Autonomy, networks and subsidiary performance

JENS GAMMELGAARD, FRANK MCDONALD, ANDREAS STEPHAN, HEINZ TUESELMANN, CHRISTOPH DÖRRENBÄCHER

Jönköping International Business School Jönköping University JIBS Working Papers No. 2012-3

Autonomy, networks performance

and

subsidiary

Jens Gammelgaard (a),* FrankMcDonald (b), Andreas Stephan (c), Heinz Tueselmann (d), Christoph Dörrenbächer (e)

(a) Copenhagen Business School, Department of International Economics and Management, Frederiksberg, Denmark (b) Bradford University School of Management, Bradford, United Kingdom (c) Jönköping International Business School, Department of Economics, Finance and Statistics, Jönköping, Sweden (d) Manchester Metropolitan University, Manchester, United Kingdom (e) Berlin School of Economics and Law, Berlin, Germany

Abstract Using network approaches to subsidiary theory, this paper investigates the impact on performance of interactions between the factors of autonomy, intra and inter-organizational network relationships. The paper analyzes both direct and indirect interactions between these factors. This study develops and extends existing research that use network based approaches to the study of subsidiary performance by considering the role of autonomy as well as network relationships. In addition the study examines changes in the interactions between the main factors rather than the levels of these factors. Examination of the interactions between changes in autonomy and networks and the subsequent impact on performance provides a framework for obtaining evidence on how these interactions affect changes in performance and thereby contributes towards a better understanding of the evolution of subsidiaries. The results of the data analysis, gathered from a survey of 350 foreign owned subsidiaries in the UK, Germany and Denmark, reveals evidence of the complex interactions between increases in autonomy and in network relationships and the subsequent impact on performance. The results also highlight the central role of interorganizational network relationships in the interaction between the factors that produce significant and positive effects.

1|P a g e

Keywords: Subsidiaries; Subsidiary Evolution; Multinational Companies; Autonomy; Network Relationships; Performance; Partial Least Squares

2|P a g e

1.0 Introduction The pressures of globalization provide incentives to change the organizational structures and strategies of multinational corporations (MNCs). From the subsidiary point of view, the main question is how to integrate effectively into the local host country environment, and simultaneously benefit from being part of the MNC network (Holm, Holmström, & Sharma, 2005). This paper develops the network based approaches to subsidiary theory (Nohria & Ghoshal, 1997) by adding autonomy (Young & Tavares, 2004) to the assessment of the impact on performance of the interaction between intra and inter-organizational networks. The paper analyzes the impact of changes in autonomy and network relationships rather than using levels, thereby permitting a more dynamic analysis of the effect of these interactions on subsidiary performance. The main objective of this paper is to develop and extend the Nohria and Ghoshal (1997) MNC network type approaches to subsidiary theory, by integrating autonomy into the assessment of the impact on performance of the interactions between intra and inter-organizational network relationships, and provide a consideration of the dynamic nature of these interactions. The paper, therefore, enhances network based approaches to subsidiary theory, and provides evidence on the nature of these interactions from a survey of foreign owned subsidiaries in Germany, Denmark and the UK. The evidence provided in this paper, therefore, helps to illuminate important aspects of the evolution of MNCs subsidiary strategy in response to global changes. This is an important element in the search for better theories on how MNCs respond to changes in the global environment (Buckley, 2009). Subsidiary strategic development is often connected to the examination of subsidiary evolution and associated with the changes in the host country environment (Birkinshaw & Hood, 1997), parent company strategy, and the allocation of production mandates (Young & Tavares, 2004), subsidiary entrepreneurship (Birkinshaw, 1998) and competence development (Cantwell & Mudambi, 2005). Other approaches embrace institutional theory to examine the impact of the parent company and host country institutions on subsidiary developments (Kostova & Zaheer, 1999). In this literature, the additional costs arising from liability of foreignness influences subsidiary evolution (Zaheer & Mosakowski 1997). This paper, however, enlarges the network based approaches to subsidiary evolution by considering performance in the light of changes in the interaction between autonomy and network relationships. The paper, therefore, has a stronger focus on evolutionary developments in subsidiaries than studies that examine interaction based on levels of autonomy and network relationships. The paper further responds to the call from Giroud and Scott-Kennel (2009) to expand the range of studies on the intensity of MNC linkages by including the quantity, scope and quality of network relationships. A number of studies have investigated autonomy and the network relationships of subsidiaries, and their influence on performance. Most of these studies, however, do not use an integrated framework of autonomy and inter-organizational and intra-organizational relationships. Vernaik, Midgley, and Devinney (2005) examine inter-organizational networking, and subsidiary autonomy, independently and exclude the effects of intra-organizational networks. Andersson, Forsgren, and Holm (2007) meanwhile investigate inter and intra-organizational networks in a federative MNC, but omit autonomy. An earlier study by Andersson, Björkman and Forsgren (2005) investigates the relationship between inter-organizational networks and subsidiary knowledge creation and control only. Other subsidiary performance-oriented studies, treat performance as an independent variable (Lovett, Pérez-Nordtvedt, & Rasheed, 2009) or focus on issues

3|P a g e

such as psychic distance (Dikova, 2009) or spillover effects (Duanmu & Fai, 2007; Giroud, 2007). One study does centre on the relationship between networks, autonomy and performance, but it tests the model using a few case studies (Birkinshaw, Hood, & Young, 2005). In this paper, we build on this study of Birkinshaw et al. to develop a model that includes interaction effects of changes on all three factors, and tests the model using a large sample from three European countries. This study is relevant for subsidiary management because it helps assess the effects on performance of the changing interactions between autonomy and inter and intra-organizational networks. The most important managerial consequence is likely to be on the indirect effects of investment in inter-organizational network relationships, on autonomy and intra-organizational network relationships, and, subsequently, on performance. The paper develops six hypotheses on the links between autonomy, inter and intra-organizational network relationships, and performance. This leads to the methodology section describing the sample and the constructs. A section on the data analyses leads to a discussion of the results and a conclusion.

2.0 Network Theory The theoretical foundation for this paper is network based theory. Different schools of thought, tradition, and analytical focus makes the term ‘network based theory’ ambiguous. Networks are social structures made up of actors that can be individuals or organizations, and the literature analyzes the connections between these actors from many different perspectives. Social network theorists focus on the social contacts of an individual, and the conflicts, power relationships, and communication patterns that arise from such social interaction (Blazejewski & Dorow, 2003). Granovetter (1973) puts this into perspective by investigating how weak tie relationships affect information seeking and innovation. Burt (1992) contributed to the discussion by investigating the value of structural holes, and how actors may profit through interacting with a range of actors that are not interrelated. Later, Granovetter (1995) developed network theory by discussing the concept of embeddedness and the benefits of strengthening relationships with other actors, which is likely to lead to trust-based relationships that decrease transaction costs. The industrial network theory, also termed the ‘Swedish network approach’, builds on this concept of embeddedness, and shows how market actors adapt to each other by adjusting attitudes, strategies, knowledge, and knowledge transfer modes resulting in modifications to products and processes (Forsgren, Hägg, Håkansson, Johanson & Mattsson, 1995). Over time, these repeated business relationships lead to routinization, and the relationships become institutionalized (IMP group, 1982). This paper analyzes networks from the position of the inter- and intra-organizational relationships of the MNC (Ghoshal & Bartlett, 2005), the differentiated network view (Nohria & Ghoshal, 1997) and the ‘heterarchy’ viewpoint (Hedlund, 1986). This literature has changed the view of the MNC from being centralized and uniform, to decentralized and differentiated in their relationships with subsidiaries. In principle, each parent company-subsidiary, or subsidiary-subsidiary relationship, is unique in terms of scale and scope of activities, resource flows, strategic mandates and autonomy. This paper builds on this research tradition by focusing on changes in the number and variety of network relationships, and the frequency of interaction. This type of network analysis links to subsidiary evolution, in the context of how the subsidiary acts as a bridge between the parent company and host country actors (Giroud & ScottKennel, 2009). According to their approach, the strategic development of subsidiaries is affected by the

4|P a g e

quantity of linkages (number of transactions and relationships), scope of these linkages (variety in the engagement in the value-chain, sectors and countries), and performance effect like learning opportunities. Number and frequency in network relationships directly links to this approach. The approach of this paper further builds on the notion of embeddedness (Granovetter 1973; 1995), as the paper encompasses embeddedness by considering the number and frequency of relationships. The embeddedness concept is, therefore, an underlying logic, where increased number of network relationship help subsidiaries to establish weaker tie relationship, whereas increased frequency facilitates strong tie establishments. Finally, we elaborate on a new research field, which investigates the micro-political struggles for power in the MNC (Bouquet & Birkinshaw, 2008; Dörrenbächer & Gammelgaard, 2010; Geppert & Dörrenbächer, 2011). Subsidiary autonomy naturally links to this discussion, but as we later argue, network relationship intensity is likely to influence the distribution of power within the MNC. Frequent relations between headquarters and a subsidiary bring the latter into a more influential position vis-à-vis other subsidiaries. Another example is a high number of external relationships, which increases the likelihood of creating subsidiaryspecific advantages (Rugman & Verberke, 2001) which again will influence performance. The following sections further consider these issues.

2.0 Definition of Inter- and intra-organizational network relationships In this paper the terms ‘network’ and ‘relationships’ addresses the two important elements in MNC linkages, which are based on reciprocal and interdependent task-related activities that are performed given a certain constellation of resources among a number of players in a specific context (Ghoshal & Bartlett, 2005). ‘Network’ defines the actors with which the subsidiary establishes a relationship. Intraorganizational networks consist of the parent company and other subsidiaries of the MNC. Interorganizational networks comprises actors outside the boundaries of the MNC such as domestic or international business entities such as customers, suppliers, and competitors. Relationships refer to the interaction between the subsidiary and network partners, including the linkages and transactions between these partners. These relationships cover the backwards linkages and transactions (such as supply and logistics), forward linkages and transactions (such as distribution and sales), and collaborative linkages (such as strategic alliances with competitors) of the subsidiary (Giroud & Scott-Kennel, 2009), which are carried out within an intra- and inter-organizational network (Jindra et al., 2009). The number of actors with which the subsidiary links characterizes network relationships. Network frequency shows how often, the subsidiary interacts with its partners. A change in networks is, therefore, an increase (or decrease) in the number and frequency of a subsidiary’s network relationships.

2.1 Subsidiary inter-organizational network relationships Subsidiary inter-organizational network relationships are associated positively to performance because of the value attached to establishing the subsidiary within local networks to develop international competitiveness (Porter, 1994) and utilize localization advantages (Dunning, 2009). Obtaining such higher value activities also arises from learning and innovation benefits (Frost, 2001) and entrepreneurial activities (Birkinshaw et al., 2005; Jindra et al., 2009; Powell, Koput, & Smith-Doerr, 1996). Following this logic, an increase in inter-organizational network relationships is likely to increase subsidiary performance by enhancing the power to the subsidiary because of increased resource dependency by other parts of the MNC (Andersson et al, 2007, Bouquet & Birkinshaw, 2008), and thereby decreasing uncertainty of further parent company investment into the subsidiary (Astley & Sachdeva, 1984).

5|P a g e

Increased number and frequency of subsidiary network relationships therefore positively affect performance. However, different logics underlie how increased number and frequency leads to increased performance. Johanson and Vahlne (2009) distinguish between liability of foreignness, which is the additional cost of internationalization caused by a lack of institutional market knowledge in terms of rules and regulations, norms and values, and cognitive structures (Zaheer, 1995; Scott, 1995), and liability of outsidership, which is lack of business market knowledge in terms of lack of knowledge about the market, its players and their mutual relationships (Eriksson, Johanson, Majkgård & Sharma, 1997). To obtain information about institutions, which is likely to be indwelled in social structures and therefore tacit, Johanson and Vahlne (2009) predict the need to form network leading to ‘relations-specific knowledge’. We believe that increasing frequency of network relationships facilitates this process, as frequency motivates trust building between partners which again increases the likelihood of the local partner revealing institutional type of knowledge. Increased frequency thus exposes the subsidiary to multiple sources of valuable assets available in the host environment and thereby increases the likelihood of exploiting such sources (Mu, Gnyawali, & Hatfield, 2007), especially in cases where knowledge is tacit or requires trustbased relationships to be grasped effectively (Madhavan & Iriyama, 2009). On the other hand, to be part of a larger coalition, i.e., increased numbers of network relationships, brings about insights of how the market operate, and the different roles of different actors, which is by Johanson and Vahlne (ibid) termed as ‘general relationship knowledge’. Thus knowledge of the best practices from ‘many’ players is likely to smooth the progress of mimicry or imitation of these practices, which improves performance (Powell & DiMaggio, 1991). Granovetter’s (1973; 1995) notion on strong and weak ties fits this discussion, as increased frequency of network relationships establishes strong ties needed to gain access to tacit knowledge structures, whereas increased number of network relationships facilitates weak ties improving general knowledge seeking processes. On average, it is likely that increased inter-organizational network relationships in terms of increased frequency and number will generate positive subsidiary performance. These arguments lead to the first hypothesis: H1: Increases in inter-organizational network relationships increase subsidiary performance

2.2 Subsidiary Intra-organizational Network Relationships Intra-organizational network relationships have been found to be positively associated to subsidiary performance (Vernaik et al., 2005), as these network relationships boost learning processes and stimulate subsidiary’s entrepreneurial efforts (Gnyawali, Singal, & Mu, 2009). Research shows that in particular the network relationship with the parent company is highly valuable (Birkinshaw & Hood, 1998). For example, knowledge exchanges with a parent company positively affect innovation and performance (Monteiro, Arvidsson, & Birkinshaw, 2008). Further, the extent of the parental international network is important for the subsidiary’s ability to establish its own international network (Elango & Pattnaik, 2007). The parental network provides the subsidiary with important connections, and lowers transaction costs. Furthermore, Gnyawali et al. (2009) argue that the network relationship with the parent company is also decisive for the underperforming subsidiary, as this network relationship reduces the subsidiary’s strategic vulnerability. Luo (2003) argues that parental support reduces the subsidiary’s dependencies on resources located in the host country, and thereby reduces the uncertainty of subsidiary operations. From an MNC network perspective, subsidiaries with frequent network relationships with the parent company benefit from a more the central position in the intra-organizational network (Ghoshal & Bartlett,

6|P a g e

2005). Centrally positioned subsidiaries control value chain operations (Astley & Zajac, 1991) or access to critical resources in the host country. In this situation, the subsidiary is powerful (Forsgren, Holm, & Johanson, 2005). The concept of network centrality, therefore indicates a high number of direct exchange relationships within MNCs are important for subsidiaries to flourish. Ghoshal & Bartlett (2005) show that subsidiaries with a higher number of exchanges with other subsidiaries or the parent company achieve more prominent positions in the MNC’s organization. Hence, on average, increased intra-organizational network relationships enhance the likelihood of the subsidiary’s ability to influence the behavior of the parent company’s strategic decision making behavior, in favour of the subsidiary, and are, therefore, positively associated with performance. This leads to the second hypothesis: H2: Increases of the intra-organizational network relationships increase subsidiary performance

2.3 Subsidiary relationships

inter-organizational

and

intra-organizational

network

Increased number and frequency, in both intra- and inter-organizational network relationships, are proposed to positively influence subsidiary performance. The question, however, is whether an increase in the number and frequency of inter-organizational network relationships will lead to a corresponding increase in intra-organizational network relationships. The main issue here is whether the subsidiary operates in two separate and independent networks, or as argued by Wang, Liu and Li (2009), that there in some cases are interdependencies between the two networks. Wang et al. (ibid) find that this depend on subsidiary role. One case is the MNCs emphasising local responsiveness on behalf of global integration, the multidomestic strategy argument follows, that increased intra-organizational network relationships will not lead to a corresponding increase in intra-organizational network relationships (Jarillo & Matinez, 1990; Prahalad & Doz, 1987). On the other hand, the same authors present the counter argument, in terms of the transnational company This paper argues that both increased number and frequency of inter-organizational network relationship is likely to increase intra-organizational relationships. However, different mechanisms cause this effect. As argued above increased frequency leads to strong tie development needed to overcome institutional and tacit knowledge barriers and it also improves the absorptive capacity of the subsidiary (Cohen & Levinthal, 1990). According to these arguments the accumulation of external knowledge builds up a knowledge reservoir in the subsidiary, which helps the subsidiary to absorb internal knowledge. This further increases the likelihood of increased intra-organizational network relationships. Following Birkinshaw & Hood (1997) absorptive capacities then facilitate the attraction of resources from both internal and external sources. Further Håkanson and Nobel (2001) show how strong intra-organizational relationships induces innovations in foreign R&D units, although dedicated integration efforts are then needed to ensure the dissemination of technologies back to the parent company. Hansen (1999) also points to the need for multiple intraorganizational network relationships to effectively diffuse tacit-type of knowledge. Increased subsidiary inter-organizational relationships, therefore, facilitate the infusion of knowledge into the MNC, and the subsequent dissemination of this knowledge requires an increase in intra-organizational interaction in terms of knowledge transfers and measures to help to diffuse the acquired knowledge. These circumstances lead to an increase in intra-organizational network relationships.

7|P a g e

An increase in the number of inter-organizational network relationships can also increase intraorganizational relationships because increasing engagement by subsidiaries with external partners may increase the parent company (or other subsidiaries) dependency on the subsidiary. In these circumstances problems for parent companies arising from geographical and institutional distances encourages the growth of intra-organizational network relationships. Increased inter-organizational network relationships is typically an outcome of subsidiary ambitions and entrepreneurial behaviour (Birkinshaw et al, 2005), and is likely to lead to subsidiary-specific advantages (Rugman & Verbeke, 2001) or to create a distinct position in the MNC network (Garcia-Pont, Canales & Noboa (2009). Boehe (2007) also demonstrates that the more the subsidiary is able to create ‘resource-dependency’ network relationships, its position within the MNC is enhanced. Participation in powerful intra-organizational networks thus requires that the subsidiary is able to build external network relationships upon which central players in the MNC depend. However, subsidiary evolution, as emphasized by Dörrenbächer and Gammelgaard (2010), requires not only the possessing of attractive resources, but these profitable exchange relationships need to be utilized effectively by the subsidiary. Thus, attractive resources acquired from inter-organizational network relationships need an effectively means of exploitation that encourages increasing appropriate intraorganizational network relationships, which again improves network centrality. These arguments lead to the third hypothesis: H3: Increases in inter-organizational network relationships increase intra-organizational network relationships

2.4 Subsidiary autonomy and performance The evidence on whether high levels of autonomy is directly related to performance is limited and mixed. Young and Tavares (2004) link the concept of autonomy to many different characteristics, but not to performance. Most studies investigate factors that are likely to lead to enhanced performance, such as marketing innovation (Vernaik et al., 2005), degree of internationalization (Fenton-O’Creevy, Gooderham, & Nordhaug, 2008) and subsidiary growth (Johnson & Medcof, 2007). The literature however seldom investigates the direct effect of autonomy on performance. A study of subsidiary roles by Birkinshaw & Morrison (1995) found that both high and low (but not medium) levels of autonomy lead to good performance. McDonald et al., (2008) found limited evidence for positive relationships between some types of autonomy and performance. Subsidiary autonomy is defined in this paper as the decision making rights that are granted by parent companies. A subsidiary possesses high autonomy when operational and/or strategic decisions are primarily made by the subsidiary. Low autonomy arises when such decisions are largely made by parent companies. From the evolution of subsidiaries perspective, autonomy is often associated with performance, because autonomy often correlates with the granting of more advanced subsidiary roles (Birkinshaw & Morrison, 1995; Enright & Subramanian, 2007; White & Poynter, 1984). Autonomy has also been associated with subsidiary entrepreneurship (Boehe, 2007) that permits a degree of freedom to engage in a variety of both intra- and inter-organizational network linkages and transactions that can involve a multitude of possible strategic and operational areas and a large number of possible outcomes. A virtuous cycle effect results where increased subsidiary performance leads to increased autonomy and so on, thus emphasizing the links between subsidiary autonomy and performance. Subsidiaries that have a

8|P a g e

low degree of autonomy will, therefore, often benefit from increased autonomy, and highly autonomous subsidiaries are likely to develop into advanced subsidiary roles, such as centres of excellences (Frost, Birkinshaw, & Ensign, 2002). There are counter arguments to this outcome. Autonomy can lead to a peripheral position in the MNC network, leading to decreasing parent company support (Phelps & Fuller, 2000), autonomy can be used by subsidiaries to engage in rent seeking behaviour (Mudambi & Navarra, 2004), or the development on location-bound resources (Rugman & Verbeke, 2001). Thus, both theoretical and empirical results on the effects of autonomy on performance are mixed. Further, decisions solely made by autonomous subsidiaries might not be as good as decisions negotiated with the parent company, since they will build on a larger knowledge base (Aghion & Tirole 1997). However, the limited empirical evidence suggests that increased autonomy, at least in some cases, positively affects performance. Moreover, there are theoretical arguments for a positive association between autonomy and performance. This view on the theoretical arguments and the existing empirical evidence leads to the fourth hypothesis: Hypothesis 4: Increases in the autonomy of the subsidiary increase subsidiary performance

2.5 Subsidiary autonomy and inter-organizational network relationships Subsidiary flexibility helps management to establish and deal with beneficial inter-organisational network relationships more successfully because of a decreased need to obtain approval from the parent company (Birkinshaw et al. 2005). This is especially the case when operating in networks where not only the relationship to a client (customer/supplier etc.) is important, but the client’s relationships to its suppliers, customers and competitors are critical resources of the subsidiary (Lindstrand, Eriksson, & Sharma, 2009). Increased autonomy, therefore, is, suggested to increase the number and frequency of the subsidiary network (Giroud & Scott-Kennel, 2009). With increased freedom to make decisions, the subsidiary can increase the number of actors with whom it interacts, and further increase the frequency of interaction when needed. In cases where a subsidiary faces resource constraints and is consequently not able to increase the number of actors in its network relationships, an increase in autonomy is likely to facilitate innovation to take advantage of its existing relationships (Verbeke & Kenworthy, 2008). A subsidiary in this position is therefore more likely to engage more intensively with local partners by increasing the frequency of interaction in order to gain access to knowledge. This reasoning lead to the fifth hypothesis: H5: Increases in subsidiary autonomy increases inter-organizational network relationships of the subsidiary.

2.6 Subsidiary autonomy and intra-organizational network relationship Increases in subsidiary autonomy are likely to lead to decreases in intra-organizational network relationships. The primary reason for this is that subsidiary autonomy, by nature, is likely to disassociate the subsidiary from other units in the organization (Phelps & Fuller, 2000). Thus, studies indicate that subsidiary autonomy has a negative effect on intra-organizational knowledge sharing. Noorderhaven & Harzing (2009) find that subsidiary autonomy leads to ‘stand alone activities’, which implies less social interaction between the subsidiary’ employees and other MNC staff members. The intra-MNC flows of knowledge from subsidiaries with high autonomy have proven to be limited (Monteiro et al., 2008). The direct effect of increasing autonomy on intra-organizational network relationships is therefore likely to be negative. As argued in hypothesis 5 however autonomy directly increases inter-organizational network relationships and hypothesis 3 postulates that increased inter-organizational network relationships directly increases intra-organizational relationships. The combined postulates of hypotheses 3 and 5 imply that

9|P a g e

increased autonomy indirectly increases intra-organizational network relationships by increasing interorganizational network relationships which in turn increases intra-organizational relationships. The direct effect of autonomy on intra-organizational network relationships is however likely to be negative as increased autonomy tends to lead to a degree of disengagement with other parts of the MNC. This line of reasoning leads to the final hypothesis: H6: Increases in subsidiary autonomy decrease the intra-organizational network relationships of the subsidiary.

3.0 Methodology To test the hypotheses a survey of foreign owned subsidiaries located in the UK, Germany and Denmark was undertaken following Tung and Witteloostuijn’s (2008) recommendation of surveying international business themes using a comparative sample. A self-administrated questionnaire sent to these subsidiaries provided data to test the hypotheses. The design, administration and procedures of the mail survey adopted the main techniques recommended by Dillman (1991), supplemented by specific steps informed by Harzing (2000) and Harzing and Noorderhaven (2006) to increase response rates of international mail surveys. The native-speaking subject specialists of the research group translated an initial questionnaire, written in English, and developed using a literature review of former surveys in this area, into Danish and German. This included back-translation, consultation with linguistics and final adjustments resulting from the pilot tests in the three countries. The target audience of the survey was the managing directors of the subsidiaries. The mailings consisted of a four page carefully designed and easy to complete structured questionnaire, a return envelope and a personalized covering letter, which included a description and the objectives of the study, the importance and relevance of the study for subsidiary managers, the ethics and confidentiality procedures undertaken, contact details for enquiries and the deadline. To incentivize participation, respondents were to receive a summary of the main findings of the survey and its implications. To increase the likelihood of responses, the home institutions of the research team administered the process in their country. The survey took place in 2007/08. The construction of the sampling frame used the Commerzbank database, the Experian database, listings of foreign chambers of commerce, Dunn & Bradstreet Lists, regional authorities, and the commercial sections of embassies. The German and British samples consisted of a random selection of 3.000 foreign owned subsidiaries and in the case of Denmark of a census of the 2.996 identified foreign owned firms. A total of 528 responses resulted after sending reminders. This consisted of 249 Danish, 155 British and 124 German replies. After deducting holding type establishments, real estate firms, registered office, non-active trading addresses, wrong addresses, establishments that moved away, and those with a change in ownership, the effective sample size reduced from 8,996 to 5,584, yielding an effective response rate of 9.5%. Although relatively low, it is not out of line with other international mail surveys (e.g. Dikova & van Witteloostsuijn, 2007; Harzing & Nooderhaven, 2006; Nooderhaven & Hazing, 2009) and is not unusual for multi-country studies with high level management as respondents (Harzing, 1996; Harzing & Noorderhaven 2006; Noorderhaven and Harzing, 2009). In terms of host country nationality, the response rate equals 15% for Denmark, 10.4% for Britain and 5.3% for Germany. There are substantial differences in the response behavior in foreign owned firms among the three host countries, previous studies point to similar patterns (Brewster & Hegewisch, 1994; Harzing, 1996; Pudelko & Harzing, 2007). Moreover, the relatively low response rate for Germany is not out of line with international business surveys in Germany (Coeuderoy &

10 | P a g e

Murray, 2008; Schwens & Kabst, 2009). Tests for representativeness along broad industry characteristics indicated no significant differences, neither for the total sample nor among the three host country settings. In light of potential non-coverage error (Dillman, 1991), tests comparing the respondents’ industry profile with official data on the number of foreign owned firms according to industry, utilizing information from the Deutsche Bundesbank, Office of National Statistics and the European Statistical Data showed no significant differences for the sample as a whole, or according to host country. Detailed controls, such as host country, home country, type of industry, size, entry mode were included in the statistical analysis nonetheless (see below).Test for non-response bias used wave analysis, based on the observation that mail survey late respondents tend to be more similar to non-respondents (Fowler, 1993). The comparison of early and late respondents, along the characteristics of broad industry, age, entry mode and nationality of managing director, did not reveal any significant differences in response behavior. The usable responses reduced to 350 because of responses with missing data due to the fact that the study’s focus is on change and this requires retrospective data relating to 5 years ago requiring the inclusion of only those subsidiaries which have been in foreign ownership for at least five years. Missing cases account for about 32% of the responses, this compares favorably to other international mail surveys (Björkman et al 2007; Dikova & van Witteloostuijn, 2007). Inspection of the missing data revealed no discernable patterns and statistical tests confirmed the randomness of the missing data and differences in the number of missing data between the three host countries were non-significant. The questionnaire collected data on the salient characteristics of the foreign owned subsidiaries, see Table 1. Over half of all subsidiaries are German, Dutch or Swedish owned. The majority of subsidiaries employ less than 100 employees, host country managers predominate, as do Greenfield sites. Most subsidiaries are more than 10 years in foreign ownership and have relatively high share of sales/distribution in their activity profile. The majority of subsidiaries operate in high/medium-high or knowledge intensive industries. The subsidiary profile is generally similar to that of other studies on foreign owned subsidiaries in western European countries (McDonald, Tüselmann, Dimitratos, & Voronkova, 2005; Tüselmann, McDonald, & Thorpe, 2006; Vitols, 2001). --------Table 1 around here-------

3.1 Testing techniques A partial least square (PLS) approach to structural equation modeling tested the hypotheses. A PLS modeling approach to subsidiary research is used by, among others, Vernaik et al., (2005), Money and Graham (1999), and Fey, Morgulis-Yakushev, Park, and Björkman (2009) and in other related areas, such as international joint ventures (Inkpen & Birkinshaw, 1994). This technique is preferable to Lisrel and AMOS techniques with smaller samples. PLS models operate with two sets of linear equations, an inner model that specifies relationships between latent variables, and an outer model analyzing relationships between the latent variables and associated manifest variables. This permits simultaneous analysis of the path coefficients between latent variables, and the path coefficients between these variables and their constructs (measurements) (Fey at al., 2009). This provides an assessment of the reliability and validity of the measurement model, followed by an assessment of the structural model (Hulland, 1999). Further, as Vernaik et al., (2005) argues, since models and measures in international business are still in the initial

11 | P a g e

stages of development, the regression based approach is more appropriate than covariance-based methods like LISREL. Finally, PLS method is effective against inadequacies such as skewed distribution of manifest variables, multi-collinearity within blocks of manifest and between latent variables, and omission of data (Cassel, Hackl, & Westlund, 1999). Using SmartPLS 2.0, t-statistics emerge by bootstrapping procedures. This technique makes the result more reliable as it uses repeated random samples (Vernaik et al., 2005). Further, total effects are calculated. This is advantageous since it includes both direct and indirect effects (Sönke, 2010). Chang, Witteloostuijin, & Eden (2010) address the problem of common method variance (CMV) which is likely to occur in self-reported questionnaire surveys where the same respondent provides information for both dependent and independent variables. The survey used in this paper is of this kind, but the respondents reported on measures ‘five years ago’ and ‘currently’, thereby allowing use a change variable that reduces the likelihood of CMV. Further, as suggested by Podsakoff & Organ (1986), the reporting on discrete event reduces the likelihood of CMV, thus in the questionnaire, the number and frequency of relationships were sought. Furthermore, the questions relating to performance appeared before questions related to relationships and autonomy. This order reduces the likelihood of the respondent estimating, for example, good performance as an outcome of the high levels of relationships. Complexity of the PLS technique further reduces the likelihood CMV (Chang et al., 2010). The Harman's one-factor test results are satisfactory, using a Principal Component Analysis, where first eigenvalue counts for 25% of the variance and 6 eigenvalues are above 1.

3.2 Construct and Measures The model has four main constructs – 1. 2. 3. 4.

‘Subsidiary autonomy’, ‘Inter-organizational network relationships’, ‘Intra-organizational network relationships’, ‘Subsidiary performance’.

Data for these constructs covered the current period and 5 years ago. Previous questionnaire work has shown that this time span provides more accurate information than longer periods (Peng & York, 2001). This five-year period establishes a mathematically explicit relationship between observed scores (or manifest variables) and latent variables (the change variable) (Borsboom, Mellenbergh & van Heerden (2003). The use of latent variables is widely used in structural equation models, and as Borsboom et al., (ibid, p. 203) write, it is: “the analysis of interindividual differences data by statistically relating co-variation between observed variables to latent variables”. To capture rich data, the constructs derive from multiple questionnaire items using five-point Likert scales. To give one example, respondents have reported the level of the number of relationships with customers (inter-organizational relations) on a 1-5 Likert scale, for five years ago and for currently. Respondents reporting a ‘2’ five years ago, and a ‘5’ currently results in the change (latent) variable of 3. The latent variable then becomes an amalgamation of the change in the number and frequency of network relationships on the one hand and in strategic and operational autonomy on the other. Using PLS, each variable gets a weight - a coefficient - that reflects how important the manifest variable is for the latent. The t-tests for the outer relations (manifest variables) indicate

12 | P a g e

whether those coefficients (weights) are significant. The coefficients for the manifest variables are determined and the R-square for the inner relation maximizes the structural model. All constructs are based on self-reports and therefore subsidiary performance includes subjective (nonfinancial) measures. This may be subject to bias, however, this method is widely used in literature, and in general, there is evidence of general reliability (Venkatraman & Ramanujam, 1986). In relation to measuring the performance of MNC operations, problems exist in relation to its multi-faceted nature of performances (Glaister & Buckley, 1999; Gong, Shenkar, Luo; & Nyaw, 2005; Miller, Lee, Chang, & Le Betron-Miller, 2009). Thus, subsidiary performance is a single measure to gauge overall performance, and to provide a rounded view of subsidiary performance. There are also well documented problems of collecting accurate and valid performance measures using questionnaires (Luo, 2007). Management’s strategic thinking and actions, however, guides their perception of company performance rather than solely by objective performance indicators, and it is the perception of facts and not facts themselves that management is likely to act upon (Thompson, 2003). In addition, at subsidiary level many objective indicators, especially financial ones, are suspect because of the reporting arrangements of MNCs (Guest, Michie, Conway, & Sheehan, 2003). Due to reservations with similar studies, objective measures of performance were not included in the questionnaire (Demirbag, Tatoglu & Glaiser, 2007). Various studies employing subjective measures of performance ask respondents to assess performance in relation to their competitors (Ellis, 2007; Griffith & Myers, 2005). This facilitates the comparison of establishments across different size categories and across diverse industries. Subsidiary performance uses a 5 items measurement with items frequently and reliably used in other studies (Birkinshaw et al, 2005; Gong et al, 2005; Lou & Park, 2004): • • • • •

‘Sales growth by volume’, ‘sales growth by value’, ‘productivity’, ‘customer satisfaction’ ‘market share’.

Respondents assessed each of these performance items compared to their market competitors on a scale of 1 (a lot better) to 5 (a lot worse). The constructs inter-organizational networks and intra-organizational networks were measured by items adapted from Holm and Pedersen (2000), measuring the number and frequency of the relationships of the subsidiary with a range of partners. Intra-organisational partners included: • • • •

‘Buyers’, ‘suppliers’, R&D and innovation centres’, ‘other units within the MNC organisation’.

Inter-organisational partners included:

13 | P a g e

• • •

‘customers’, ‘suppliers’, ‘competitors’.

Both inter and intra organisational relationships were measured by the number of relationships ranging from 1 (none) to 5 (many), and frequency of contact with networks on a scale 1 (low) and 5 (high). Measurement of autonomy followed the approach of Young and Tavares (2004) with strategic decision making (policy decisions) and operational decision making (tactical decisions). The measurement for strategic and operational decision making authority uses approaches and measurement scales adapted from Birkinshaw and Hood (2000), and Taggart and Hood (1999). The items of strategic decision making authority were: • • • • • •

‘market areas supplied’, ‘product range’, ‘R&D and new product development’, ‘production of goods or services’, ‘financial control’ ‘human resource management’.

Areas of operational decision making were: • • • • •

‘marketing activities’, ‘R&D and new product development activities’, ‘activities involved in producing goods or services’, ‘financial management practices’ ‘human resource management practices’.

For both the strategic and the operational decision making items, respondents assessed the extent of their decisions making decision making autonomy, using a measurement scale ranging from 1 (exclusively by headquarters) to 5 (exclusively by the subsidiary). Figure 1 illustrates the tested model. --------------------Figure 1 about here --------------------Control variables are (a) host country, (b) home country, (c) size (number of employees), (d) type of industry (as defined in Note 1 under Table 1) and (e) entry mode (Greenfield, acquisition). This type of control variables has been used in other PLS test (Fey et al., 2009).

14 | P a g e

4.0 Results The results of the PLS test are reported in Table 3 and Table 2 provides the composite reliabilities, Cronbach alpha values, and R-squares. The Goodness-of-Fit is 0.996, and Standardized Root Mean Square Residual is 0.080. Table 2 and Table 3 around here The results confirm hypothesis 1 suggesting a positive relationship between increases in interorganizational network relationships, and increases in subsidiary performance. There is some support for hypothesis 2 because there is a significant relationship without the controls. The results confirm the relationship between inter-organizational and intra-organizational network relationships, thus supporting hypothesis 3. There is some support for hypothesis 4 with a significant link between increased autonomy and increased performance with controls. There is strong support for hypothesis 5, confirming the positive relationship between autonomy and inter-organizational network relationships. The results support an indirect effects route of autonomy on performance as autonomy positively affects inter-organizational network relationships and they in turn positively relate to performance. There is no support for hypothesis 6, as the relationship between autonomy and performance has the correct sign, but it is not significant. There is some evidence for an indirect effect of autonomy on performance via intra-organizational network relationships. Autonomy positively relates to inter-organizational network relationships and these relationships are associated with intra-organizational network relationships, which significantly and positively relate to performance without controls. Table 4 summarizes the results. Table 4 around here The controls affect the triangle of intra-organizational network relationships, autonomy, and their relations to performance. Using a Multivariate Analysis of Variance in a two-step procedure, where introduction of control variables reveals that there are no home country and industry effects. Host country, size, and entry mode do however affect the results. These three factors have a negative effect of intra-organizational relationships on performance, and the effect of autonomy on intra-organizational relationships. These controls positively affect the relationships between autonomy and performance. To test the effects of the host country, division of the sample into Denmark, Germany and UK respectively reveals, in relation to hypotheses 1, 2, 3, and 5, that host countries makes no difference in relation to the reported findings for the total sample. Host country effects are evident however for hypothesis 4 where the effects of autonomy on performance reveal that in Denmark and the UK, there is a positive and significant relationship between autonomy and performance and a significant and negative for Germany. Finally, regarding the relationship between autonomy and intra-organizational relationships (hypothesis 6), the results of Denmark and Germany confirm the results from the total model, whereas UK reports negative total effects. Confirmation of the robustness of the data analysis arises from use of Generalized Structured Component Analysis that reveals similar results. Finally, the causal relationships between factors were tested. In the main model, we assume that autonomy affects network relationships, and that inter-organizational network relationships affect intra-organizational network relationship. An alternative model, where intraorganizational network relationships affect inter-organizational network relationships, shows similar pathcorrelations; however, this model is weak, showing lower R-squares. Another alternative model, testing how network relationships affect autonomy, produces the same path correlations and similar R-squares as

15 | P a g e

the main model. There is, therefore, a reverse causation effect regarding the relation between autonomy, network relationships and performance, expressed as the ‘virtuous circle effect’ in discussions leading to hypotheses 4

5.0 Conclusions and discussions This study sheds light on the complex interactions between increases in autonomy, inter- and intraorganizational network relationships and performance. The results provide strong evidence of direct and positive linkages between increases in; inter-organizational network relationships and performance; autonomy and inter-organizational network relationships; and inter- and intra-organizational network relationships. The evidence is less strong that direct and positive links exist between increases in; autonomy and performance; and intra-organizational network relationships and performance. The study also finds support for indirect effects between autonomy and performance, by the affect of increases in autonomy leading to increases in inter-organizational network relationships and thereby impacting on performance. Evidence also emerges of indirect links between increases in autonomy and in intra-organizational network relationships. The route for this link is that increased autonomy leads to increased inter-organizational network relationships, which in turn is associated with increased intra-organizational network relationships. There is no evidence, for the whole sample, of a direct negative effect between increasing autonomy and intra-organizational network relationships. There are however country specific evidence that a negative relationship exists in this relationship. Overall the results suggest that increasing autonomy has perhaps only a weak direct influence in reducing intra-organizational network relationships, and the indirect effects of interactions between the autonomy and inter-organizational network relationships tend to increase intra-organizational network relationships. The paper provides new insights on some key issues about the effects on performance of the evolution of autonomy and inter- and intra-organizational network relationships in subsidiaries. This study contributes to the differentiated network model of MNCs (Nohria & Ghoshal, 1997). In particular, the inclusion of autonomy extends and develops the network based literature on subsidiary strategy based on intra-and inter-organizational networks (Andersson et al., 2002 and 2007). The study enhances the literature on the role of subsidiary autonomy in MNC strategies (Young and Tavares, 2004) by the approach and results of this study. The paper also develops and deepens the work of Birkinshaw et al., 2005 on the interactions between autonomy and inter and intra-organizational networks and performance. The results from the alternative causation model in this study further develops these network based approaches to subsidiary strategy by indicating that a two-way directional relationship is likely to exist in the relationships between the key factors. The paper thereby highlights the complex nature of interactions between key factors in subsidiary development, and provides evidence on the significance of direct and indirect effects. In addition the study indicates that two-way causation may lead to virtuous cycles in the interactions between these key factors. This study contributes to helping us to increase our understanding on the interaction between network connections and autonomy in the context of embeddedness in both the MNC (internal links) and the host location (external links). A major implication of the study is that increasing embeddedness in the host location (or external embeddedness) appears to be at the core of subsidiary evolution. However, external embeddedness requires the establishment of internal embeddedness, if subsidiaries intend to utilize its resource dependency type of power to bring them into a more central position in the MNC

16 | P a g e

network. Further, increased absorptive capacity is likely to smoothen the dissemination of external gained knowledge. Increases in autonomy further provide freedom to interact in these networks. The paper, therefore, sheds light on how a careful evolution of autonomy, inter- and intra-organizational network relationships can generate performance benefits, which do not necessarily drive a wedge between subsidiaries and parent companies. Indeed the paper suggests that evolution centred on increasing interorganizational network relationships can lead to subsidiaries strengthening intra-organizational network relationships and autonomy. The focus of the study on changes highlights some major issues on the evolution of subsidiary strategy. The results indicate that complex processes are at work whereby both direct and indirect routes exist in the interactions between changes in autonomy and in inter-and intra-organizational network relationships and the subsequent effect on performance. The study does not provide evidence on what determines these complex interactions in terms of issues such as, subsidiary roles, parent company/subsidiary power systems, and resource constraints and competencies in host locations and in the MNC. Furthermore, the study does not focus on the drivers of such developments in terms of the attitudes, characteristics and behaviour of key players in this game. However, the findings of this paper put an emphasis on further investigations of how liability of foreignness and outsiderships affect subsidiary evolution. Clearly, an increased frequency of interaction is needed for subsidiaries to obtain institutional knowledge that is likely to be indwelled in complex social structures. However, qualitative studies at the individual level must reveal such mechanisms. Increased number of external players with whom the subsidiary interact, is also likely to decrease outsiderships. However, future research must develop proper measurements for this concept, in order to test it in large samples. The paper further indicates that future investigation of evolutionary paths in subsidiary development in the context of these issues should take into account the complex direct and indirect paths of the interactions between changes in autonomy and inter- and intra-organizational network relationships and performance. Furthermore, the central role of changes in inter-organizational network relationships in these interactions should be included in such studies. Some implications for subsidiary management emerge from the study. The results imply, for both parent company management, and subsidiary management, a need to coordinate strategies to manage effectively all three factors simultaneously, instead of focusing on the factors individually. The causal relationships between the three factors identified in this study need to be further developed, at subsidiary, level to obtain rich data on these interactions in a dynamic context and thereby provide detailed guidance to managers on how best to develop strategically the subsidiary. An important implication of the study is that inter- and intra-organizational network relationships are not mutually exclusive. A connected implication is that in the case of resource constraints, the increased use of the external links of the subsidiary appears to lead to a higher payoff that supports the internal links. This indicates a need for careful management of external and internal network relationships to achieve performance benefits. Subsidiary managers should also be aware that the rationale for increased autonomy is complex. The case that increased autonomy will directly lead to enhanced performance is not strong, but increasing autonomy does appear to have indirect and positive effects on performance. Links to inter-organizational network relations are probably at the centre of the complex interactions that lead to enhanced performance therefore increasing autonomy is primarily a facilitator for managing inter-organizational relationships, which can subsequently influence performance. Managing increasing autonomy to boost performance achieved by increasing interorganizational network relationships needs careful management to avoid harming intra-organizational

17 | P a g e

network relationships. The results of this study indicate that it may be possible to achieve the difficult task of the correct balance between increases in autonomy necessary to achieve required increases in interorganizational network relationships while also accomplishing the correct intra-organizational network relationships. The results are limited due to the sample and the conceptual focus of network theory taken in this study. The results derive from Western European countries that are geographically close and they are developed economies. Even although the three host countries represent a cross-institutional setting (Fenton-O’Creevy et al., 2008), and a variation in country size (Gammelgaard, McDonald, Tüselmann, Dörrenbächer, & Stephan, 2009), the sample restrictions mean that the conclusions for host locations with different characteristics are limited. Future research needs to encompass a wider range of countries to include issues such as the economic strength, and comparative factor endowments of the host country (Kogut, 1985), and the value chain distribution in relation to location (Mudambi, 2008). From a conceptual point of view, future research on subsidiary evolution should consider the condition under which autonomy is an obstacle to achieve overlapping network benefits between intra- and inter-organizational networks. The question, more specifically, is under which conditions increased subsidiary autonomy hinders the subsidiary from benefiting from external knowledge sources. Future research in this direction, using the type of interaction processes developed in this paper, would help to reveal the conditions when autonomy disconnects the subsidiary from its MNC contexts, and if this creates a hindrance to intra-MNC knowledge transfers. Another area of possible area for research is the conditions under which increased autonomy, in the context of the interactions processes identified in this paper, facilitates entrepreneurial activities that help to make subsidiaries centres of excellence.

18 | P a g e

References Aghion, P. & Tirole, J. (1997). Formal and real authority in organizations. Journal of Political Economy. 105(1), pp. 1-29. Andersson, U., Björkman, I., & Forsgren, M. (2005).Managing subsidiary knowledge creation: The effect of control mechanisms on subsidiary local embeddednes. International Business Review, 14(5), 521-538. Andersson, U., Forsgren, M., & Holm, U. (2002). The strategic impact of external networks: Subsidiary performance and competence development in the multinational corporation. Strategic Management Journal, 23(11), 979-996. Andersson, U., Forsgren, M., & Holm, U. (2007). Balancing subsidiary influence in the federative MNC: A business network view. Journal of International Business Studies, 38(5), 802-818. Astley, W. G., & Sachdeva, P. S. (1984). Structural sources of intraorganizational power: A theoretical synthesis. Academy of Management Review, 9(1), 104-113. Astley, W. G., & Zajac, E. J. (1991). Intraorganizational power and organizational design: Reconciling rational and coalitional models of organization. Organization Science, 2(4), 399-411. Birkinshaw, J. & Hood, N. (1997). An empirical study of development processes in foreign-owned subsidiaries in Canada and Scotland. Management International Review, 37(4), 339-364. Birkinshaw, J., & Hood, N. (1998). Multinational subsidiary evolution: Capability and charter change in foreign-owned subsidiary companies. Academy of Management Review, 23(4), 773-795. Birkinshaw, J. & Hood, N. (2000). Characteristics of foreign subsidiaries in industry clusters. Journal of International Business Studies, 31(1): 141-154. Birkinshaw, J., Hood, N., & Young, S. (2005). Subsidiary entrepreneurship, internal and external competitive forces, and subsidiary performance. International Business Review, 14(2), 227-248. Birkinshaw, J., & Morrison, A. J. (1995). Configurations of strategy and structure in subsidiaries of multinational corporations. Journal of International Business Studies, 26(4), 729-753. Björkman, I., Fey, C. & Park, H.J. (2007). Explaining MNC Subsidiary HRM Practices: Evidence from a Three-Country Study. Journal of International Business Studies. 38(3), 430-46. Blazejewski, S. & Dorow, W. (2003). Managing organizational politics for radical change: the case of Beiersdorf-Lechia S. A., Poznan. Journal of World Business, 38, 204-223. Boehe, D. M. (2007). Product development in MNC subsidiaries: Local linkages and global interdependencies. Journal of International Management, 13(4), 488-512. Borsboom, D., Mellenbergh, G. J. & van Heerden, J.(2003). The theoretical status of latent variables. Psychological Review, 110(2), 203-219. Bouquet, C., & Birkinshaw, J. (2008). Managing power in the multinational corporation: How low-power actors gain influence. Journal of Management, 34(3), 477-508. Brewster, C. and Hegewisch, A. (1994). Policy and Practice in European Human Resource Management. London: Routledge Buckley, P J. (2009). Internalisation thinking: From the multinational enterprise to the global factory. International Business Review, 18(3), 224-235. Burt, R. S. (1992). Structural Holes. Cambridge MA: Harvard University Press. Cantwell, J. & Mudambi, R. (2005). MNE competence-creating subsidiary mandate. Strategic Management Journal, 26, 1109-1128. Cassel, C., Hackl, P., & Westlund, A. H. (1999). Robustness of partial-least squares method for estimating latent variable quality structures. Journal of Applied Statistics, 26(4), 435-446. Chang, S-J., Witteloostuijn, A. van, & Eden, L. (2010). From the Editors: Common method variance in international business research. Journal of International Business Studies, 41, 178-184. Coeurderoy, R., & Murray, G. (2008). Regulatory environments and the location decision: Evidence from the early foreign market entries of new-technology-based firms. Journal of International Business Studies, 39(4): 670–687. Cohen, W. M. & Levinthal, D. A. (1990). Absorptive capacity: A new perspective on learning and innovation. Administrative Science Quarterly, 35(1), 128-152. Demirbag, M., Tatoglu, E., & Glaiser, K. (2007). Factors influencing the perceptive performance: The case of Western FDI in an emerging market. International Business Review, 16, 310-36. Dikova, D. (2009). Performance of foreign subsidiaries: does psychic distance matter? International Business Review, 18(1), 38-49. Dikova, Desislava & Arjen van Witteloostuijn (2007). Foreign direct investment mode choice: Entry and establishment modes in transition economies. Journal of International Business Studies 38(6): 1013-1033. Dillman, D.A. (1991). Design and administration of mail surveys. Annual Review of Sociology, 15, 225-249. DiMaggio, P. J. & Powell, K. (1983). The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields. American Sociological Review, 13(2), 147-160. Dörrenbächer, C. & Gammelgaard, J. (2010). Multinational corporations, inter-organizational networks and subsidiary charter removals. Journal of World Business, 45, 206-216. Dörrenbächer, C. & Geppert, M (2011). Politics and Power in the Multinational Corporation: The Role of Institutions, Interests and Identities. Cambridge: Cambridge University Press

19 | P a g e

Duanmu, J-L., & Fai, F.M. (2007). A processual analysis of knowledge transfer: from foreign MNEs to Chinese suppliers. International Business Review, 16(4), 449-473. Dunning, J. (2009). Location and the multinational enterprise: John Dunnings’s thoughts on receiving the Journal of International Business Studies 2008 Decade Award. Journal of International Business Studies, 40(1), 20-34. Elango, B. & Pattnaik, C. (2007): Building capabilities for international operations through networks: a study of Indian firms. Journal of International Business Studies, 38:4, 541-555. Ellis, P. (2007). Distance, dependence and diversity of markets: Effects on market orientation. Journal of International Business Studies, 38(3), 374-386. Enright, M. J., & Subramanian, V. (2007). An organizing framework for MNC subsidiary typologies. Management International Review, 47(6), 895-924. Eriksson, K., Johanson, J., Majkgård, A. & Sharma, D. D. (1997). Experiental knowledge and cost in the internationalization process. Journal of International Business Studies, 28(2), 337-360. Fenton-O’Creevy, M., Gooderham, P., & Nordhaug, O. (2008). Human resource management in US subsidiaries in Europe and Australia: Centralisation or autonomy? Journal of International Business Studies, 39(1), 151-166. Fey, C. F., Morgulis-Yakushev, S., Park, H. J., & Björkman, I. (2009). Opening the black box of the relationship between HRM practices and firm performance: A comparison of MNE subsidiaries in the USA, Finland, and Russia. Journal of International Business Studies, 40(4), 690-712. Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement errors. Journal of Marketing Research, 18, 39-50. Forsgren, M., Hägg, I., Håkansson, H., Johanson, J., & Mattsson, L-G. (1995). Firms in Networks - A New Perspective on Competitive Power. Uppsala: Acta Universitatis Upsaliensis Vol. 38. Forsgren, M., Holm, U., & Johanson, J. (2005). Managing the embedded multinational: A business network view. Chelterham: Edward Elgar. Fowler, Jr, F.J. (1993). Survey Research Methods. Newbury Park, CA: Sage Frost, T. (2001). The geographic sources of foreign subsidiaries’ innovations, Strategic Management Journal, 22, 101-123. Frost, T. S., Birkinshaw, J. M., & Ensign, P. C. (2002). Centers of excellence in multinational corporations. Strategic Management Journal, 23(11), 997-1018. Gammelgaard, J., McDonald, F., Tüselmann, H., Dörrenbächer, C., & Stephan, A. (2009). Subsidiary role and skilled labour effects in small developed countries. Management International Review, 49(1), 27-42. Garcia-Pont, C., Canales, J. I., & Noboa, F. (2009). Subsidiary strategy: The embeddedness component. Journal of Management Studies, 46(2), 182-214. Glaister, K., & Buckley, P. (1999). Performance relationships in UK international alliances. Management International Review, 39(123-47). Ghoshal, S., & Bartlett, C. A. (2005). The multinational corporation as an inter-organizational network, in S. Ghoshal, S., & D. E. Westney, (Eds.), Organization theory and the multinational corporation (pp. 68-92). Houndmills, Bas.: Palgrave MacMillan:, 68-92 Gong, Y., Shenkar, O., Luo, Y., & Nyaw, M-K. (2005). Human resources and international joint venture performance: A system perspective. Journal of International Business Studies, 36, 505-18. Giroud, A. (2007). MNEs vertical linkages: The experience of Vietnam after Malaysia. International Business Review, 16(2), 159-176. Giroud, A., & Scott-Kennel, J. (2009). MNE linkages in international business: A framework for analysis. International Business Review, 18, 555-566. Gnyawali, D., Singal, M., & Mu, S. (2009). Knowledge ties among subsidiaries in MNCs: a-multi-level conceptual model. Journal of International Management, 15, 387-400. Granovetter, M. S. (1973). The strength of weak ties. American Journal of Sociology, 78(6), 1360-1380. Granovetter, M. S. (1985). Economic action and social structure: The problem of embeddedness. American Journal of Sociology, 91(3), 481-510. Griffith, D. & Myers, M. (2005). The performance implications of strategic fit of relational norm governance strategies in global supply chain relationships. Journal of International Business Studies, 36(3), 254-269. Guest, D. E., Michie, J., Conway, N., & Sheehan, M. (2003). Human resource management and corporate performance in the UK. British Journal of Industrial Relations, 41(2): 291-314. Hansen, M. T. (1999). The search-transfer problem: The role of weak ties in sharing knowledge across organization subunits. Administrative Science Quarterly, 44(1). 82-111. Harzing, A. W. (1997). Response rates in international mail surveys: Results of a 22-country study. International Business Review, 6(6), 641-665. Harzing, A.W. (2000). Cross-national Industrial Mail Surveys: Why do Response Rates Differ Between Countries, Industrial Marketing Management, 29, (3), 243-254 Harzing, A.W. and Noorderhaven, N.G. (2006). Knowledge flows in MNCs: An empirical test and extension of Gupta & Govindarajan's typology of subsidiary roles, International Business Review, 15(3), 195-214. Hedlund, G. (1986). The Hypermodern MNC – A Heterarchy? Human Resource Management, 25(12), 9-35.

20 | P a g e

Holm, U., Holmström, C., & Sharma, D. (2005), Competence development through business relationships or competitive environment?: Subsidiary impact on MNC competitive advantage. Management International Review, 45(2), 197-218. Holm, U., & Pedersen, T. (2000). The emergence and impact of MNC centres of excellence – A subsidiary perspective. Houndsmill, Bas.: MacMillan Press. Hulland, J. (1999). Use of partial least squares (PLS) in strategic management research: A review of four recent studies. Strategic Management Journal, 20, 195-204. Håkanson, L. & Nobel, R. (2001). Organizational characteristics and reverse technology transfer. Management International Review, 41(4). 395-420. Inkpen, A. C., & Birkinshaw, J. (1994). International joint ventures and performance: an interorganizational perspective. International Business Review, 38(3), 201-217. Jarillo, J. C. & Martínez, J. I. (1990). Different roles for subsidiaries: The case of multinational corporations in Spain. Strategic Management Journal, 11(7), 501-513. Jindra, B., Giroud, A., & Scott-Kennel, J. (2009). Subsidiary roles, vertical linkages and economic development: Lessons from transition economies. Journal of World Business, 44(2), 167-179. Johanson, J. & Vahlne, J-E. (2009). The Uppsala internationalization process model revisited: From liability of foreignness to liability of outsidership. Journal of International Business Studies, 40(9), 1411-1431. Johnson, W. H. A. & Medcof, J. W. (2007). Motivating proactive subsidiary innovation: Agent-based theory and socialization models in global R and D. Journal of International Management, 13(4), 472-487. Kogut, B. (1985). Designing global strategies: Profiting from operational flexibility. Sloan Management Review, 27(1): 27-38. Kogut, B,. & Zander, U. (1992). Knowledge of the firm, combinative capabilities, and the replication of technology. Organization Science, 3(3), 383-397. Kostova, T. & Zaheer, S. (1999). Organizational legitimacy under conditions of complexity: The case of the multinational enterprise. Academy of Management Review, 24(1), 64-81. Lindstrand, A., Eriksson, K., & Sharma, D. D. (2009). The perceived usefulness of knowledge supplied by foreign client networks. International Business Review, 18(1), 26-37. Lovett, S. R., Pérez-Nordtvedt, L., & Rasheed, A. A. (2009). Parental control: A study of U.S. subsidiaries in Mexico. International Business Review, 18(5), 481-493. Luo, Y. (2003). Market-seeking MNEs in an emerging market: How parent-subsidiary links shape overseas success. Journal of International Business Studies, 34, 290-309. Luo, Y. (2007). An integrated anti-opportunism system in international exchange. Journal of International Business Studies, 38, 855877. Madhavan, R., & Iriyama, A. (2009). Understanding global flows of venture capital: human networks as the “carrier wave” of globalization. Journal of International Business Studies, 40(8), 1241-1259. Marginson, P., Edwards, P., Edwards, T., Ferner, A. and Tregaskis, O. (2010). Employee Representation and Consultative Voice in Multinational Companies Operating in Britain. British Journal of Industrial Relations, 48: 151–180. McDonald, F., Tüselmann, H., Dimitratos, P. and Voronkova, S. (2005) ’The strategic development of foreign owned subsidiaries and employment in host locations in the UK’, Environment and Planning C: Government and Policy, 23(6), 867-882. McDonald, F., Warhurst, S., and Allen, M. ( 2008). Autonomy, Embeddedness and the Performance of Foreign Owned Subsidiaries, Multinational Business Review, 16(3), 73-92 Miller D., Lee. I., Chang, S., & Le Betron-Miller, I. (2009). Filling the institutional void: The social behaviour and performance of family vs. non-family technology firms in emerging markets. Journal of International Business Studies, 40, 802-817. Money, R. B., & Graham, J. L. (1999). Salesperson performance, pay, and job satisfaction: Tests of a model using data collected in the United States and Japan. Journal of International Business Studies, 30(1), 149-172. Monteiro, L. F., Arvidsson, N., & Birkinshaw, J. (2008). Knowledge flows within multinational corporations: Explaining subsidiary isolation and its performance implications. Organization Science, 19(1), 90-107. Mu, S.C,, Gnyawali, D.R., & Hatfield, D. E., (2007). Foreign subsidiaries’ learning from local environments: An empirical test. Management International Review, 47(1), 79-102. Mudambi, R. (2008). Location, control and innovation in knowledge-intensive industries. Journal of Economic Geography, 8, 699725. Mudambi, R. & Navarra, P. (2004). Is knowledge power? Knowledge flows, subsidiary power and rent-seeking within MNCs. Journal of International Business Studies, 35, 385-406. Nohria, N. & Ghoshal, S. (1997). The Differentiated Network - Organizing Multinational Corporations for Value Creation. San Francisco: Josey-Bass Publishers.. Noorderhaven, N., & Harzing, A-W. (2009). Knowledge-sharing and social interaction within MNEs. Journal of International Business Studies, 40(5), 719-771. Peng, M. W. E. & York, A. S. (2001). Export intermediary firms: A note on export development research. Journal of International Business Studies, 32(2), 327-346. Podsakoff, P. M., & Organ, D. W. (1986). Self-reports in organizational research: Problems and perspectives. Journal of Management, 12(4), 531-544. Porter, M. E. (1994). The role of location in competition. Journal of the Economics of Business, 1(1), 35-39.

21 | P a g e

Powell, W., Koput, K., & Smith-Doerr, L. (1996). Inter-organizational collaboration and the locus of innovation: Networks of learning in biotechnology. Administrative Science Quarterly, 41(1), 116–145. Prahalad, C. K. & Doz, Y. L. (1987). The Multinational Mission: Balancing Local Demands and Global Vision. New York: The Free Press Pudelko, M.; Harzing, A.W. (2007). How European is Management in Europe? An Analysis of Past, Present and Future Management Practices in Europe, European Journal of International Management, 1(3),206-224. Rugman, A. M. & Verbeke, A. (2001). Subsidiary-specific advantages in multinational enterprises. Strategic Management Journal, 22, 237-250. Schwens, C. and Kabst R (2009). How early opposed to late internationalizers learn: experience of others and paradigms of interpretation. Journal of International Business Studies 40(3), 509–522. Scott, R.W. (1995). Institutions and Organizations. Thousand Oaks: Sage. Sönke A. (2010). PLS and success factor studies in marketing, chap. 18, V. Esposito Vinzi, W.W. Chin, J. Henseler, & H. Wang (Eds.), Handbook of Partial Least Squares. Berlin: Springer. Taggart, J. H., & Hood, N. (1999). Determinants of autonomy in multinational corporation subsidiaries. European Management Journal, 17(2), 226-236. Thompson E. (2003). A grounded approach to identifying national competitive advantage: a preliminary exploration. Environment and Planning A, 35, 631 – 657. Tung, R. L., & van Witteloostuijn, A. (2008). From the Editors: What makes a study sufficiently international? Journal of International Business Studies, 39(2), 180-183. Tüselmann, H., McDonald, F. and Thorpe, R. (2006) ’The emerging model of employee relations in German overseas affiliates: A role model for international operation?’, Journal of World Business, 41(1), 66-80. Venkatraman, N., & Ramanujam, V. (1986). Measurement of business performance in strategy research: a comparison of approaches. Academy of Management Review, 11(4), 801-814. Vernaik, S., Midgley, D. F., & Devinney, T. M. (2005). Dual paths to performance: The impact of global pressures on MNC subsidiary conduct and performance. Journal of International Business Studies, 36, 655-675. Vitols, S. (2001) “Unternehmensführung und Arbeitsbeziehungen in deutschen Tochtergesellschaften großer ausländischer Unternehmen”. Forum Mitbestimmung und Unternehmung: Bertelsmann Stiftung: Gutersloh. Young, S., & Tavares, A. T. (2004). Centralization and autonomy: Back to the future. International Business Review, 13, 215-237. Zaheer, S. & Mosakowski, E. (1997). The dynamics of the liability of foreignness: A global study of survival in financial services. Strategic Management Journal, 18(6), 439-464. Wang, J., Liu, X., and Li, X. (2009). A dual-role typology of multinational subsidiaries. International Business Review, 18, 578-591. White, R. E., & Poynter, T. A. (1984). Strategies for foreign-owned subsidiaries in Canada. Business Quarterly, 49(2), 59-69. Zaheer, S. (1995). Overcoming the liability of foreignness. Academy of Management Journal, 38, 341-363.

22 | P a g e

Figure 1: Conceptual Model

Changes in Number of relationships within corporation

Changes in Frequency of relationships within corporation

Intra-org

Changes in strategic decision making rights

Networks H6

H2 H4

Performance

Autonomy Changes in operational decision making rights

H3 H5

Inter-org

H1

Networks

Changes in Number of relationships outside corporation

23 | P a g e

Changes in frequency of relationships outside corporation

Changes in performance (compared to market competitors)

Table 1: Subsidiary Profile (%) _____________________________________________________________________________________ Activity1

Host Country Britain Germany Denmark

33.8 23.4 42.1

Production of Goods or Services Sales / Distribution Ancillary Service Functions R&D / New Product Development Others

19.1 15.4 13.1 8.2 6.3 6.2 5.4 26.3

Years in Foreign Ownership

Home Country Germany Netherlands Sweden Switzerland USA Japan France Others

5 – 10 11 – 20 21 – 30 >30

32.1 31.8 16.9 17.8 2

Type of Industry

Size (Employment) 1 – 10 11 – 100 > 100

25.6 42.6 17.9 3.3 10.4

High / Medium High Tech Manufacturing Medium Low / Low Tech Manufacturing Knowledge Intensive Service Industries Less Knowledge Intensive Service Industries

25.7 53.2 21.1

46.0 24.1 17.8 12.1

Nationality of Managing Director Entry Mode Greenfield Investment Acquisition

Host Country National Home Country / Third Country National

62.8 37.2

76.5 23.5

___________________________________________________________________________________________________________ Notes: 1. Distribution of employment according to activity 2. Based on 2 digit NACE classification and collapsed into industry based on OECD technology and knowledge intensity industry classification.

Table 2: Composite Reliabilities, Cronbach Alphas and R squares Composite

Cronbachs

R squares

Reliabliity

Alphas

Autonomy

0.94

0.93

-

Inter-organizational Networks

0.84

0.79

0.16

Intra-organizational Networks

0.87

0.82

0.20

Performance

0.84

0.75

0.19

Composite should be above 0.70 for each construct (Fornell and Larcker, 1981). Cronbach alpha values should be above 0.70 (Hulland, 1989). When using PLS technique, one variable is ‘locked’ and R-squares are reported in relation to this variable.

24 | P a g e

Table 3: PLS test Path Coefficient Mean

Total Effect Mean

t-statistics Total Effects

H1: inter- organizational network relationships AND performance

0.36

0.40

7.00***

H2: Intra-organizational network relationships AND performance

0.08

H3: inter- organizational network relationships AND Intra-organizational network relationships

0.49

H4: autonomy performance

0.07

AND

(7.15)*** 0.07

1.18 (1.82)*

0.49

6.84*** (6.99)***

0.23

3.17*** (2.56) ***

H5: autonomy AND : interorganizational network relationships

0.40

H6: autonomy AND Intraorganizational network relationships

- 0.08

0.40

6.29*** (5.30)***

0.11

1.37 (0.53)

T values in parentheses are without controls Means are computed via a Bootstrapping procedure, 1000 samples are constructed and the model is estimated 1000 times. The Mean is the mean of the 1000 coefficients that have been obtained in the 1000 estimations. And standard error (deviation) comes also from these 1000 coefficients. Standard deviations are in all cases between 0.06 and 0.08.

25 | P a g e

Table 4: Summary of the results of the tests of hypotheses

Hypotheses

In whole sample

In countries

H1 – inter-org networks and performance

Confirmed with and without controls

Same result as in whole sample

H2 – intra-org networks and performance

Confirmed without controls

Same result as in whole sample

H3 – inter and intra-org networks

Confirmed with and without controls

Same result as in whole sample

H4 – autonomy performance

and

Confirmed with controls

Supported in Denmark and UK but not for Germany

H5 – autonomy inter-org networks

and

Confirmed with and without controls

Same result as in whole sample

Not confirmed

Supported for Denmark and Germany but not for UK

H6 autonomy and intra-org networks

26 | P a g e