each other and how they together are able to show the total CDBA.
In the Urban and Globalization literature an enormous amount of work has already been conducted to illustrate the large variety between cities (Knox & Taylor, 1995; Sassen, 2002; Beaverstock, Smith & Taylor, 1999). Using the ‘concept of cities’ derived from the Urban and Globalisation literature and the ‘disadvantages of internationalization’ from the IB literature, this paper combines both literature streams to gain deeper insight in MNEs’ complete disadvantages at city level. More specifically, this paper attempts to find out what the total CDBA is and how these can vary across cities. Since the subnational variation is greater in developing countries (Chan et al., 2010) this paper focuses on cities in emerging markets.
Based on the above the following research question has been developed:
2. Theoretical Foundation
The following chapter will turn the focus to literature on the disadvantages of internationalization and the concept of geographic conceptualization. Firstly, the knowledge on the different types of disadvantages, LOF, LOR, LON and LOM will be reviewed. Next, the shift of location choice in IB literature from national to subnational and city level of analysis will be presented, and finally literature on cities, global vs. non-global city, in the urban and globalization literature will be reviewed.
2.1 MNEs and the total Costs of Doing Business Abroad
In the first stage of modern IB analysis, Vernon (1966) and Dunning (1998) recognize that countries can be endowed with certain resources that contribute to country-specific advantages (CSAs). In order to gain an advantageous position abroad, MNEs seek such country factors and combine them with firm-specific advantages (FSAs) (Rugman et al., 2011). The underlying idea is that MNEs can benefit from internationalizing, by expanding to geographic countries that attain CSAs and that can be matched with FSAs.
However, as Hymer (1960) correctly argues, there are always barriers that can add up to foreign firm disadvantages in a host-country environment. These disadvantages are broadly known as the costs of doing business (CDBA) (Hymer, 1960). The idea is that foreign firms have a disadvantage with respect to their domestic competitors. There are four types of foreign disadvantages he identifies: (1) foreign firms have a lack of information and acquire information at a high cost, (2) the government in the host-country can have disadvantageous or discriminatory policies to foreign firms to protect the domestic market, (3) the home market can levy taxes on MNEs that engage in certain activities abroad, and (4) foreign exchange currencies can jeopardize payments or receipts that domestic firms would not have otherwise (Hymer, 1960). Since his theory on CDBA, multiple researchers have focused on how to overcome these barriers. One of the first scholars that performed an empirical analysis to address this issue was Zaheer (1995), who introduced the ‘Liability of Foreignness’, and drew attention towards the disadvantages of multinationality (Ramachandran & Pant, 2010).
Yet, according to multiple authors (Eden & Miller, 2004; Sethi & Judge, 2009), there is some confusion about the concepts of CDBA and LOF. Sethi & Judge (2009), for example, find
11 that LOF together with the concept of ‘Liability of Multinationality’ (LOM) adds up to the total CDBA, indicating that LOF is just an element the total costs abroad. Eden & Miller (2004) have a similar view as they point out that LOF stresses only the social costs abroad, whereas CDBA stresses economic and social costs. Some authors argue that even more elements should be added. Ramachandran & Pant (2010) for example argue that LOF in itself is inadequate to describe the disadvantages and thus introduced the concept of ‘Liabilities of Origin’ (LOR). Others (Lu & Beamish, 2004; Stinchcombe & March, 1965) find that the ‘Liability of Newness’ (LON) should be added since new firms have lower levels of legitimacy to compete against older established organizations abroad.
Based on these foundations Sethi and Judge (2009) stipulated that the two different concepts LOF and LOM are separate elements of CDBA. However, looking at LOF, LON, and LOR, certain commonalities can be found. Figure 1 shows all liabilities together, which presents the total CDBA; the arrows represent the overlap between the concepts. An explanation on the basis of this figure will be presented in the following sections, which discusses each liability and tries to understand how LOF, LOR and LON theoretically might be related or different to each other.
Figure 1: Total Costs of Doing Business abroad (Source: Author)
2.1.1 Liability of Foreignness (LOF)
Zaheer’s (1995) concept of LOF states that costs can be incurred from four different sources: (1) costs from spatial distance (geographical distance), (2) firm-specific costs (unfamiliarity of the environment), (3) costs from the host-environment (cultural differences, lack of legitimacy), and (4) costs from the home-environment (home governmental restrictions). These costs or challenges can be related to new operations, such as establishing new networks, new target markets and purchasing staffing abroad (Lu & Beamish, 2004).
Interestingly, Luo, Shenkar and Nyaw (2002) contend that LOF is strongest in emerging markets, because they have unique industrial and institutional environments, which makes them unstructured, unpredictable and challenging.
In order to internationalize successfully, it is in a MNE’s best interest to benefit from advantages abroad while keeping LOF at a minimum (Lamin and Livanis, 2013). Different literature streams therefore have sought to identify how MNEs can reduce their unfamiliarity abroad most optimally. According to the resource-based views of strategy, MNEs need to overcome these challenges by importing capabilities and resources from parent enterprises and create a sustainable competitive advantage (Barney, 1991). The premise is that MNEs are able to outperform local firms by implementing a unique value creating strategy based on firm-specific resources. In order to do so, Barney (1991) suggests that firms gain a sustainable competitive advantage when firms’ resources are valuable, rare, inimitable, non-substitutable and above all exploitable. This work is consistent with Hymer’s (1960) work in which he argues that MNEs need to possess some kind of firm specific advantage that outweighs the disadvantages.
Institutional theorists on the other hand argue that MNEs need to adapt in order to have similar organizational practices that meet the demand of the local host country environments, in turn reducing the unfamiliarity of MNEs in a host environment (DiMaggio & Powell, 1991). LOF is thus attributed to the MNEs lack of isomorphism in the host-country environment (Sethi & Judge, 2009). MNEs can adapt to local practices through either: (1) coercive isomorphism, (2) normative isomorphism or (3) mimetic isomorphism (DiMaggio & Powell, 1991). This work is to some extent similar to Porter’s industry-based view, where he states that the industry is more likely to shape the practices of the company in order to reduce the unfamiliarity of a host country (Porter, 1979).
Based on these two perspectives Zaheer (1995) examined whether MNEs were better in overcoming LOF through importing capabilities from the parent or by local isomorphism. Her results suggest that firm-specific advantages embodied in a MNEs practices are more useful to overcome LOF. All in all, there is a great deal of work on overcoming LOF. The essence is that LOF arises from the institutional and cultural differences relative to local firms; LOF is thus always expected to decline over time when MNEs gain better understanding of the
13 institutional environments (Remachandran & Pant, 2010; Lu & Beamish, 2004; Sethi & Judge, 2009).
However, looking at the concept as discussed by Zaheer (1995), Sethi and Judge (2009) find that LOF is actually a subset of the broader CDBA and that LOF itself consist of two different categories; incidental costs and discriminatory costs. Discriminatory costs would be those that arise from regulations exclusively targeting MNE subsidiaries to protect the indigenous firms or costs that arise from economic nationalism and prejudice specifically targeted at foreign MNEs. Incidental costs, on the other hand, arise from costs of learning such as: costs of lack local information, costs of cultural differences and not being isomorphic with the institutional environment (Sethi & Judge, 2009). The most notable difference between these two types of categories is that incidental costs are easier to be learned away over time compared to discriminatory costs (Sethi & Judge, 2009). This indicates that LOF might not always be expected to decline completely over time.
2.1.2 Liability of Newness (LON)
According to Lu and Beamish (2004) there is another liability that arises in geographic diversification. This has been described as the Liability of Newness (LON). This liability exists since MNEs have a lack of information in the host country, examples can consist of challenges for a new subsidiary in staffing, establishing new networks and establishing new distribution channels (Lu & Beamish, 2004). This concept has initially been founded by Stinchcombe and March (1965) in organizational theory, and has been later used in the IB literature to explain the disadvantages of multinationality. According to his work older organization are advantageous over younger organizations as they have cumulated experience in workforce, and have stronger external and internal relationships (Stinchcombe & March, 1965). Because subsidiaries face a greater risk of failure of being new relative to older firms, they depend on cooperation of strangers and are therefore unable to compete effectively against older established organizations (Lu & Beamish, 2004; Stinchcombe & March, 1965).
LON, however, does not necessarily have to apply to all organizations in the same population and is contingent to the external environment on factors such as external legitimacy (Singh, Tucker and House, 1986). These authors contend that LON strongly decreases once an organization acquires more legitimacy. Earning legitimacy proves to be essential when a new
firm wants to earn social capital as they give an example where it is necessary in Asian cultures to build relationships, earn social capital and create legitimacy (Carlisle & Flynn, 2005; Luo et al., 2002). This indicates that LON has some overlap with the incidental (I) element of LOF, which is visualized with the arrows in figure 1. In the incidental LOF, described by Sethi and Judge, (2009) it was explained that disadvantages mostly resulted from lack of information in cultural and institutional different environments, and could be learned away over time. This is consistent with Lu and Beamish (2004) and Rabbiosi and Santangelo (2013) as they contend that both LOF and LON are learned away over time once organizations get more familiar with the local environment by learning, improving legitimacy and accumulate knowledge stock. On the other hand, differences are to be found in the concepts’ mechanism where LOF is measured to local firms (Ramachandran & Pant, 2010) and LON is measured relative to older firms (Stinchcombe & March, 1965).
Based on the above, both commonalities and differences can be found between the two concepts. However, what is not known is where the boundaries are and to what extent they can be seen as different concepts or sub-categories of one another. This thesis therefore seeks to understand how these concepts, empirically, can be related to each other.
2.1.3 Liability of Origin (LOR)
Next to LOF and LON, Ramachandran and Pant (2010) argue that the disadvantages faced by emerging MNEs in international markets, includes a concept that is called the Liability of Origin (LOR). This concept argues that the national origins can contribute to disadvantages in international markets within three different contexts: (1) the home context, (2) the host context and (3) the organizational context (Ramachandran & Pant, 2010). Looking at the home context, these authors contend that MNEs’ disadvantages from the home country are created as a result of the underdevelopment of their domestic institutional environment rather than the institutional distance within the host-country. However, their focus is on EM MNEs, which might have different effects on MNEs from developed economies that expand to emerging markets.
The basic argument of LOR is that every MNE has an identity, which is shaped by home country imprinting (Ramachandran & Pant, 2010; Bartlett & Ghoshal, 1989). For this reason, it is argued that experiential learning and local isomorphism would not help in order to
15 decrease the LOR because the disadvantages can be traced to the home country rather than to institutional distance per se (Ramachandran & Pant, 2010). This indicates that LOR can be homogenous across the host country as national origins are always embedded in foreign MNEs, which in turn cause discriminatory practices. As LOR derives from the host-country institutions targeting foreign MNEs, some theoretical similarities with the discriminatory (D) element of LOF can be found. The arrow in figure 1 represents the theoretical commonalities between the two concepts. Ramachandran & Pant (2010) however stress that there is still a notable distinction that should be made. That is, while LOF can be discrimination targeted at foreign MNEs because of where they are not from LOR, would be discrimination targeted at foreign MNEs because of where they are from (Ramachandran & Pant, 2010).
As can be seen, and consistent with Ramachandran and Pant (2010) and Sethi and Judge (2009) it can be argued that the total costs of doing business abroad comprises of more factors than just the LOF. Although there is overlap between the concepts LOF-LON and LOF-LOR, they also have their differences, meaning that we have to be more nuanced with respect to these concepts. Empirically, this thesis aims to find how these concepts are related and to what extent they are not.
Now that the overlap and differences between these concepts have been presented, the next sub-section will take a look at LOM which, compared to the other liabilities, does not derive from the host-country context but from an multinational context.
2.1.4 Liability of Multinationality (LOM)
Opposed to the disadvantages derived relative to local firms in the host-country environment, there are costs that derive relative to local firms outside the host-country environment (Eden & Miller, 2004; Sethi & Judge, 2009). This is consistent with the work by Sethi and Judge (2009), who explain that the costs of doing business abroad is much richer than just the LOF as geographic diversification includes coordination costs between parent and subsidiaries as well as between subsidiaries. According to these authors, LOM can be distinguished from LOF when: (1) costs arise within the parent-subsidiary network, (2) financial exposure to transactions between subsidiaries increase, (3) subsidiaries are bound to the central decision making from the parent, and (4) it interacts with transnational economic institutions (Sethi & Judge, 2009). It can be seen that LOF is described from the host-country context, while LOM
has a multinational context. Other examples of costs relating to LOM are costs that arise from different time zones, exchange rate fluctuations (Sethi & Judge, 2009; Hymer, 1960) and the failure to effectively transfer firm-specific advantages from the parent to subsidiary (Cuervo-Cazurra et al., 2007).
LOM can differ across MNEs as the type of transactions between subsidiaries or between subsidiaries and the parent office depend on subsidiaries’ role (Sethi & Judge, 2009). Moreover, due to the globalization in the marketplace inter-firm transactions have been increasing significantly, while LOF for each subsidiary might stay overall stable in the host-country environment (Sethi & Judge, 2009). This has made it difficult for MNEs to cope with LOM as it remains rather variable and unpredictable. Thus, as opposed to, learning away LOF through organizational learning and creating legitimacy (Zaheer, 1995), MNEs need to cope with LOM by creating a global mind-set and increase the efficiency processes within MNEs intra-network (Sethi & Judge, 2009).
2.1.5 Total Costs of Doing Business Abroad from national to subnational level
Now that literature has shown that there is theoretical overlap and differences between concepts it is also seen that these liabilities have been focussing on country level or multinational context. Zaheer (1995), Ramachandran and Pant (2010) and Sethi and Judge (2009) are all explaining liabilities at host-country level, home country level or multinational level. Goerzen et al. (2013) however, argue that the unit of analysis can be turned to a deeper level as they find that MNEs’ LOF can be explained within the context of global cities.
Subsequently, this thesis aims to find how each of the above spoken elements of total CDBA, which might comprise of LOF, LOM, LON and LOR arises when MNEs attempt to expand to a foreign country. More specifically, this thesis aims to do this at city level as IB literature has been shifting from country level to city level of analysis with respect to location choices. Now let us take a look at the earlier conceptions of location choice and how it has developed over time.
2.2 Location choice
17 multinational enterprises (MNEs) is more likely not a phenomenon related to an incremental process of internationalization, but rather chosen on the basis of a cost-benefit analysis, or so to say, a rational choice. This is consistent with Dunning’s (1998) eclectic OLI paradigm, in which it is posited that foreign direct investment typically results from Ownership, Location and Internalization advantages. His work exemplifies that ‘location’ is an essential element of foreign direct investment as it illustrates that countries have country-specific advantages (CSAs) in terms of resources, demand conditions or other factors.
For this reason, many studies have attempted to find and understand the relationship between location and firm performance. In an empirical research by Makino, Isobe and Chan (2004) evidence is found that national contextual factors influence firm behaviour and economic performance. They find that country endowments and industry factors have great influence in foreign affiliates’ performance and that these effects matter more in less advanced countries than advanced countries.
2.2.1 Within-country spatial variation
Interestingly, previous studies developed a sophisticated understanding of MNEs’ location strategies using Dunning’s OLI framework and using spatial dimensions on the country level of analysis. However, by using this level of analysis, many studies ignored the enormous variety of spatial dimensions within a particular country. Beugelsdijk & Madumbi (2013) argue that subnational spatial variation should be added to the analysis and that borders should be viewed as discontinuities in space, which thereby acknowledges that spatial heterogeneity within national borders exist. In doing so, clustering and agglomeration effects of subnational regions can provide deeper insight in the determinants of location choice and the effects of within-country location choices.
Various scholars have attempted to change the nature of the existing distance dimensions. Chan, Makino and Isobe (2010) have looked whether subnational regions instead of differences between countries could explain variations in performance. They find that subnational regions have explanatory power with regard to foreign affiliate performance, thereby indicating that this level of analysis is of added value. Furthermore, they found that the regional variations tend to be larger in emerging economies than in developed countries. Similar research has been conducted by Ma, Tong and Fitza (2013) who find that besides
subnational regional effects, their interactions with industry, corporate parent, and home-country effects are also significant to explain variation of subsidiary performance. Moreover, they find that the subnational region effects tend to be stronger in the country’s less-developed subnational regions, indicating that the institutional contexts can be different within countries. Ma et al. (2013) do argue, however, that their study examined location effects at subnational region (province) level, whereas it might be that industrial clusters reside in certain cities or areas within a state.
Goerzen et al. (2013) therefore attempted to look deeper into these clusters by moving to the analysis of global city level. More specifically, they explain the attraction of MNEs to global cities and find that MNEs are attracted towards these cities since they provide micro-locational advantages that mitigate negative effects of a LOF. They posit that it is the international connectedness, producer services and cosmopolitan environment that alleviate the costs of doing business that is associated with uncertainty, discrimination and complexity in a foreign country. Mitigating the negative effects of LOF is in line with previous research conducted by Mariotti, Piscitello and Elia (2010) and Lamin and Livanis (2013) who explain that the choice of location on spatial behaviour is to locate near other MNEs to overcome uncertainty by imitating other MNEs through information externalities and generate a net benefit of knowledge spillovers. This might be a possible explanation why MNEs locate near other firms, next to their cosmopolitan environment and international connectedness.
However drawing further on the work by Goerzen et al. (2013) one of their limitations is that they have been focussing on the commonalities of global cities and thereby ignoring the variance between global cities and most of all the variance of non-global cities. They posit that global cities might vary in their industrial traditions and that future studies need to explore this. As mentioned previously, it is stipulated that within a country spatial dimensions can vary enormously in emerging markets and that is essential to explain location choices. Variance can exist between non-global cities and global cities, so let us take a look at the conceptualization of cities within the literature.
2.2.2 Cities
Cities have been studied intensively in the Globalization and Urban literature over the past three decades. In the globalisation literature, cities have been typically categorized as world cities or global cities (Taylor, 2005). Sassen (2002) describes that global cities are those cities
19 in which global processes are partly embedded in strategic concentrations of resources and infrastructure. These cities are those that are location sites for direct transactions with the global economy (Sassen, 2002). Leading examples are financial district cities such as London, New York and Tokyo (Sassen, 2001). What typifies these global cities from other cities mainly has to do with their embeddedness into the globalized economy as essential spatial nodes (Sassen, 2002). This is somewhat consistent with Taylor’s (2005) argument as he explains that world cities can be identified in their context to other cities around the world, also known as the world city network.
Building on these authors we can see that global cities do have commonalities, which Goerzen et al. (2013) used in their research. However, as previously explained these authors did not look at the variance that exists between cities. Shatkin (2007) shed lights on this matter by arguing that global cities in emerging countries are not converging, thus acknowledging that variance exist between global cities. Taylor (2005), however clearly states that a difference must be made between global cities and world cities. He argues that world cities can also have a globalisation process but that they are less embedded in the global economy than global cities. Furthermore Knox & Taylor (1995) argue that world cities in itself are heterogeneous. New York, London and Toyo might have the similar characteristics as metropolitan areas but there are significant differences as world cities (Knox & Taylor, 1995). According to their research the variance between cities largely results from their cultural, political and economic structures.
Despite much work in the literature there has been some confusion in terminology within the literature. There is overlap between world cities, mega cities and global cities, but there was no clear conceptualization to identify classes of cities. Calcutta might for instance be a mega city but not a world city, whereas Zurich is a world city but not a mega city. For this reason, Beaverstock et al. (1999) categorize world cities based on four key world city services: accounting, advertising, banking and law. Based on these four categorizations, the authors ranked the scores and created a typology to categorize cities. They categorize three different types of cities: i) Alpha world cities, ii) Beta world cities and iii) Gamma world cities. Data from Globalization and World Cities (GaWC) categorizes leading cities like, London, New York, Tokyo and Shanghai as Alpha cities. Examples of Beta cities are Guangzhou, Santiago and Rio de Janeiro and finally Lahore, Glasgow and Leipzig are examples of Gamma cities.
Although these city types will not be used in this research, it provides a clear overview on what the literature already knows about differences between cities.
Looking more in depth at internationalization theory and cities, some studies show that MNEs aim at cosmopolitan characteristics rather country characteristics (Riefler, Diamantopoulos & Siguaw, 2012). This indicates that the degree of cosmopolitan consumers influences and drives an MNE for international activities (Riefler, et al., 2012). According to their research they find that a key of cosmopolitans is their openness towards foreign influences. Tourism, travelling and living and working abroad are central in fostering these cosmopolitan characteristics (Cannon & Yaprak, 2002). Goerzen et al. (2013) show that it is indeed these cosmopolitan characteristics of cities together with, the global interconnectedness, and abundance of advanced producer services that make global cities so interesting a location.
Relating this to the enormous spatial variety in emerging markets, this thesis argues that cities vary in their cultural, political and economic structures but also in their key services in accounting, advertising, banking and law. Especially in emerging markets such as China, foreign direct investment (FDI) is playing a large role to turn Chinese metropolises into global cities (Wei and Leung, 2005). As explained in the previous sections, the liabilities of an MNE can consist of LOF, LON, LOR and LOM. Looking at these liabilities and the subnational variation of emerging countries and more specifically cities, it is my argument that these liabilities are highly likely to differ per city type. The next chapter will elaborate on the mechanism of each liability and categorization of cities to developing working propositions that can be tested in this study.
21
3. Conceptual framework
As discussed in the previous chapter, it is highly likely that the institutional and cultural environment varies in emerging countries (Chan et al., 2011). Furthermore, in the literature on cities, global cities and world cities it is indicated that cities can differ within a national context in terms of their political, economic, cultural and cosmopolitan characteristics (Knox & Taylor, 1995; Goerzen et al., 2013). Urban literature indicates that cities can be heterogeneous to one another, which adds a new dimension within the IB literature to analyse MNEs internationalization process. Adding cities as a unit of analysis might result in the development of City-Specific Advantages (City-SAs) instead of Country-Specific Advantages (CSAs). Hence, not only advantages but also disadvantages/liabilities can be conceptualized at city level. The following chapter presents the working propositions that will be tested in this study.
Relating cities to the concept of LOF, we have seen that LOF arises from four different sources (Zaheer, 1995). Emphasizing the source ‘the costs from the host-country environment’, we see that LOF is measured relative to local firms that are operating in the same national context, which indicates that LOF can be bounded to a country and thus location. Using the argument of heterogeneity between cities from the urban literature where political, economic and cultural characteristics differ per city (Knox & Taylor, 1995), it is expected that LOF will differ across emerging market cities, since emerging markets tend to have a large degree of subnational variation (Chan et al., 2011). This argument has led to the following working proposition.
Proposition 1: LOF is bounded to location and is thus expected to be heterogeneous across emerging country cities.
As has been discussed in the previous sections LON shows some theoretical overlap with the incidental category of LOF indicating that parts of LON can be learned away. Furthermore, LON is measured relative to older firms within the same national context, which shows that this concept is measured within the host-country context and can be seen as bounded to country or location. Combining this concept with the urban literature where cities can be categorized as heterogeneous (Knox & Taylor, 1995), and the fact emerging markets tend to have a large degree of subnational variation (Chan et al., 2011), this thesis argues that LON is
bounded to location and expected to be different across cities. Following this argument, the following working proposition has been developed.
Proposition 2: LON is bounded to location and is thus expected to be heterogeneous across emerging country cities.
With respect to LOR, it is expected that there is some overlap with the discriminatory category of LOF, but with a slight difference in that LOR results from discrimination towards MNEs from the country they are from instead of where they are not from (Ramachandran & Pranay, 2010). This indicates that discriminatory LOF would be more likely to be location bound while LOR would be homogenous across locations within a country because of similar views on a MNEs national origin. Relating this to the heterogeneity of city characteristics within national context (Knox & Taylor, 1995), this study argues that it is expected that LOR will most likely be similar across cities, since MNEs are targeted because of their national origins and not because of where they are located. Following this reasoning, the following proposition has been developed.
Proposition 3: LOR is not bounded to location and is thus expected to be homogeneous across emerging country cities.
Lastly, looking at LOM several authors agree that this concept is primarily a disadvantage borne at the level of MNE headquarters and less at the subsidiary level (Sethi & Judge, 2009; Ramachandran & Pant, 2010). This indicates that LOM is most likely to be an act of multinationality itself (Ramachandran & Pant, 2010) being less likely to be inherent to the host-country environment. Relating the concept of LOM to the localization of MNEs in various city types, this thesis argues that because LOM is less likely inherent to the host-country environment, LOM is not bounded to location and thus expected to be similar in various emerging country cities within the same national context. Following this argument, the following proposition has been developed.
Proposition 4: LOM is not bounded to location and is thus expected to be homogeneous across emerging country cities.
Now that the working propositions have been developed, the next chapter will turn the focus to the development of the research design, in which is explained how the working propositions can be tested (Saunders, Saunders, Lewis, & Thornhill, 2011).
23
4. Methodology
The following chapter will present the research design of this thesis. In the first sub-section the research philosophy will be explained that is maintained throughout this thesis. Then, the case study design is explained, which is followed by the quality criteria section. Subsequently, the fourth and fifth sub-section presents the case selection and data collection respectively. Finally, the data analysis strategy is explained.
4.1 Research philosophy
There are two ways of thinking about research philosophy, which are ontology and epistemology. These two views indicate how the research views the world, in which the assumptions lead to a specific research design (Saunders et al., 2011). Positivism, one of the epistemological views, is being used most often in management studies (Gephart, 2004). However, according to Ryan (2006) the positivist approach attempts objectively to discover the truth, whereas post-positivists try to disrupt the predictability by not finding the ‘truth’ but rather interpret findings and emphasize on meaning, seeing and experience. In a post-positivist epistemology the researcher takes a stance that allows them to create their own epistemologies (Ryan, 2006). In other words, using this stance, it is acknowledged that our own view on the world affects us as a researcher. Also, it should be noted that using this approach, we are in a learning role rather than a testing role, implying that this research focuses on ‘learning with people rather than conducting research on them’ (Wolcott, 1990: p.19). Thus, it contents that researchers should create their own view on social constructions. Furthermore, the post-positivist approach explains that researchers cannot find the total reality but only probabilistically (Gephart, 2004). Lastly, Ryan (2006) contends that the post positivist approach focuses on theory generation and extension. As this research is looking to extend theory on complete disadvantages and finding how each liability differs across cities, this research takes the post-positivist stance.
4.2 Qualitative multiple case study research design
Qualitative research is often being confused with collecting qualitative data (Eisenhardt, 1989). As opposed to quantitative research, qualitative research provides more in-depth and richer insights as opposed to research statistical analyses (Gephart, 2004).
As this research is trying to build theory using qualitative data, this research will make use of a multiple-case study strategy on the basis of Eisenhardt’s (1989) case study approach. More specifically this case study will be a ‘type 4’ or a multiple case study with embedded units of analysis (Yin, 1989). Such approaches allow the researcher to gain understanding of phenomena as they are studied in their real life context (Yin, 1989). According to Yin (2003), evidence from multiple case studies is more robust compared to single-case studies. Furthermore, the advantage in multiple-case studies is that it can link rich qualitative data to mainstream deductive research (Eisenhardt, 1989). In its broadest sense, case studies are rich and help research in the theory-building process (Eisenhardt, 1989; Ridder, Hoon & McCandless, 2009).
Yin (1989) proposes that each case should be chosen carefully as they can follow different replication logic, thus this research aims for theoretical sampling. The replication logic of the sampled cases can be separated into two distinct categories, which are theoretical and literal replication logic (Yin, 2003). In literal replication logic it is expected to find similar results, whereas theoretical replication logic would ‘predict contrasting results but for anticipatable
reasons’ (Yin, 2003, p.4). This research aims for literal replication, with the challenge that it
difficult for researchers to have a theoretical sample that is appropriate. However, as Eisenhardt (1989) posits, the purpose of qualitative research is to develop theory and not to test it, explaining that cases do not necessarily have to be representative to a population. Thus, ideally if possible, multiple cases should allow researchers to do within-case as well as cross-case analyses in order to find whether some phenomena are idiosyncratic to a single cross-case or consistently replicated in multiple cases (Eisenhardt, 1989).
As this study is focused on finding how the liabilities vary across cities, the multiple case study design helps as it allows the researcher to find multiple cases that have business in a variety of cities. The multiple case study will consist of five case companies with two embedded managerial units of analysis per case. Section 4.4 will further elaborate on the case selection criteria.
4.3 Quality criteria
According to Yin (1989) there are four tests that should be done in order to determine the reliability and validity of the case study research. Eisenhardt (1989) also stresses the
25 importance validity and reliability as she contends that even if there is no validity, researchers still have the opportunity to refine and extend theory. To make sure that the quality of this case study will be increased each of the four tests will be taken into consideration (Gibbert & Ruigrok, 2010).
The first test is construct validity which ensures that what we measure is what we intended to measure (Saunders et al., 2011). According to Gibbert and Ruigrok (2010) construct validity, which can be achieved through data triangulation, is essential as it is leads to accurate findings (Denzin and Lincoln, 1994). For this reason, it was chosen to interview two units per case and find supporting quotes from newspapers and articles retrieved from the LexisNexis database. It should be noted, however, that this research is looking for in-depth information from managers making it difficult to find and support all the working propositions with multiple methods.
The second test is internal validity, which explains the causal relationship between variables and results (Saunders et al., 2011). These authors contend that this test is essential in order to explain plausible relationships. In order to ensure internal validity, multiple strategies can be used (Gibbert & Ruigrok, 2010), such as, pattern matching that is achieved by comparing results with previous research (Denzin & Lincoln, 1994; Eisenhardt, 1989). In this research this has been ensured, by building working propositions in the literature based on previous research and comparing them with the findings in the within-case analysis and discussion of findings chapter.
The third test is external validity, which is the belief that theories can be replicated in other settings than the setting in which they are studied in (Gibbert & Ruigrok, 2010). The key is to find generalization, however the authors do argue that it is difficult to find external validity in multiple case studies. Eisenhardt (1989) argues that a multiple case study should ideally consist of 4 to 10 cases and Gibbert and Ruigrok (2010) explain that researchers should provide a clear case study selection. For these reason, this multiple case study consists of five cases with two embedded units of analysis and a set of criteria to select cases, which in turn ensures external validity. The case selection can be found in section 4.4.
Finally, the last test is reliability, which refers to the consistency of results (Gibbert & Ruigrok, 2010). These authors explain that reliability can be ensured through audiotaping,
transcribing and presenting extract of data in the results. Furthermore, reliability is increased through transparency, which has been safeguarded by developing an interview protocol that has been consistent in all cases. The interview protocol can be found in the appendix. Lastly, as Gibbert and Ruigrok (2010) contend that reliability can be ensured through audiotaping and transcribing, all interviews have been transcribed and recorded and stored in the researchers’ database.
Once both the validity and reliability in this research were established, Eisenhardt (1989) contends that researchers should propose a research question. This is essential in building theory from case studies since researchers need to have a focus in order to prevent being overwhelmed by the enormous amount of data collected (Eisenhardt, 1989). For this reason, the following research question has been developed:
How do the total liabilities that MNEs face vary across emerging market cities?
4.4 Case selection – MNEs in emerging country markets
This research’s focus is on MNEs that are located in emerging country markets, with the goal to gain insights into how within country spatial variety is applicable to the degree to which MNEs face costs of doing business and whether the costs of doing business are affected by this spatial variety. More specifically, the MNEs LON, LOR, LOF and LOM will represent the total costs of doing business, while the spatial variation will be represented across emerging country cities. The reason for choosing emerging country markets is because as Ma et al. (2013) explain that subnational region effects can be expected to be relatively larger in less developed countries. Thus the emerging market context has the potential to accentuate the presence of the conceptualized patterns in the results of the data compared to developed countries.
Table 1 shows a detailed description of the selected cases. In the selection procedure five criteria have been used, to ensure the cases can be analysed according to the research question. The first criteria is that a case MNE subsidiary needs to be located or is operating in at least two cities within an emerging country as subnational variation is largest and disadvantages tend to be most intense these type of markets (Luo et al., 2002; Ma et al., 2013). Secondly, the interviewees or embedded units of analysis needed to have their origin from a western culture in order to make sure that they can relate their experience in emerging markets to their experience from their home developed countries.
27 Company
(Country) Company A (China) Company B (China) Company C (China) Investment motive Market-seeking Resource-seeking Market-seeking Company description
Company A is a privately owned Austrian MNE headquartered in Fuschl Am See and is operating in the non-alcoholic beverages industry.
Company A produces, markets, and distributes non-alcoholic licensed products and is sold in 167 countries.
The subsidiary age of this company in China is 2 years and operates in three cities.
Company B is a privately owned Dutch MNE headquartered in Amsterdam and is a global garment sourcing company that sources garment for large retail chains.
Company B, sources garment from factories and distributes and sells to large retailers. Company B operates in 5 countries.
The subsidiary age of this company in China is 11 years and operates in two cities.
Company C is a public Dutch MNE headquartered in Amsterdam and is operating in the recruitment consulting industry.
Company C specialises in temporary and long-term employment, consultancy and project management in the heavy industries and operates in 41 countries.
The subsidiary age of this company in China is 1 year and operates in four cities.
Company
(Country) Company D (Indonesia) Company E (Indonesia) Investment
motive Market-seeking Market-seeking
Company description
Company A is a privately owned Austrian MNE headquartered in Fuschl Am See and is operating in the non-alcoholic beverages industry.
Company A produces, markets, and distributes non-alcoholic licensed products and is sold in 167 countries
The subsidiary age of this company in Indonesia is 1 year and operates in four cities.
Company E is a public UK MNE headquartered in London and is operating in the alcoholic beverages industry.
Company E produces, markets, and distributes non-alcoholic licensed products and is sold in 180 countries
The subsidiary age of this company in Indonesia is >25 years and operates in three cities.
The third criteria was that the case MNE subsidiary needed to have multiple subsidiaries globally so that LOM could be measured. This is because LOM are costs that can derive from a MNEs inter-network coordination (Sethi & Judge, 2009). The last criteria that was maintained in selecting the cases was to ensure that the cases’ embedded units of analysis were in a management function and at least responsible for location choices. This group of interviewees are also known as experts, which consist of a unique source that allows the researcher to gather ‘inside’ information (Dorussen, Lenz & Blavoukos, 2005). Although there are issues stated in the literature regarding expert interviews, Dorussen et al. (2005) still found that there is a sufficient inter-expert agreement on issues. This thus implies that experts can be a valuable source and is thus chosen in this research.
4.5 Data collection & Data analysis
With regards to data collection, primary data is collected using semi-structured interviews (Saunders et al., 2011). According to Eisenhardt (1989) case studies typically combine data collection methods such as archives, interview and observations. For this reason it is also chosen to gather secondary data via the Lexis Nexis database. This triangulation has also been used in order to safeguard the credibility of the collected data (Saunders et al., 2011). They describe that triangulation is the use of two or more independent sources of data or data
Case MNE subsidiary (country)
Embedded unit of analysis’
job title Main emerging country city
operations
Company A
(China) General Manager Operations Manager
Shanghai, Beijing, Guangzhou
Company B
(China) General Manager Operations Manager
Hong Kong, Shanghai, Hefei, An Qing, Tong Ling, Jiangsu, Nan Tong
Company C
(China) General Manager Operations Manager
Shanghai, Hong Kong, Shenzhen, Beijing
Company D (Indonesia)
Country Manager Marketing Manager
Jakarta, Bali, Surabaya, Batam
Company E (Indonesia)
Country Manager Marketing Manager
29 collection methods to corroborate research findings within the research. It must be noted, however, that considering the depth of this research the difficulty of finding supporting quotes from newspapers and news articles was quite substantial. On the contrary, per case two embedded units of analysis have been interviewed, which provided a rich source of data that helped to saturate the data collection. Table 2 shows the job title of these units and the cities the MNE has its main operations/office locations.
Semi-structured interviews have been chosen, as it was preferred to hold natural conversations in which relevant known topics were discussed. The semi-structured interviews proved to be an efficient method of primary research as it also added some flexibility to the interviews so that other potential related topics could be discussed (Saunders et al., 2011). The interview questions can be found in in the appendix and the related working proposition of each question can be found in table 3 and 4.
With regards to collecting the data, the researcher has used the snowball sampling method in order to retrieve information from experts in each case. Snowball sampling, a traditional
‘nonprobabilistic form of sampling in which persons initially are chosen for the sample are used as informants to locate other persons’ (Bailey, 2008, p. 520) has been used in order to
create a sample. According to Cohen & Arieli (2011) snowball sampling is used in qualitative research to access potential interviewees. They also explain that snowball sampling can be used when interviewees are hard to reach because of bureaucratic obstacles or social status, which in this research proved to be an efficient method as the interviewees were all influential decision making units.
Initially six emails have been sent to Company A, B, and C that were within the researchers own network. These persons were asked the request to conduct interviews with the subsidiaries’ general manager and operational manager. As these contacts were within the researcher’s on network, positive reactions were received. For both Company A, B and C the general manager and operations manager of the subsidiary have been interviewed.
After conducting the interviews at Company A, B, and C, informants within these subsidiaries have helped the researcher to conduct interviews at Company D and E. The general manager of Company A has send out request to subsidiaries within their network in other emerging countries to make the sample larger, thus indicating the snowball sampling technique as
Q# Question WP1 WP2 WP3 WP4
1
Could you start by explaining your role as a country manager/middle manager in an emerging country?
Could you tell me something about your experience in a Western multinational
in an emerging country?
3 In which countries are you currently operating/located?
4 How long is your MNE currently been operating in the focal country?
5 In which cities are you currently operating/located? X X X X
6
So far, do you feel that institutional differences or cultural differences have had
an impact on your performance? If so could you please explain? X
7
Referring back to the cities in which your MNE is located, are the
institutional/cultural differences different across cities? If so, how? X
8
Could you tell me something about your relationship with the corporate parent?
X
9 In which phase of multinationality are you currently in? X
10
How important are relationships with local partners to you? (Stinchcombe &
March, 1965) If so, could you give me an example? X
11
How do you deal with the disadvantages of being a younger multinational an a
relatively new environment? X
12
With respect to city types, are you currently looking to do operations or build new locations in other cities?
X
13
Do cities differ in terms of ease in creating new operations, finding new clients or setting up new distribution channels?
X
14 As a western MNE, what kind of identity does your firm have? X
15
Do customers, partners or other stakeholders know usually know where your
MNE is from? X
16 In general, does the origin of your brand impact daily operations? X
17
Referring back to the cities in which you are located/having operations, do you feel that so far this impact has been similar across cities?
X
31 proposed by Bailey (2008). This resulted in two interviews at Company D, one with the general manager and one with the marketing manager, who personally contacted the researcher by email. Then, the general manager of Company C was able to get in touch with an alcoholic spirit distributor in Indonesia and via this distributor contact eventually was made with Company E in which both the general manager and marketing manager have been interviewed.
With respect to post-interview activities, all interviewees received a personal disclosure consent form, which listed the levels of anonymity. On the basis of the researcher’s advice all companies have chosen to use a company pseudonym but allowed to use job titles. In total, 10 semi-structured interviews have been conducted with decision-making units in five MNEs. In order to safeguard consistency and efficiency all interviews were transcribed within 48 hours. All interviews were recorded using two audio recorders; smartphone and laptop, to make sure that the collected data had a backup. The interviews ranged from 48:38 minutes to 1:10:59 minutes and were transcribed using Otranscribe.com. Next to the data retrieved from the conversation, the researcher also made personal notes on a separate form that were noted during the interview. In addition to primary data collection, also secondary data collection
Themes Sub - codes Working
proposition
Definition
Liability of Foreignness (LOF)
Culture differences WP1 LOF is bounded to location and is thus expected to be heterogeneous across emerging country cities. Discrimination Governmental restrictions Liability of Newness (LON) Age
WP2 LON is bounded to location and is thus expected to be heterogeneous across emerging country cities.
Partner relationship
Liability of Origin (LOR) MNE identity
WP3 LOR is not bounded to location and is thus expected to be homogeneous across emerging country cities. National origin Liability of Multinationality (LOM) Coordination costs WP4
LOM is not bounded to location and is thus expected to be homogeneous across emerging country cities.
Parent-subsidiary relationship
has been used in order to triangulate data and corroborate findings. Via the LexisNexis database news articles and papers were found on the topics that included the case MNE subsidiary name or cities in which they were located.
After all the transcriptions were finished, the transcriptions were copied into NVivo and analysed using thematic coding, to make sense of the qualitative data (Ryan & Bernard, 2003). These authors contend that thematic coding can be done by establishing themes and sub-codes and linking them to the theoretical framework. Firstly, general themes were identified within NVivo using the themes ‘LOF’, ‘LOM, ‘LON’ and ‘LOR’. Table 3 represents each of these themes and links them to the working propositions derived from the theoretical framework. Subsequently, for each of these themes, sub-codes have been assigned that help determine or describe themes when developing them from the large amount of data (Ryan & Bernard, 2003). After text passages were identified in the transcriptions the next step was to do a manual coding in Excel where they could be linked to each of the themes and sub-codes, this has been done for every text passage. After all related text passages were found, a within-case analysis was conducted per case, after which a cross-case analysis and discussion of within-case findings followed in order to identify patterns across cases (Eisenhardt, 1989).
Now that the research philosophy, quality criteria, case selection, data collection, data analysis and strategy has been presented, the next chapter will turn to the results. In the results chapter a within-case analysis is presented of each of the above spoken case MNE subsidiaries.
33
5. Within-case analysis
The following chapter presents the results of the analysed data. The first section presents the within case analysis of the five case MNEs. Every case has two managers as embedded unit of analysis: General Manager and Marketing Manager/Operations Manager. The two managers will be individually analysed and compared to complete the within case analysis. These cases are analysed on the basis of the four liabilities (LOF, LOM, LON and LOO). This analysis will first consider the degree to which the focal case MNE subsidiaries experience liabilities or total CDBA and then evaluate the degree to which this experience varies across cities. The resulting case specific patterns (Yin, 1989) are then compared across the cases to enable the findings to be generalised to theory and the implications of this analysis discussed.
5.1 Within-case analysis Company A
Company A is an MNE from Austria and is globally known for its beverages in more than 167 countries. The General Manager (GM) of this company is responsible for setting up the China branch. The GM has more than 15 years’ experience in setting up new businesses globally, including countries such as South-Africa and Japan and has worked in marketing roles in countries like the Philippines, Malaysia and Indonesia. In his current role he is also responsible for Company A’s strategic approach and thus for the choice of cities to establish businesses in as well. The Operations Manager (OM) is responsible for all operations in China and has worked in the Netherlands before moving to China as an expatriate manager.
5.1.1 Company A’s total CDBA
Currently Company A is operating in the large cities Beijing, Guangzhou and Shanghai and is based in Shanghai. According to both managers Shanghai has been chosen as the headquarter location because of its international and westernized characteristics and large potential customer base. Since its first year in 2013, the company has faced many challenges of being a foreign company in China. With respect to Zaheer’s (1995) four sources of LOF; costs from spatial distance, firm-specific costs, costs from the host-environment, and costs from the home-environment, results showed that the disadvantages of Company A mostly resulted from only two sources. First of all, results show the company had faced firm-specific costs, as it has been difficult for the company to change their practices and routines that were in the
