Fighting the corporate immune system: a process study of subsidiary initiatives in multinational corporations

This paper examines subsidiary initiatives, which are entrepreneurial activities carried out by foreign owned subsidiaries in multinational corporations; and the forces that resist them that we refer to as the ‘corporate immune system’. The study is based on case study analyses of 44 initiatives, both successes and failures, undertaken by Canadian subsidiary managers. The key finding is that subsidiary initiatives take two different forms. Internally-focused initiatives are based on opportunities identified within the corporation, and are pursued through a traditional bottom-up process. Externally-focused initiatives are based on opportunities in the external marketplace. The most successful ones typically gain allies outside the corporation early on, and only confront the corporate immune system once they are relatively well established.  1999 Elsevier Science Ltd. All rights reserved. This paper is concerned with subsidiary initiatives, which are entrepreneurial activities carried out by the foreign subsidiaries of multinational corporations. Our interest in studying subsidiary initiatives arises from two simple observations. First, it is widely acknowledged in the literature on international management that foreign subsidiaries should take initiative (e.g. Bartlett & Ghoshal, 1986; Vernon, 1979) as a means of tapping into new opportunities as they arise around the world. Second, the empirical evidence for the phenomenon is extremely limited. Despite the large * Corresponding author. 0969-5931/99/$ see front matter  1999 Elsevier Science Ltd. All rights reserved. PII: S0969 -5931(98)00 043-2 150 J. Birkinshaw, J. Ridderstråle / International Business Review 8 (1999) 149–180 number of empirical studies done in this area (see Birkinshaw & Morrison, 1995 for a review), only a few papers offer tangible examples of subsidiary initiative (e.g. Birkinshaw, 1995; Delany, 1998; Science Council of Canada, 1980) and these tend to be rather anecdotal in nature. We therefore decided to conduct a detailed study of subsidiary initiatives in order to shed light on this apparent mismatch between theory and empirical evidence. Organization theory provided some clues as to why subsidiary initiative is a relatively rare phenomenon. First, it is widely recognized that activities of exploitation tend to ‘drive out’ activities of exploration or creation (Hedlund & Ridderstråle, 1992; Kanter, 1986; March, 1991; March & Simon, 1958). Reasons for this include that the feed-back loops and returns of creation activities are less certain and more remote in time (March, 1991). Second, subsidiary units are typically low in power, which limits both their degrees of freedom in enacting change and the visibility of their initiatives in head office. The established architecture of the organization tends, therefore, to favor ideas developed in high-influence parts of the organization at the expense of those from the periphery (Burgelman & Grove, 1996; Hamel, 1996). These arguments provide a basic explanation as to why the organization (in its totality) will often view subsidiary initiatives with suspicion or hostility. It seems there are disparate forces at various levels and locations within the organization that act to suppress subsidiary initiatives with the result that many efforts probably do not come to fruition. One way of viewing this process, we suggest, is to see the initiative as an alien body that the ‘corporate immune system’ seeks to destroy. This is because the merits of any given initiative cannot be known in advance, so the expectations of actors within the organization of its likely value is such that they would prefer to make a type I error (reject a promising initiative) than a type II error (let through a rogue initiative). This may, in fact, be the correct side to err towards if one is going to err at all, but it does mean that many promising initiatives are probably lost. The purpose of this paper, then, is to examine the subsidiary initiative process in detail with a view to understanding: (a) the nature of the corporate immune system, i.e. the structures, systems and predispositions that resist subsidiary initiatives, and (b) the strategies pursued by subsidiary managers in their attempts to circumvent or fight the corporate immune system. The rest of the paper is organized as follows. In Section 1 we provide some theoretical background and specify the research questions for the empirical study. In Section 2 we describe the research methodology. In Section 3 we describe the principal findings, in terms of the nature of the corporate immune system and the strategies used by subsidiary managers to circumvent or fight it. Finally, we discuss the implications of the study for theory and for practise. 1 Though obviously there are exceptions to this general rule, e.g. Philips’ US subsidiary or NCR’s Scottish subsidiary. 151 J. Birkinshaw, J. Ridderstråle / International Business Review 8 (1999) 149–180 1. Theoretical background 1.1. The initiative process Subsidiary initiative is a specific form of corporate entrepreneurship. In keeping with Kanter (1982); Miller (1983), we define initiative as a discrete, proactive undertaking that advances a new way for the corporation to use or expand its resources. As in a number of previous studies (e.g. Burgelman, 1991; Sathe, 1985), the initiative process becomes the fundamental unit of analysis in this research. Subsidiary initiative simply refers to any initiative that occurs outside the home country of the multinational corporation. In this study, all cases were of initiatives in Canadian subsidiaries of U.S. multinationals. A variety of models of the initiative process have been put forward in studies of corporate entrepreneurship and innovation. A common approach is to see the process in three phases, whereby the initiative is conceived through the identification of an opportunity, it gathers support and impetus as it is pushed through the socio-political organizational system, and it is then implemented. At any stage along the way the initiative can fail to gather the necessary support or resource commitments and fail. This basic process can be seen in the study by Schön (1971) of the emergence of public policies, Marquis (1969), for the framework for the innovation process, Bower (1972) for the decision process model, and a number of more recent studies (e.g. Burgelman, 1983; Howell & Higgins, 1990; Maidique, 1980). While this three-phase model captures the essence of the innovation process, it has over the years been refined and enhanced in a number of directions. One important observation is that the initiative process is not sequential. Phases of development are often overlapping (Imai, Nonaka & Takeuchi, 1985); they may be reciprocal (Galbraith, 1982); they may occur as parallel or overlain activities (Burgelman, 1983); and they may even occur in a non-linear sequence (Van de Ven & Garud, 1995). In its extreme form, such non-linearity begins to resemble the garbage can model of organizational decision making (Cohen, March & Olsen, 1972). A second issue is that ‘success’ can be measured on a number of different dimensions. One important measure of success is the commitment of resources to a project (e.g. Bower, 1972). Another is the commercial success or market approval of the resultant business activity (e.g. Day, 1994). However, there is also evidence that legitimacy is important, both from corporate management (Burgelman, 1983) and from other functional areas (Dougherty & Heller, 1994). If an initiative is to lead eventually to a new business activity for the corporation it seems logical that it needs to gain resource allocation, market approval and organizational legitimacy. However, the order in which these three outcomes are achieved, and their relative importance, is likely to vary enormously depending on the nature of the initiative, the structure and systems in the organization, and the industrial environment. In sum, it would appear to be rather limiting to put forward a single model of the initiative process, given the almost infinite variety of forms it appears to take. One key point that must, however, be borne in mind is that success is more than just a matter of resource 152 J. Birkinshaw, J. Ridderstråle / International Business Review 8 (1999) 149–180 allocation. A second point, which we deal with subsequently, is that resistance is likely to be encountered throughout the process. 1.2. Power and influence in the initiative process Because the initiative process is driven to a large degree by the relative power of the subsidiary unit vis-à-vis the corporate headquarters, it is useful to briefly review the relevant literature in this area. Power is defined in a multitude of ways, but we can identify several key elements. First power is typically associated with influence over people’s behavior and decision outcomes. Thus, Pfeffer (1981) argues that it is “a force sufficient to change the probability of an individual’s behavior from what it would have been in the absence of the force”, Galbraith (1983) outlines power as “the possibility of imposing one’s will on the behavior of others”, and March (1995: 141) describes it as “the capability to get what you want or to fulfill your identity”. Second, power is relative, in that any actor has a unique power position in relation to each of the various external parties. Third, power is situational or domain specific (March, 1995), since the specific task or process in question may limit the power of some actors while others become more powerful. Fourth, power is socially constructed or enacted (Berger & Luckman, 1967; Weick, 1979) in that it is subject to different interpretations depending on the perspective of a single individual. Where does power come from? In the literature a large number of different sources of power have been identified, but these can with some simplification be organized under two headings: (1) structural power; and (2) resource-based power. Structural power in a corporation is ultimately a legally granted authority (Weber, 1947), in that the legal charter of an organization allows shareholders to grant a CEO and the CEO may push this authority further down the organization. Such structural power is typically manifest through the administrative system, or more specifically in a units’ ability to control three primary integration mechanisms (Thompson, 1967): (1) standardization of work processes, output, or norms and belief systems; (2) planning, which refers to the different types of objectives that guide units’ activities; and (3) mutual adjustment, which refers to more informal systems such as meetings, agendas, and information systems (Crozier, 1964; Hickson, Hinnings, Lee, Schenck & Pennings, 1971; Perrow, 1979; Pettigrew, 1973). Resource-based power accrues to those who have control over valuable assets, because they represent assets on which others depend (March, 1995; Pfeffer & Salancik, 1974, 1978). Relatedly, knowledge is also a source of power, in the form of skills tied to certain individuals or supra-individual routines (Cyert & March, 1963; Nelson & Winter, 1982). Moreover, it is also important to recognize that the subsidiary’s external network is an asset from which power can be derived, if it consists of relationships that can only be accessed through the subsidiary. Thus, by finding allies and tapping into their assets the power of the subsidiary unit may increase in relation to other actors within the organization. One further element of resourcebased power is reputation, which is gained from performance in historical actions (Pfeffer, 1992). A subsidiary unit with a strong reputation will have its suggestions 153 J. Birkinshaw, J. Ridderstråle / International Business Review 8 (1999) 149–180 listened to with more respect by other corporate units, and will typically be favourably treated in resource allocation decisions (Beerlew & Hall, 1966). These two basic sources of power provide a strong rationale for distinguishing between ‘core’ subsidiaries and ‘peripheral’ subsidiaries. Peripheral subsidiaries are situated away from the centre, both hierarchically and geographically, and are thus by definition low in structural power. Their resource-based power will also be low in the early years of their existence, though over time this situation is likely to change as the subsidiary develops unique resources or relationships of its own (Prahalad & Doz, 1981). Indeed, as this paper will show, many peripheral subsidiaries are very good at leveraging their limited resource-based power, such as a strong reputation or good external relationships, as a means of pursuing their initiatives. 1.3. The corporate immune system From either a corporate entrepreneurship or a power perspective, it can be expected that subsidiary initiatives will encounter significant resistance from the existing power bases within the corporation. Resistance can take a multitude of forms, from strict funding criteria to bureaucratic inertia to subtle political manoeuvring. Surprisingly, however, most existing research has given limited explicit attention to the nature of such resistance. The preference among scholars has been to focus on the active process, as manifest in the actions of the initiative champion, rather than the passive process as manifest in the action, or inaction, of other corporate actors. Resistance to corporate entrepreneurship is, of course, absolutely appropriate. Corporations typically have well-defined procedures for assessing funding proposals in order that the most financially rewarding and strategically viable projects gain funding while others do not. Similarly, managers develop informal procedures or heuristics to help them evaluate and choose between the multitude of initiatives that they are asked to lend their support to. In a vastly oversimplified world where bounded rationality constraints and politically-motivated actions were absent, we could imagine a process that accurately distinguished between promising and unpromising initiatives. However, in reality it seems likely that errors are made in the selection of initiatives. While these may be both type I and type II errors, our expectation is that they will be predominantly of the type I variety (i.e. the rejection of a promising initiative) rather than type II (approving a rogue initiative). Our approach therefore is to model the resistance from other organizational actors as a ‘corporate immune system,’ which we see as the set of organizational forces that suppress the advancement of creation-oriented activities such as initiatives. The immune system in the body is built into the cells in the bloodstream. It has the task of eliminating or neutralizing any alien bodies that find their way into the system. 2 Important exceptions are “ten rules for stifling innovation” (Kanter (1985): 101), the discussion of Van de Ven (1986) of the management of attention, and the discussion of Burgelman (1983) of strategic context definition. 3 The Corporate Immune System metaphor has also recent been used by De Geus (1997), though he uses the term somewhat differently, e.g. to explain the problems encountered in mergers and acquisitions. 154 J. Birkinshaw, J. Ridderstråle / International Business Review 8 (1999) 149–180 This system acts to prevent alien substances from affecting the body in a harmful way. However, as in the case of the rejection of an organ transplant, it is possible that the immune system may reject an alien body that is to its long-term benefit. By analogy, our belief is that most initiatives, and subsidiary initiatives in particular, face a corporate immune system that views them as alien and potentially harmful bodies. It is important to note that the preceding argument is essentially a new way of conceptualizing an age-old observation about corporate entrepreneurship, namely that new ideas can expect to encounter resistance from the established power bases. Two lines of theory are relevant in this regard. First, it is well established that creationoriented activities tend to be driven out by exploitation-oriented activities (Hedlund & Ridderstråle, 1992; March, 1991; March & Simon, 1958). Creationoriented activities offer less certain and more remote returns on investment, they threaten existing power bases within the corporation, and they challenge institutionalized routines and behaviours. Moreover, it has been shown that the dominating coalition of the firm often base their control on competence in the areas in which the firm is currently pursuing exploitation (Leonard Barton, 1992). Second, the ability to gain support for an initiative has been found to vary directly with the power base of the championing individual or unit. As studies by Bower (1972); Day (1994); Rothwell (1977) have shown, the level of influence of the sponsoring unit is more critical to initiative success than the (proposed) technical or financial implications of the underlying project. Our belief, then, is that the corporate immune system is fundamentally conservative. Individuals within the system prefer to work within existing routines, throw their support behind low-risk projects, and resist ideas that challenge their own power base. Rather than risk allowing a potentially-harmful initiative to gain currency, the corporate immune system would prefer to regard all such initiatives as harmful even if that means a few worthwhile initiatives are rejected. This discussion suggests the following research questions: Research question 1. What is the nature of the ‘corporate immune system’ as it applies to subsidiary initiatives? Research question 2. What are the strategies used by subsidiary unit managers to circumvent or fight the corporate immune system in order to achieve initiative success? 1.4. The multinational corporation One final layer of theoretical background should be discussed, and that is the particular qualities of the multinational corporation (MNC). The MNC is frequently modeled as a network of semi-autonomous subsidiary units in different sovereign jurisdictions (Forsgren & Johanson, 1992; Ghoshal & Bartlett, 1990). Traditionally most MNCs gave their foreign subsidiaries exploitational roles, usually marketing and selling of the corporation’s products in the foreign market. Increasingly academic thinking has suggested that subsidiary units should take on more value-added roles, and even take responsibility for building those roles themselves (Birkinshaw, 1995; 155 J. Birkinshaw, J. Ridderstråle / International Business Review 8 (1999) 149–180 Bartlett & Ghoshal, 1986; Hedlund, 1986). The empirical evidence for this phenomenon is, however, mixed. While many studies have documented cases of subsidiary units taking on higher value-added roles (e.g. Bartlett & Ghoshal, 1986; Gupta & Govindarajan, 1994; Jarillo & Martinez, 1990), it is usually assumed that such roles have arisen through head-office assignment. A more interesting, and obviously more uncommon, situation is that in which the subsidiary ‘assumes’ such a role through its own initiative (Birkinshaw, 1995; Hagström, 1994). As noted in the introduction, though, the empirical evidence for such initiative is limited (Birkinshaw, 1995; Delany, 1998; Science Council of Canada, 1980). Two further issues should be mentioned here in terms of the multinational corporation. First, subsidiary initiatives have to be pursued with the objective of changing the charter of the subsidiary unit. Charter refers to the business or elements or a business in which the subsidiary participates and for which it has responsibility (Galunic & Eisenhardt, 1996). Thus we see initiative as the pursuit of opportunities that have implications for the rest of the multinational corporation, rather than limited-scope projects that are of interest only to the subsidiary unit. Second, multinational corporations like many other large organizations often function as internal markets, so that the ‘customer’ for a product produced by one subsidiary unit may well be another subsidiary unit within the boundaries of the corporation. This is important because one of the criteria for initiative success is market approval. ‘Market approval’ can therefore be either internal or external to the corporation. 2. Research methodology The sample for the study was drawn from the population of Canadian subsidiaries of U.S.-owned multinational corporations, which represents more than two-thirds of the total stock of foreign owned companies in Canada. Following an exploratory study, we drew up a list of potential research sites (i.e. in which initiative appeared to have transpired), based on our discussions with subsidiary managers, recent academic and popular-press articles, CD Rom databases, and discussions with industry groups and government bodies. The list was restricted to subsidiaries with a head office in the U.S., participation in a manufacturing industry, and annual revenues over $200 million, all to restrict variance. Forty companies were selected in this way. Thereafter we approached companies on a random basis. Two of the initial group of four were both appropriate (in terms of enough cases of successful and failed initiatives) and willing to grant us access. Subsequently a further five companies were added. We continued adding companies until we satisfied the redundancy criterion (Yin, 1984), that is until we stopped gaining any new substantial insights into the initiative process. The final sample, then, consisted of seven subsidiary companies and a total of 44 initiatives, 26 of which yielded sufficiently detailed accounts that we could conduct qualitative analysis on them. Table 1 provides some basic data about the sample. Table 2 lists the 26 initiatives. 156 J. Birkinshaw, J. Ridderstråle / International Business Review 8 (1999) 149–180 Table 1 Characteristics of sample companies Company Principal industry Approximate Number of ...of which Number of size (1993 initiatives detailed interviews revenues) studied accounts: Alpha Industrial and consumer $600m* 10 4 20 products Beta Chemicals $450m 7 6 18 Gamma Computer $500m 8 1 11 hardware/software Delta Computer $800m 6 3 18 hardware/software Epsilon Industrial products $420m 5 4 12 Zeta Engineering systems $420m 3 3 11 Theta Electrical and industrial $1100m 5 5 8