Coordination structures and innovative performance in global R&D labs

Coordination structures and innovative performance in global R&D labs

Ajax Persaud


Recently, multinational corporations have been globalizing their research and development (R&D) activities as a way to harness and leverage the scientific and technological capabilities residing in their subsidiary labs around the world and to tap into the knowledge base of overseas locations. Prior research suggests that proper coordination and control mechanisms are critical for the realization of this goal. An empirical study was undertaken to ascertain the extent to which four coordination and control mechanisms-autonomy, socialization, formalization, and communication-enhance the innovative capabilities of multinationals. R&D executives from 79 research facilities of 27 North American, European, and Japanese multinational corporations participated in this study. The results from multivariate regression and factor analysis indicate that innovative capability is influenced by the level of autonomy of the labs, the extent of socialization, and the effectiveness of in-person communication between the HQ and subsidiary labs.


Recemment, les corporations multinationales ont mondialise leurs activates de recherche et developpement afin d’exploiter et de retenir les capacites scientifiques et technologiques au coeur de leurs laboratoires auxiliaires autour du monde. Des recherches anterieures suggerent qu’une bonne coordination et qu’un bon controle est critique pour la realisation de ce but. Une recherche empirique a ete effectue pour evaluer jusqu’a quel point les quatre mecanismes de coordination et de controle-autonomie, socialisation, formalisation, et communication-pourraient ameliorer les capacites innovatrices des compagnies multinationales. Les directeurs de recherche et developpement parvenant de 79 laboratoires de recherche de 27 corporations nordamericaines, europeennes, et japonaises ont participe a cette etude. Les resultats de regressions multivariees et d’analyse de facteurs indiquent que les capacites innovatrices sont influencees par le niveau d’autonomie dans les laboratoires, les limites de la socialisation, et par l’efficacite de l’intercommunication entre le siege social de l’entreprise et les laboratoires auxiliaires.

Over the last 15 years an increasing number of multinational corporations (MNCs) from around the world have been decentralizing their research and development (R&D) operations in order to leverage the innovative capabilities residing in their worldwide subsidiary labs. The importance of building technological capabilities and ensuring effective performance has been discussed extensively in the innovation literature (e.g. Bartlett & Ghoshal, 1989; Tushman & Anderson, 1997). It has been argued that effective cross-border coordination and integration is a critical element in harnessing globally distributed scientific and technological knowledge for new and rapid innovations (Bartlett & Ghoshal, 1989; Chiesa, 1996; Medcof, 1998). A substantial amount of research has been published on the organizational structures employed by MNCs to coordinate, control, and integrate globally distributed R&D facilities (Asakawa, 1996; Brockhoff & Schmaul, 1996; Chiesa, 1996). However, very little research exists on the extent to which these structural elements influence the innovative capabilities of MNCs. This research contributes to fulfilling this need by examining empirically the extent to which the mechanisms used to coordinate and control the R&D activities of globally distributed labs influence the innovative capabilities of the labs. Based on a comprehensive review of the literature on the organization of global R&D and borrowing from classical organizational theory research, four coordination mechanisms are investigated in this study. These constructs are (a) autonomy of R&D labs in decision-making, (b) formalization of decision-making, (c) socialization of decision-making, and (d) communication among R&D labs. Communication among R&D labs is decomposed into communication between subsidiary labs and the headquarters (HQ) and communication among subsidiary labs themselves (inter-subsidiary labs communication). Justification of the selected constructs is provided in the literature review and conceptual model sections of this paper.

The paper is divided into six sections. The next section provides a review of the literature on the organization of global R&D; the third section presents the conceptual model of the study; the fourth section describes the methodology used in collecting and analyzing the data; the fifth section presents the statistical results; and the final section discusses the research and managerial implications of these results and offers the conclusion.

Literature Review

Previous research on the globalization of R&D focused on issues such as the reasons for globalization of R&D (De Meyer, 1993; Dunning, 1992; Granstrand, Hakanson, & Sjolander, 1992; Pearce, 1989), the establishment processes of R&D activities in foreign countries (De Meyer & Mizushima, 1989; Gerybadze & Reger, 1999; Hakanson, 1992; Kuemmerle, 1997; Reger, 1997), the types of activities performed by foreign R&D labs (Casson, Pearce, & Singh, 1992; Florida, 1997; Hewitt, 1980; Pearce, 1989; Ronstadt, 1977), and the organizational structures of global R&D labs (Brockhoff & Schmaul, 1996; Chiesa, 1996; Gassmann & Zedwitz, 1999). This review focuses on the last strand of research, the organization and management of global R&D, since this is the most relevant aspect in terms of the present study.

Research on the organization and management of global R&D focused on two interdependent issues. Developing an understanding of the structural relationships between the HQ lab and overseas labs, and among overseas labs themselves, and developing an understanding of the role overseas labs play within the broad range of R&D activities undertaken by the MNC group. Medcof (1998) provides an excellent summary of this body of research, highlighting the contradictions and suggesting ways to resolve them.

In an attempt to develop a deeper understanding of how MNCs organize their global R&D, researchers have focused on the structural elements characterizing the complex relationships among worldwide R&D labs. Three interdependent issues dominate this body of research. One issue concerns the level of autonomy granted to overseas subsidiary labs to make strategic decisions and how autonomous labs relate to the rest of the labs within the MNC group. The second issue concerns the coordination structure used to integrate and control globally distributed R&D labs. The third issue relates to the communication patterns among the labs. The research on these issues is discussed next.

A thorny question in the literature is which factors determine the extent of autonomy and strategic significance of the R&D tasks observed among the labs? Bartlett and Ghoshal (1989) and Nohria and Ghoshal (1997) suggested that the labs’ administrative heritage, resource levels, and the complexity of the environment in which they operate are possible explanations. They also observed that the level of autonomy of overseas labs seems to vary with the innovation tasks performed. R&D labs that only created innovations had the greatest autonomy while those that created, adopted, and diffused innovations had intermediate amounts of autonomy. De Meyer and Mizushima (1989) identified three factors that determine the amount of local autonomy given to overseas R&D labs. These are (a) a company’s orientation towards centralization, that is, companies that were highly centralized tended to centralize their global R&D as well; (b) time pressure (the greater the time pressure to complete an R&D project, the greater the tendency towards centralization); and (c) size of the R&D lab (the smaller the overseas R&D lab, the stronger the corporate control exercised over it). Behrman and Fischer (1980) identified four management styles in relation to foreign R&D: absolute centralization, participative centralization, supervised freedom, and total freedom. Participative centralization and supervised freedom were the two most commonly used management styles. Firms with a home-market orientation (emphasize the home market) appeared to be more centralized than firms with a host– market or global market orientation (emphasize the foreign or world market). Casson et al. (1992) characterized the relationship between the HQ R&D facility and overseas R&D labs that performed basic/original research as “supervised freedom”. The research also found that labs perform different roles ranging from listening posts to new product development, to basic research (e.g. Chiesa, 1996; Florida, 1997; Gerybadze & Reger, 1999). These authors point out that although the labs perform different roles, each role is an important aspect of the internationalized innovation process, therefore, they are expected to contribute to the innovation process in varying degrees.

One outcome of establishing globally dispersed R&D labs is that coordination becomes more difficult due to geographic, time, and cultural differences. Advances in information and communications technologies have helped to mitigate some of the coordination difficulties. Nonetheless, significantly more time and resources are needed to achieve effective coordination, compared to the traditional approach in which R&D is centralized at the headquarters.

Nohria and Ghoshal (1997) observed that MNCs use a combination of formalization and socialization mechanisms to coordinate and control the activities of decentralized subsidiaries. Many researchers argue that coordination and control of globally dispersed R&D labs are achieved more effectively through socialization or normative cultural control rather than by rules and edicts (Baliga & Jaeger, 1984; Kanter, 1988; Ouchi, 1980). According to Coleman (1990), socialization is important for building trust and shared values, learning the organization’s code of conduct, and monitoring in order to ensure compliance with the norms. Both trust and shared values reduce coordination costs within organizations (Ring & van de Ven, 1992). Nohria and Ghoshal (1997) argued that socialization allows the multinational to leverage its worldwide pool of resources much more effectively than through the formal structure. Hence, as the need to tap into the innovative potential of its overseas R&D labs increases, the multinational must devote more resources to building its social capital-the individual’s network of contacts.

Within globally dispersed R&D labs, communication must be managed both between the HQ and overseas labs and among the overseas labs. The importance of internal communications for innovation is well established in both the theoretical and empirical literature on innovations (Tushman & Anderson, 1997). Stock, Greis, & Dibner (1996) examined the communication patterns between HQ and subsidiaries of European and Japanese biotechnology MNCs in the U.S. with respect to processing technical information. They observed that a global setting substantially increases the complexity and difficulty of such processing; however, the effectiveness of such processing greatly influences innovative performance. Nohria and Ghoshal (1997) observed a connection between HQ-subsidiary communication patterns and innovations within their four typologies of innovations. Local-for-local innovations are characterized by high-density communication within subsidiaries. Local– for-global innovations have high-density communication both within and among subsidiaries. Centre-for-global innovations are characterized by high-density communication between HQ and subsidiaries. Global-for-global types have high-density communication within subsidiaries, among subsidiaries, and between HQ and subsidiaries. Bartlett and Ghoshal (1989) found that subsidiaries with higher levels of inter-lab communications were more effective in creating innovations.

Strong inter-subsidiary communication is facilitated through organizational directives from corporate HQ and the use of advanced communication technologies. Hakanson and Zander (1988) noted that extensive intersubsidiary communication, particularly lateral information flows among subsidiaries, is a primary contributing factor for achieving global coordination among foreign subsidiaries. Marschan (1996) observed that the adoption of less-hierarchical structures has different consequences on communication patterns at different organizational levels. For instance, top management communication patterns improved, but middle and operating level managers experienced barriers in exchanging information. Communications at the top management level became less formal, while at the middle and operating levels became more formal.

Conceptual Model

From the preceding literature review, it is clear that the four key constructs used to analyze the organization and management of global R&D labs are autonomy, socialization, formalization, and communication among R&D labs. These studies also suggest a link between these constructs and the creative potential of the labs. In addition, organizational theory research provides theoretical justification for these four constructs as coordination and control mechanisms in complex organizations. Since the landmark study of the Aston Group (Pugh, Hickson, & Turner, 1968; Pugh, Hickson, & Hinings, 1969), centralization and formalization have become central constructs in the analysis of internal relations in complex organizations. Similarly, since the studies of Edstrom and Galbraith (1977) and Ouchi (1980), normative integration or socialization has been treated as another primary structural element in the analysis of multi-unit organizations. Studies by Lawrence and Lorsch (1967), Burns and Stalker (1961), Aiken and Hage (1968), Rogers (1995), and Burt (1982) emphasize the importance of effective communications across units for the creation of innovations. Thus, it can be plausibly asserted that the four constructs-autonomy, formalization, socialization, and communication-together constitute a fairly comprehensive characterization of the nature of the relationships among globally dispersed R&D labs of MNCs.

The model depicted in Figure 1 suggests that synergistic innovative capabilities in MNCs may be enhanced through the establishment of appropriate coordination and control mechanisms as reflected in the four constructs. The model also suggests that the impact of the relationship between coordination structures and synergistic innovative capabilities may be moderated by uncertainty of environments in which the labs operate, the resource levels of the labs, cultural diversity, and the level of trust among the labs working together. Each of the constructs shown in Figure 1 is described below.

Dependent Construct-Synergistic Innovative Capability

The word synergy is derived from the Greek word synergos, and means to work together. It connotes combined effects or the outcomes of cooperative interactions. Though it is often associated with the slogan “the whole is greater than the sum of its parts,” it would be more accurate to say that synergy refers to effects that the parts cannot achieve alone, effects that are interdependent. Extrapolating from this conceptualization of synergy, synergistic innovative capability is defined as the ability to create new knowledge or to recombine existing knowledge to create new products, processes, and technologies more efficiently by exploiting the unique capabilities of worldwide R&D labs. In this study, synergistic innovative capability is measured by ascertaining the outcomes in each R&D lab that can be attributed to the labs working interdependently. A number of such possible outcomes are discussed below.

First, through collaboration, the labs could participate in a larger number of projects or even major projects in terms of resources, complexity and newness, which might otherwise not be possible if the labs were working independently of each other. In addition, some labs may be able to broaden the scope of their research efforts by applying R&D knowledge to expand their product range in terms of new or related products. For example, in one company, a lab with expertise in polymer and fibre technology has teamed up with another lab with expertise in chemical technology to develop a range of agricultural products. Hence, both labs were able to enter new lines of business because they combined their existing knowledge capabilities.

Second, advances in information and communication technologies (ICT) substantially increase the opportunities for virtual collaboration among globally dispersed R&D labs, which could result in more efficient utilization of R&D resources. For example, discussions with a global R&D manager of a Canadian high technology firm revealed that during the day the company’s Canadian R&D lab uses its facilities to work on projects and downloads the results to a database at the end of the day. At night, while the Canadian lab is closed, the company’s R&D lab in India accesses the results on the database, uses the Canadian data to continue working on the project, and then downloads its result to the database which is accessed the next day by the Canadian lab. This type of collaboration could result in a reduction in development time for projects.

Third, by working interdependently it is easier for labs to share complementary know-how, skills, and technologies, as well as access the external networks that each lab has within their country of operation. For example, the Ottawa-based R&D lab of a Canadian company could access the external knowledge networks its R&D lab in India has with other Indian companies, universities, and government research institutes. The multiplier effect of such access could be significant to the MNC in terms of generating new innovations, developing new products faster, entering new markets more easily, and improving internal production and administrative processes.

Since no previous measure of synergistic innovative capability was found in the literature, the construct was developed by piecing together various indicators of innovations from previous studies (Tushman & Anderson, 1997). This construct was measured on a 5-point scale consisting of 13 items as shown in Box 1 of the Appendix.

Independent Constructs

Autonomy is defined as the degree to which a subsidiary R&D lab has control over the strategic decisions affecting its direction and operations. It is the obverse of centralization where decision-making is centralized. As noted in the literature review, autonomous labs are more likely to perform work of greater strategic significance and, therefore, may be able to enhance their innovative capabilities more than labs that perform adaptive R&D. Autonomous labs also have the flexibility to cooperate with other R&D labs within their group and to share information and expertise, thereby improving their innovative capabilities. Thus, within the multinational system autonomy is expected to have a positive effect on innovative capabilities (Gates & Egelhoff, 1986; Kanter, 1988). Additionally, Thompson (1967) and Aiken and Hage (1968) contended that a high degree of bureaucratic control inhibits creativity and innovation.

Generally, managers of autonomous labs have greater freedom to establish formal or informal arrangements with other labs to work on research projects that match their technical and organizational expertise. The incentive to collaborate would be greater where the labs have complementary skills, the cost of the project is more than what a single lab could afford, or the risk is too much for a single lab to undertake. By working together, it is possible that the participating labs’ synergistic innovative capability could be enhanced substantially. On the other hand, because autonomy shifts the focus of power asymmetrically in favour of subsidiary labs, it can lead to situations where some labs may want to act opportunistically, pursuing their own research projects independently rather than working with other labs. Such behaviours could reduce the potential contribution to synergistic innovative capability because the lab may pursue projects with limited applications elsewhere within the MNC group (Gassman & Zedwitz, 1999; Gerybadze & Reger, 1999). Using the approach advanced by De Bodinat (1975), autonomy of subsidiary was measured on a 5-point scale consisting of 15 items as shown in Box 2 of the Appendix.

Formalization refers to decision-making based on formal systems, established rules, and prescribed procedures. It provides a structured context for collaboration among labs (Burgleman, 1983). However, formalization also reduces the power of both the HQ and subsidiary labs as it constrains the exchange relation to an impersonal set of rules that may be used to curtail unwanted behaviours (Hedlund, 1981). In complex, uncertain, and rapidly changing environments, highly formalized behaviours as control and coordination mechanisms could inhibit flexibility and social interactions, which may be necessary to promote innovations (Mintzberg, 1979). In this sense, formalization may be negatively related to innovative capabilities (Gates & Egelhoff, 1986; Kanter, 1988). Adapting the approach developed by Aiken and Hage (1968) and modified by Nohria and Ghoshal (1997), formalization was measured on a 5– point scale consisting of 6 items as shown in Box 3 of the Appendix.

Socialization emphasizes the creation of common and shared understandings of goals, values, and practices to influence both how subsidiary labs perceive their interests and how they act. It is the process by which organizational members learn the value system, norms, and required behaviours of the organization (Schein, 1968). By having shared values, it is expected that organizational members would act in the interest of the overall corporation rather than their own individual units (Ouchi, 1980). In this manner, socialization facilitates innovations not only by motivating subsidiaries to be entrepreneurial but also by enhancing the headquarters’ responsiveness to subsidiaries’ needs and initiatives (Kanter, 1988). This construct was measured as the proportion of labs with goals, values, and mission similar to that of the HQ.

Communication among research labs is decomposed into HQ-subsidiary communications and intersubsidiary communications. Communication between the HQ and subsidiary labs represents the vertical flow of information. Communication with the HQ could be multilevel involving managers, scientists, and engineers, and multidimensional ranging from routine reporting to extensive project collaboration. Communications may be effected through electronic media, face-to-face, or both. Nohria and Ghoshal (1997) observed that dense HQ-subsidiary communication facilitated innovations by the subsidiary. Following Marschan (1996), HQ-subsidiary communication was measured on a 5-point scale using 6 items as shown in Box 4 of Appendix 2.

Inter-subsidiary labs communication represents the horizontal flow of information which connects the network of labs and has the potential to foster greater informal personal contacts. Effective inter-subsidiary communications promote a better understanding of the strengths, limitations, and complementarities of labs, which could lead to more efficient utilization of resources due to enhanced coordination and reduced duplication. Lateral networking encourages the sharing of knowledge, which could have a positive effect on innovative capability (Gassman & Zedwitz, 1999; Gerybadze & Reger, 1999; Hakanson & Zander, 1988). This construct was measured in the same way as in Box 4 of the Appendix.

Moderating Variables

The international management literature has identified several factors that influence the performance of subsidiaries. It is believed that some of these factors may have a moderating effect on the relationship between the four coordination mechanisms investigated in this study and synergistic innovative capability. Of these, the four most frequently cited factors are selected for further analysis on synergistic innovative capability. These factors are cultural diversity among R&D labs, the level of trust among managers and R&D staff of the labs, the resource levels of the labs, and the environment in which the labs operate. The impact of these factors on the creation of synergistic innovation capability is discussed next.

Cultural diversity. Because R&D activities are, by their very nature, open-ended and involve a lot of creativity and experimentation by highly qualified professionals, the success of most R&D collaborations depends heavily upon the cultural context within which they are executed (Neff, 1995). This is particularly significant when the collaboration involves R&D labs from many culturally diverse countries even though the labs may belong to the same MNC group. Cultural differences are a source of misunderstandings and conflicts in global R&D collaborations (Maccoby, 1995; Shane, 1994). Maccoby (1995) noted that cultural values could interfere with communications and cause distrust unless they are well understood. In a study of 4,000 managers from 32 countries in 8 organizations, Shane (1994) found that the styles used by managers to champion innovations vary significantly across cultures and that culturally appropriate styles are not always the ones that promote the most innovations. He argued that using a championing style in a country where the overarching cultural values are different could create conflicts that could reduce the innovativeness of organizations.

Cultural differences in management styles and the work ethics of scientists and engineers could also hinder the effectiveness of R&D collaborations and hence the ability of the MNC to create synergistic innovative capability. For example, some R&D staff and managers may avoid working with labs where the culture is different from their own because they feel uncomfortable or are unwilling to learn about another culture. Language is another cultural artifact that may influence the outcome of collaborations among labs that speak different languages because of the potential for misunderstandings (Gwynne, 1995). Linguistic differences could discourage some labs from participating in collaborative R&D. Opportunities for misunderstandings and miscommunications are increased when the communication is remote and the information is very technical. Viewed from this angle, linguistic differences could conceivably reduce the synergistic effects from networking among R&D labs.

Hambrick and Mason (1984) and Finkelstein and Hambrick (1996) show a link between the top management team diversity and a variety of organizational processes and outcomes including innovations. Top management team diversity has been shown to influence the frequency, quantity, and ease of communication, reduce misunderstandings, and foster innovations due to greater variance in ideas (Smith et al., 1994; Wagner, Pfeffer, & O’Reilly, 1984). Top management team diversity was found to be a strong indicator of the impact of cultural diversity in MNCs (Nohria & Ghoshal, 1997). Thus, in this study, top management team diversity is used as a proxy measure for cultural diversity. The measures for this item are shown in Box 5 of the Appendix. A diversity index was computed from these measures.

Trust among R&D labs. Kuemmerle (1997) argued that the HQ played a pivotal role in fostering collaboration among R&D labs by creating an environment in which managers feel comfortable to establish informal relationships. A restrictive posture by managers could severely restrict the willingness of labs to network with each other. HQ managers could encourage networking by, for example, sponsoring joint research projects, organizing and coordinating cross-cultural teams, organizing joint planning sessions and research seminars, and providing incentives to labs that engage in cooperative R&D (Neff, 1995). R&D managers have a powerful influence on the long-term research agenda and performance of their labs. Consequently, managers who feel secure enough to allow their scientists and engineers to collaborate with their counterparts in other labs could play a positive role in enhancing the creative capability of their labs. On the other hand, managers who are uncomfortable with collaborative relationships could inhibit the creation of synergistic innovative capability because they may be less supportive of joint innovative projects (Neff, 1995). A wide range of measures for trust have been proposed in the literature, however, for reasons of parsimony, the influence of trust on synergistic innovative capability is measured by the five items shown in Box 6 of the Appendix.

Resource levels. Several studies have shown that the level of resources available within an organization for creative activities is critical for innovations (Bartlett & Ghoshal, 1989; Nohria & Ghoshal, 1997). Subsidiaries with higher levels of resources tend to register higher performance than their counterparts with lower level resources. The level of R&D resources available to a lab could be expressed in two ways. First, whether or not the available resources are above or below the average for similar labs within the MNC group, and second, the amount of slack resources a lab possesses. Slack resources refer to the pool of resources within the lab that are in excess of the minimum necessary to produce a given level of output (Nohria & Ghoshal, 1997). Some theorists argue that slack resources play a crucial role in allowing organizations to innovate because they permit organizations to experiment with new innovations that might not ordinarily be approved in a more resourceconstrained environment (Galbraith, 1973). Opponents of slack resources argue that slack simply promotes undisciplined investment in R&D activities that rarely yield economic benefits (Antle & Fellingham, 1990). Slack is viewed as organizational inefficiency. Nohria and Ghoshal (1997) sought to reconcile these opposing views by showing that the relationship between innovation and slack resources is a curvilinear, inverse Ushaped one.

Based on the preceding arguments, we infer that slack could have a positive effect on synergistic innovative capability because labs with slack may be more likely to enter into collaborative partnerships. Also, through collaboration among labs, the MNC could harness its worldwide slack resources which could have a positive effect on the synergistic innovative capability of the organization. Labs facing tight resources may not be willing to enter into collaborations out of concern that existing projects may suffer. The resource levels of labs were measured as shown in Box 7 of the Appendix.

Environmental complexity. Environmental complexity refers to rapidly changing or very unstable environments in terms of resource availability, innovations, and markets. Lawrence and Lorsch (1967), Thompson (1967), Mintzberg (1979), and others have argued that organizational units operating in complex environments require greater flexibility and freedom to make decisions in order to be able to respond to changes in a timely manner. Lack of freedom could lead to inflexibility and the inability to respond quickly, which may result in missed opportunities. Nohria and Ghoshal (1997) view environmental complexity as a stimulus for greater collaboration because of the increased vulnerability of the subsidiaries to rapid changes in the environment. Labs that cooperate with the HQ and other subsidiaries will be better equipped to manage these complexities. Through interdependence and collaboration, subsidiaries can share information, personnel, and other resources to their mutual benefit thereby reducing their vulnerability to environmental instability. Thus, it seems that labs operating in more complex environments are likely to be more innovative because they are “forced” by environmental pressures to be innovative in order to maintain their competitive advantage. In less complex environments, the pressure to innovate may not be as great and they tend to be less innovative. Environmental complexity was measured as shown in Box 8 in the Appendix.

Reliability of Constructs

Since several of the measures used in the study consisted of multiple items, a reliability analysis using Chronbach’s Alpha is appropriate. Table 1 shows the Chronbach’s Alpha for the scales variables. Usually, a Chronbach’s Alpha of 0.70 or greater is considered an acceptable benchmark for reliability. In every case, the reliability exceeded this threshold, indicating a high level of reliability of the constructs.

Research Methodology

The study is based on responses obtained from 79 R&D labs belonging to 27 North American, Japanese, and European research-intensive MNCs in the electrical, electronics, pharmaceutical, chemical, and automotive industries. Respondents were mainly R&D vice presidents, directors, and managers. Data were collected by a survey questionnaire. In addition, qualitative data were obtained from several participants through telephone interviews. Table 2 provides descriptive statistics profiling the participating R&D labs.

Factor analysis and multivariate regression analysis were employed to investigate the relationship between the dependent variable synergistic innovative capability and five independent variables (autonomy, socialization, formalization, HQ-subsidiary communication, and intersubsidiary communication) and four moderating variables (resource levels, environmental uncertainty, trust, and cultural diversity). Essentially, the focus is on the extent to which the combined effects of the independent variables provide adequate explanation of variations in the dependent variable.

In light of the exploratory nature of the study and the fact that many of the constructs consisted of multiple items, some of which are used for the first time in this study, a decision was made to employ a conservative data analysis approach. Essentially, a two-step data analysis strategy was employed in the study.

In the first step, the relationship between the dependent, independent, and moderating variables were assessed using regression analysis. These regressions were executed under the assumption that all the scale variables were unidimensional constructs in the sense that the simple arithmetic mean of all the items for the respective constructs was used as the appropriate measure for the constructs. These are referred to here as main model and moderating variables regressions. Due to the relatively small sample size, the interaction terms were analyzed sequentially.

In the second step, the dimensionality of the constructs was investigated through factor analysis with varimax rotation. Based on the new dimensions emerging from the factor analysis, another set of regression runs was executed. The variables in the regressions based on the factor analysis were measured using the simple arithmetic mean of the items making up the components. The results from the two phases of the data analysis are presented and discussed below.

Statistical Analysis and Results

Main Model and Moderating Variable Regressions

Table 3 presents the results of the main model and moderating variables regressions. In the main model regressions, preliminary regression runs indicate that greater explanation is obtained when the two communication variables are grouped according to the communication media involved, that is, electronic and face-toface. Also, the formalization variable was dropped because its incremental contribution to r’ is negligible and its impact on the magnitude and direction of the remaining variables is minimal. This does not mean, however, that formalization is not an important variable, but rather that, from a statistical perspective, it does not add explanation over and above that provided by the other variables.

In the main model regression, the results indicate that all the variables except for ECOMSU are statistically significant at the 0.05 level. The sign of the estimates for all variables in the model are as expected except for IPCOMHQ and ECOMSU. The model yielded an R^sup 2^ of 32% and a significant F-test of overall model fit. The fact that thebeta values for IPCOMHQ and IPCOMSU are so very similar in magnitude but opposite in sign suggests that they may be collinear. However, an examination of the variance inflation factor, an index of multicollinearity, does not reveal significant multicollinearity since they were all substantially below 10, the threshold level for multicollinearity (Stevens, 1992). The interpretation, therefore, is that these two variables enter as a contrast between headquarters and subsidiary communications. These results are consistent with our expectations of the relationships. The negative sign of IPCOMHQ may indicate that subsidiary labs perceive in-person visits to their labs as a form of monitoring analogous to “big brother is watching.” Alternatively, it may be interpreted as more HQ than subsidiary communication has one effect and more subsidiary than HQ communication has the other effect. The approach employed to assess the impact of the four moderating variables involves adding various interaction terms to the main model regressions (e.g., socialization*trust, socialization*IPCOMHQ, etc.). As it is impractical to include all the results emerging from this second approach, the results of four models (i.e., one for each moderating variable) are also presented in Table 3. The results indicate that no interaction term was statistically significant in any of the regressions. Thus, there is no evidence that the four moderating variables have a moderating effect on innovative capability. This may, however, be an artifact of the data.

In summary, we found that variations in innovative capability are explained by the level of autonomy of the labs, the extent of socialization, and the effectiveness of in-person communication between the HQ and subsidiary labs. In-person communication among subsidiary labs registered a higher level of statistical significance than electronic communication. Overall, the results obtained from both the main model and moderating variables regressions are consistent in terms of the significant variables, the r^sup 2^, and magnitude and direction of the parameter estimates.

Factor Analysis

In the preceding regressions, the scale variables were treated as unidimensional constructs since the simple arithmetic mean of all the items for the respective constructs were used. In order to determine the dimensionality of the scales, factor analysis with varimax rotation was undertaken. The results are given in Tables 4 to 7 and discussed below.

Factor analysis on the dependent variable synergistic innovative capability yielded four components based on items with eigenvalues greater than 1. These four components and the items they comprise are consistent with theoretical constructs in the management of technology area of research. The loadings of the items that make up each component are identified in bold in Tables 4-7. Together, these four components explain 68% of the variance in the construct.

The first component is named knowledge creation and management because its four items pertain to knowledge creation, transfer, and management within organizations (Inkpen, 1997). The second component is named managerial and operational efficiency because its items deal with the efficient utilization of managerial and operational resources. The third component indicates the increasing participation of subsidiary labs in a larger number of complex R&D projects as opposed to their traditional roles of simply adapting HQ-generated products to local market conditions, thus it is described as strategic R&D (Brockhoff & Schmaul, 1996; Chiesa, 1996; Howells, 1990; Medcof, 1998). The fourth component describes the capability of subsidiary R&D labs to develop innovations faster at lower costs. Together, these two dimensions reflect the competencies of the labs in generating new and successful innovations proficiently. Thus, it is called innovative proficiency.

Autonomy. Three components emerged with eigenvalues greater than 1 and together these three components explained close to four-fifths of the variance in the construct. The results are shown in Table 5. The items of these three components are closely aligned with constructs in the management of technology literature. The first component (autonomy in project management) deals with a lab’s freedom to make decisions regarding its project management approach. These have been widely used in the project management literature (Balachandra & Friar, 1997; Kumar, Kumar, & Persaud, 1996). The second component (autonomy in inter-lab collaboration) consists of items relating to collaboration with other labs within the MNC group. The third is autonomy in managing knowledge workers since the items pertain to decisions regarding the company’s human resources.

Trust and formalization. Factor analysis on trust indicates that the concept as measured in this study has two components (Table 6). The first relates to the technological aspects of collaborations while the second relates to the social aspects of collaborations. Thus, the components could be named technology-based collaboration and relationship-based collaboration. Since both components explain similar proportion of variance, they are retained as two distinct constructs. The construct formalization yielded two components (Table 7), which do not lend themselves to sound theoretical characterization, therefore formalization is retained as a single dimension construct.

Application of Factor Analysis Results

In light of the results from the factor analysis, it was decided to re-run the regression analysis with these new larger sets of variables. Thus, there are now nine independent variables in the main model regression-autoproj, autocollab, autoHR (the components of autonomy), formalization, socialization, IPCOMHQ, IPCOMSU, ECOMHQ, and ECOMSU. Regressions were run with these nine independent variables and each of the four components of synergistic innovative capability sequentially. The regression results from this analysis, as displayed in Table 8, show that the regression with knowledge creation and management as the dependent yielded three statistically significant variables at the 0.1 level. All three variables pertain to communication among research labs. We observed that in-person communication among subsidiary labs has a positive effect on knowledge generation and management whereas in-person communication between the HQ and subsidiary labs has the opposite effect. However, technologically supported communication between HQ and subsidiary labs has a positive effect on knowledge generation and management.

These results seem to suggest that knowledge creation and transfer among subsidiary labs is accomplished much more effectively in face-to-face settings than by technologically supported media. In particular, technical knowledge, a large part of which is considered tacit knowledge, is transferred more effectively when scientists and engineers meet in person. At this time, it seems that technologically supported communication may be more effective in communicating explicit knowledge. The positive impact of electronic communication between HQ and subsidiary labs (ECOMHQ) could be interpreted as the HQ being quite effective as a clearinghouse for information generated throughout the MNC.

Table 8 also shows that four variables-AutoHR, ECOMHQ, ECOMSU and IPCOMSU-are having a positive and significant impact on strategic R&D (at the 0.1 level). These results suggest that close communication among subsidiary labs facilitates the sharing of information, which tends to increase the participation rates of labs in more strategic R&D projects. The role of HQ in disseminating information regarding new projects, available technical capabilities within the MNC group, and so on seems to create greater awareness of existing opportunities, thereby encouraging greater participation in strategic R&D projects. Freedom of subsidiary labs to recruit and decide on the career paths of scientists, engineers, and other technical personnel means that the labs can now hire and retain professionals with the required expertise.

Table 8 shows that three variables are significant in explaining the innovative proficiency of R&D labs. Two of the variables relate to autonomy and one to socialization, indicating that the freedom to share information, personnel, and other resources within the context of a social network positively influences innovative proficiency. The autonomy of the labs to decide on their project portfolios and priorities or to change an existing production process appears to have a negative effect on innovative proficiency. A closer examination of the ratings on these items shows that the HQ generally takes the lead in these decisions in order to foster greater coordination and integration of R&D activities across the MNC. However, subsidiary labs perceive this lack of freedom as constraining their ability to innovate in a timely and cost effective manner.

Discussion and Implications

The significant effect of autonomy, socialization, and in-person communication on innovative capability can be viewed as being mutually reinforcing. Socialization leads to the creation of informal social networks which, in turn, leads to more frequent communication among the labs, which may increase the willingness of the labs to share information, expertise, and technology or even work collaboratively on joint projects. The freedom to exploit collaborative opportunities and to engage in a wider range of projects means that the labs now have access to a wider pool of resources, which they may be able to access quickly and at lower costs. This may lead to increased operational efficiency and participation in a wider range of innovative activities. In fact, many of the labs described several incidences where social networks have positively influenced their ability to complete projects faster and cheaper, and they have engaged in R&D activities of greater strategic significance.

The negative effect of inter-subsidiary labs’ electronic communication and the positive effect of HQ-subsidiary labs’ electronic communication on innovative capability seem somewhat surprising, particularly since considerable emphasis is given to technologically supported communication. The seemingly counterintuitive result with respect to inter-subsidiary communication may be due to the fact that since electronic communication is much quicker and cheaper, there is a tendency for overuse. For example, one respondent remarked that it is not uncommon for him to receive in excess of 50 e-mails a day in addition to voice messages and faxes. But a more fundamental explanation may be that since the collaboration among subsidiary labs involves mostly tacit knowledge as opposed to explicit knowledge, electronic media are not appropriate. The positive effect of electronic HQ-subsidiary communication is a reflection of the information clearinghouse role that the HQ plays which has a positive impact on the labs. Generally, HQ makes greater use of technologically supported communication to communicate approvals, administrative procedures, general information, and routine reporting with subsidiary labs than to communicate tacit knowledge. Thus, it seems that technologically supported communication may be very efficient when transmitting knowledge even of a technical nature but is less effective when tacit knowledge is involved.

From a practical point of view, the findings of this study suggest that managers may improve the innovative capabilities of their R&D organization by implementing policies aimed at fostering greater informal networking, allowing labs greater freedom in decision-making, and encouraging greater face-to-face communication. Faceto-face communication should be encouraged particularly when the intent is to leverage tacit knowledge. Providing a structured context for collaboration and communication could intensify the positive impacts of these policies. The role of HQ as a facilitator of this process and as a clearinghouse for information of interest to the labs could have a positive impact.

From a research perspective, this study has made several important contributions to existing research on this issue. First, this study not only confirms the importance of effective coordination in order to leverage global knowledge and learning, but also extends our knowledge by showing that the specific coordination mechanisms are linked to specific innovative outcomes. For example, deep socialization coupled with effective communication, both in-person and technologically supported, are critical for fostering strategic R&D and knowledge creation and management. With this kind of knowledge, managers may be able to design a mix of coordination strategies that is more reflective of the goals they want to achieve for respective facilities. This finding is also consistent with previous research (e.g., Chiesa, 1996; Florida, 1997; Gassman & Zedwitz, 1999; Gerybadze & Reger, 1999; Howells, 1990; Kuemmerle, 1997) which showed that different R&D units tend to perform different roles ranging from listening posts to new product development and, therefore, need to be managed differently. The findings of this study present a first view of how managers may go about this task.

The results from the factor analysis indicate clearly that the concept of synergistic innovation is a multidimensional one with each dimension consisting of specific items that are grounded in the innovation literature. This finding lends itself to the development of a more rigorous and parsimonious conceptual analytical framework to guide future research. The lack of such a framework in this particular research area has been discussed by several authors (e.g., Gerybadze & Reger, 1999; Howells, 1990; Medcof, 1998). However, further studies along the lines of this one, with different samples, would help to refine and validate the findings reported here. The four dimensions of the synergistic innovative capability construct could be used by managers to operationalize and measure the extent to which the various units are contributing to the achievement of corporate goals in these terms.

Despite the contribution of the study there are a few limitations. First, the relatively small sample size requires some caution in interpreting the results reported here, particularly in cases where the results seem counterintuitive or contradict previous research. Second, the focus on large MNCs from selected industrial sectors would require caution in generalizing the results to all MNCs regardless of sectors. Third, this study did not consider the impact on synergistic innovative capability of external collaborations, that is, with other MNCs, universities, research institutions, and government agencies. These activities are quite substantial in the modern MNC. Fourth, the focus was on coordination variables and did not include other variables that may impact the synergistic innovative capability (e.g., research intensity, existing knowledge base, international experience). These limitations in themselves represent fruitful avenues for future research.

In conclusion, this paper examines the extent to which four coordination and control mechanisms used by MNCs to organize their global R&D activities affect their ability to enhance their innovative capability. Based on a sample of 79 research facilities belonging to 27 research-intensive multinationals, we conclude that three mechanisms are critical: the level of autonomy of overseas research facilities, the level of socialization among the labs, and the effectiveness of in-person communication between the HQ and subsidiary labs. It was observed that technologically supported communication among subsidiary labs had a limited effect.

Although the four components reported in this study are consistent with existing theories in the area of management of technology, further testing of the constructs using the same scale items is needed to validate the statistical properties of the constructs. Until such studies become available, it can be inferred from this study that the concept of synergistic innovative capability is a multidimensional construct. Also, the scale items used in this study provide a starting point for further investigation concerning the reliability and dimensionality of the concept.


1 Using Blau’s (1977)

Diversity Index given as:

formula omitted


Aiken, M., & Hage, J. (1968). Organizational interdependence and intraorganizational structure. American Sociological Review, 33, 912-930.

Antle, R., & Fellingham, J. (1990). Resource rationing and organizational slack in a two-period model. Journal of Accounting Research, 28 (1), 1-24.

Asakawa, K. (1996). External linkages and overseas autonomy-control tension: The management dilemma of Japanese R&D in Europe. IEEE Transactions on Engineering Management, 43 (1), 24-32.

Balachandra, R., & Friar, J. H. (1997). Factors for success in R&D projects and new product innovation: A contextual framework. IEEE Transactions on Engineering Management, 44 (3), 276-287

Baliga, B.R., & Jaeger, A.M. (1984). Multinational corporations: Control systems and delegation issues. Journal of International Business, 15, 25-40.

Bartlett, C.A., & Ghoshal, S., (1989). Managing across borders. Boston: Harvard Business School Press.

Behrman, J.N., & Fischer, W.A. (1980). Overseas R&D activities of transnational companies. Cambridge, MA: Oelgeschlager, Gunn and Hain.

Brockhoff, K., & Schmaul, B. (1996). Organization, autonomy, and success in internationally dispersed R&D facilities. IEEE Transactions on Engineering Management, 43 (1), 33-40.

Burgleman, R.A. (1983). A model of the interaction of strategic behavior, corporate context, and the concept of strategy. Academy of Management Review, 8, 61-70.

Burns, T., & Stalker G.M. (1961). The management of innovation. London: Tavistock.

Burt, R.S. (1982). Toward a structural theory of action. Orlando, FL: Academic Press.

Carson, M., Pearce, R., & Singh, S. (1992). Business culture and international technology: Research managers’ perceptions of recent changes in corporate R&D. In 0. Granstrand, L. Hdkanson, & S. Sj6lander (Eds.), Technology management and international business. Chichester, UK: John Wiley & Sons Ltd.

Chiesa, V. (1996, September). Strategies for global R&D. Research Technology Management, 19-25.

Coleman, J.S. (1990). Foundations of social theory. Cambridge, MA: Harvard University Press, Belknap Press. De Bodinat, H. (1975). Influence in the multinational corpora

tion: The case of manufacturing. Ph.D. Thesis, Harvard University.

De Meyer, A. (1993, July). Internationalizing R&D improves a firm’s technical learning. Research Technology Management, 42-49.

De Meyer, A., & Mizushima, A. (1989). Global R&D management. R&D Management, 19, 135-146.

Dunning, J.H. (1992). Multinational enterprises and the globalization of innovatory capacity. In 0. Granstrand, L. Htkanson, & S. Sj6lander (Eds.), Technology management and international business. Chichester, UK: John Wiley & Sons Ltd.

Edstrom, A., & Galbraith, J.R. (1977). Transfer of managers as a coordination and control strategy in multinational organizations. Administrative Science Quarterly, 22, 248-263.

Finkelstein, S., & Hambrick, D.C. (1996). Strategic leadership: Top executives and their effects on organizations. St. Paul, MN: West.

Florida, R. (1997). The globalization of R&D: Results of a survey of foreign-affiliated R&D laboratories in the USA. Research Policy, 26, 85-103.

Galbraith, J.R. (1973). Designing complex organizations. Reading, MA: Addison-Wesley.

Gassman, O., & Zedwitz, M. (1999). New concepts and trends in international R&D organization. Research Policy, 28, 231-250.

Gates, S.R., & Egelhoff, W.G. (1986). Centralization in parent headquarters-subsidiary relationship. Journal of International Business, 17 (2), 71-92.

Gerybadze, A., & Reger, G. (1999). Globalization of R&D: Recent changes in the management of innovation in transnational corporations. Research Policy, 28, 251-274.

Granstrand, O., HAkanson, L., & Sjolander, S. (1992). Technology management and international business. Chichester, UK: John Wiley & Sons Ltd.

Gwynne, P (1995, January). Managing ‘multidomestic’ R&D at ABB. Research Technology Management, pp. 30-33. HAkanson, L. (1992). Locational Determinants of Foreign

R&D in Swedish Multinationals. In 0. Granstrand, L. Hfkanson, & S. Sjolander (Eds.), Technology management and international business. Chichester, UK: John Wiley & Sons Ltd.

HAkanson, L., & Zander, U. (1988). International management of R&D: The Swedish experience. R&D Management, 18 (3), 217-226.

Hambrick, D.C., & Mason, P.A. (1984). Upper echelons: The organization as a reflection of its top managers. Academy of Management Review, 9, 193-207.

Hedlund, G. (1981). Autonomy of subsidiaries and formalization of headquarters-subsidiary relations in Swedish MNCs. In L. Otterbeck (Ed.), The management of headquarters-subsidiary relations in multinational corporations. Aldershot, UK: Gower.

Hewitt, G. (1980). Research and development performed abroad by U.S. manufacturing multinationals. Kyklos, 33. Howells, J. (1990). The location and organization of research

and development: New horizons. Research Policy, 19, 133-146.

Inkpen, A.C. (1997). The management of knowledge in international alliances: The role of collaborative process. Conference Proceedings of the CBI, USA.

Kanter, R.M. (1988). When a thousand flowers bloom: Structural, collective, and social conditions for innovation in organizations, In B.M. Staw & L.L. Cummings (Eds.), Research in organizational behavior. Greenwich, CT: JAI Press.

Kuemmerle, W. (1997, March). Building effective R&D capabilities abroad. Harvard Business Review, 61-70.

Kumar, U., Kumar, V., & Persaud, A. (1996). To terminate or not an ongoing R&D project: A managerial dilemma. IEEE Transactions on Engineering Management, 43 (3), 273-284.

Lawrence, PR., & Lorsch, JX (1967). Organization and environment: Managing differentiation and integration. Boston: Graduate School of Business Administration, Harvard University.

Maccoby, M. (1995, September). Human engineering leads to operating principles for global management. Research Technology Management, 58-60.

Marschan, R. (1996). New structural forms in multinational: Decentralization at the expense of personal communication networks? International Journal of Technology Management, 11 (1&2), 192-206.

Medcof, J. (1998). Strategic contingencies and power in networks of internationally dispersed R&D facilities. Academy of Management Annual Meeting, Boston.

Mintzberg, H. (1979). The structure of organizations: A synthesis of the research. Englewood Cliffs, NJ: PrenticeHall, Inc.

Neff, P. (1995, May). Cross-cultural research teams in a global enterprise. Research Technology Management, 15-19.

Nohria, N. & Ghoshal, S. (1997). The differentiated network: Organizing multinational corporations for value creation. San Francisco: Jossey-Bass.

Ouchi, W.G. (1980). Markets, bureaucracies and clans. Administrative Science Quarterly, 25, 129-141.

Pearce, R.D. (1989). The internationalization of research and development by multinational enterprises. London: Macmillan.

Pugh, D.S., Hickson, D.J., & Hinings, C.R. (1969). The context of organizational structure. Administrative Science Quarterly, 14, 91-114.

Pugh, D.S., Hickson D.J., & Turner, C. (1968). The dimensions of organization structure. Administrative Science Quarterly, 13, 65-105.

Reger, G. (1997). Benchmarking the internationalization and co-ordination of R&D of Western European and Japanese multinational corporations. International Journal of Innovation Management, 1 (3), 299-331.

Ring, P.S., & Van de Ven, A.H. (1992). Structuring cooperative relationships between organizations. Strategic Management Journal, 13, 483-498.

Rogers, E. (1995). Diffusion of innovations (411 ed.). New York: Free Press.

Ronstadt, R.C. (1977). Research and development abroad by U.S. multinationals. New York: Praeger.

Schein, E.H. (1968, Winter). Organizational socialization and the performance of management. Industrial Management Review, 1-16.

Shane, S. (1994, July). Championing innovation in the global corporation. Research Technology Management, 29-35. Smith, K.G., Smith, K.A., O’Bannon, D.P., Olian, J.D., Sims,

MR, & Scully, J. (1994). Top management team demography and process: The role of social integration and communication. Administrative Science Quarterly, 39, 412438.

Stevens, J. (1992). Applied multivariate statistics for the social sciences (2″nd ed.). Hillside, NJ: Lawrence Erlbaum Associates.

Stock, G.N., Greis, N.P., & Dibner, M.D. (1996). Parent-subsidiary communication in international biotechnology R&D. IEEE Transactions on Engineering Management, 43 (1), 56-68.

Thompson, J. D. (1967). Organizations in action. New York: McGraw-Hill.

Tushman, M. & Anderson, P (1997). Managing strategic innovation and change. New York: Oxford University Press.

Wagner, G.W., Pfeffer, J., & O’Reilly, C.A. (1984). Organizational demography and turnover in top management groups. Administrative Science Quarterly, 29, 74-92.

Ajax Persaud

University of Ottawa

Uma Kumar

Vinod Kumar

Carleton University

Address all correspondence to Dr. Ajax Persaud, School of Management, University of Ottawa, 136 Jean-Jacques Lussier Street, P.O. Box 450, Stn. A, Ottawa, ON, Canada K1N 6N5. E-mail:

The authors would like to thank the area editor and three anonymous reviewers for their feedback on the manuscript.

Copyright Administrative Sciences Association of Canada Mar 2002

Provided by ProQuest Information and Learning Company. All rights Reserved