In this milestone, you will write an 8- to 10-page

Effective Coordination of Shared Leadership in Global Virtual Teams

EMMA S. NORDBÄCK AND J. ALBERTO ESPINOSA

EMMA NORDBÄCK ([email protected]; corresponding author) is a Postdoctoral Researcher at the Department of Management Studies at Aalto University School of Business, Finland. She received her doctoral degree from Aalto University School of Science. Her research focuses on virtual work arrangements ranging from globally distributed teams to workplace flexibility, with a special emphasis on technology, leader- ship, and boundary-spanning practices for innovation. Dr. Nordbäck’s work has been published in such journals as Journal of Organization Design, Journal of Computer- Mediated Communication, Journal of Applied Communication Research, Journal of Virtual Worlds Research, and in various leading academic conference proceedings.

J. ALBERTO ESPINOSA ([email protected]) is a Professor of Information Technology and Analytics at the Kogod School of Business, American University. He holds a Ph.D. in Information Systems from the Tepper School of Business at Carnegie Mellon University. He has co-authored two books on work coordination across time zones, and on big data and analytics for service delivery. He has published in leading journals, including Management Science; Organization Science; Information Systems Research; the Journal of Management Information Systems; IEEE Transactions on Software Engineering, IEEE Transactions on Engineering Management; Communications of the ACM and others. He also has many years of experience as a senior manager for global organizations.

ABSTRACT: In this study we investigate how shared leadership is coordinated in global virtual teams and how it relates to team effectiveness. Based on 71 interviews with team members and leaders from eight teams from two global software development companies, we found that shared leadership had a more positive effect on team effectiveness when shared leadership was coordinated both implicitly and behavio- rally. Implicit leadership coordination is about members sharing same perceptions or cognitive schemas regarding who has leadership over what, and influences whether leadership actions are acted upon. With a mix of national cultures in the team, members are less likely to share the same leadership expectations, which may make shared leadership less effective. In turn, behavioral leadership coordination is associated with the explicit actions aimed at coordinating the leadership activities taking place in the team. This behavioral coordination increases in importance with a higher degree of shared leadership. Our findings contribute to theory and practice by showing that when leadership is highly shared in the team and uncoordinated, it may actually lead to detrimental effects in terms of lower team effectiveness. In contrast, shared leadership may reap its potential benefits if it is well coordinated.

KEY WORDS AND PHRASES: global virtual teams, team leadership, team effectiveness, shared leadership, online leadership.

Journal of Management Information Systems / 2019, Vol. 36, No. 1, pp. 321–350.

Copyright © Taylor & Francis Group, LLC

ISSN 0742–1222 (print) / ISSN 1557–928X (online)

DOI: https://doi.org/10.1080/07421222.2018.1558943

Suddenly I am like “hey what it this?”… and then Mikko (pseudonym) is working on something completely different than we agreed upon and tells me that “Benjamin (pseudonym) [the formal team leader] told me that I should do this now,” and I am like “hey why did I not get informed?.” This is very confusing and not a sustainable solution!

This quote from a global team member illustrates an instance of uncoordinated shared leadership in which a member has been re-directed by the formal leader in another direction than what was agreed upon with an emergent leader, which in his opinion, has therefore led to uncoordinated action. This situation may arise in teams with multiple leaders, a leadership structure commonly referred to as shared leadership (e.g., [60]). While leadership entails “a process whereby intentional influence is exerted over other people to guide, structure and facilitate activities and relationships in a group or organization” [74, p. 2], shared leadership distributes this influence process over multiple individuals. Shared leadership has been found to be particularly important for team effectiveness [32, 56] in global virtual teams (GVTs) in which members collaborate through technology over spatial, temporal, and cultural bound- aries [47], partly because such boundaries hinder communications. At the same time, the visibility of leadership actions in GVTs will be generally lower than in co-located teams, potentially leading to misunderstandings and uncoordinated actions. Given these barriers for coordination, we argue that it is important to develop a nuanced understanding of how GVT leaders coordinate their shared leadership actions, which will lead to team effectiveness and achievement of desirable outcomes. The extant research on shared leadership has paid virtually no attention to how GVTs coordinate their leadership activities (or fail to do so). Our study aims to fill this gap. Organizations are increasingly relying on GVTs to perform their core work

activities [26], which come with many leadership challenges. For instance, indivi- dual leaders have reduced ability to exert direct influence on team members due to the diminished communication opportunities that come along with increased vir- tuality [1, 55]. Also, GVT members may vary in their expectations for leadership because of cultural differences, which may lead to diverse leadership expectations across locations [76], making vertical (i.e., single) leadership less effective. Especially when the degree of virtuality is high, shared leadership may be more effective than vertical leadership [32]. Previous research on shared leadership has focused primarily on its relationship to

team effectiveness and several studies have found a positive association (see [15, 69] for recent meta-analyses), particularly with GVTs [32, 54]. On the one hand, studies have formulated this association between shared leadership and team effectiveness through mediators such as: enhanced participation and information sharing; increased team cohesion and team consensus; and better team functioning [15, 69]. On the other hand, other studies have found opposite effects (e.g., [6, 51, 63]), indicating that shared leadership may as well have negative effects on team effectiveness. For instance,

322 NORDBÄCK AND ESPINOSA

Robert [63] found shared leadership to decrease team performance in GVTs and Carte et al., [10] found that leadership behaviors, such as producer behaviors, aimed at motivating completion of the group’s task, led to lower team performance when GVT leadership was shared. Together, these contradicting findings suggest that there might be some interaction effects at play, which may explain this seemingly inconsistent association between shared leadership and team performance. For instance, Mehra and colleagues [51] identified the importance of formal and emergent leaders to perceive each other as leaders for shared leadership to be beneficial to performance. Drawing on Mehra’s study and the extant literature, we argue that for shared leadership to be beneficial to team performance, it must be effectively coordinated, above and beyond traditional task coordination. That is, we investigate whether coordination of the leader- ship itself can explain why and how shared leadership is effective in some cases and detrimental in others, helping us fill this gap in the research literature. Our rationale follows from coordination theory [44, 45], which we discuss next. Task coordination has been found to mitigate the negative effects of global bound-

aries (e.g., spatial and temporal) [14, 17] on GVT performance, particularly when the task activities have dependencies. In fact, task coordination is defined precisely as the management of these task dependencies [45]. In a similar vein, we argue that shared leadership generates additional coordination needs. Thus, consistently with coordina- tion theory we define shared leadership coordination as the management of dependen- cies among leadership activities, above and beyond task coordination.While leadership in general has been regarded as a key mechanism for overcoming task coordination challenges faced by GVTs [43], shared leadership coordination is not about task coordination, but about the coordination that the multiple leaders themselves and their followers need to achieve in order for the overall leadership to work as a cohesive whole. When a single leader enacts influence on the whole team (vertical leadership, see Figure 1), only task-dependencies needs to be coordinated. But when multiple leaders enacts influence on the team (shared leadership, see Figure 1), leader- ship dependencies needs to be coordinated, in addition to task dependencies. While it has been argued that shared leadership influences team effectiveness in

GVTs positively through an increase in task coordination [54], it has also been argued that shared leadership decreases team effectiveness in GVTs, through increased coordination problems [63]. However, we lack empirical evidence and a nuanced understanding about the suggested relationship between shared leader- ship and coordination, and we still know very little about how shared leadership creates additional leadership coordination needs, above and beyond the task. Given this gap in our knowledge, we set out to address the following research questions:

Research Question 1: How do global virtual teams coordinate shared leadership?

and

Research Question 2: How does this leadership coordination influence global virtual team effectiveness?

IMPLICIT AND BEHAVIORAL LEADERSHIP COORDINATION 323

By answering these questions, our work advances organizational theory in several important ways. First, we address the previously under-theorized relationship between shared leadership and team effectiveness in GVT’s by uncovering how leadership coordination acts as an important intervening factor to influence the effect of shared leadership on GVTeffectiveness. Second, while the extant literature is inconclusive on the effect of shared leadership on team effectiveness, our study examines the role of shared leadership coordination as the explanation for these prior contradicting results. Moreover, we explore two main forms of shared leadership coordination: implicit and behavioral; thus providing a more nuanced understanding of how shared leadership operates in teams. In doing so, we integrate shared leadership theory with coordination theory. None of these perspectives have been adequately investigated yet. We first present our theoretical foundations. We then describe our methods,

followed by our results. Last, we discuss our results and the implications and limitations of the study.

Theoretical Foundations

Shared Leadership in GVTs

There is a general consensus in the literature that leading GVTs successfully is challenging, yet vital for team effectiveness (e.g., [27]). In the context of teams (including GVTs), leadership is about fulfilling team needs, which may entail motivating and monitoring team processes, with the ultimate goal of enhancing team effectiveness [32, 53]. Prior research has commonly focused on leader traits and behaviors, as well how various situational variables influence leadership effectiveness [38, 74]. While the majority of prior research has focused on vertical solo leadership, more

recently, scholars have begun to question these top-down, hierarchical and formal leadership roles. However, the notion that leadership does not reside in a single individual is not new. Gibb [23, p. 884] articulated in the 1950s that: “leadership is

Figure 1. A completely vertical vs. a completely shared leadership structure. Note: An arrow from one node to another means that leadership moves in that direction.

324 NORDBÄCK AND ESPINOSA

probably best conceived as a group quality, as a set of functions which must be carried out by the group.” Gibb was hence among the first to pave the way for the emergence of the concept of shared leadership. Later, Pearce and Conger [60, p. 1] defined shared leadership as “a dynamic, interactive influence process among individuals in groups for which the objective is to lead one another to the achieve- ment of group or organizational goals.” While other definitions exist (e.g., [9, 15]), they all collectively suggest that leadership responsibilities are shared and distrib- uted over more than one person in the team. However, which of these leadership responsibilities are being shared varies from one study to another. Research on shared leadership has focused primarily either on the aggregation of

leadership contributed by leaders or on the specific leadership behaviors integrated into the entire collective leadership (see [69] for a meta-analysis). Within a behavioral approach, leadership has been conceptualized using a wide variety of actions aimed at satisfying team needs with the goal of enhancing team effec- tiveness [75]. For example, Yukl and colleagues [75] classified leadership behaviors into three categories: task-oriented (e.g., providing directions and monitoring per- formance), relations-oriented (e.g., providing support and encouragement), and change-oriented (e.g., proposing a new strategy or vision). All of these behaviors have been found to predict GVT success (e.g., [40, 59]). Consequently, this behavioral approach provides a useful theoretical lens to investigate how leadership is shared and coordinated in GVTs. Consistent with this approach, we study shared leadership manifested through task, relations, and change-oriented leadership beha- viors, as well as through cumulative leadership influence [69]. For the most part, studies on shared leadership in GVTs thus far have been

theoretical, offering propositions and predictions (e.g., [31, 42, 55, 56]). These studies have suggested that shared leadership increases team effectiveness in GVTs. A few recent empirical studies [32, 34, 54] also found that shared leadership leads to increased team performance in GVTs, while a few other studies showed the opposite effect [63], especially with some leadership behaviors [10]. Robert [63] offers reasons such as: having multiple members in charge resulting in no one being in charge; too much focus on trying to accommodate everyone; and potential coordination problems. In contrast, Muethel et al. [54] theorized that shared leader- ship leads to increased task coordination and improved communication practices in teams, which in turn affects team performance positively. Hoch and Kozlowski [32] reasoned further that shared leadership: creates stronger bonds among team mem- bers; facilitates trust, cohesion, and commitment; and mitigate disadvantages of GVTs, for example by helping members to overcome communication challenges [3, 60], leading to team performance. Based on these conflicting accounts, it is difficult to draw a unified conclusion on the effect of shared leadership on team effectiveness in GVTs. In this study, we argue that employing a coordination perspective may help reconcile these seemingly contradictory findings. In line with coordination theory [45], if we decompose leadership into its multiple func- tions, and distribute these functions across various members, either by adopting

IMPLICIT AND BEHAVIORAL LEADERSHIP COORDINATION 325

a shared leadership structure or by letting shared leadership emerge informally, we anticipate that these leadership functions will need to be tightly coordinated in order to be effective. In other words, we argue that how well shared leadership enhances team effectiveness is contingent on how well the shared leadership is coordinated within the team. We now discuss the role of coordination in shared leadership.

Coordination in GVTs

When team member activities can be accomplished independently, there is a minimal need to coordinate. But when such activities are interdependent, team members need to coordinate their work. In fact, coordination has been defined as the management of such dependencies [45]. Our focus in this research is on coordination processes associated with shared leadership in GVTs. There are two streams of research on coordination. The first stream is from the classic organiza- tional literature going back several decades, which argued that coordination is carried out mechanistically (e.g., plans, programs, schedules, procedures) or orga- nically (e.g., feedback, communication, mutual adjustment) [46]. These two types of coordination are referred to as “behavioral” [68], because they are based on what people do to coordinate. Behavioral coordination of workflows has been shown to be especially important for GVT performance [48]. The second stream comes from the team cognition literature in the psychology

field. While there is an abundance of team cognition labels and constructs in the literature, they are all based on some form of knowledge team members share about the task and each other [7]. Team cognition thus refers to the collective knowledge structure that enables team members to acquire and share knowledge within the team. Team cognition has been deemed important for coordination because it helps teams coordinate implicitly through a better “synchronization of member actions based on unspoken assumptions about what others in the group are likely to do” [71, p. 129]. Once team members have developed intra-group knowledge [12] through prior communication and working together [61], this familiarity helps them anticipate each other’s actions more accurately [62, 71]. Prior research has shown that behavioral and implicit coordination are particularly

important for GVTs, partly because of the communication barriers caused by global boundaries [19]. Compared to vertical leadership, shared leadership requires addi- tional coordination within GVTs above and beyond what is necessary for pure task coordination because global boundaries generate the need for local leadership in each location, effectively breaking down the team’s leadership structure from a single individual into multiple individuals (see Figure 1). This creates dependencies between the multiple leaders’ actions. Hence, when leadership is enacted by several individuals in the team in a decentralized coordination structure [44] the individual leadership activities need to be coordinated into a coherent whole in order to be effective. In line with Malone’s decentralized coordination structure [44], shared leadership

is likely to create high coordination costs when a large number of team members

326 NORDBÄCK AND ESPINOSA

participate in the leadership, requiring most leaders to interact frequently, which has been linked to reduced GVT performance [17]. On the other hand, shared leader- ship may decrease task production costs because team members may themselves engage in leadership, rather than consulting a formal leader who may not be available when needed, thus reducing the overall time to complete their respective task activities. Consistently, Malone suggests that decentralized coordination sys- tems have least vulnerability costs in the event of a task failure because task activities can be quickly reassigned to another member, minimizing disruptions [44]. This also applies to shared leadership activities because leaders can substitute for each other as needed. In sum, based on coordination theory, we expect that shared leadership will increase some coordination related costs and decrease others, which will have differential impacts on team effectiveness, and we argue that shared leadership coordination can explain these differences. We also argue that the need for leadership coordination is influenced by the global

boundaries spanned by the team. GVTs need to bridge multiple boundaries, such as time zones, geographic distance, functional, organizational, and national [30]. As more global boundaries are bridged by team members, the collaboration environment becomes more complex [18, 41] making task coordination and shared leadership more difficult [40]. For instance, an individual’s implicit view of leadership is likely to be related to cultural specific values, such as power distance [29], which can be defined as the extent to which a person accepts and endorses authority, inequality in power, and status privileges [8, 35]. Team members in high power distance cultures are more likely to accept unequal distribution of power in organizations [35] and accept their social status as followers [5], making them less equipped and less likely to participate in the team’s leadership (e.g., [11, 29, 55]). On the other hand, team members from low power distance cultures are more likely to attempt to minimize inequalities and favor less centralized leadership approaches [11, 29, 55]. Therefore, having members with differing power distances in the GVT, may hamper the development of shared mental models [64] about the shared leadership in the GVT. Moreover, the distribution of leadership across global boundaries may cause leadership actions to go unnoticed because communication is hindered.

Methods

We conducted a qualitative multi-case study to investigate how eight GVTs coor- dinate their leadership activities or fail to do so, and how this coordination impacts team effectiveness. Using multiple cases is a recommended method to develop theoretical constructs and propositions from case-based empirical evidence [16]. We treated the eight cases, ranging from high to low levels of shared leadership, and excellent to poor team effectiveness, as a series of “natural experiments,” each case serving to confirm or disconfirm the inferences drawn from the others [72]. Yet, the study was designed to be open-ended and to allow new themes to emerge. This inductive approach has been regarded particularly suitable for the study of

IMPLICIT AND BEHAVIORAL LEADERSHIP COORDINATION 327

social influence processes such as leadership, and for the study of phenomena that are not well understood [52, 58]. Therefore, we applied interpretive research methods for answering our research questions.

Cases and Data

This study is based on 71 in-depth interviews conducted with members and leaders of eight GVTs of two organizations. Both organizations developed software and provided support to their customers worldwide but operated on different markets. TechAlpha (pseudonym) employed 170 workers and TechBeta (pseudonym) had 600 employees at the time of the study. While both organizations provide technol- ogy-mediated support to their customers, TechAlpha also provides installations and on-site support to customers. The organizations have headquarters in Finland and area offices throughout Europe, Asia, and the United States. Table 1 provides information on team composition, communication media, and temporal distribution for the participating teams. All teams had one formal team leader, while the rest of the teams’ leadership was more informal. Team A’s task was to provide technology support and on-site training to custo-

mers, whereas Team B focused on delivering products and supporting customers during the initial usage period of the product. The sub-locations in Teams A and B operated quite independently within their specific geographical areas providing services to local customers in their native languages, but team members shared resources and provided support to each other, working together in a moderately interdependent manner. Teams A and B used a common information and customer management tool, which brought transparency to the team both in terms of provid- ing information about customer cases as well as information on who was working on what. In addition, the teams followed a series of work processes that guided their work. Team B had weekly global meetings, while Team A lacked meeting routines and rarely gathered for formal meetings. Teams C, D and E (from TechAlpha) developed software and worked together in

a highly interdependent manner across sub-locations. They followed the agile work process Scrum, which is a development process for team tasks consisting of short iterative rounds, where the team is given significant autonomy to carry out their tasks in whatever way they find necessary [65]. Scrum consists of a series of meetings, such as daily status, planning and retrospective meetings, so it requires intensive communication. In addition, the teams we studied used an issue tracking management tool (commonly used in Scrum), listing all the team’s tasks and each member’s current task, to enable task sequencing and delegation. The teams had two assigned roles: Product Owner (PO) and Scrum Master (SM). The SM facili- tated teamwork by removing obstacles, keeping the team focused on the task, and ensuring that the team adheres to team rules. The PO represented the voice of the customer and was ultimately responsible for team success or failure. In our sample, the PO also functioned as the formal team leader.

328 NORDBÄCK AND ESPINOSA

T ab le

1. T ea m

C om

po si ti on

In fo rm

at io n

O rg .

T ea m

T ea m

si ze

T ea m

T yp

e L oc at io ns

(a nd

N um

be rs ) of

In te rv ie w ed

M em

be rs

C om

m un

ic at io n te ch no

lo gy

T em

po ra l

di st ri bu