Corps de l’article

One of the great challenges of multinational corporations (MNCs) is to balance global integration and local responsiveness (Ghoshal and Nohria, 1989; Prahalad and Doz, 1987). Local managers of MNCs are then expected to cultivate a global mindset (Gupta and Govindarajan, 2002), which for some authors consists in the ability of managing the duality between “thinking globally” and “acting locally” (Levy et al., 2007). Based on this notion of duality between global vs. local pressures, a large body of research has considered MNCs as contested and political spaces (Clegg et al., 2018; Geppert, 2003) where global and local expectations may compete (Saka-Helmhout et al., 2016). Consequently, researchers have studied the range of potential responses to conflicting institutional demands, such as compliance, compromise, manipulation or non-compliance (Kostova and Roth, 2002; Pache and Santos, 2010). However, this stream of institutional research tends to overlook the role of actors and consequently offers very few insights into actors’ practices (Jarzabkowski and Lê, 2017). Moreover, the notion of duality tends to consider that subsidiaries operate in an OR logic, which means that they have to make a choice (Smith and Tracey, 2016). We consider that the diffusion of some global decisions or practices may be a source of paradoxical tensions for local managers since they can generate simultaneous and contradictory demands for them (Ambos et al., 2020). Paradox theory enables us to understand the practices that managers developed to cope with simultaneous contrary and interdependent needs (Jarzabkowski et al., 2013; Knight and Paroutis, 2017). Little attention has been paid to the process of responding to conflicting demands and how tensions and responses may evolve over time (Holm et al., 2017; Jarzabkowski et al., 2013). To overcome the limitations of the extant literature we propose to combine a micro-political and a paradoxical perspective to study the process and the practices through which local managers of MNCs deal with simultaneous global vs. local contradictions. Our aim in this article is therefore to examine the following question: how do local managers of MNCs navigate local vs. global paradoxical tensions? In order to provide relevant insights, we conduct an in-depth case study of a French entity of IBM located in Montpellier (France). Since the early ’90s, local managers of IBM Montpellier (IBM Mtp) frequently struggled with various global decisions and expectations that seemed incompatible with their local objectives and even threatened their survival. Through the analysis of two specific periods of tensions, we build a grounded process model of how local managers deal with paradoxical tensions over time. Our study contributes to previous work on MNCs by providing insights on the overlooked practices of local managers and how such practices may influence, negotiate, and even deviate from the implementation of global demands. We bring to light two new kinds of paradoxical responses to competing demands that combine elements of compliance and non-compliance: transgressive alignment and obedient deviance. In the first part of the article, we review the existing institutional research on responses to competing demands within MNCs, before highlighting how paradox theory may provide new insights and thus enrich the literature. The second part of the article describes our empirical context and methodology before presenting and discussing our main findings. Finally, in the last section, we present and discuss the process model that emerged from our data analysis.

Theoretical Background

Local Managers’ Responses To Institutional Duality And Competing Global Vs Local Demands

The literature on MNCs has long been dominated by a globalized and convergent discourse that stresses the top-down diffusion of global policies and best practices from headquarters (HQ) to-ward subsidiaries (Guler et al., 2002). The works of authors such as Ghoshal, Bartlett, Nohria, Prahalad and Doz consider MNCs as an interorganizational network in which each subsidiary op-erates in a specific local environment and develops unique networks, resources and knowledge (Ghoshal and Bartlett, 1990; Ghoshal and Nohria, 1989). In the mid-1990s, Birkinshaw and col-leagues showed that subsidiaries do not limit their role to the passive execution of HQ mandates but are able to actually have a more active position in the decision-making process of the MNC through their local initiatives (Birkinshaw et al., 1998; Birkinshaw, 1999). One of the greatest challenges for MNCs is, then, to develop a global vision while simultaneously responding to local specificities (Prahalad and Doz, 1987) which implies that local managers of subsidiaries must have the abilities to balance “thinking globally” and “acting locally” (Levy et al., 2007; Ambos et al., 2020).

Much of the research on HQ-subsidiaries’ relationships stresses that local entities of MNCs face “institutional duality” (Greenwood et al., 2010; Meyer et al., 2011; Narula, 2014) as “each foreign subsidiary is confronted with two distinct sets of isomorphic pressures and a need to maintain legitimacy within both the host country and the MNC” (Kostova and Roth, 2002, p. 216).

Local managers of MCN are thus confronted with multiple logics (Schildt and Perkmann, 2017) and are expected to fulfil different and often conflicting roles: they are supposed to act in the MNC’s interest, but at the same time in that of their local entity (Ambos et al., 2020; Beddi, 2015; Saka-Helmhout et al., 2016).

Based on this notion of duality, a large body of research has considered MNCs as contested and political spaces where local and global logics or expectations may diverge, giving rise to bargaining and conflicts (Clegg et al., 2018; Dörrenbächer and Gammelgaard, 2011; Dörrenbächer and Geppert, 2006; Geppert et al., 2016). When global and local demands compete, it can be especially challenging to manage both sets of expectations, and subsidiaries’ managers may adopt a wide range of responses. These responses may vary from “active” adoption (Kostova and Roth, 2002), conscious obedience (Oliver, 1991), or simple acquiescence (Tempel et al., 2006) to non-compliance (Pache and Santos, 2010) or deviance (Jonsson et al., 2009, Spreitzer and Sonenshein, 2004) via compromise, avoidance, or negotiation (Greenwood et al., 2010; Holm et al., 2017; Tempel et al., 2006). While this institutional line of research has made significant contributions to the study of political relationships between HQ and subsidiaries, it presents some limitations. First, institutional approaches consider that the decision of local managers to adopt, negotiate or refuse HQ demands will be largely dependent on institutional pressures and on various environmental conditions (Kostova and Roth, 2002; Pache and Santos, 2010). The actions and practices of actors who adopt and implement responses to tensions remain overlooked (Jarzabkowski and Lê, 2017). Consequently, research on local responses to competing demands in the MNC tells us very little about the dynamics of negotiations and conflicts at the local level of the MNC (Geppert and Williams, 2006). Second, in line with Smith and Tracey (2016), we believe that the notion of duality implies that tensions may be resolved by a choice. In the MNC context it means, for instance, that local managers of subsidiaries may have to choose one response to global decisions: acquiescence OR deviance, act local OR act global. This OR logic implies a kind of dichotomy that leaves little room for eventual ambivalent or paradoxical responses toward HQ decisions. In line with Morgan and Kristensen (2006), we believe that more paradoxical responses are likely to exist within MNCs and that the literature lacks studies that explore the more complex picture between acceptance and deviance.

In order to fill this gap, we propose to use paradox as a conceptual lens to study local managers’ responses to the global diffusion of decisions and practices. We believe that paradox theory may well complement the research stream on micro-politics since it allows us to consider that, when confronted by conflicting demands, local managers of MNCs should not just make a choice but have to find a dynamic balance to combine such paradoxical demands (Smith and Lewis, 2011). We thus believe that this theoretical prism may allow us to shed light on more complex and paradoxical responses to global diffusion that may combine elements of compliance and deviance.

Using A Paradox Lens To Study Local Managers’ Responses To Competing Demands In The MNC

According to Farjoun (2010), a duality refers to two tendencies that are oppositional but fundamentally interdependent, complementary, and also mutually enabling. Dual elements might thus be potentially compatible (Putnam et al., 2016). However, when elements are difficult or even impossible to simultaneously conjugate, they become paradoxical (Lewis, 2000). In the context of organizations, paradoxes can occur when actors’ positions are at the interplay between interrelated and conflicting demands (Cuganesan, 2017) and where they have no choice but to cope with them (Knight and Paroutis, 2017). In line with O’Brien et al. (2019), we consider that in MNCs, a local manager’s position may be compared to that of middle managers since they play a decisive role at the interface between top management (HQ) and their local entity. Like middle managers, local managers’ organizational lives are pervaded by contradictions and inconsistencies (Lüscher and Lewis, 2008; Ambos et al., 2020). Paradoxes are embedded in the everyday actions and decisions of local managers and the way they deal with these paradoxical tensions will have a decisive impact on their entity’s performance and even survival (Jarzabkowski et al., 2013; Smith and Lewis, 2011). Paradox theory identifies four kinds of paradoxical tensions. The learning paradox, which refers to tensions between the old and the new, and the need to destroy but also to build upon the past in order to envisage and invent the future (Smith and Tushman, 2005). The organising paradox, which surfaces when actors struggle with competing structures, conflicting processes or divergent forces such as control and flexibility or integration and differentiation (Lewis, 2000). The belonging paradox that arises from the tension between the ego and the other in an organizational context. It occurs when actors experience inconsistencies between the values and beliefs of their group and those of the wider organization (Jarzabkowski et al., 2013; Lüscher and Lewis, 2008). For example, at the firm Lego, Lüscher and Lewis (2008) empirically explore a belonging paradox arising from the situation that managers might relate to the management team and to their own production teams. This produced belonging tensions identified as creating unity and emphasizing diversity. Finally, the performing paradox, which refers to the plurality of objectives of internal and external stakeholders. It may emerge when actors are expected to perform multiple and competing roles or tasks (Smith and Lewis, 2011). Scholars investigate the organizing paradox of exploration and exploitation by addressing the ambidexterity paradox and the role of managers in reaching a balance between exploratory and exploitative activities (Knight and Paroutis, 2017). Paradox theory enables researchers to study ambidexterity by going beyond the capacity of a firm to pursue high levels of exploration and exploitation concurrently, but by managing tradeoffs to find the most appropriate balance between the two (Papachroni et al., 2015). If scholars on paradox theory have widely studied the different way of coping with paradoxes, studies on MNCs remain scarce on tensions between local and global competing demands and especially on the practices local managers employed to deal with these tensions (Ambos et al, 2020). This is the gap we aim at starting to fill through this study.

Theories of paradoxes used to oppose defensive vs. active responses to tensions (Jarzabkowski et al., 2013; Smith and Lewis, 2011). Defensive responses are defined with a short-term perspective since they enable to only temporarily cope with tensions (Jarzabkowski et al., 2013). Such coping strategies consist in avoiding, denying (Derksen et al., 2019) or splitting tensions (Poole and Van de Ven, 1989) which results in separating the two poles of the tensions and avoid interaction. In contrast, active responses may drive longer-term success (Smith and Lewis, 2011). Such responses may include acceptance, which consists in embracing the two poles causing the tension (Poole and Van de Ven, 1989; Smith and Tushman, 2005); or transcendence, which consists in going beyond the tension and using its complexity to propose new alternatives (Lewis, 2000).

An important specificity of paradoxes is that they persist over time (Smith and Tracey, 2016). According to Miron-Spektor et al. (2017, p. 4), since paradoxes are persistent, whatever response individuals choose to adopt, “the underlying tension is likely to remain and resurface again”. Moreover, paradoxical tensions are not only persistent, but they can also evolve over time (Pant and Ramachandran, 2017). One tension may give rise to new challenges and lead to other tensions (Jarzabkowski et al., 2013; Tuckermann, 2018). Consequently, many paradox studies assert that the main utility of this theory is not to investigate the nature of tensions or responses per se, but to examine the process of responding to competing demands (Jarzabkowski et al., 2013; Smith and Lewis, 2011; Tuckermann, 2018). However, as highlighted by Birkinshaw et al. (2016)firms frequently have to manage strategic dualities--pairs of imperatives that are equally important but to some degree in conflict with one another. While there is a sizable empirical and theoretical literature in this area, research that seeks to understand how firms manage strategic dualities in practice is underdeveloped. We address this gap in knowledge through a detailed longitudinal analysis of one firm, Softcorp, conducted in real time. Softcorp’s executives were faced with the well-known global integration/local responsiveness duality, which they sought to resolve by creating an unusual Dual Headquarters that oriented employees across the firm toward a dual focus on Europe and Asia. We describe the sequence of changes (which we label counterweight, hybrid engine, and flywheel, the literature on paradoxes lacks processual studies. Consequently, it lacks insights on “the intentions and actions of managers, and the consequences of those actions […]” (Birkinshaw et al., 2016, p. 52). The aim of our study is thus to combine a micro-political and a paradoxical perspective to cover the process, as well as the practices that allow local managers of MNCs to cope with competing demands. We therefore propose addressing the following research question: how do local managers of MNCs navigate local vs. global paradoxical tensions?

Research Methods

A Case Study Of IBM Montpellier, France (1993-2009).

The setting for our study is a French entity of IBM, established in 1965. In studying the history of this facility, our attention was particularly drawn to two specific events that we will describe in detail in the findings section: (1) The first is the unprecedented crisis experienced by IBM in 1993 and the subsequent restructuring of the company. (2) The second is when, for reasons of financial optimization, IBM decided in 1997 to open a second European manufacturing facility in Dublin (Ireland) and to transfer part of the Montpellier plant’s activity and competences to this new location. These two events generated many changes and threats for IBM Mtp’s management. Each time, local French management had to carry out substantial restructuring while also looking for new business opportunities that might ensure the survival of the subsidiary. As already highlighted by several authors (Jarzabkowski et al., 2013; Lüscher and Lewis, 2008), during contexts of change such as restructuring, middle managers often experience confusion and anxiety caused by contradictions. Such periods consequently offer researchers good opportunities to observe paradoxes. We decided to focus on these two main episodes of IBM Mtp’s history since they were likely to provide rich information on contexts that trigger tensions and how local managers deal with tensions over time to ensure their entity’s survival. As pointed out by Smith and Tracey (2016, p. 461), paradox theory lacks insights about “real-world experiences of tensions”, which means that researchers need to “explore the everyday experience of tensions in organizations by the actors who inhabit them.” Our study aims to fill this gap by an in-depth exploration of a rich single case study to extract abundant and informative data. By focusing on informants’ perceptions and interpretations of past events, we seek to gradually bring out new elements of understanding. We therefore decided to conduct a qualitative, inductive research using the methodology proposed by Gioia et al. (2013) since such a grounded methodology is particularly suited to the exploration of complex and dynamic social phenomena (Suddaby, 2006). Anchoring the analysis in the empirical data, this approach allows us to progressively and iteratively bring out concepts and more abstract categories from which researchers may develop a grounded theoretical model about complex and dynamic processes.

Data Collection

This study is based on data collected between 2009 and 2013 as part of a more general historical study on the practices developed by local managers to ensure the survival of the French subsidiary between 1965 and 2009. Between 2009 and 2012, the first author was given an office onsite where he spent three or four days a week. Between 2012 and 2013, the first author went back several times onsite in order to collect further data or to get some precisions on data previously collected. Besides promoting good knowledge of the case under study, this immersion allowed the first author to soon be treated as an active member, which finally provided easier access to data and above all a solid basis of trust, which is vital for this kind of study (Balogun and Johnson, 2004; Lüscher and Lewis, 2008). Key retired informants were directly contacted by managers in position at the time of the study to introduce the first author and the research conducted. This form of cooptation allowed us to obtain 100% of positive responses to our demands for interviews. We conducted 54[1] semi-structured interviews, averaging 1.5 hours, and each was recorded and transcribed. The final data corpus contains more than 85 hours of interviews with 41 different informants resulting in 317 transcribed pages (see Table 1).

The questioning was open ended and encouraged information about the history of the French site, the challenges faced by local managers, and the way they dealt with these challenges, namely the practices they adopted. By focusing on the tensions and the practices elaborated by the local managers of IBM Mtp, our study is at an organizational level of analysis. The aim was to understand how informants make sense of the local organization’s past (Boyce, 1995) and to capture their perspective on what has occurred (Arnaud et al., 2016). In addition, we used four supplementary sources of data: (1) annual reports and company internal archives; (2) newspapers; (3) books on corporate history[2] and (4) extensive personal field notes from non-participant observations and informal conversations that occurred during the immersion of the first author. Data extracted from these sources were very useful in developing our knowledge of the company, gaining an in-depth understanding of the various events and challenges encountered, cross-checking data collected during interviews, and enriching the contextual part of our theorization.

In this research, we consequently propose to combine archival and retrospective data to study the history of IBM Mtp and explore how local managers navigated local vs global tensions. Retrospective data have been widely used in organizational research as “popular tools for learning about the past” (Miller et al., 1997, p. 189). Retrospective accounts allow us to collect information about past phenomena, to reconstruct past events, and to study how managers behaved (Eisenhardt and Schoonhoven, 1990). However, authors such as Golden (1992) or Huber and Power (1985) have pointed out the pitfalls of retrospective reports. They warn about the risk of an inaccurate recall of key informants due to memory lapses, inappropriate rationalization, or a tendency to present a favorable image of themselves or their organizations. To increase the reliability of retrospective accounts, we adopted data-gathering procedures that follow the recommendations of Huber and Power (1985) and Miller et al. (1997): we collected data from multiple informants selected from different organizational levels and we asked about concrete facts, events and practices in order to limit potential cognitive biases of past opinions. As shown in Table 1, for each event we interviewed at least 36 participants which means that we have at least 36 different interpretations of the context, the local responses, and the outcomes of what occurred at that time. The amount and the variety of stories collected allowed us to cross-check data, to have confidence in the accuracy of elements that appeared in several interviews and to discard those which seemed more isolated. In addition, archival data (reports, speeches, newspapers, books) enabled us to cross-check information collected during interviews and to increase their reliability. While most historical research relies only on data archives, we believe that, when possible, combining archival data with interviews on concrete past events can limit the pitfalls of such studies.

Table 1

Data collection

Data collection

-> Voir la liste des tableaux

Data Analysis

Data were analysed using an inductive content analysis approach, with the aim of gaining understanding of the context, the practices and the consequences of what occurred (Corbin and Strauss, 2008; Gioia et al., 2013). We consider that people who have lived the situation under study are “knowledgeable agents” (Gioia et al., 2013). Our aim is thus to give voice to informants and to focus on their interpretations and perceptions of past events without seeking to transcribe an objective picture of what really happened. In the early stages of data analysis, we consequently stayed really close to the informant’s voice and then progressively and iteratively gained abstraction to finally present a grounded model of their sensemaking. A model that allows us to transcribe a participant’s experience in theoretical terms. Our data analysis included three main steps: (1) We began with the first-order analysis (Gioia et al., 2013). We identified initial concepts in the data that we grouped in categories (first-order terms) without limiting us in the number of codes and categories assigned. During this first-order analysis, we used “open” codes to stay as close as possible to informants’ terms and to keep a distance from our own experience and assumptions—similar to open coding (Corbin and Strauss, 2008). (2) We then engaged in the second-order analysis (Gioia et al., 2013). We went through the many emergent categories several times, seeking similarities and differences that allow us to regroup them into second-order themes to provide a more abstract level of representation. Each emergent second-order theme was compared to the existing literature in order to determine whether prior theories might help us to understand what the informants were describing and to finally distil second-order themes into more conceptual aggregate dimensions (see Figure 1). This second-order analysis allowed us, for instance, to compare the second-order theme that we called “Act local AND Act global” (Figure 1) with the literature on paradox, and to notice that it can be assimilated into a paradoxical tension known in the literature as a belonging tension (Lewis, 2000). It also allowed us to highlight that the second-order theme that we called “transgressive alignment” has no equivalent in the existing literature on MNCs. (3) The final step consisted in seeking relationships between our different second order themes to finally construct the grounded process model. We finally proposed a process model (see Figure 2) that depicts the tensions experienced by local managers, their triggers, and the responses the managers developed to deal with them. These steps of analysis were not linear. They instead formed an iterative process that continued until we had a clear picture of the emerging themes, a clear grasp of relationships between themes, and until we achieved theoretical saturation, which means that additional interviews didn’t reveal a new theme or new relationships (Gioia et al., 2013; Glaser and Strauss, 1967). Results of the data analysis process are summarized in the data structure (see Figure 1). Data structure provides a representation of how our analysis progressed from first-order categories to second-order themes and to aggregate dimensions. In the findings section we provide a narrative story of what occurred during the two events under study while explaining each emergent second-order theme.

In order to increase the rigor of our study we relied on the four criteria for trustworthiness identified by Guba (1981): credibility, transferability, dependability and confirmability. The first author’s prolonged immersion in the field enabled her to continually confront and check data and interpretations with members of the case under study, strengthening the credibility of our findings (Lincoln and Guba, 1985). Furthermore, the multiplicity of informants and data sources enabled us to triangulate, to cross check, and to ensure the stability and dependability of our data. Confirmability of our analysis was established by the second author who did not intervene during data collection and consequently kept an objective distance with the field but participated and discussed the coding and the emerging model. Finally, by maximizing the range of information collected and by providing thick descriptions of contexts we increased the transferability of our findings (Guba, 1981).

Findings

First Breakpoint (2ot.1 In Figure 1): 1993, Ibm Announces $8 Billion In Losses And Appoints A New Ceo To Rectify The Situation

After many years of growth and stability, the early ’90s marked an important turning point for IBM: the IT market was shifting, competition was increasing, and sales of mainframes were declining. Between 1990 and 1993, IBM’s profits fell by more than 200% and in 1993 the company announced more than $8 billion in losses. To cope with this crisis, a new chief executive officer was appointed, Lou Gerstner. His arrival at the helm of IBM is associated with decisive technological, strategic and organizational changes: the company’s geographical and multi-domestic organization was abandoned for a more integrated and vertical structure; the bipolar technology that had always been used in IBM mainframes was replaced by the more integrated and efficient CMOS technology (complementary metal oxide semiconductor); and the new CEO decided to transform IBM’s core business and make a major shift from just manufacturing hardware to delivering integrated customer services and solutions. These transformations had a huge impact on IBM since they involved the closure of many plants and the displacement of 125,000 jobs worldwide between 1991 and 1994. Locally, these changes also had significant impacts on the French entity: its production system became obsolete, its structure oversized, and its manufacturing business was no longer in line with IBM’s strategy and ambitions.

Local Managers Under Paradoxical Tensions

Act Global AND Act Local (2OT.3)

Data collected reveal that, on the one hand, local managers were marked by a strong “corporate mindset”, reflected in their belonging and their commitment to the IBM Corporation. And on the other hand, these same actors also have a strong local embeddedness, a strong local cohesion, and a real mission for local management: to protect people who were working there. This first paradoxical tension may be illustrated in the following quotes, in which a former director first evoked his global anchorage: “There was a deep attachment to the IBM company and a very strong culture of global commitment”. But later on, during the same interview, he also explained that “there was an obvious local anchoring. I didn’t want to shut down the site, so I had to find solutions to preserve the people who were working here” (Inf 37).

Restructure AND Explore (2OT.4)

The significant losses of IBM in the early ’90s, as well as the technological breakthrough decided by the new CEO, had huge consequences for IBM Mtp. Indeed, for several consecutive years, the local management faced pressures from HQ to drastically cut costs while maintaining their manufacturing activity: “Production plans remained the same, but we had to dramatically reduce our expenses—by about 30% per year! These were extremely tough goals!” (Inf 37). While the local management was largely occupied with downsizing, it simultaneously perceived the threat presented by IBM’s strategic shift. The new CEO wanted IBM to evolve towards services and asked each entity to “come up with solutions”. The director of IBM Mtp explained that even if they managed to meet global expectations in terms of downsizing, it was just a short-term solution. According to him, if IBM Mtp remained just a manufacturing facility, it had no chance of surviving much longer in the MNC. “We had to find a lifeline to protect the subsidiary, to prevent it from sinking, to avoid it being snatched from Montpellier. And the way for us to anchor it locally was to develop a local competitive advantage here, onsite” (Inf 37). This former plant director was convinced that the only solution for IBM Mtp to survive was to align with IBM’s global strategy. It had to diversify and explore new business opportunities oriented towards customer services and solutions.

Figure 1

Data structure

Data structure

-> Voir la liste des figures

The data collected reveal a second tension in the objectives of local managers: reduce costs while at the same time diversify activities in order to ensure longer term survival. “Everything got more complicated since we had to do everything at the same time: downsizing, retaining the production plans and trying to make sure that the site not only remained a production site” (Inf 7).

Local Practices Developed To Deal With Paradoxical Tensions

Align Global Strategy Through Transgressive Practices (2OT.6)

To cope with these tensions, local teams, supported by the facility manager, initiated the development of a new service-oriented activity: the customer demonstration centre. The idea of this “subsidiary initiative” (Birkinshaw et al., 1998) was to use mainframes produced in Montpellier as well as the expertise of local engineers and technicians to invite customers to come and test, onsite, the solutions proposed by IBM’s sales teams.

In 1994, returning from the annual meeting at HQ, I was thinking about the CEO’s discourse. He spoke about solutions, about customer orientation […] and then it struck me… I realized that in IBM Mtp we know how to meet such expectations. We know how to set up a structure that allows the customer to come and simulate their solution onsite, and not only to simulate on today’s computers but also on those of tomorrow since the latest machines were assembled here in Montpellier!

Inf 37

However, the objective assigned to IBM Mtp by the parent company at that time was to reduce expenses, not to develop a new business. For more than two years, therefore, this activity was secretly developed in Montpellier.

We identified our best engineers, we isolated them, and we hid them. They are paid by the official manufacturing activity even though they are working to develop the new mission. We gave them a large and luxurious area, we lent them mainframes coming from our manufacturing lines and we kept all this secretly until the suitable moment: until the moment we were able to prove that we know how to do it and that it is profitable for the company.

Inf 9

“We hid it at first because we were trying to develop an activity that our global leaders did not want us to do. They wanted us to make big systems, at minimum cost, with the smallest buildings possible and the minimum of resources … we were not asked to do anything else” (Inf 7). Several informants used the term “perruque” when describing the way this new activity was developed: a local initiative, hidden from HQ, where human and material resources normally used for the manufacturing activity were diverted and used for the development of this new service-oriented activity. For nearly four years, this activity was developed without informing head office. The demonstration centre soon proved its usefulness and its profitability by making it possible to conclude major sales contracts. In 1996, on the occasion of a visit by the CEO, the local management team presented the customer demonstration centre and IBM Mtp officially became IBM’s European customer centre. The French entity thus reinforced its position within the MNC by formalising its new orientation towards customer services.

Second Breakpoint (2ot.2): 1997, Ibm Opens A Second European Manufacturing Facility in Dublin

The end of the ’90s marked the rebirth of IBM and the end of several years of crisis and restructuring. IBM’s revenues exceeded $80 billion in 1998 and $100 billion in 2008. The company continued to pursue the strategy initiated at the beginning of the 2000s: it gradually decreased its manufacturing activity and pursued its transformation into a service business. At the same time, the focus was placed on value chain optimization and new manufacturing facilities were established in low-cost countries offering advantages in terms of labor costs or taxes, and the internal competition between IBM subsidiaries increased. In 1997, IBM HQ decided to take advantage of the economic attractiveness of Ireland and announced that a second European plant would be set up in Dublin to produce mainframes for the EMEA market. “The parent company decided that the new Irish plant would produce mainframes and therefore asked Montpellier’s teams to transfer part of their manufacturing activity” (Inf 4). As explained by this former middle manager, when the new plant opened in Dublin, IBM Mtp had to transfer part of its activity to Ireland. In addition, HQ also instructed local teams to train and support Dublin teams in the launch of their manufacturing activity. In other words, the French subsidiary had to transfer its competences and part of its business to Dublin.

Local Managers Under Paradoxical Tensions

Act Global AND Act Local (2OT.3)

As during the analysis of the previous event, our findings highlight that local managers once again faced a global/local paradoxical tension. On the one hand the analysis of data collected regarding this second breakpoint revealed that the previous period of substantial restructuring is still on the minds of local managers and has accentuated their local attachment and their commitment to safeguarding the site. “Moments of crisis such as we experienced in 1993 gave rise to a local sense of community and solidarity. It created a strong site culture” (Inf 7). On the other hand, interviews with these same informants also still revealed this strong “corporate mindset” and this culture of global commitment. This local/global paradoxical tension may be illustrated by the following quote in which the site manager at the time of the facts first explains his local anchorage and its associated pressures:

At Montpellier, we had brilliant, visionary and charismatic directors. When you are a young manager, these people inspire you and become a model […] When they pass the baton, when it’s your turn to become a director, you feel you have been charged with a mission: you must continue to safeguard the plant. You must ensure its survival. I didn’t want to be the one who was going to shut down the site.

Inf 6

He then also mentions his global belonging and its associated pressures of commitment: “You can’t belong to an organization like IBM and go against its strategy. […] we belong to a global governance; we must not forget it” (Inf 6).

Transfer activity AND avoid restructurings (2OT.5)

The analysis of data collected regarding this second breakpoint reveals that local managers were confronted by another paradoxical tension: they perceived global pressures to transfer activity to the new site and to support its launch but at the same time they also feel that their local mission is to retain as much activity as possible onsite to avoid relocation and job losses. The following quotes, extracted from interviews with two local managers, illustrate this tension by showing how informants are confronted by two contradictory objectives: commit to global demand by transferring their activity and also meet their local mission by avoiding restructuring and the relocation of activity. “They (HQ) ask us to transfer our manufacturing activity to Dublin, to train them and help them to get started; we cannot reverse this decision […] We didn’t want to proceed to local restructuring which means that we couldn’t just watch our business go to Dublin” (Inf.6); “The head office asks us to be the sponsor of Dublin, their guardian. It was painful, it’s like telling us to ‘cut our own arm’ but we had no choice than to execute […] We had to prevent the whole activity from going to Dublin” (Inf 19).

Local Practices Developed to Deal With These Paradoxical Tensions

Deviate from Global Demand Through Obedient Practices (2OT.7)

“We knew that if we refused to transfer any business to Dublin, we were dead. We also knew that if we let all the business go, we were just as dead. We therefore decided to participate in the transfer and to support Dublin, but also to keep the control. This allowed us to keep hundreds of jobs.” (Inf 6). As explained by this former director, local management decided not to resist the global decision to transfer business and competences to the new plant. On the contrary, they chose to accompany this change and even to be a key agent in the transfer. Different interviewees linked the strategy adopted by Montpellier managers at that time to a “judoka attitude”: “I believe that in Montpellier we became masters of judo: the art of accompanying the movement and not opposing it. Because sometimes, when you go against the movement, you get crushed” (Inf 7). According to them, in this kind of situation, fighting against change might be fatal and one good way to resist is actually to follow the movement and understand how to take advantage of it. “There’s no point in fighting rearguard action or opposing decisions that are irreversible. […] We must support the transfer and take advantage of the situation to prove our usefulness and increase our added value.” (Inf 7). Local management consequently decided to act as a “big brother” of the Dublin facility, to support the transfer and to keep control of it by only giving them a minimum of autonomy. “We explained to the parent company that we were able to support Dublin and to ensure IBM’s European manufacturing operations. We were somehow their life insurance!” (Inf 33). Citing their experience and know-how, local management team finally convinced HQ to let them keep the most complex and value-added parts of the production process such as component and module testing, as well as launching of new products. “Dublin had to start with the minimum of responsibility and autonomy. We wanted to prove to HQ that IBM Mtp was a decisive support, a real backup. We wanted to show that we were reliable and essential in complex missions such as launching new products” (Inf 22). This response ultimately allowed IBM Mtp’s management to cope with the threat of relocation to Dublin and to maintain some manufacturing activity and jobs in Montpellier.

Discussion and contributions

The last step of our methodology consisted in building a process model that shows how local managers of MNCs may face and deal with paradoxical tensions over time (Figure 2). Each element of this model will be explained and discussed in this section.

Figure 2

Process model

Process model

-> Voir la liste des figures

A latent “act locally/act globally” belonging tension.

Regardless of the period or the context of the study, our findings reveal that local managers were facing a latent tension: they appeared to constantly feel concerned about the interests of the MNC but also those of their local entity. This tension between act globally/act locally echoes the widely studied concept of institutional duality (Kostova and Roth, 2002): “the concurrent, and contradictory, institutional pressures experienced by the multinational subsidiary and emanating from the two intersecting institutional contexts inhabited by the subsidiary—the global organizational environment of the MNE and the local host country environment” (Pant and Ramachandran, 2017, p. 665). However, in our study this identity duality did not appear to be predetermined by local and/or global institutional pressures. It appears that, owing to their background and history within the organization, local managers developed this dual identification and dual attachment to both the parent company and the local entity. What we observed is more in line with the notion of a belonging tension (Jarzabkowski et al., 2013; Lewis, 2000), which in this case involves contradictions between the interests of their group (local entity) and those of the wider organization (the MNC) (Jarzabkowski et al., 2013; Lüscher and Lewis, 2008). Hence, as our model shows, it seems that there is a persistent belonging tension in being a local manager in an MNC due to “the juxtaposition of the contradictory, interdependent, and persistent characteristics of the ‘global’ and the ‘local’ in the subsidiary’s identity” (Pant and Ramachandran, 2017, p. 1). However, we contend that this is a latent tension (Smith and Lewis, 2011; Tuckermann, 2018). This means that in the absence of triggering events, which will be described below, these “act globally vs. act locally” contradictions are not perceived by actors as fundamentally incompatible. Yet, the existence of this pervasive belonging tension does not systematically trigger local managers’ action since it can remain “unperceived or unnoticed by organizational actors for years” (Knight and Paroutis, 2017, p. 404).

Events that trigger a paradox’s saliency.

In our study, we observed that between 1993 and 2009, the history of IBM Mtp was marked by two key events: the first one in 1993 when IBM went through an unprecedented crisis, and the second in 1997 when IBM opened a second European manufacturing facility in Ireland. Both events engendered profound changes for IBM Mtp and a feeling that its position within the MNC was threatened. Because of the changes and ruptures that they caused, we consider these two events as “breakpoints”, which can be defined as “transitions between cycles of activities” (Van de Ven and Poole, 2000, p. 42). As our model shows, these breakpoints seem to trigger the paradox’s saliency, which means that when such events occur, the latent belonging duality gives birth to salient paradoxical tensions. Indeed, we observed that following the breakpoints of 1993 and 1997, local managers suddenly experienced paradoxes: they perceived that compliance with global HQ and protection of local interests were no longer compatible. In sum, our model echoes and develops the work of Smith and Lewis (2011) and Knight and Paroutis (2017) since it gives empirical evidence that major changes such as breakpoints may trigger paradox saliency.

The emergence of belonging-performing paradoxical tensions.

As illustrated in our process model, at the two breakpoints previously mentioned, we observe that the latent belonging tension caused local managers to suddenly struggle with two new tensions: restructure vs explore; transfer activity AND avoid restructuring. Based on Smith and Lewis’ (2011) classification, we propose to call these salient tensions “belonging-performing tensions” since in both situations local managers were confronted with competing objectives (performing tension) because they wanted to simultaneously act locally and globally (belonging tension). These empirical observations thus confirm that tensions coevolve over time and that one tension may trigger another one (Jarzabkowski et al., 2013; Tuckermann, 2018).

Paradoxical responses of local managers.

Our model shows that local managers cope with these salient tensions by adopting paradoxical responses that blend elements of compliance and non-compliance. We identified and characterized two kinds of paradoxical responses that we call “transgressive alignment” and “obedient deviance”.

Transgressive alignment: When looking at local managers’ response to the first belonging-performing tension that arose in 1993, one is struck by the subversive nature of this response. Indeed, when local managers decided to secretly develop the new customer demonstration business, they transgressed global directives and adopted rather deviant practices, which echoes de Certeau’s (1984) notion of “la perruque”. This term refers to marginal and illegal practices where employees use their working time to do something that they are not supposed to do. However, looking at the local managers’ intentions at that time, we see that their goal was to develop a new service-oriented business, which was perfectly aligned with the strategic shift taken by IBM in the early ’90s. They consequently adopted transgressive and subversive practices to ultimately align and comply with the company’s global orientations and expectations (transgressive alignment). This kind of response echoes previous works on misbehaviour (Ashcraft, 2005; Richards, 2008) showing that deviance is not always harmful to organizations and that some acts of disobedience “should be viewed as ethical, proper, and ultimately deliver effective outcomes for organizations” (Richards, 2008, p. 656)(Coetsee, 1999).

Obedient deviance: Finally, as our model shows, when the second belonging-performing tension became salient, local managers adopted another kind of paradoxical response that we call obedient deviance. In contrast to the previous response, local managers’ aim was not to comply with IBM’s global orientation but to deviate, influence and even manipulate (Oliver, 1991) HQ’ decisions in order to avoid a complete relocation of their manufacturing activity and thus to safeguard local jobs. However, while local managers’ goals may be perceived as a kind of deviance, the practices and means adopted to achieve those objectives were, on the contrary, rather obedient. They did indeed comply with global demands and expectations since, rather than opposing change, they decided to be a key agent of it. By accompanying change and by achieving HQ’ objectives, local managers sought to prove their usefulness and to modify the initiated changes in their favour. We propose to qualify this second local response as obedient deviance: a way of dealing with belonging-performing paradoxes by adopting obedient practices in order to deviate from or counter global changes and expectations.

To sum up, as illustrated in the process model, our study reveals that there is a pervasive latent belonging tension in being a local manager in an MNC. This tension may remain dormant unless a break occurs that renders it salient. Local managers then perceive their local and global identifications as conflicting, which leads them to struggle with a new set of tensions. By responding to these tensions in a paradoxical manner, local managers temporally deal with the tensions and return them to a dormant state.

Theoretical contributions

By providing insights into the process of navigating paradoxical tensions within MNCs, we contribute to existing research that mainly adopts a static vision of local responses to competing demands (Greenwood et al., 2010; Kostova and Roth, 2002; Pache and Santos, 2010). Our study also contributes to the literature since it sheds light on local managers’ practices in dealing with competing demands. We highlight two new kinds of responses that we call transgressive alignment, which consists in aligning with global expectations through transgressive practices, and obedient deviance, which consists in deviating from or countering change while engaging in compliant practices. We thus go beyond the compliance/non-compliance dichotomy by showing that some responses to competing demands may be paradoxical since they simultaneously contain elements of compliance (alignment/obedience) and of non-compliance (transgression/deviance). The identification of such paradoxical responses allows us to stress that deviant behaviours of subsidiaries’ local managers are not systematically destructive for MNCs but may be more ambiguous (Ashcraft, 2005; Kidwell and Martin, 2005) and sometimes assimilated into positive deviance (Richards, 2008; Stashevsky and Weisberg, 2003).

Finally, our study contributes to paradox literature since it reveals that, while managers are responding to competing local vs. global demands, the tensions they experienced may transform, subside, and resurface later and elsewhere (Schad et al., 2016). The tensions do not disappear but alternate between latent and salient states. Furthermore, in line with Tuckermann (2018), our model shows the evolution of paradox saliency and latency by highlighting the role of breakpoints as triggers between the states.

Managerial contributions

Regarding managerial implications, we believe that our study provides two useful insights for practitioners. First, previous works have shown, based on Freudian psychology, that defensiveness is a natural reaction to the anxiety that paradoxes generate (Lewis, 2000; Vince and Broussine, 1996). When struggling with global vs. local tensions, local managers of MNCs may naturally try to resolve and suppress tensions. However, our study reveals that, on the contrary, accepting and embracing tensions may improve their chances of developing new lines of business and ensuring organizational survival. Our study thus makes a managerial contribution by encouraging local managers of MNCs to “value paradoxes as a vital ingredient of high performance” (Lewis et al., 2014, p. 63) instead of considering them as problematic situations that need to be resolved. Indeed, sustaining strategic paradoxes within complex contexts can contribute to organizational sustainability and foster creativity (Smith and Lewis, 2011). Paradoxes should be seen as a free space allowing local managers room to manoeuvre so they can open up the field of possibilities and create new options.

Second, our study shows that when working under paradoxical tensions, local managers of MNCs may adopt behaviours that deviate from the standards and expectations of their top management while generating positive impacts on their organization. These findings consequently provide contributions for the top management of MNCs since they show that local deviant practices should not consistently be avoided. Instead of systematically looking for local managers’ acquiescence and adoption of global norms and practices, the top management of MNCs may choose to leave them some degree of liberty in order to foster creativity and local constructive initiatives.

Limitations

The contributions of this study should be considered in light of some limitations. The first one concerns data collection since, as mentioned in the research method section, we asked participants to tell a story of events that happened more than 20 years in the past. Consequently, even if we tried to limit the well-known pitfalls of such methodology, there is still a risk of faulty memories or attempts to modify history in order to give a more desirable social image of what occurred. The second limitation is related to our choice of carrying out a single case study in immersion: our findings are situated, hardly generalizable and the inquirer did not maintain an objective distance from the object. However, our approach may be assimilated into a naturalistic inquiry (Guba, 1981) and consequently it asserts that there are multiple realities and that “the inquirer and the respondent […] are interrelated, with each influencing the other.” (Guba, 1981) We finally explained how we achieved the four criteria of trustworthiness (Lincoln and Guba, 1985) to ensure the rigor of our qualitative analysis.

Future research

As previously highlighted, the paradox theory is very little used in works on MNCs. It would therefore be interesting for future works to further investigate paradoxical tensions in MNCs and the way local managers deal with them. Such studies could consequently confirm the kind of tensions and responses that emerged in our study but also highlight new forms of paradoxes and new ways of navigating them. Future research may thus confirm, modify and enrich the process model proposed in this article. Second, as previously discussed, when local management is confronted with paradoxical tensions, it may adopt paradoxical responses that can be associated with positive or constructive deviance (Kidwell and Martin, 2005; Richards, 2008). Future research should focus on this kind of response and on the way the top management of MNCs perceive and manage (or should manage) such paradoxical local behaviours. Do they seek (or must they seek) to control and avoid these kinds of behaviours? Or do (or should) they instead stimulate such positive deviance in order to encourage local managers’ initiative and creativity?

Conclusion

Our study reveals that in a dynamic and hypercompetitive organizational context, subsidiary managers often struggle with situations and decisions that will require major restructuring or even threaten their business and their future within the MNC. To cope with these threats and survive within the MNC, local managers must get to grips with the belonging and performing tensions that pervade their role: act locally/act globally; explore/exploit; accept/refuse change, etc. We observe that tensions persist and evolve over time and that dealing with them consequently results in a process of navigating paradoxical tensions. We show that local managers may respond to paradoxes by adopting paradoxical responses—transgressive alignment, obedient deviance—that simultaneously combine elements of compliance and non-compliance.