Levina & Vaast/Status and in Offshore Collaboration MIS Quarterly Vol. 32 No. 2, pp. 307-332/June 2008 307 SPECIAL ISSUE INNOVATING OR DOING AS TOLD? STATUS DIFFERENCES AND OVERLAPPING BOUNDARIES IN OFFSHORE COLLABORATION1 By: Natalia Levina Information Systems Group/IOMS Stern School of Business New York University 44 West Fourth Street, Suite 8-78 (KMC) New York, NY 10012 U.S.A. nlevina@stern.nyu.edu Emmanuelle Vaast School of Business, Public Administration and Information Sciences Long Island University 1 University Plaza���H700 Brooklyn, NY 11201 U.S.A. emmanuelle.vaast@liu.edu Abstract Increasingly, firms source more complex and strategic as well as harder to codify information technology projects to low- cost offshore locations. Completing such projects success- fully requires close collaboration among all participants. Yet, achieving such collaboration is extremely difficult because of the complexity of the context: multiple and over- 1 This paper was recommended for acceptance by Associate Guest Editor Kate Kaiser. lapping boundaries associated with diverse organizational and national contexts separate the participants. These boundaries also lead to a pronounced imbalance of resources among onshore and offshore contributors giving rise to status differences and inhibiting collaboration. This research adopts a practice perspective to investigate how differences in country and organizational contexts give rise to boundaries and associated status differences in offshore application development projects and how these boundaries and status differences can be renegotiated in practice to establish effec- tive collaboration. To illustrate and refine the theory, a qual- itative case study of a large financial services firm, which sourced a variety of high-end IT work to its wholly owned subsidiaries (���captive centers���) and to third party vendors in multiple global locations (in particular, to India and Russia), is presented. Using a grounded theory approach, the paper finds that differences in country contexts gave rise to a num- ber of boundaries that inhibited collaboration effectiveness, while differences in organizational contexts were largely mediated through organizational practices that treated vendor centers and captive units similarly. It also shows that some key onshore managers were able to alleviate status dif- ferences and facilitate effective collaboration across diverse country contexts by drawing on their position and resources. Implications are drawn for the theory and practice of global software development and multiparty collaboration. Keywords: Offshore software development, outsourcing, collaboration, qualitative methods, boundaries, status, power, Bourdieu, practice theory, cross-cultural teams, distributed teams, virtual teams, middle managers

Levina & Vaast/Status and in Offshore Collaboration 308 MIS Quarterly Vol. 32 No. 2/June 2008 Introduction Companies are sourcing increasingly complex parts of their software development process to their own subsidiaries in low-cost countries (sometimes called ���offshore captive units���) or to third party offshore vendors (referred to as ���offshore outsourcing���) (Stack and Downing 2005). High-technology firms like Microsoft, Yahoo!, Google, and IBM have moved parts of their strategic research and development activities to India, China, and Russia. An innovative software application like ���Google Finance,��� for example, was, for the most part, developed by Google���s Indian subsidiary (Padmanabhan 2006). Whether or not firms find actual cost savings in offshoring, complex and poorly codifiable work has been the subject of intense debate (e.g., Pallatto 2005). Whatever the case may be, this practice is on the rise and firms are realizing that it requires organizational capability to undertake multiparty collaboration spanning geographic and temporal distances as well as organizational, national, and professional boundaries (Couto et al. 2006). Multiparty collaboration (co-laboring) occurs when parties with different practices, interests, and competencies engage in joint work. However, because of complex internal and external dynamics, effective collaboration cannot be measured by objective outcomes alone (e.g., whether the project was completed on time or on budget). Rather, the work of Hardy et al. defines effective collaboration as a process that (1) leverages the differences among participants to produce innovative, synergistic solutions and (2) balances divergent stakeholders��� concerns (2005, p. 58). This process is faci- litated by the existence of shared identity and practices (Hardy et al. 2005 Levina and Vaast 2005) but is impeded by status differences among participants which inhibit open dialogue (Hoegl and Gemuenden 2001 Levina 2005 Metiu 2006). Social boundaries and physical distance separate participants on multiparty projects and make it difficult to establish shared identity and practices (Levina and Vaast 2005). Boundaries can also be used by participants as sources of distinction, thus creating status inequalities (Levina 2005). Effective collaboration is difficult to achieve in global off- shoring projects as there are often multiple boundaries that must be bridged simultaneously (Espinosa et al. 2003 Hinds and Bailey 2003). Some of these boundaries have been studied separately in the past. For example, researchers looked at how cultural (Cramton and Hinds 2007 Krishna et al. 2004 Lam 1997), organizational (Espinosa et al. 2003 Lam 1997 Srikanth 2007), and functional (Birnholtz and Finholt 2007 Espinosa et al. 2003) boundaries each pose challenges to distributed teams. It remains to be seen, though, which of these or other boundaries become salient in affecting the collaboration and how different boundaries interrelate (Espinosa et al. 2003). While, to date, researchers have tended to focus on physical distance (temporal or spatial) or institutionalized social boundaries (cultural, organizational, or functional), it has been argued that the most salient boundaries are often situated in the practices of collaborating parties (Cramton and Hinds 2007 Walsham 2002). It is well known that one of the key obstacles in achieving effective collaboration in distributed teams is the embeddedness of knowledge in local, situated practice and the resulting difficulty of establishing a common ground (Cramton 2001 Hinds and Mortensen 2004 Kumar et al. 2004 Lam 1997 Metiu 2006). Research also shows that status differences arising from the context of collabo- ration itself (e.g., who owns the code) may become key impediments for effective collaboration (Metiu 2006). Practice theory has shed much light on how obstacles for effective collaboration and related power relations emerge in organizations (Carlile 2004 Levina and Vaast 2005). It argues that boundaries set agents apart on the basis of their practices and thus become salient or stop mattering as prac- tices evolve (Bourdieu 1977). Status differences are seen as a result of differential access to resources, which, in turn, are produced or recognized as relevant or irrelevant through the everyday actions of actors (Bourdieu 1977). Thus, on the basis of practice theory we can address the following research questions: How do differences in country and organizational contexts give rise to boundaries and associated status dif- ferences in offshore software development projects, and how can boundaries and status differences be renegotiated in prac- tice to establish effective collaboration? Drawing on Bourdieu���s writings we use an in-depth case study of one firm���s offshore information systems development (ISD) projects to address these questions. This paper is organized as follows. First, we introduce our theoretical perspective. Next follows the details about the case study method and the research site. Then, the findings section identifies and describes the salient boundaries, explores how specific boundaries gave rise to, and were reinforced by, status differences and discusses how certain actors helped alleviate these differences. The discussion section then expands our theoretical model on the basis of these findings. We conclude with the implications of our study for the theory and practice of global distributed teams and sourcing management.

Levina & Vaast/Status and in Offshore Collaboration MIS Quarterly Vol. 32 No. 2/June 2008 309 Background: Boundaries and Status Differences in Offshore Collaboration A Practice Perspective on Offshore Collaboration Practice theory attributes the emergence, institutionalization, and transformation of socio-structural properties to the micro social interactions of people within the context of their everyday practices (Bourdieu and Wacquant 1992 de Certeau 1984 Giddens 1984).2 Enacted structural properties constrain social activity, but they can also be transformed through agents��� actions. Status differences arise as agents do not share equal access to three fundamental types of capital (resources), including economic capital (money, time, access to technology), intellectual3 capital (professional expertise, education, ownership of information), and social capital (networks of interpersonal relations upon which an agent can draw). There is also symbolic capital, which refers to agents��� differential ability to classify any other resource as valuable (Bourdieu and Wacquant 1992). Through their practices, agents are constantly engaged in shaping fields of practices as well as the boundaries that separate these fields. Boundaries delimit fields and arise from differences in practices that are differentially recognized and rewarded across fields (Levina and Vaast 2006). Simul- taneously, fields of practice emerge as constellations of agents who share unique sets of practices and interests while producing their own unique forms of capital (Bourdieu and Wacquant 1992). Agents within fields have different accu- mulations of relevant types of capital and, hence, are divided into ���haves��� and ���have-nots������a distinction often differen- tiating newcomers from old-timers within a field. Fields of practice may become institutionalized, fragmented, and/or overlap according to the changing practices of agents and the resources they acquire in the process (Abbott 1995). Thus, practices, boundaries, and fields are mutually constructing none is given theoretical dominance (Silber 1995). These fields are dynamic and their boundaries can be rene- gotiated in practice. In the context of ISD projects, the pre- existing differences in backgrounds of project participants will become more or less salient in producing status dif- ferences depending on the composition of the team and the context of work (Levina 2005). Moreover, because the practices surrounding the development of a new information system can produce a unique distinction between agents who know and control the design and those who do not, the situated development effort may give rise to a new field and associated boundaries (Levina 2005 Levina and Vaast 2005). Each of the preexisting and emergent fields will unite agents on the basis of their joint interests and practices and set them apart when those differ. Thus, fields and boundaries produce both sharedness and differences. The sharedness is a necessary condition for achieving effective collaboration as without joint engagement and at least a minimal common understanding, collaboration is not possible (Carlile 2002 Cramton 2001 Levina and Vaast 2005). However, even when people share some interests and practices, differences among them can be used as status markers and impede collaboration (Metiu 2006). They can impact whether participating parties��� knowledge is ���transformed��� to become an integrated part of a synergistic solution (Carlile 2004) or whether it is merely combined or even completely ignored (Levina 2005). Prior literature has discussed how the differences in interests, identities, knowledge, and language create impediments for effective collaboration and how they can be mediated in practice in some cases (Carlile 2002 Levina and Vaast 2005 Orlikowski 2002). In this paper, we focus on the status dif- ferences as those are particularly pronounced in offshoring contexts. Situated Boundaries in Offshore Collaboration The challenges of offshore collaboration can be interpreted using a practice perspective. For example, it has been argued that differences in national culture are among the key chal- lenges to offshore teams (Carmel and Agarwal 2001 Carmel and Tjia 2005 Cramton and Hinds 2007 Cusumano 2006 Dibbern et al. 2008). From a practice perspective, societal characteristics (such as the distribution of educational degrees and economic resources as well as taste, communicative patterns, treatment of authority, etc.) are institutionalized practices that have emerged over time from interactions among unequal agents (Bourdieu 1984). Every society is characterized by its own relative distribution of economic, cultural, social, and symbolic capital as well as by its relation 2 See Orlikowski (2002) for a more in-depth discussion on the relevance of practice theory for the organization and information systems field. The scholars we cite have developed different theories based on the practice perspective, but all agree on its fundamental tenets. 3 According to Bourdieu, cultural capital refers to the ability to produce new representations of practice or information. In organizational theory, this kind of resource is typically termed intellectual capital. We will use this term to avoid confusion between the cultural capital of specific nations (Indian or United States) and Bourdieu���s notion of cultural capital, which can exist within societies, organizations, industries, professions, etc.

Levina & Vaast/Status and in Offshore Collaboration 310 MIS Quarterly Vol. 32 No. 2/June 2008 to other nations (Bourdieu 1998). Different cultural norms that are usually discussed in cross-cultural research such as attitude toward authority (Hofstede 1980) are produced and reproduced through actions and reinforced through the symbolic interpretations of these actions. In turn, they alter how agents work and interact (e.g., the degree of formalism in business communications). Indeed, these symbolic re- sources are frequently used by high-status agents to maintain their positions of power. A priori differences in status, however, do not entirely deter- mine how collaboration unfolds on any given project: human agency is not negated and often facilitates the renegotiation of differences. Walsham (2002), for example, shows how dif- ferences in cultural norms (a symbolic difference) were renegotiated on a cross-cultural project to arrive at a shared set of norms. Cramton and Hinds (2007) have illustrated how a personnel selection practice helped render traditional differences in norms and expectations between German and Indian cultures (e.g., attitude to authority) less relevant. Using practice theory to understand the role of the organi- zational boundaries, we can see that differences in organi- zational affiliations set agents apart on the basis of their identities, interests, and practices (Jarzabkowski 2004 Levina and Vaast 2005). Scholars have argued that organizations exist, in part, to create institutionalized practices which are easily spread internally but hard to copy externally (Kogut and Zander 1993). This is achieved by building capital within organizations that facilitates the creation of network ties, interpersonal relations, and shared systems of meaning (Nahapiet and Ghoshal 1998). Also, using economic perspec- tives such as property rights theory (Hart and Moore 1990), boundaries of the firm are used to align interests and facilitate particular investment patterns in shared economic capital. When organizations come to work together in an outsourced relationship, the lack of shared economic, intellectual, social, and symbolic resources may define the organizational boun- daries between them, leading to power dynamics that under- mine collaboration (Allen et al. 2002 Nicholson et al. 2006). Yet, these sorts of difficulties are not insurmountable, as prior accounts of successful alliances (Dyer and Singh 1998) and outsourcing relationships (Helper et al. 2000) attest. More- over, it is still unclear if the lack of organizational boundaries plays any role in facilitating effective collaboration or the concomitant creation of knowledge (Grandori and Kogut 2002 Merali 2002). Even inside the boundaries of globally operating firms geographically dispersed units may become isolated and fail to gain or share knowledge with other units (Moteiro et al. 2007). In the offshore context, the role of organizational boundaries is even less clear as internal and external (outsourced) rela- tionships often have relatively little history (3 or 4 years). Thus, little shared capital (economic, intellectual, and/or social) is accumulated in internal or external relationships alike. To confound things further, given the high status of large Western client firms (their higher symbolic significance and economic wealth), vendor employees may identify more readily with the client organization as opposed to the offshore vendor company (Ravishankar and Pan 2006). Also, the usual model of outsourced relationships assumes the vendor has greater expertise (Goles 2003). In offshoring, however, onshore clients may have more relevant technical expertise pertaining to their industry and systems than the vendor (Dibbern et al. 2008). Differences in professional and industry practices may form other important boundaries. Professional and industry fields have been studied as examples par excellence of social spaces with highly institutionalized practices, significant boundaries precluding easy socialization of new members, and strong power relations differentiating participants (Bourdieu 1984 Carlile 2004 DiMaggio 1991 Montgomery and Oliver 2007). In the offshore context, these boundaries may be particularly pronounced, at least initially. For example there are signifi- cant differences in professional training (e.g., more mathe- matically trained IS developers in Russia versus the United States), vendor practices (e.g., more CMM-54 certified vendor firms in India than in Western Europe), and vendor industry age (younger in offshore countries as compared to the onshore countries according to Arora and Gambardella 2006). Differences in IT staff���s professional training may result in knowledge hoarding by high status parties (Espinosa et al. 2003 Metiu 2006), thereby reinforcing institutionalized boundaries. Yet, for projects like Google Finance (Pad- manabhan 2006) and other examples of effective offshore collaboration (Helper and Khambete 2004) to be realized, some renegotiation and transformation of these status differences must have occurred. Next we present the method we used to undertake the investi- gation of how status differences were established and renego- tiated in practice on projects offshored to India and Russia, describing our research site and data analysis approach. 4 CMM-5 refers to the high level of standardization and maturity in software development practices (Paulk et al. 1993).

Levina & Vaast/Status and in Offshore Collaboration MIS Quarterly Vol. 32 No. 2/June 2008 311 Methods We used an interpretive case study approach (Walsham 1995) to understand which boundaries produced status differences and which practices helped alleviate those differences. An interpretive approach was essential in so far as collaborative processes and the emergence of relevant boundaries are deeply embedded in actors��� subjective understanding of their work. This research analyzes a firm that sourced its application development work to Asia and Eastern Europe while making use of captive centers and third-party providers in each of these locations. Rarely used together, this set of sourcing policies allowed for a unique opportunity to better understand the role of national and organizational contexts in collaboration. Site Description The study was conducted by this paper���s first author (hence- forth, ���the field researcher���) at Global Bank (pseudonym), a large, multinational financial services firm with various divisional headquarters in the United States and Western Europe. Starting in 1989, the bank established relationships with several Indian vendors for providing onshore sub- contractors for the bank���s operations in local (offshore) markets. In 1996, the bank started offshoring its application maintenance projects to its newly established Indian captive center and to several vendors due to IT labor shortages. Offshoring picked up in 2001 during the downturn in U.S. financial markets. The first venture into Russia dates back to 2001. By 2005 (the beginning of the study), Global Bank had captive centers in both India and Russia and long-term relationships with a number of local and global vendors in both regions. The amount of money spent on off-shored projects (in IT and other business processes) was estimated to be a quarter of a billion U.S. dollars in 2005. This was in addition to IT work sourced by Global Bank to IT services vendors in its onshore locations which included the United States, Western Europe, Australia, Singapore, and Japan (Figure 1). In terms of size, the captive unit in Russia reached 230 people in 2005 and Global Bank���s dedicated offshore development center (ODC) set up by the primary Russian vendor (hence- forth Russian Vendor ODC) reached approximately 150 people. In India, the size of the ODC with one of the top tier vendors (the one who purchased the captive center established earlier) was over 2,000 people (henceforth Primary Indian ODC). Operations with the other local and global vendors ranged from 30 to 500 people. There was also a large captive facility in India used primarily for financial services opera- tions. In ODCs, the contracts were typically two-year, renewable, billed as time and materials with a capped budget. Fixed price contracts have become more popular in recent years, but the old contracts were still delivered by ODC���s dedicated employees. By 2005, Global Bank was sourcing a variety of IT work to all of these locations. This was largely due to independent decision making by middle managers with respect to sourcing. For many years, they were able to choose where and how to source with relative autonomy. Providers were chosen for a variety of reasons, including Global Bank���s prior relationship with the provider, the provider���s prior experience with a given type of application, or attractive rates. New geographic areas were added when they offered qualified developers at competitive rates. As a result, the so-called ���high-end��� pro- jects (application development or migration projects that involved complex design, hard to specify requirements, and a fair degree of business knowledge) were sourced throughout these diverse settings. The breakdown of functions between the onshore and offshore locations was also very similar across settings (see Appendix A for a description of some of the projects and Appendix B for a breakdown of work activities). Finally, there was almost no interaction between the Russian and the Indian offshore staff because different middle managers sourced their projects independently to either Russia or India. Data Collection and Analysis Data collection spanned May 2005 to May 2006, with some follow-up interviews conducted until May 2007. The CIO of one of the bank���s divisions provided entry and widespread access to the firm. The CIOs and the participants were told that the focus of the study was on understanding collaborative practices on offshored projects. Participants were motivated to contribute because they felt it was difficult to collaborate effectively and wanted to reflect on their frustrations and/or successes. In-depth, semi-structured interviews lasting from 40 minutes to 2 hours constitute the majority of the data, sup- plemented by business press accounts. The data was collected in the United States, Western Europe, India, and Russia, mostly through face-to-face interviews. A total of 69 inter- views were conducted and recorded. Over 40 projects (ranging in size from five to 80 participants and averaging 15) were discussed. The study was limited to the practices concerning offshore and nearshore vendors, and excluded onshore vendors.