Value Creation in Alliance Portfolios: The Benefits and Costs of Network Resource Interdependencies

Authors


Ulrich Wassmer, John Molson School of Business, Concordia University, 1455 Guy Street, MB 13-355, Montreal, Quebec, Canada H3G 1M8, Tel: 514 848 2424 xt 2921, Fax: 514 848 4292. E-mail: uwassmer@jmsb.concordia.ca

Abstract

We draw on the resource-based view of the firm and the alliance benefits and costs literature to advance a model of value creation in firms that access network resources through multiple simultaneous strategic alliances with different partners. Our model suggests that value creation on the alliance portfolio level is a function of the benefits created through synergistic combinations of network resources accessed through alliances with different partners and of the costs generated by the substitutability of resource combinations between the focal firm and its alliance partners. We specify the conditions under which firms can leverage their network resources accessed from alliances with different partners to create benefits above and beyond the benefits they create at the level of any individual alliance. We conclude that the value creating potential of network resources should not only be evaluated on the basis of each individual alliance but also from an alliance portfolio perspective.

Introduction

Value creation1 is a central topic in the strategic management literature (Brandenburger and Stuart, 1996; Bowman and Ambrosini, 2000; Lepak et al., 2007). With the increasing importance of interfirm alliances as a strategic device to create value by developing and accessing strategically critical resources (Eisenhardt and Schoonhoven, 1996; Das and Teng, 2000; Gulati, 2007), researchers have developed value-based perspectives on interfirm alliances by taking both alliance benefits and costs into account (Zajac and Olsen, 1993; Madhok and Tallman, 1998). Prior research has also extended the traditional resource-based view of the firm (RBV), a key theory used to explain how firms can achieve a sustainable competitive advantage and create value (Wernerfelt, 1984; Barney, 1991; Peteraf, 1993), to what has come to be known as ‘interconnected firms’, namely, firms that participate in interfirm alliances (Dyer and Singh, 1998; Lavie, 2006). Authors in this line of research have specified the conditions under which interfirm alliances can be a source of competitive advantage and value creation (Dyer and Singh, 1998), and identified the types of rents that interconnected firms can create and appropriate (Lavie, 2006).

In today's business landscape, most firms no longer rely on a single alliance: many firms maintain entire alliance portfolios comprised of multiple simultaneous strategic alliances with different partners in order to access a broad range of resources (Parise and Casher, 2003; Hoffmann, 2005, 2007; Lavie, 2007; Wassmer, 2010). For the purpose of this paper, a firm that maintains an alliance portfolio is referred to as an alliance portfolio maintaining firm. A central issue concerning alliance portfolios is the portfolio effect created by the sub- or super-additive interdependencies between individual alliances in an alliance portfolio that make the overall portfolio value greater or smaller than the sum of the values created by each individual alliance (Vassolo et al., 2004). Although the extant literature addresses the issue of resource interdependencies on the single alliance level, namely, between a focal firm and its alliance partner in a given alliance (Chung et al., 2000; Das and Teng, 2000; Rothaermel, 2001), and also acknowledges resource interdependencies beyond the single alliance level (Lavie, 2007), it has not formalized how interdependencies between network resources accessed through multiple simultaneous alliances with different partners affect value creation in alliance portfolios. One empirical study that takes an important step towards addressing the portfolio effect is Vassolo et al.'s (2004) analysis of biotechnology alliances as portfolios of investment options. This study combines real options logic with the RBV to examine how the portfolio effect makes the value of an alliance portfolio greater or smaller than the sum of the individual alliance values.

Although existing theoretical RBV extensions take an important step towards better understanding value creation in strategic alliances (Dyer and Singh, 1998; Lavie, 2006), their applicability to alliance portfolio maintaining firms is yet limited. Indeed, they do not account for, nor formalize, the effect that interdependencies between the network resources such firms access through their multiple simultaneous alliances with different partners have on the overall value of their alliance portfolio. These developments essentially provide theory that extends from the single alliance level of analysis to the alliance portfolio level of analysis in an additive way, namely, equate the value of the alliance portfolio with the sum of the values created by each individual alliance in the portfolio, and thus leave aside the portfolio effect component. Similarly, the existing value-based interfirm collaboration frameworks take the individual alliance as the unit of analysis and view the value created by an alliance as a function of both its benefits and costs (Zajac and Olsen, 1993; Madhok and Tallman, 1998). Such a perspective does not capture benefits and costs, and thus value creation, from an alliance portfolio perspective.

The objective of this paper is to contribute to filling the gap the extant literatures have left in examining the role of network resource interdependencies in value creation in alliance portfolios by providing a systematic theoretical analysis concerning the process of value creation (i.e., how value is generated) in alliance portfolios (Lepak et al., 2007). Here, the purpose of this theorizing is not to determine the factors for value creation on the individual alliance level of analysis but rather to identify alliance portfolio level factors that affect the benefits and costs of interfirm collaboration. Since the traditional RBV and its extensions to interconnected firms put more emphasis on the resource related benefits than costs, we combine insights from three literatures to provide a more integrated perspective. First, we draw on the general value creation literature (Bowman and Ambrosini, 2000; Lepak et al., 2007), and the value focused alliance frameworks (Zajac and Olsen, 1993; Madhok and Tallman, 1998), to develop an overall value perspective on alliance portfolios that takes both benefits as well as costs into account. Second, we draw on insights from the RBV in order to understand how network resource interdependencies affect alliance portfolio benefits. Lastly, we incorporate insights from the literature on control (Hennart, 1988; Williamson, 1991; Dyer, 1997), cooperation (White and Lui, 2005), and competitive (Koh and Venkatraman, 1991; Park and Zhou, 2005), costs of alliances to examine how network resource interdependencies affect the costs, and thus the value, of individual alliances in an alliance portfolio.

More specifically, we identify and discuss three issues that determine value creation in alliance portfolios: (1) the opportunity to leverage supplementary or complementary network resources accessed through multiple simultaneous alliances with different partners to create benefits above and beyond the benefits that these resources create on the individual alliance level; (2) the conditions under which alliance portfolio maintaining firms can exploit such network resource supplementarity or complementarity to enhance alliance portfolio benefits; and (3) the substitutability of resource combinations between an alliance portfolio maintaining focal firm and its existing alliance partners as a mechanism that increases the costs, and thus reduces the value, of individual alliances in the focal firm's alliance portfolio.

In terms of how this study positions vis-à-vis existing work, it is worthwhile pointing out two things. First, our intended contribution is different from Vassolo et al. (2004) in the sense that we do not focus on the option value but rather apply a value perspective concerning the active resource combinations already in the alliance portfolio. Second, we do not address the issue of value appropriation/capture on the level of individual alliances, namely, how alliance benefits are split among partners. Value appropriation and capture is frequently confused with the concept of value creation (Lepak et al., 2007), and represents a complex research issue that merits theorizing in its own right (Bowman and Ambrosini, 2000; Gulati and Wang, 2003; Dyer et al., 2008). More specifically, in the alliance context, value appropriation is a research issue that takes the individual alliance (be it dyadic or multi-partner) as the unit of analysis and involves non-RBV relevant factors such as bargaining power (Brandenburger and Stuart, 1996), contractual safeguards (Hennart, 1988), investments in transaction-specific assets (Dyer et al., 2008), relative absorptive capacity (Lane and Lubatkin, 1998), and others. However, since we are interested in understanding network resource interdependencies on the alliance portfolio level of analysis we clearly separate value creation on the alliance portfolio level from value appropriation on the level of the individual alliance.

The remainder of this paper is structured as follows. In the next section, we review the relevant background literature. We then develop a model of value creation in alliance portfolios. Lastly, we conclude the paper with a discussion of the implications of the proposed theoretical model and highlight the limitations and avenues for future research.

Value creation, firm resources, and resource combinations

As a first step in our theory building, we identify and review fundamental conceptual underpinnings of value creation as well as relevant theoretical concepts related to within and beyond firm boundary resources. We then draw on these ideas to establish the cornerstones of our theoretical model.

Value creation and firm resources

Value creation as a concept has a content (i.e., what is valuable), and process (i.e., how value is generated), dimension (Bowman and Ambrosini, 2000; Lepak et al., 2007). In their definition of value creation, Bowman and Ambrosini (2000, p. 2) conceptually distinguish between what they call use value (i.e., customers' perceptions concerning the benefits of a product or service related to their needs), and exchange value (i.e., the monetary amount paid by a buyer to a seller for the use value). Lepak et al. (2007) point out that a positive use value does not always guarantee a positive exchange, that is, monetary, value. Indeed, Lepak et al. (2007, p. 182) posit that in order to create a positive monetary value, two conditions must be fulfilled: (1) the monetary amount paid by the buyer must exceed the costs of the seller for producing the value; and (2) the monetary amount paid by the buyer is determined by the perceived performance differential between the buyer's use value and the buyer's second best alternative. Similarly, Brandenburger and Stuart (1996) point out that value is created if a buyer's willingness to pay for a product or service exceeds the opportunity cost.

From an RBV perspective it has been argued that benefits are rarely created by single resources: it usually requires entire combinations of resources to render a benefit creating service (Grant, 1991; Amit and Shoemaker, 1993; Miller and Shamsie, 1996; Teece et al., 1997; Eisenhardt and Martin, 2000). A benefit creating service can be thought of as a service rendered by a resource combination that enables a firm to reduce its costs through improving its operational efficiency and/or enhance its revenues through creating additional product and service offerings (Barney, 1997). Consequently, a resource combination creates monetary value for a firm when: (1) it renders such a benefit creating service; and (2) the benefits exceed the costs incurred for producing them (Brandenburger and Stuart, 1996; Lepak et al., 2007). A resource combination destroys monetary value when its costs exceed its benefits.

The latest landmark developments within the RBV suggest that, because all resources2 can be categorized into either scale-free or non-scale-free resources, the issue of resource scalability and capacity constraints should be incorporated into RBV research (Levinthal and Wu, 2010). Scale-free resources can be thought of as resources, such as a brand name, that face ‘limits on the breadth of its fungibility (i.e., how broadly fungible is a given brand name) but not on its extent of application (i.e., the number of markets in which a given brand can be applied for a given level of fungibility)’ (Levinthal and Wu, 2010, p. 784). Non-scale-free resources, such as a plant, on the other hand are resources that can only render a limited amount of services, and consumption by one user of a given quantity of such non-scale-free resources reduces the amount of the considered resources available for use by others. Because resources can either be scale-free or non-scale-free (Levinthal and Wu, 2010), any resource combination is essentially a function of these two resource types. Consequently, a resource combination is complete when it contains all resource elements required to render a benefit creating service. In other words, to be complete, a resource combination has to contain: (1) all required types of scale-free resources; and (2) the minimum quantity of non-scale-free resources required to produce a benefit. By ‘minimum quantity’, we mean the quantity required to produce the minimum possible benefit to the focal firm. In other words, even though a resource combination may have the minimum quantity of one or more non-scale-free resources to produce the minimum benefit, a greater benefit is prevented by a resource constraint, namely, the limited available quantity of the non-scale-free resource/s deployed in the combination. Consequently, greater available quantities of such non-scale-free resources in such a combination could lead to greater benefits, if they relieve the capacity constraint affecting the extent of services rendered by the combination. It thus follows that a resource combination comprised of scale-free and/or non-scale-free resources is incomplete in the sense that it fails to render a benefit creating service when it lacks either: (1) one or more scale-free resources; or (2) the minimum required quantity of one or more non-scale-free resources needed to produce the minimum possible benefit creating service, or both.

We will return to this argument later when we examine the conditions under which an alliance portfolio maintaining firm can exploit supplementary or complementary network resources accessed through multiple simultaneous alliances with different partners. Before we examine the concepts of resource supplementarity and complementarity, we first discuss the role of within and beyond firm boundary resources in alliance portfolio maintaining firms.

Within and beyond firm boundary resources

From the perspective of the RBV, firms have a sustainable competitive advantage and achieve superior performance when they possess a stock of valuable, rare, imperfectly imitable, and non-substitutable resources (Barney, 1991). It has also been argued that such strategically relevant firm resources are generally non-tradeable and thus cannot be acquired in strategic factor markets (that are considered inefficient). Instead, they can only be accumulated internally by choosing appropriate paths of resource flows over a period of time (Dierickx and Cool, 1989), or accessed through non-traditional market mechanisms, in particular strategic alliances used to overcome these market inefficiencies (Eisenhardt and Schoonhoven, 1996; Ahuja, 2000; Das and Teng, 2000). More recently, Makadok (2001) suggested that firms can achieve a competitive advantage not only by picking critical resources but also by building critical capabilities that help them deploy their resources more productively.

While the traditional RBV has mainly focused on resources that reside within firm boundaries and are owned and controlled by a focal firm (Wernerfelt, 1984; Barney, 1991), recent theoretical extensions have relaxed this condition and have also included resources that firms access through interfirm alliances in the analysis of competitive advantage and rent generation (Dyer and Singh, 1998; Lavie, 2006). For example, Lavie (2006) distinguishes between a focal firm's non-shared resources, a partner firm's non-shared resources and the shared resources which are contributed by both the focal firm and its partner/s to an alliance. Gulati (2007, p. 8) refers to such beyond firm boundary resources as network resources which ‘encompass resources that a firm's partners may possess and are available to a focal firm through its connections with those firms.’ More recently, Lavie (2008) suggested that viewing network resources as ‘assets that are owned by the firm's partners but can potentially be accessed by the firm through its ties to these partners’ offers the greatest opportunities to further enhance management theory.

We build on Gulati's (2007) and Lavie's (2008) work in order to analyze how alliance portfolio maintaining firms can create value in their alliance portfolios and conceptually divide such firm resources into: (1) resources residing within firm boundaries, that is, resources that are owned and controlled by the alliance portfolio maintaining focal firm; and (2) resources residing beyond firm boundaries but to which the alliance portfolio maintaining focal firm has access through its alliances. Henceforth, we shall respectively refer to these resources as the firm's (1) own resources and (2) network resources (Gulati, 2007; Lavie, 2008). Moreover, we incorporate the latest insights from the RBV that stress the importance of incorporating the issue of resource scalability into RBV research (Levinthal and Wu, 2010). Because all resources can be categorized into either scale-free or non-scale-free resources (Levinthal and Wu, 2010), an alliance portfolio maintaining firm's own and network resources can, thus, be either of a scale-free or non-scale-free nature.

Figure 1 depicts the two resource types that make up an alliance portfolio maintaining firm's resource stock. Here, a focal alliance portfolio maintaining firm i possesses four own resources: Ri1, Ri2, Ri3, and Ri4. Moreover, focal firm i is engaged in two simultaneous alliances (i.e., alliances ih and ij) with two different partners (i.e., firms h and j). Through alliance ih with partner h, focal firm i obtains access to h's resource Rh1, resulting in the jointly deployed resource combination Ri4Rh1. Alliance ij with partner j, grants focal firm i access to j's resource Rj1, resulting in the jointly deployed resource combination Ri3Rj1. Thus, focal firm i's network resources are represented through Rh1 and Rj1.

Figure 1.

Resource stock of an alliance portfolio maintaining firm

Resource supplementarity and complementarity and value creation in strategic alliances

In this study we view a strategic alliance as an arrangement between firms for the joint deployment and coordination of one or more resource combination/s as part of a project or business operation (Dyer and Singh, 1998; Dussauge et al., 2000). As mentioned above, a number of studies have applied an economic value focused perspective to interfirm collaboration and highlighted the importance of focusing on the transaction value created by interfirm collaboration, as opposed to solely the transaction costs (Zajac and Olsen, 1993; Madhok and Tallman, 1998). Zajac and Olsen (1993) define such value as the difference between the benefits and costs created by the collaboration. Madhok and Tallman (1998) take a similar approach and define the potential economic value of interfirm collaboration as the difference between the potential rents to specialized resources and costs or expenditures associated with the collaboration. Thus, each alliance formation decision is essentially a trade off decision that considers the ex-ante benefits and costs of the envisioned alliance (Buckley and Casson, 1988; Koh and Venkatraman, 1991; Madhok and Tallman, 1998).3

The extant RBV-alliance literature has identified resource supplementarity and complementarity between alliance partners as important explanatory factors for alliance formation but also for value creation on the individual alliance level of analysis. Resource combinations comprised of supplementary partner resources essentially bring together identical resources in the same product or geographic market domain (Das and Teng, 2000). To illustrate, in order to develop a new aircraft engine Pratt and Whitney and GE bundled supplementary R&D resources in the form of aerospace R&D engineers because both companies did not have enough R&D resource capacity available that could be dedicated to the project (PR Newswire, 1996). Alliances based on resource supplementarity logic have sometimes been referred to as ‘scale alliances’ (Hennart, 1988; Dussauge et al., 2000). In contrast, resource combinations comprised of complementary partner resources essentially bring together different and non-overlapping resources (Hill and Hellriegel, 1994; Gulati, 1995; Dyer and Singh, 1998; Chung et al., 2000; Das and Teng, 2000; Rothaermel, 2001; Gimeno, 2004). To illustrate the concept of resource complementarity between alliance partners, consider the Fuji-Xerox alliance (McQuade and Gomes-Casseres, 1991). In order to manufacture and distribute photocopiers in Japan, alliance partners Xerox Corp. and Fuji Photo Film bundled complementary resources in the form of Xerox's photocopying technology and Fuji's access and knowledge of the Japanese market. Alliances based on resource complementarity logic have often been termed ‘link alliances’ (Hennart, 1988; Dussauge et al., 2000).

While there seems to be little ambiguity around the concept of resource supplementarity, the concept of resource complementarity is however not totally unambiguous due to a vast number of different conceptualizations and definitions in the economics and strategy literatures (Ennen and Richter, 2010). Some of the existing definitions explicitly focus on the complementarity of resources (Wernerfelt, 1984; Harrison et al., 2001), while others focus on the complementarity of assets (Richardson, 1972; Teece, 1986; Amit and Shoemaker, 1993, Tripsas, 1997), products and markets (Ansoff, 1965; Lewellen, 1971; Haspeslagh and Jemison, 1991; Chung et al., 2000), or activities (Milgrom and Roberts, 1990). In his seminal article, Wernerfelt broadly defines complementary resources as resources ‘which combine effectively with those you already have’ (1984, p. 175). In the context of innovating firms, Teece (1986, p. 288) refers to complementary assets or capabilities as those that are to ‘be utilized in conjunction with other capabilities or assets’ needed to make an innovation successful. In a more economics-based perspective, a number of researchers stress a mutually reinforcing process and posit that strategic assets, namely, resources and capabilities (Amit and Shoemaker, 1993), or activities (Milgrom and Roberts, 1990), are complementary when having more of one element raises the value of the other element and when a combination of these elements together is worth more than the sum of the individual values of having each element separately.

For the purpose of this study, we define supplementary alliance partner resources as identical resources (Das and Teng, 2000). By identical we mean the same resource type in the same product or geographic market domain, for example, manufacturing knowledge in pharmaceuticals or marketing knowledge in China. Our definition of complementary partner resources refers to resources that: (1) can be combined effectively, namely, used jointly in order to advance the resolution of a problem or contribute to the production of new and valuable services (Wernerfelt, 1984); and are either (2) of a different, that is, non-overlapping, type in the same product or geographic market domain and thus on a different stage of the value chain; or (3) of the same type, namely, on the same value chain stage, but in a different product or geographic market domain. To illustrate the first case, R&D knowledge in electronics and manufacturing knowledge in pharmaceuticals are different functional resource types, but due to the difference in product market domains they are not combinable. To illustrate the second case, in their study of alliance formations in the US investment banking industry, Chung et al. (2000) measure resource complementarity through a number of variables, including banks' strengths in particular geographic areas. Thus, two banks may have complementarity resources of the same type, for example, a marketing resource in the form of access to particular regional markets, but with different attributes, namely, bank i in geography k and bank j in geography l. By combining their complementary geographic presence, two banks may be able to offer customers better service or to reduce their operating costs. Thus, these complementary resources (i.e., geographically complementary branch networks), when combined offer a higher level of valuable service in the form of better quality or lower costs.

Overall, much of the extant literature views the combination of supplementary or complementary partner resources not only as a necessary but also as a sufficient condition for value creation in strategic alliances (Hill and Hellriegel, 1994; Gulati, 1995; Dyer and Singh, 1998; Chung et al., 2000; Das and Teng, 2000; Rothaermel, 2001). The assumption that combinations of either supplementary or complementary resources automatically create value is, however, problematic for two reasons. First, it ignores the fact that resources may be supplementary or complementary, as defined above, but as a combination fail to render a benefit creating service. To illustrate, an R&D resource in a certain product domain, for example a patent for an innovative technology and a manufacturing resource in the same domain, may be complementary in the sense that they can be combined effectively in order to contribute to the creation of a valuable service (Wernerfelt, 1984), but the resource combination as a whole fails to render a benefit creating service due to the lack of a marketing resource, for example, access to a distribution channel through which the end product can be distributed (Teece, 1986). Second, this assumption also ignores the second key condition for the creation of value: the benefits of the resource combination must exceed the costs for producing them. In other words, a resource combination comprised of either supplementary or complementary resources may render benefit creating services but may fail to create a monetary net value if the costs outweigh the benefits.

For the purposes of our theorizing, we view resource supplementarity or complementarity and the rendering of a benefit creating service by a resource combination as two necessary but individually insufficient conditions for value creation. We argue that a resource combination of supplementary or complementary resources that is enabled through a strategic alliance creates value only when: (1) it renders a benefit creating service; and (2) the resource combination's rendered benefit exceeds the costs of producing it.

To summarize, we have now clarified when resource combinations create value, identified the issue of completeness in resources combinations, and clarified the role of resource complementarity and supplementarity.

Network resource interdependencies and value creation in alliance portfolios

After having established the theoretical anchors for our theory building, we now move on to further explore how interdependencies between network resources affect value creation in alliance portfolios. More specifically, we discuss how these interdependencies affect both benefits as well as costs related to alliance portfolios.

Levels of value creation in alliance portfolios

For the purpose of this paper, we define an alliance portfolio as all of an alliance portfolio maintaining firm's active strategic alliances at a given point in time. From an external resource acquisition strategy point of view, the alliance portfolio is the result of wider corporate development decisions: it is a function of having chosen alliances over other external resource acquisition modes such as mergers and acquisitions which has been addressed by a separate stream of research (Dyer et al., 2004; Capron and Mitchell, 2009).

A key issue in defining alliance portfolios is to choose a definition that is in close alignment with the theoretical lens applied (Wassmer, 2010). Given our interest in interdependencies between network resources accessed through multiple simultaneous alliances with different partners, our main lens is the RBV and our focus lies on active resources and resource combinations involving network resources, that is, resources held by current partners; past alliances are non-problematic to our argument. Thus, we exclude inactive and past alliances, namely, past and inactive resource combinations that have ceased to earn rents, from the analysis. It is also important to point out here that we use the term ‘portfolio’ to refer to a set of alliances. This is not related to the portfolio concepts used in the finance literature which refers to a portfolio of investments or in the corporate diversification literature which uses the term ‘portfolio’ to refer to a set of businesses a firm is engaged in.

Moreover, an important assumption in our analysis is that alliance portfolio maintaining firms are boundedly rational. Recent research has indeed highlighted that alliance portfolios are often the result of a cumulative process through which multiple alliances accumulate over time in a somewhat haphazard way leading essentially to alliance proliferation and to a sub-optimally configured alliance portfolio (Wassmer et al., 2010). Thus, the state of an alliance portfolio at any given point in time represents the result of a process that is to a great extent rationally bound. In fact, it appears that the bounds of rationality are greater on the alliance portfolio level than on the level of the individual alliance, because the emergence of a firm's alliance portfolio involves multiple individuals and multiple organizational units and is spread out over long periods of time. Indeed, many firms treat the decision to form alliances as stand-alone decisions that are essentially left to the unit most directly concerned by the considered partnership, rather than as part of an integrated decision process that equally monitors value creation within individual alliances as well as at the entire alliance portfolio level (Wassmer et al., 2010).

As prior value creation research has suggested, being explicit about the level of analysis on which value creation is studied (Lepak et al., 2007), and about the examined value creation dimension (i.e., content versus process), is critical (Bowman and Ambrosini, 2000). In this study, we take a process perspective of value creation in alliance portfolios. More specifically, we focus on two levers that affect value creation in alliance portfolios: the effect of network resource interdependencies on: (1) alliance portfolio benefits; and (2) costs of individual alliances inasmuch as they are affected by the alliance portfolio.

As mentioned earlier, a core issue concerning value creation in alliance portfolios is whether the overall net value created by an alliance portfolio is greater or smaller than the sum of the values created by the individual alliances in the portfolio. If an alliance portfolio's overall net value is greater than the sum of the values created from all individual alliances in the portfolio then a positive alliance portfolio effect prevails. Consequently, when the alliance portfolio effect is negative, an alliance portfolio's overall net value is smaller than the sum of the values created from all individual alliances in the portfolio. For our theorizing it is important that we distinguish conceptually between two different components that determine the overall net value created by an alliance portfolio. First, there is the value that is created on the level of the individual alliances in an alliance portfolio. The second component is the net value of the alliance portfolio effect. To summarize, the overall value created by an alliance portfolio can be expressed through the following value function:

image

where: VAP-Overall is the overall net value created by an alliance portfolio; VAi is the net value created through individual alliance I; and VAP-Effect is the net value of the alliance portfolio effect.

We argue that the net value of the alliance portfolio effect (i.e., VAP-Effect), is determined through all alliance portfolio level benefits (i.e., benefit creating services rendered by network resources beyond their benefits rendered within the individual alliances in the portfolio), and the incremental costs concerning individual alliances in the portfolio that are negatively affected by interdependencies between network resources the focal firm accesses through multiple simultaneous alliances with different partners. The value of the alliance portfolio effect can be expressed as:

image

where: BAj is the benefits rendered by combination j of network resources accessed through multiple simultaneous alliances with different partners; and CAj is the costs of individual alliance i caused by the interdependencies of network resources accessed through multiple simultaneous alliances with different partners.

To sum up, we argue that an alliance portfolio maintaining firm's network resources can enhance as well as reduce value in alliance portfolios, that is, above and beyond the value that they create on the individual alliance level. Our theoretical model suggests that the benefits created in alliance portfolios are determined by the synergistic combinations of network resources accessed through multiple simultaneous alliances with different partners. Moreover, our model lays out the conditions under which an alliance portfolio maintaining firm can implement and exploit such synergistic combinations of network resources. Finally, our model suggests that alliance portfolio value is negatively affected by the substitutability of resource combinations between the focal firm and its existing partners. We now discuss the effect of network resource interdependencies on alliance portfolio benefits, the conditions to realize the benefit potential present in alliance portfolios, and the effect of network resource interdependencies on alliance-level costs in more detail.

The effect of network resource interdependencies on alliance portfolio benefits

Much of the extant alliance literature that focuses on value creation adopts a single alliance perspective in the sense that it views resource supplementarity and complementarity between a focal firm and its alliance partner as critical determinants for value creation (Arora and Gambardella, 1990; Chung et al., 2000; Das and Teng, 2000; Rothaermel, 2001; Mowery et al., 2002). Although this single alliance perspective has significant explanatory power, it does not tell the entire story because it assumes away any type resource supplementarity or complementarity on the alliance portfolio level of analysis. Lavie (2007, p. 1192) notes that a focal firm ‘may create value by combining network resources of distinct partners, and thus enjoy synergies that are unavailable to individual partners in its alliance portfolio. For example, a firm that specializes in systems integration may combine the hardware platforms of one partner with the software development expertise of another partner in the course of system implementation projects.’ In other words, if an alliance portfolio maintaining firm benefits from network resource supplementarity or complementarity, then the single alliance perspective and the additive approach of the existing RBV frameworks are insufficient to examine value creation on the alliance portfolio level, because they do not account for value creation through synergistic resource combinations made possible through the focal firm's network resources accessed from multiple partners.

We posit that network resource supplementarity and complementarity in alliance portfolios can occur in the following two ways. First, and in line with the extant literature that takes a single alliance perspective, one or more network resources accessed from one specific partner can be supplementary or complementary to one or more of the focal firm's own resources (Arora and Gambardella, 1990; Chung et al., 2000; Rothaermel, 2001; Mowery et al., 2002; Gulati, 2007). Second, and representing an extension to the concepts of resource supplementarity and complementarity beyond the single alliance level, network resources that are accessed through multiple simultaneous alliances with different partners may also be supplementary or complementary. Figure 2 depicts these two types. When taking a single alliance perspective, namely, considering alliances ih and ij individually, Figure 2 shows that focal firm i's network resources Rh1 and Rj1 are synergistically aligned to i's own resources Ri4 and Ri3 through alliances ih and ij respectively. When taking an alliance portfolio perspective, that is, considering alliances ih and ij simultaneously, focal firm i's network resources Rh1 and Rj1 may also be supplementary or complementary and thus provide opportunities for value creation.

Figure 2.

Types of resource linkages in alliance portfolio maintaining firms

Considering the alliance portfolio maintaining focal firm i's own resources (Ri3 and Ri4) and network resources (Rj1 and Rh1), firm i's alliance portfolio can render benefit creating services through the following resource combinations. First, from a single alliance perspective, bi-lateral resource combinations Ri3Rj1 and Ri4Rh1 can render benefit creating services on the individual alliance level. Moreover, by taking an alliance portfolio perspective there is additional benefit potential by exploiting the focal firm i's network resources Rj1 and Rh1, given they are supplementary or complementary. In such a case, the potential combination of network resources Rj1 and Rh1 gives the focal firm the opportunity to create additional benefits in its alliance portfolio. Such benefits are above and beyond the benefits that are created by the bi-lateral resource combinations Ri3Rj1 and Ri4Rh1.

Proposition 1:The more network resources accessed through multiple simultaneous alliances with different partners are supplementary or complementary, the higher the benefit potential of the alliance portfolio, above and beyond the benefits produced by each individual alliance in the portfolio.

To illustrate, consider the following example. On-demand service provider http://Salesforce.com has an alliance with Google through which it jointly offers a business productivity application that allows business professionals to communicate, collaborate, and work together in real time over the web. http://Salesforce.com also maintains an alliance with http://Amazon.com to offer its customers more computing power and allow them to host public web sites and company intranets over its http://Force.com platform. By leveraging the complementary nature of these two alliances, http://Salesforce.com is able to offer a more comprehensive service to its customers and thus extract benefits from its alliance portfolio that go beyond the benefits of each individual alliance in the portfolio (Wassmer et al., 2010).

Realizing benefit potential in alliance portfolios

Although a focal firm's alliance portfolio may possess high benefit potential, a key question is the extent to which the focal firm can realize this benefit potential created in its alliance portfolio by supplementary or complementary network resources accessed through multiple simultaneous alliances with different partners. From the focal firm's perspective, the issue with realizing such benefit potential in its alliance portfolio is that the network resources are owned and controlled by its partners. Although the focal firm has preferential access to these network resources, the resource access condition (Dyer and Singh, 1998; Lavie, 2006), will not suffice to assemble a benefit-creating resource combination because the alliance partners, if aware of the opportunity, can bypass the focal firm and exploit such a supplementary or complementary network resource combination alone. Whether an alliance portfolio maintaining firm is able to exploit such an opportunity for realizing additional alliance portfolio level benefits made possible by synergistic network resources and prevent its partners from going alone essentially depends upon whether the focal firm is able to affect the benefits and/or costs associated with such a combination. From a benefit perspective, it depends on whether the focal firm is able to make a resource contribution that enhances the benefits created through the combination. Moreover, the focal firm's partners may refrain from exploiting the opportunity if they are not able to cost effectively realize the additional benefit offered by such a combination and thus fail to create a monetary value. Thus, from a cost perspective the focal firm may be able to realize such an opportunity if its participation helps to implement the combination more cost effectively and thus enhances the overall value created by the combination.

At first sight it appears that the focal firm's ability to exploit such an opportunity can also be explained by Burt's (1992) theory of structural holes. According to a structural hole logic, the alliance partners would be unaware of this opportunity if there were a structural hole in the network of actors that left these partners unconnected. Moreover, if the focal firm occupies a network position allowing it to bridge that hole and broker the relationships between these unconnected partners it could exploit this opportunity. However, this is only partially true because even if these partners were connected it is still not guaranteed that the resource combination by the partners would be complete and thus benefit creating. Our argument, therefore, complements the structural hole perspective in this context.

Thus, a necessary condition for affecting the effectiveness of the resource combination is the existence of a gap in the combination which can be filled by the focal firm's resource contribution. Such a gap can be thought of as a missing resource element in the combination – namely, the lack of either one or more scale-free resources or the minimum required quantity of one or more non-scale-free resources. The point here is, however, that from the focal firm's perspective, a supplementary or complementary alignment of network resources accessed through multiple simultaneous alliances with different partners is a necessary but insufficient condition to create value on the alliance portfolio level because the focal firm faces the risk of being bypassed if it cannot contribute to the combination.

Conceptually, there are two different scenarios. First, network resources in the focal firm's alliance portfolio may be complementary or supplementary but constitute an incomplete resource combination, that is, once combined they fail to render a benefit-creating service. Second, network resources in the focal firm's alliance portfolio may be complementary or supplementary and constitute a complete resource combination, that is, once combined they render at least the minimum possible benefit creating service because they possess at least the minimum required quantity of non-scale-free resources to produce some valuable service, but more value creation is prevented by the lack of greater quantities of such non-scale-free resource elements. These two types of resource combinations and their respective benefits are shown in parts (a) and (b) in Figure 3.

Figure 3.

Gaps in combinations of network resources and focal firm contributions

Consequently, there are two conceptually different types of gaps depending on whether the resource combination is incomplete or complete but with limited resource capacity. The resource gap prevailing in incomplete combinations prevents benefit creation altogether. The resource gap in complete resource combinations with limited resource capacity limits the amount of benefit that is created through that combination. Figure 3 depicts the two conceptually distinct resource combination types (see row a), the benefit that is created upon combination of the resources (see row b), the two conceptually distinct types of gaps that can occur in the respective resource combinations (see row c), the focal firm's contribution (see row d), and the outcome (see row e). In the next sub-section we discuss how the focal firm can fill these gaps through the contribution of one or more resources it owns, and how this contribution affects benefit creation on the alliance portfolio level.

Realizing benefits in incomplete combinations of synergistic network resources.  This scenario refers to the situation in which the focal firm's network resources accessed from multiple simultaneous alliances with different partners are supplementary or complementary but, if they were combined, would not constitute a complete and thus benefit creating combination. In other words, what prevents the combination of network resources from being complete is a resource gap which can either be: (1) a missing scale-free resource; (2) missing capacity of one or more non-scale-free resources deployed in the combination; or (3) both. Row (c) of Figure 3 shows these three gaps that can occur for incomplete combinations. Consequently, there are three ways the focal firm can exploit the benefit creating potential of such these network resources: (1) by contributing the missing scale-free resource; (2) by contributing additional quantities of the non-scale-free resource/s; or (3) by contributing both. Row (d) in Figure 3 shows the three potential contributions.

However, simply contributing the missing resource element/s is an individually insufficient condition for realizing the benefit that these network resources offer to the focal firm. In fact, if the missing resource element is not valuable, rare, inimitable, and non-substitutable (i.e., the VRIN [valuable, rare, inimitable, and non-substitutable] conditions; Barney, 1991),4 the focal firm may be bypassed by its partners and excluded from that combination. In other words, if the missing element does not fulfill the VRIN conditions, the partners could develop it internally, acquire it on the respective factor market, or access it through another alliance and thus fill the gap themselves (Dierickx and Cool, 1989). Our argument is in line with the relational view of the firm (Dyer and Singh, 1998), that posits that interfirm alliances can be a source for competitive advantage and that a firm's VRIN resources can reside in interfirm alliances. More specifically, if the missing element is rare, inimitable, and non-substitutable partner firms are unlikely to possess it.

Proposition 2a:To realize the benefit potential, created in its alliance portfolio by an incomplete combination of synergistically aligned network resources, the focal firm must contribute valuable, rare, inimitable, and non-substitutable own resource/s that complete the combination.

To illustrate, consider the passenger air transportation industry in which most airlines maintain a portfolio of code sharing agreements. For example, Lufthansa maintains code sharing agreement with Spanair and Air New Zealand. To realize the alliance portfolio level benefit potential created by these two alliances (i.e., offer routes from Spain to New Zealand), Lufthansa must contribute a resource, namely, here a flight segment. Thus, for certain departure times the itinerary Madrid to Auckland is comprised of the segments Madrid to Frankfurt (operated by Spanair), Frankfurt to Los Angeles (operated by Lufthansa), and Los Angeles to Auckland (operated by Air New Zealand).

Enhancing benefits in complete combinations of synergistic network resources.  This scenario refers to the situation in which the focal firm's network resources accessed from multiple simultaneous alliances with different partners are not only supplementary or complementary but, if they were combined, would also constitute a complete combination that could render at least the minimum benefit creating service. As discussed earlier, in such a combination, more benefit creation is essentially prevented by the limited quantity of the non-scale-free resource/s deployed in the combination. Thus, more benefit creating services in such a combination can be created if the focal firm is able to add more capacity of one or more non-scale-free resources deployed in the combination. Adding more capacity to the resource combination, for example, may allow the collaborating firms to operate more efficiently and thus reduce costs or may make it possible for them to enhance revenues through additional product and service offerings (Barney, 1997). As stated earlier, if the focal firm's partners owning and controlling these network resources have full information about this opportunity and are able to fill the gap by providing the required resource capacity, they could partner directly and exclude the firm from such a potential combination. Consequently, if the non-scale-free resource/s that is only available in limited quantity does not fulfil the VRIN conditions (Barney, 1991), the focal firm would also be bypassed by its partners and excluded from that combination because the two partners could develop it internally, acquire it on the respective factor market or access it through another alliance and thus fill the capacity gap themselves (Dierickx and Cool, 1989).

Proposition 2b:To enhance the benefit potential created in its alliance portfolio by a complete resource combination of synergistically aligned network resources, the focal firm must contribute additional quantities of at least one valuable, rare, inimitable, and non-substitutable non-scale-free resource.

To illustrate, we can consider an example that is very similar to the one provided above to illustrate Proposition 2a. Air France maintains code sharing agreement with China Southern and Aeromexico. To realize the alliance portfolio level benefit potential created by these two alliances (i.e., offer routes from Chengdu in China to Cancun in Mexico, via Guangzhou, Paris and Mexico City), Air France must either contribute a resource to complete the combination (i.e., one of the required flight segments), as discussed in Proposition 2a or additional quantities of a non-scale free resource, in this case additional capacity on one of the flight segments. If either the Guangzhou to Paris flights on China Southern or the Paris to Mexico City flights on Aeromexico are fully booked, the Air France operated flights on these two routes then allow the resource combination to render additional services and thus create more value for all involved. In this case, though Air France contributes no unique resource needed to complete the combination, the additional capacity on a capacity constrained resource does increase the overall value of the combination and allows Air France not to be bypassed by its partners.

The effect of network resource interdependencies on alliance-level costs

While we have so far examined the role that network resource interdependencies can play in enhancing alliance portfolio benefits, in this section we focus on how they may also increase alliance costs. More specifically, by integrating insights from the RBV and the literature on the costs of alliances (Hennart, 1988; Koh and Venkatraman, 1991; Williamson, 1991; Dyer, 1997; Gulati and Singh, 1998; Park and Zhou, 2005; White and Lui, 2005), we argue that interdependencies between network resources accessed from multiple simultaneous alliances with different partners can negatively affect the value of individual alliances in the portfolio by increasing their costs.

Depending on the life cycle stage of an alliance (Dyer et al., 2001), firms incur different types of costs. Generally, alliance related costs can be categorized into control (Hennart, 1988; Williamson, 1991; Dyer, 1997), cooperation (White and Lui, 2005), and competitive costs (Koh and Venkatraman, 1991; Park and Zhou, 2005). Control costs are incurred in the partner assessment and selection phase as well as contract negotiation and governance set-up phase of the alliance life cycle. More specifically, control costs mainly include search costs related to searching for suitable partners and contracting costs related to negotiating and writing a contract (Williamson, 1985; Dyer, 1997). Control costs are mainly emphasized in the transaction cost economics alliance literature which is primarily concerned with the selection of appropriate alliance governance structures to preempt opportunistic behavior by alliance partners (Hennart, 1988; Williamson, 1991; Hennart and Zeng, 2005). Cooperation costs, on the other hand, are essentially the ongoing alliance management and coordination costs that originate from handling joint tasks and ensuring social integration between alliance partners (Gulati and Singh, 1998; White and Lui, 2005). Such ongoing cooperation costs include costs related to communicating with the partner, resolving conflicts, monitoring the contract fulfillment, and enforcing sanctions on the partner if the contractual obligations are not fulfilled (Dyer, 1997; Gulati and Singh, 1998; White and Lui, 2005). Ongoing alliance cooperation costs are essentially constituted by the managerial time and effort spent on an alliance (White and Lui, 2005). Competitive costs include costs related to the weakening of the competitive position vis-à-vis the other partner due to the leakage of resources and capabilities (Koh and Venkatraman, 1991; Park and Zhou, 2005). Before we examine how alliance portfolio level resource interdependencies can increase the control, cooperation, and competitive costs of individual alliances in an alliance portfolio, we first identify and conceptually distinguish between two types of network resource interdependencies.

Types of resource interdependencies.  From an RBV perspective, the performance of a firm is determined by the simultaneous interaction of the firm's resources and the resources of its competitors (Conner, 1991). The RBV also posits that a key condition for protecting competitive advantage and rent streams is the non-substitutability of resources (Barney, 1991). When two firms deploy resource combinations that are substitutes for one another, for example, bus travel can substitute air travel and vice versa, the two firms are essentially competitors in the same market (e.g., traveling from point A to point B), by offering substitutable products or services.

Conceptually, resource combinations deployed through an alliance in a firm's alliance portfolio can substitute resource combinations of the focal firm's alliance partners in two ways. First, they can substitute a resource combination that is privately deployed by one of the focal firm's existing partners (henceforth we refer to this scenario as the substitutability of partner private resource combinations). Second, they can substitute a resource combination that is jointly deployed through another alliance by the focal firm with another partner (henceforth we refer to this scenario as the substitutability of within alliance portfolio resource combinations). Figure 4 illustrates both cases.

Figure 4.

Types of resource combination substitutability

Figure 4 shows focal firm i that maintains an alliance portfolio comprised of two simultaneous alliances (i.e., alliances ih and ij), with two different partners (i.e., firms h and j). Focal firm i deploys resource combination Ri4Rh1 through alliance ih with partner h and resource combination Ri3Rj1 through alliance ij with partner j. The substitutability of partner private resource combinations case refers to the situation in which resource combination Ri4Rh1 that focal firm i deploys through alliance ih substitutes partner j's privately deployed resource combination Rj2Rj3. In other words, both resource combinations render substitutable products or services. The substitutability of within alliance portfolio resource combinations case refers to the situation in which resource combination Ri4Rh1 that focal firm i deploys through alliance ih substitutes resource combination Ri3Rj1 that focal firm i jointly deploys with partner j through alliance ij. Here, alliances ih and ij are essentially substitutes to one another. We now discuss both cases of resource substitutability in more detail and clarify how they affect alliance control, cooperation, and competitive costs of other active alliances in the alliance portfolio and thus reduce the value of these alliances.

Substitutability of partner private resource combinations.  The substitution of a resource combination privately deployed by one of the focal firm's partners exacerbates the competitive intensity between the focal firm and that partner, increasing the likelihood of conflict between the two firms, opportunism by the partner, and resource and capability leakage. These three issues negatively affect the cooperation, control, and competitive costs the focal firm incurs in its cooperation with this partner. We now discuss each of them in more detail.

With cooperation costs, prior research has suggested that conflict between alliance partners on the single alliance level of analysis can arise through opportunistic behavior or by defecting from a mutual agreement and can negatively affect alliance performance (Hamel, 1991; Parkhe, 1993). According to this view, the cause for conflict is internal to the alliance because the opportunistic behavior or defection from the agreement relates to activities performed in the alliance. However, it may also be that causes for conflict between alliance partners are external to the alliance, in the sense that they are not related to the activities performed in the alliance, but may nonetheless spill over into the relationship between the alliance partners. For example, such externally instigated conflict between alliance partners can occur when one partner undertakes a competitive move vis-à-vis the other partner by imitating that partner's technology or entering into that partner's market domain and thus increases the market domain overlap between the two firms (Kogut, 1989).

The substitutability of a resource combination privately deployed by a focal firm's partner through a resource combination enabled by one of the focal firm's alliances with another partner represents essentially such an external source for conflict between the two firms. In other words, an individual alliance in the focal firm's alliance portfolio may be an external source for conflict between the focal firm and a partner in another active alliance. Thus, substitutable resource combinations deployed through one alliance with a specific partner can alter the competitive positioning of the alliance portfolio maintaining focal firm vis-à-vis other existing alliance partners and may therefore cause disturbance in the relationships with these partners.

Such conflict and disturbance may lead affected partners to make certain threats or undertake defensive moves such as to withdraw from the alliance with the focal firm (Park and Russo, 1996). Indeed, it has been argued that the greater the market domain, namely, resource combination, overlap between two firms, the greater the competitive intensity between them and the higher the likelihood of a defensive move (Chen and MacMillan, 1992; Chen and Miller, 1994; Baum and Korn, 1996). To address such a threat, mitigate the risk of potential alliance termination, and resolve the conflict, the focal firm will have to spend more managerial and legal time and effort on the alliance in order to re-establish trust and goodwill with the partner (White and Lui, 2005), or even grant additional rewards to the affected partner. Consequently, the focal firm will incur higher cooperation costs associated with this particular alliance due to the additional time and effort spent on the alliance. These additional cooperation costs unfavorably change the cost/benefit ratio of the alliance leading to reduced overall value (Zajac and Olsen, 1993; Madhok and Tallman, 1998).

In the case of control costs, the substitutability of private partner resource combinations also increases the risk of opportunistic behavior by the partner towards the focal firm (Brandenburger and Nalebuff, 1996). Consequently, this may lead the focal firm to preempt such opportunistic behavior through the implementation of more complex governance structures (Hennart, 1988; Williamson, 1991). More complex governance structures increase the control, that is, transaction costs of the alliance and also unfavorably changes the cost/benefit ratio of the alliance leading to reduced overall value (Zajac and Olsen, 1993; Madhok and Tallman, 1998).

As explained above with competitive costs, the substitutability of private partner resource combinations exacerbates the competition between the focal firm and its partner, leading the two firms to compete away each other's rents (Conner, 1991). Since the relationship between the focal firm and this partner can be described as coopetitive (Brandenburger and Nalebuff, 1996), the focal firm faces an increased risk and thus competitive costs related to the leakage of intangible and knowledge-based resources (Hamel, 1991; Miller and Shamsie, 1996), and capabilities (Koh and Venkatraman, 1991; Park and Zhou, 2005), that can weaken its competitive position vis-à-vis the partner.

To summarize, when one alliance in the focal firm's alliance portfolio creates resource combinations that substitute resource combinations that existing partners deploy privately, the focal firm will incur increased costs that essentially reduce the overall value of the focal firm's alliance with that partner.

Proposition 3a:Substitutability between a resource combination deployed through one of the focal firm's alliances and a resource combination privately deployed by one of the focal firm's partners increases the focal firm's alliance costs with that partner, and thus, reduces the value of the alliance with that partner.

To illustrate, consider Lufthansa's agreement with Air Canada to code share on the route Vancouver–Frankfurt. Lufthansa's existing partner United Airlines competes in the market Vancouver–Frankfurt via an indirect service by combing two of its segments (e.g.,Vancouver–Chicago and Chicago–Frankfurt). Lufthansa's alliance with Air Canada thus overlaps with United Airlines' privately deployed combination.

Substitutability of within alliance portfolio resource combinations.  Similar to the above described case, the substitutability of two resource combinations deployed through two different alliances with different partners exacerbates the competitive intensity between the focal firm and those partners, thus increasing the likelihood of conflict with the partners, opportunism by the partners, and resource and capability leakage towards the partners. These three issues negatively affect the cooperation, control, and competitive costs the focal firm incurs in the two affected alliances.

In addition to the above described argument, the focal firm's cooperation costs are increased by the redundancy of the two alliances: the focal firm is led to duplicate investments and costs in order to pursue essentially the same benefits. In other words, if two resource combinations that are deployed through two of the focal firm's alliances with different partners are substitutable then the focal firm may experience a loss of value at the intersection of the targeted product or market domains, because the substitutability of resource combinations cannibalizes the benefits the focal firm derives from both alliances, in the sense that it reduces the common benefits that the focal firm receives from these two alliances (Khanna, 1998; Khanna et al., 1998; Roberts and McEvily, 2005). Thus, the joint value function of the two alliances is characterized by reduced benefits due to cannibalization and enhanced costs due to increased cooperation, control, and competitive costs as well as the additional redundancy costs.

Proposition 3b:Substitutability between the resource combinations deployed in two distinct alliances with different partners in the focal firm's alliance portfolio negatively affects the benefits and costs and thus, the value of both alliances.

To illustrate, consider the case of Danone's strategic partnership with Hangzhou Wahaha Group. When Danone set up competing joint ventures with other local Chinese firms such as Robust, Aquarius, Mengliu Dairy or Bright Dairy and Food, Wahaha dragged several Danone officials to court for conflict of interests because of their simultaneous membership in the boards of the Wahaha-Danone joint venture and of other competing joint ventures. As a result, the relationship further deteriorated and the two partners ended up suing each other in about 40 law suits filed on four different continents. In fall 2009 Danone had no option but to sell its 51% stake in the alliance which accounted for 80% of the French firm's sales in China. Though the value of the transaction was not officially disclosed, it was rumored to be much less than the actual value of Danone's share in the JV (Wassmer et al., 2010).

Despite the duplication of costs, the investment, time and effort to set up and coordinate two redundant alliances, there may be benefits to redundancy in alliance portfolios that exceed such costs. Overlap in scope between individual alliances and the substitutability of the resource combinations deployed by these alliances may sometimes be created deliberately to spread risks and pursue various simultaneous options, such as, for instance, to invest in the development of competing technologies. While such a strategy creates additional costs for all the reasons described above, the benefits of spreading risks and maintaining multiple options may create sufficient benefits to more than outweigh the costs.

Proposition 3c:Substitutability between the resource combinations deployed in two distinct alliances with different partners in the focal firm's alliance portfolio creates value only if the benefits associated with maintaining redundant alliances exceed the redundancy costs.

To summarize, we argued that synergistic combinations of network resources and substitutability of resource combinations between the focal firm and its partners are critical determinants for value creation in alliance portfolios. Moreover, we extended the existing concepts of resource supplementarity and complementarity beyond the single alliance level and argued that supplementary or complementary network resources accessed through multiple simultaneous alliances with different partners is not a sufficient condition for benefit creation in alliance portfolios. We showed that certain conditions have to be met in order to leverage network resources into benefit-creating combinations. We argued that the substitutability of resource combinations between the focal firm and its alliance partners negatively affects the costs, and thus the value, of individual alliances in the focal firm's alliance portfolio.

Discussion and conclusion

This paper started by highlighting the need to formalize interdependencies between network resources accessed through multiple simultaneous alliances with different partners in order to better understand value creation in alliance portfolios. Our theoretical model has a number of implications for the RBV and capabilities literature. First, our framework helps to better understand which types of network resources are most valuable. Here, one can assess the benefits and costs and thus the value creation potential of a network resource on two distinct levels: (1) the individual alliance level through resource combinations in which the network resource is deployed with the focal firm's own resources; and (2) the alliance portfolio level through any resource combination in which the network resource is deployed with other network resources that the focal firm accesses from other partners. As stated earlier, however, resources rarely generate benefits on their own but rather as parts of resource combinations. From a focal firm perspective, network resources that can be deployed in combinations on both levels are therefore particularly valuable. Therefore, an important issue for the focal firm to deal with is the fit of a network resource with both its own resources as well as other network resources accessed from partners of other alliances. This is especially important to understand the mechanics of resource recombination, namely, the process within a firm of synthesizing existing resources and reconfiguring resource linkages in order to create novel and rent generating resource combinations (Galunic and Rodan, 1998). In alliance portfolio maintaining firms, resource recombination refers not only to the synthesis of existing own but also existing network resources accessed through multiple simultaneous alliances with different partners, and reconfiguring linkages between these resources in novel and value-generating ways. Thus, access to network resources through multiple simultaneous alliances with different partners provides additional recombination opportunities to create novel and rent generating combinations. In other words, such firms have an opportunity to further exploit the rent generating capacity of their own, as well as their network resources. Our model shows, however, that the exploitation of benefit creating opportunities through recombining network resources depends on a number of conditions that have to be met simultaneously to exploit these opportunities.

Second, our theoretical model also has implications for the way in which alliance portfolios evolve over time, that is, how firms form new alliances to add new strategically critical resources (Ahuja, 2000; Chung et al., 2000), and dissolve existing alliances to remove network resources that have ceased to render rent generating services (Reuer and Zollo, 2005; Makino et al., 2007). Consequently, alliance portfolio evolution and network resource reconfiguration (Capron et al., 1998; Karim and Mitchell, 2000), go hand in hand. A key issue when adding new network resources through alliance formations is, therefore, the identification of partners that can provide network resources with the highest value creating potential. Consequently, when dissolving existing alliances and deleting network resources from its resource base, alliance portfolio maintaining firms will have to consider the overall net value that the resource is creating before undertaking such action. In other words, by viewing the alliance portfolio as an integrated resource system, alliance portfolio maintaining firms can optimize the overall alliance portfolio value rather than optimizing value on the level of each individual alliance. This is especially important whenever a firm changes its alliance portfolio's configuration through either forming a new alliance or terminating an existing one.

Finally, our proposed theoretical framework also has implications for how firms trade off resource picking and capability building strategies (Makadok, 2001). From a resource picking perspective, our model suggests that firms that are superior at picking network resources that provide not only a good fit with their own resources but also with the other networks resource in their alliance portfolios will be able to create more value than firms that are less skilled at resource picking. Thus, a firm can maximize value by picking the right network resources that create the maximum value on both the individual alliance as well as alliance portfolio levels. From a capability perspective it can also be argued that firms that build the appropriate organizational capabilities that enable them to better exploit the rent generating potential of their network resources on both the single alliance as well as the alliance portfolio levels will be able to create more value. The extant literature has shown that alliance capability is an important firm level capability relevant for the success of alliance portfolio maintaining firms (Anand and Khanna, 2000; Kale et al., 2002; Heimeriks and Duysters, 2007; Heimeriks et al., 2007). Therefore, considering the resource picking and capability building strategies for value creation (Makadok, 2001), an implication of our model is that a firm's alliance capability does not only reflect the firm's capacity to manage alliances; it also encompasses its ability to pick the network resources that have an optimal fit with its alliance portfolio as well as to build the appropriate capabilities, such as a specific alliance portfolio management capability (Heimeriks and Duysters, 2007; Heimeriks et al., 2007), and organizational processes and systems such as a dedicated alliance function (Kale et al., 2002). Only through such capabilities, systems, and processes will an alliance portfolio-maintaining firm be able to effectively monitor and coordinate its alliance portfolio in a deliberate way so that it can exploit all synergies between network resources and avoid competitive overlap with existing partners. In fact, recent research has shown that alliance portfolio management capability also affects the alliance portfolio's capital, namely, the value residing in alliance portfolios, but this relationship is contingent upon the existence of a dedicated alliance function and the diversity of the alliance portfolio (Sarkar et al., 2009).

To sum up, all of the above points suggest that network resources accessed through strategic alliances should be evaluated in the context of an alliance portfolio and not as standalone resources. We propose that optimally configuring a firm's alliance portfolio is a process that is subject to bounded rationality. Indeed, because alliance formation and dissolution decisions tend to be made in a decentralized way by the directly involved units of the firm, there is a huge potential for sub-optimization and conflict within the resulting portfolio of alliances (Wassmer et al., 2010). It is therefore important to develop a full understanding of how value is created and destroyed at the alliance portfolio level in order to minimize the undesired effects of alliance conflict and maximize the synergy effects within the portfolio. We suggest our modeling efforts provide the necessary basis for such an understanding.

This study has a number of limitations. First, we do not apply a dynamic perspective in the sense that we do not theorize about how alliance portfolio evolution and changes over time affect value creation (Hoffmann, 2007). However, our theory is dynamic in the sense that we take interactions between resources on different levels of analysis into account. Second, because the RBV served as our main theoretical lens, a shortcoming is that this perspective does not allow problematizing any structural characteristics within alliance portfolios. Therefore, the integration with social network theories may provide additional insights on how the focal firm can exploit its network position to create value in its alliance portfolio. Furthermore, our theory does not address value appropriation (Lepak et al., 2007), and thus, does not allow us to draw inferences about how much value is actually appropriated by the focal firm from its alliance portfolio (Lavie, 2007). Therefore, greater integration with transaction cost economics and the literature on bargaining power could help to develop some more fine grained theory on value appropriation in alliance portfolios. Lastly, our theory does not formalize the exact flow of different types of rents in interconnected firms (Lavie, 2006). Thus, future research should incorporate these rent types into any formalized model.

We also suggest that the proposed model should be tested empirically. We believe that it should be tested in industry settings in which strategic alliances are used as key strategic devices to access critical resources and in which firms maintain entire alliance portfolios (Contractor and Lorange, 2002). Previous empirical studies have primarily examined value created at the individual alliance level (Chen et al., 1991; Koh and Venkatraman, 1991; Chan et al., 1997; Das et al., 1998; Merchant and Schendel, 2000; Kale et al., 2002), while those that have focused on interconnected firms have mainly studied the link between overall alliance portfolio characteristics and their performance impact (Baum et al., 2000; Stuart, 2000; Bae and Gargiulo, 2004). Empirically disentangling the value impact of individual alliances from the portfolio effect of the network resources they contribute to is likely to be a challenging endeavor; we are convinced nevertheless that future empirical research would make a valuable contribution by empirically testing parts of the proposed model. An additional promising avenue for future research is to further refine the theory through systems dynamics-based simulation methods, namely, modeling the alliance portfolio maintaining firm as a resource system in which both own and network resources interact in a value creating as well as destroying way (Davis et al., 2007).

We submit that the contribution of our paper rests in two areas. First, we drew on three important streams of literature central to research concerning interconnected firms and developed an integrated perspective on value creation in alliance portfolios. Second, we extended some of the conventional value focused and resource-based concepts from the individual alliance level of analysis to the alliance portfolio level. We believe that our analysis contributes to three important strategy research streams, namely research on value creation (Brandenburger and Stuart, 1996; Bowman and Ambrosini, 2000; Lepak et al., 2007), the RBV and its recent extension to interconnected firms (Wernerfelt, 1984; Barney, 1991; Dyer and Singh, 1998; Lavie, 2006) as well as to the performance consequences for firms participating in strategic alliances (Gulati, 1998). We are confident that this study provides a useful perspective and furthers understanding on some of the issues that alliance portfolio maintaining firms face.

Acknowledgements

We are grateful to Editor in Chief Alfonso Gambardella and two anonymous reviewers for their helpful comments and suggestions. We would also like to thank Jay Anand, Xavier Castañer, Rodolphe Durand, Prashant Kale, Anoop Madhok, and Marco Tortoriello for their helpful comments on earlier versions of this article. Pierre Dussauge acknowledges support from the HEC Air France-Edenred-SNCF Chair on Service Management.

Footnotes

  • 1

    For the purpose of this paper, the creation of economic, that is, monetary, value (e.g., increases in stock price, additional cash flows, etc.) is referred to as ‘value creation’.

  • 2

    Levinthal and Wu (2010) use the terms ‘resources’ and ‘capabilities’ interchangeably. Knowledge can be placed in either category depending on whether it is codified/explicit or tacit.

  • 3

    We frame monetary value created by alliances within an expected net present value (NPV) logic to capture the time value of resources as well as uncertainties related to the future. Thus, rationally behaving firms engage in alliances only when their net present value is positive.

  • 4

    The argument of rareness is similar to what Brandenburger and Stuart (1996) refer to as scarcity.

Ancillary