Introduction
Firms that engage in multiple alliances manage them as a portfolio, rather than managing each alliance in isolation (Wassmer, Reference Wassmer2010; Faems, Janssens, & Neyens, Reference Faems, Janssens and Neyens2012). Managing a portfolio of alliances is a challenge, requiring substantial managerial attention and skills (Heimeriks & Duysters, Reference Heimeriks and Duysters2007; Schilke, Reference Schilke2014). Over the last two decades, alliance portfolio has emerged as a unit of analysis, and researchers have started addressing various important questions central to this emerging research stream (Wassmer, Reference Wassmer2010; Faems, Janssens, & Neyens, Reference Faems, Janssens and Neyens2012). Researchers have studied various aspects of alliance portfolio, such as origin and evolution (Hoffmann, Reference Hoffmann2007; Lavie & Singh, Reference Lavie and Singh2012), characteristics and configuration (Rothaermel & Deeds, Reference Rothaermel and Deeds2006; Andrevski, Brass, & Ferrier, Reference Andrevski, Brass and Ferrier2013), and management of alliance portfolio (Heimeriks & Duysters, Reference Heimeriks and Duysters2007; Schilke, Reference Schilke2014). While the questions of performance impact of alliance portfolio characteristics have attracted a great deal of attention from researchers, questions pertaining to antecedents of alliance portfolio have received relatively less attention (Wassmer, Reference Wassmer2010; Lee, Kirkpatrick-Husk, & Madhavan, Reference Lee, Kirkpatrick-Husk and Madhavan2014). Moreover, most of the studies are limited in generalizability because these studies are conducted in narrow industrial contexts (e.g., Beckman, Bird Schoonhoven, Rottner, & Kim, Reference Beckman, Bird Schoonhoven, Rottner and Kim2014) and/or have considered only specific type of alliances, such as technology alliances (e.g., Jacob, Belderbos, & Gilsing, Reference Jacob, Belderbos and Gilsing2013; Castro, Casanueva, & Galán, Reference Castro, Casanueva and Galán2014).
In particular, researchers, concerned with alliance portfolio diversity (henceforth APD), have focused primarily on the APD-performance linkage (Wassmer, Reference Wassmer2010; Lee, Kirkpatrick-Husk, & Madhavan, Reference Lee, Kirkpatrick-Husk and Madhavan2014); and, only a few researchers have attempted to examine the antecedents of APD. These researchers have taken various theoretical perspectives such as social capital and relational perspectives (Beckman et al., Reference Beckman, Bird Schoonhoven, Rottner and Kim2014; Golonka, Reference Golonka2015), resource- and knowledge-based views (Duysters & Lokshin, Reference Duysters and Lokshin2011), real options (Tao, Jiang, & Santoro, Reference Tao, Jiang and Santoro2014), and organizational learning perspective (Jacob, Belderbos, & Gilsing, Reference Jacob, Belderbos and Gilsing2013; van Beers & Zand, Reference van Beers and Zand2014). The studies that draw on organizational learning perspective have focused mainly on one dimension of learning, i.e. learning from similar experiences (Jacob, Belderbos, & Gilsing, Reference Jacob, Belderbos and Gilsing2013; van Beers & Zand, Reference van Beers and Zand2014). A study by Jacob, Belderbos and Gilsing (Reference Jacob, Belderbos and Gilsing2013), on a sample of 2,488 European firms, found that firms build on their prior international alliance experience to enhance the geographical diversity of their alliance portfolio. Similarly, a study of 12,811 Dutch and foreign innovating firms found that prior alliance experience and patenting are key antecedents of APD (van Beers & Zand, Reference van Beers and Zand2014). However, organizations learn not only from similar experiences, i.e. prior alliance experience, but also from other diverse experiences, such as experience in diverse product-markets and/or geographic markets. The effect of learning from diversity on focal firm’s APD is less researched (Barkema & Vermeulen, Reference Barkema and Vermeulen1998; Schulz, Reference Schulz2002).
This study attempts to address such gaps in the literature, which focuses primarily on learning from similar experiences perspective while examining the antecedents of APD. In particular, it integrates these two streams of learning – learning from similar experiences and learning from diversity – perspectives with regard to antecedents of APD (Barkema & Vermeulen, Reference Barkema and Vermeulen1998; Jacob, Belderbos, & Gilsing, Reference Jacob, Belderbos and Gilsing2013; van Beers & Zand, Reference van Beers and Zand2014). Moreover, considering that alliance behavior of firms from emerging markets may be different than those from developed markets (Hitt, Ahlstrom, Dacin, Levitas, & Svobodina, Reference Hitt, Ahlstrom, Dacin, Levitas and Svobodina2004), this study conducted in the Indian context would complement prior studies which were mostly based in the context of developed countries.
The main objective of this study is to examine experiential antecedents of APD. Drawing primarily on organizational learning perspective, it examines the effects of prior product and international diversification experience, prior alliance experience, and alliance experience heterogeneity (henceforth AEH) on focal firm’s APD. Findings of this study suggest that learnings from both similar and diverse experiences enhance a firm’s ability to form and maintain diverse alliances. In particular, results of the study provide support for the hypotheses that firm’s prior product and international diversification experience, alliance experience, and AEH are positively related to APD. This study makes some valuable contributions to the learning from diversity literature by examining the effects of diverse experiences on a firm’s APD. Additionally, this study contributes to the alliance portfolio literature by examining the antecedents of APD, which remains relatively less researched.
The rest of the paper is organized as follows. First, a summary of the literature on conceptualization and operationalization of APD construct is presented; and, then, the rationale behind considering APD a two-dimensional construct is stated. Next, the theoretical background and arguments are presented to develop hypotheses. In the following section, the details of data and methods are discussed. Next, results are presented; and, theoretical and managerial implications of the findings are discussed. Finally, the paper is concluded by mentioning its limitations and providing directions for future research.
APD: Conceptualization and Operationalization
Prior studies have mostly conceptualized APD as variety of information or knowledge present in the portfolio (Harrison & Klein, Reference Harrison and Klein2007; Lee, Kirkpatrick-Husk, & Madhavan, Reference Lee, Kirkpatrick-Husk and Madhavan2014). Researchers have considered various dimensions of APD, including partner type diversity, functional diversity, technological diversity, and partners’ national diversity (Jiang, Tao, & Santoro, Reference Jiang, Tao and Santoro2010; Phelps, Reference Phelps2010; Mouri, Sarkar, & Frye, Reference Mouri, Sarkar and Frye2012; Lee, Kirkpatrick-Husk, & Madhavan, Reference Lee, Kirkpatrick-Husk and Madhavan2014; Tao, Jiang, & Santoro, Reference Tao, Jiang and Santoro2014). These dimensions can broadly be categorized into partner attributes and alliance attributes (Wassmer, Reference Wassmer2010).
In terms of partner attributes, APD has been operationalized based on partners’ – (i) type (i.e., competitor, customer, supplier, etc.) (Duysters & Lokshin, Reference Duysters and Lokshin2011); (ii) resource (or industry) (Cui, Reference Cui2013); (iii) nation of origin (Knoben & Oerlemans, Reference Knoben and Oerlemans2012); (iv) technology (Phelps, Reference Phelps2010); (v) organizational type, such as private or public (Jiang, Tao, & Santoro, Reference Jiang, Tao and Santoro2010); and (vi) repeatedness, i.e. repeated partnership with existing alliance partner, versus first time alliance with a new partner (Sivakumar, Roy, Zhu, & Hanvanich, Reference Sivakumar, Roy, Zhu and Hanvanich2011).
In terms of alliance attributes, APD has been operationalized as: (i) functional diversity – based on functions to be performed by the alliances, such as joint R&D, joint manufacturing, joint marketing, or combination (Hora & Dutta, Reference Hora and Dutta2013; Hoehn-Weiss & Karim, Reference Hoehn-Weiss and Karim2014); (ii) portfolio technology diversity – based on different technology areas in which alliances in a portfolio are present (Andrevski, Brass, & Ferrier, Reference Andrevski, Brass and Ferrier2013; Wuyts & Dutta, Reference Wuyts and Dutta2014); (iii) alliance industry diversity – in terms of different industries in which alliances in a portfolio are present; (iv) governance diversity – based on governance structure of alliances, i.e. equity joint ventures (JVs) or contractual.
A majority of the studies have considered these dimensions separately, emphasizing that each of these dimensions have independent effects on the outcome variables (Jiang, Tao, & Santoro, Reference Jiang, Tao and Santoro2010; Rogbeer, Almahendra, & Ambos, Reference Rogbeer, Almahendra and Ambos2014). However, a few researchers have argued that since APD dimensions are not independent, it is important to combine them to understand the nature of the latent construct ‘APD’ (van Beers & Zand, Reference van Beers and Zand2014; Castro, Roldán, & Acedo, Reference Castro, Roldán and Acedo2015). This study considers APD a two-dimensional construct that broadly reflects the partner attributes of the portfolio (de Leeuw, Lokshin, & Duysters, Reference de Leeuw, Lokshin and Duysters2014; van Beers & Zand, Reference van Beers and Zand2014; Castro, Roldán, & Acedo, Reference Castro, Roldán and Acedo2015). APD is operationalized as a formative construct comprising two partner-related dimensions: one dimension capturing diversity in terms of partner types and the other dimension capturing diversity in terms partners’ geographic region of origin. These two dimensions adequately capture the variety, in terms of partner attributes, present in the alliance portfolio (Harrison & Klein, Reference Harrison and Klein2007; Lee, Kirkpatrick-Husk, & Madhavan, Reference Lee, Kirkpatrick-Husk and Madhavan2014). A portfolio with different partner types (i.e., supplier, customer, competitor, similar-business, government, or unrelated-business w.r.t. focal firm) is suggestive of the balance of firm’s access to supplementary or complementary resources (Jiang, Tao, & Santoro, Reference Jiang, Tao and Santoro2010). Additionally, partner type diversity also indicates a focal firm’s capabilities and status (Stuart, Reference Stuart2000), and reflects how the firm balances its ties with similar and dissimilar partners (Castro, Casanueva, & Galán, Reference Castro, Casanueva and Galán2014). On the other hand, geographic diversity indicates whether partners belong to similar or diverse cultural or institutional contexts. A firm that partners with other firms from various socio-cultural contexts gains access to variety of knowledge sources (de Leeuw, Lokshin, & Duysters, Reference de Leeuw, Lokshin and Duysters2014; Castro, Roldán, & Acedo, Reference Castro, Roldán and Acedo2015).
Theoretical Background and Hypotheses
Do firms benefit from APD?
Prior literature on alliance portfolio suggests that APD positively affects a firm’s innovation (Duysters & Lokshin, Reference Duysters and Lokshin2011; Knoben & Oerlemans, Reference Knoben and Oerlemans2012; Wuyts & Dutta, Reference Wuyts and Dutta2014) and performance (Pangarkar & Wu, Reference Pangarkar and Wu2012; Hora & Dutta, Reference Hora and Dutta2013). A focal firm may be willing to maintain diverse alliances to get benefits from APD. However, managing a diverse portfolio is a challenging task, requiring substantial managerial resources and capabilities. Researchers suggest that, at a very high level of APD, the enhanced coordination costs may offset the positive effects of APD (Hoffmann, Reference Hoffmann2007; Duysters & Lokshin, Reference Duysters and Lokshin2011; Golonka, Reference Golonka2015). Nonetheless, the diminishing positive effect of APD on firm performance is contingent on a firm’s capabilities to manage diverse alliance portfolio. For example, Duysters, Heimeriks, Lokshin, Meijer and Sabidussi (Reference Duysters, Heimeriks, Lokshin, Meijer and Sabidussi2012) found that portfolio management capabilities that enable a firm in managing diverse alliance portfolio moderates the relationship between APD and portfolio performance such that when alliance capabilities are higher alliance portfolio performance is maximized at a higher level of APD. Similarly, Cui and O’Connor (Reference Cui and O’Connor2012) argued that alliance management function positively moderates the relationship between portfolio resource diversity and firm innovation. In other words, these researchers suggest that management capabilities mitigate the negative effects and enhance the positive effects of APD on a firm’s financial and innovation performance.
Since capabilities, developed through experiential learning, enable a firm to get the intended benefits from a high level of APD, one can expect that an experienced and capable firm would make conscious decisions to maintain or expand portfolio diversity. Put differently, this study argues that an experienced firm may decide to expand its portfolio diversity as it becomes capable of reaping the benefits and mitigating the disadvantages of the high level of APD. A more experienced firm is better equipped to reap the benefits of APD than a less experienced firm.
This study considers prior diversification and alliance experience as important sources of learnings. Diversification literature suggests that diversification experience enriches the existing knowledge base of the firm (Kogut & Zander, Reference Kogut and Zander2003; Zollo & Singh, Reference Zollo and Singh2004), and may help the firm develop capabilities to manage diversity (Bartlett & Ghoshal, Reference Bartlett and Ghoshal1993, Reference Bartlett and Ghoshal1999; Augier & Teece, Reference Augier and Teece2007). Similarly, alliance portfolio researchers have examined the influence of alliance experience on the development of alliance management capabilities (Gulati, Reference Gulati1999; Rothaermel & Deeds, Reference Rothaermel and Deeds2006; Duysters et al., Reference Duysters, Heimeriks, Lokshin, Meijer and Sabidussi2012). In this study, the effects of prior diversification and alliance experience on a focal firm’s APD have been examined. Two dimensions of firm’s diversification experience – product and international – are considered. The examination of the effect of alliance experience has been further refined by considering the influence of AEH on APD.
Product diversification experience and APD
Firm’s prior diversification experience enhances its knowledge-base and capabilities (Delios & Beamish, Reference Delios and Beamish1999; Mayer, Stadler, & Hautz, Reference Mayer, Stadler and Hautz2014). A firm’s presence in multiple product-markets can be an important source of learning as the firm learns through its interactions with diverse sets of suppliers, rivals, customers, and partners (Barkema & Vermeulen, Reference Barkema and Vermeulen1998). Such learning facilitates a firm’s subsequent entries into related markets (Gang, Reference Gang2013). From a process perspective, a firm’s presence in multiple markets improves executives’ capabilities to manage multiple businesses, which in turn may change the dominant logic of the firm (Prahalad & Bettis, Reference Prahalad and Bettis1986), thereby influencing performance of the firm (Chang & Thomas, Reference Chang and Thomas1989; Chakrabarti, Singh, & Mahmood, Reference Chakrabarti, Singh and Mahmood2007). Organizations bring changes in structures and routines to meet the challenges posed by diversification (Chandler, Reference Chandler1982; Markides & Williamson, Reference Markides and Williamson1996). The refined structures, routines, and mental models may help firms better manage their business units (Doz & Prahalad, Reference Doz and Prahalad1991; Bartlett & Ghoshal, Reference Bartlett and Ghoshal1993). Overall, learnings from product-market diversity improve a focal firm’s capabilities to manage complexities (Doz & Prahalad, Reference Doz and Prahalad1991; Bartlett & Ghoshal, Reference Bartlett and Ghoshal1993). Such learning may enhance a focal firm’s abilities to manage complexities of diverse alliance portfolio. Thus, a firm having diverse experiences would be more capable of getting the intended benefits from diverse alliance portfolio than a less experienced firm.
Additionally, drawing from the observations of a few case studies, which suggest that product diversification leads to greater partner diversity (Lavie & Singh, Reference Lavie and Singh2012, p. 799), this study argues that a diversified firm’s interactions with various customers, suppliers, and rivals from multiple product-markets may create opportunities for the firm to form alliances with different partner types. Taking the above arguments together, this study suggests that a firm’s diversification experience provides the firm opportunities to form diverse alliances and enable it to maintain or expand its portfolio diversity successfully. Thus, it is hypothesized that:
Hypothesis 1: A focal firm’s product diversification experience is positively related to its APD.
International diversification experience and APD
Firms learn from their prior international diversification experience and adjust their organizational structures, routines, and practices accordingly (Yu, Reference Yu1990; Bartlett & Ghoshal, Reference Bartlett and Ghoshal1999). An internationally diversified firm is exposed to customers, rivals, suppliers, and partners from various country- and market-contexts (Barkema & Vermeulen, Reference Barkema and Vermeulen1998). Exposure to such diverse parties and circumstances helps a focal firm enhance capabilities not only to manage and coordinate operations, but also add an array of competitive tactics (Bartlett & Ghoshal, Reference Bartlett and Ghoshal1999; Nadolska & Barkema, Reference Nadolska and Barkema2007). Researchers, drawing on organizational learning perspective, have found that prior international experience influences managerial decisions concerning firm’s strategic scope (Bowen, Baker, & Powell, Reference Bowen, Baker and Powell2014), growth of the firm (Mayer, Stadler, & Hautz, Reference Mayer, Stadler and Hautz2014), and choice of foreign entry mode (Barkema & Vermeulen, Reference Barkema and Vermeulen1998). Based on prior international experience, a firm may choose alliances or JVs as modes of foreign entry compared with other modes (Kogut & Singh, Reference Kogut and Singh1988; Hennart & Reddy, Reference Hennart and Reddy1997).
Additionally, a firm’s prior international diversification experience helps it develop abilities to collaborate with partners from diverse institutional and cultural contexts (Kogut & Singh, Reference Kogut and Singh1988; Barkema & Vermeulen, Reference Barkema and Vermeulen1998). The firm may transfer its experience to future decisions such as forming alliances with partners from similar institutional or cultural context (Kogut & Singh, Reference Kogut and Singh1988; Hennart & Reddy, Reference Hennart and Reddy1997). Such prior exposures may also create cross-border partnering opportunities for the focal firm. Moreover, one may expect that a firm with international diversification experience is more able to reap the benefits from higher level of APD, compared with a less experienced firm. Additionally, a firm with prior international experience would be able to mitigate the negative concerns of high level of APD, as the accrued learning would enhance firm’s abilities to manage complexities (Duysters & Lokshin, Reference Duysters and Lokshin2011). Taking together the above lines of arguments, this study expects that a firm’s international diversification experience should positively influence its APD.
Hypothesis 2: A focal firm’s international diversification experience is positively related to its APD.
Alliance experience and APD
Prior experience in forming and managing multiple alliances helps a firm develop alliance portfolio management capabilities and enhances the likelihood of alliance formation by the firm in future (Gulati, Reference Gulati1999; Rothaermel & Deeds, Reference Rothaermel and Deeds2006; Duysters et al., Reference Duysters, Heimeriks, Lokshin, Meijer and Sabidussi2012). Firms with prior alliance experience develop routines and mechanisms to reduce the risk of misappropriation of knowledge resources by partner firms (van Beers & Zand, Reference van Beers and Zand2014). Collaborative know-how developed by a focal firm through its prior alliances enables it to devise better conflict resolution mechanisms (Das & Teng, Reference Das and Teng2001; Nielsen & Nielsen, Reference Nielsen and Nielsen2009). Moreover, firms develop learning capabilities in their prior involvements in exploration alliance, enabling the firm to invest in multiple technological alliances in future (Gulati, Reference Gulati1999).
Prior experience with multiple partners enhances firms’ social capital, as firms become more central in their respective partnership networks (Al-Laham & Amburgey, Reference Al-Laham and Amburgey2010). Focal firm’s prior ties improve its trustworthiness, making the firm an attractive partner for future relationships (Ahuja, Reference Ahuja2000a; Castro, Roldán, & Acedo, Reference Castro, Roldán and Acedo2015). Moreover, multiple prior engagements with the same partner enhance focal firm’s trustworthiness with the other firms, who may not have direct linkage with the focal firm (Das & Teng, Reference Das and Teng2002). From learning perspective, a focal firm having prior experience with multiple partners might have developed capabilities that would enable the firm to manage diverse alliances and coordinate with different types of partners in future (Duysters et al., Reference Duysters, Heimeriks, Lokshin, Meijer and Sabidussi2012). A capable firm may choose to design a diverse portfolio in order to get the intended benefits from higher level of APD (Cui & O’Connor, Reference Cui and O’Connor2012). Thus, it is expected that:
Hypothesis 3a: A focal firm’s alliance experience is positively related to its APD.
AEH and APD
Learning from diversity literature suggests that a firm’s diverse experiences enrich its learning (Barkema & Vermeulen, Reference Barkema and Vermeulen1998; Schulz, Reference Schulz2002). Although homogeneous experience may make a firm more efficient through standardization of routines and practices, it is the diversity of experience that makes a firm capable of facing diverse circumstances effectively (Reuer, Park, & Zollo, Reference Reuer, Park and Zollo2001). The learning attained from prior diverse experiences saves a firm from falling into competency trap (Levitt & March, Reference Levitt and March1988).
Firm’s prior involvement with heterogeneous partners provides the firm opportunity to learn and develop alliance management capabilities (Gulati, Reference Gulati1999), which enable the firm to benefit from high level of APD (Rothaermel & Deeds, Reference Rothaermel and Deeds2006). Additionally, a firm’s learning with and from foreign partners from various institutional and cultural contexts not only enhances its ability to gain economic benefit from collaboration, but also creates future cross-border partnering opportunities. Prior studies suggest that firms build on their prior international alliance experience to increase the geographic diversity of alliance portfolio (Jacob, Belderbos, & Gilsing, Reference Jacob, Belderbos and Gilsing2013). Drawing on these findings and above line of arguments, it is hypothesized that AEH will positively influence focal firm’s APD.
Hypothesis 3b: A focal firm’s AEH is positively related to its APD.
Data and Methodology
Data and sample
A panel data set of alliance portfolio of Indian firms for the period 2004–2014 was prepared. However, since some of the predictor variables of this study are related to alliance experience, alliance data were collected for the period 1990–2014. The main reason behind choosing 2004–2014 as the period of this study was the unavailability of the data prior to 2004 from some of the sources. In particular, this study relied heavily on corporate annual financial reports to complement the data gathered from other sources. These annual reports were gathered from corporate websites. The majority of the firms in our sample had not archived annual reports prior to 2004. Thus, the start period of the study was chosen as 2004. The data set for this study was prepared by referring multiple sourcesFootnote 1 such as Securities Data Company (SDC) Platinum database, Prowess, India Business Insight database (IBID), and annual reports of firms.
SDC is the primary reference for information related to alliances and JVs. Although, SDC is comprehensive in its coverage of alliance formation and deal specifics, it does not provide complete information related to termination of alliances (Cui, Reference Cui2013). In order to get further information about the status of alliances, corporate annual reports and IBID were referred. Because of difficulties in gathering information regarding contractual alliancesFootnote 2 , this study included only JVs in the data setFootnote 3 . The operationalization of alliance portfolio in terms of JVs is in line with prior studies (Reuer & Ragozzino, Reference Reuer and Ragozzino2006; Collins & Riley, Reference Collins and Riley2013; Cui, Reference Cui2013). Furthermore, only bilateral JVs were considered. Multilateral JVs varied in terms of the number of participants, with minimum three and maximum eight partners. In order to maintain homogeneity among all the alliance units in terms of number of partners, multilateral JVs were excluded from the data set. Additionally, only the firms that were involved in more than one JVs during 1990–2014 were included in the data set (Jiang, Tao, & Santoro, Reference Jiang, Tao and Santoro2010; Cui & O’Connor, Reference Cui and O’Connor2012). Furthermore, as for additional data on alliances this study relied on publicly disclosed information, JVs formed by private entities (including business group holding firms) were excluded. Additionally, JVs formed by state agencies and state-owned enterprises were also excluded. Consequently, the data set contained 424 JVs by 99 publicly-listed firms during 1990–2014.
Furthermore, annual reports of these 99 firms and the IBID were searched to find additional information related to formations and terminations of JVs by these firms. Search revealed that 277 JVs in the data set were terminated at different points in time during 1990–2014. Additionally, 209 new JVs were found, which were formedFootnote 4 by these firms but not mentioned in the SDC database. Consequently, the final data set comprised 633 JVs formed by 99 Indian firms since 1990. However, nine of these firms had all their JVs terminated before 2004. These nine firms were excluded from the panel data set. Additionally, while preparing the panel data set those firm-year observations were removed in which the respective firms had zero number of active JVs (Phelps, Reference Phelps2010; Tao, Jiang, & Santoro, Reference Tao, Jiang and Santoro2014). Some firm-year observations for which the data for some control or predictor variables were not available were further removed. Consequently, the final unbalanced panel data set is comprised of 776 firm-year observations for 90 Indian firms for the time period 2004–2014.
Variables and measurementsFootnote 5
Dependent variable (DV): APD
APD is conceptualized as a two-dimensional construct that captures APD in terms of partner type and partners’ geographic region of origin (de Leeuw, Lokshin, & Duysters, Reference de Leeuw, Lokshin and Duysters2014; van Beers & Zand, Reference van Beers and Zand2014; Castro, Roldán, & Acedo, Reference Castro, Roldán and Acedo2015). The composite construct APD is measured by taking average of the Blau’s index of heterogeneity of these two dimensions: partner type diversity and partner geographic diversity (Blau, Reference Blau1977; Golonka, Reference Golonka2015). The Cronbach’s (Reference Cronbach1951) α for the composite construct is 0.82. The value of composite construct ranges between 0 and 1, where 0 signifies the least diversified while 1 signifies the most diversified alliance portfolio.
Partner type diversity indicates a firm’s ability to balance its access to supplementary or complementary resources (Jiang, Tao, & Santoro, Reference Jiang, Tao and Santoro2010). Partner type diversity is measured by calculating the heterogeneity index (Blau, Reference Blau1977) of focal firm’s alliance portfolio in terms partner types (Duysters & Lokshin, Reference Duysters and Lokshin2011; de Leeuw, Lokshin & Duysters, Reference de Leeuw, Lokshin and Duysters2014). Partner geographic diversity dimension indicates whether partners belong to similar or diverse cultural or institutional contexts. Following prior studies (Duysters & Lokshin, Reference Duysters and Lokshin2011; Bahlmann, Reference Bahlmann2014), partner geographic diversity is measured as heterogeneity index (Blau, Reference Blau1977) based on the geographic region of origin of the focal firms. Table A2 (Appendix A) presents the categories of partner types and geographic regions used to measure the heterogeneity index (Blau, Reference Blau1977).
Independent variables
Product diversification experience is measured as the entropy index of firm’s sales in different four-digit standard industrial classification (SIC) industries in the year preceding to the focal year (Jacquemin & Berry, Reference Jacquemin and Berry1979; Palepu, Reference Palepu1985; Mayer, Stadler, & Hautz, Reference Mayer, Stadler and Hautz2014).
International diversification experience indicates a firm’s exposure to different cultures and geographic markets. Following prior studies (Tallman & Li, Reference Tallman and Li1996; Gaur & Kumar, Reference Gaur and Kumar2009), international diversification experience is measured as the ratio of foreign sales to the total sales (FSTS) in the year preceding to the focal year. FSTS is a better measure of international diversification experience than the other measures, such as the count of the focal firm’s foreign subsidiaries or the count of the countries in which the focal firm has its subsidiaries, as FSTS accounts for the firm’s exposure to foreign market not only through subsidiaries, but also through exports (Sullivan, Reference Sullivan1994).
Following prior studies (Hoang & Rothaermel, Reference Hoang and Rothaermel2005), alliance experience is measured by counting the total number of alliances formed by a focal firm since 1990 to the start of the focal year. This measure captures the experience accumulated over time, reflecting the effect of learning (Anand & Khanna, Reference Anand and Khanna2000).
AEH measures the diversity of alliance experience. Firms that involve in alliances with different types of partners, over the time, are exposed to diverse sources of learning. Such experience might help firm in developing capabilities to form and manage alliances (Gulati, Reference Gulati1999; Rothaermel & Deeds, Reference Rothaermel and Deeds2006). Like APD, AEH is measured as a composite construct that reflects a firm’s heterogeneity of alliance experience in two dimensions: partner type (AEH partner type) and partners’ geographic region (AEH partner geographic region). The operationalization of these two dimensions of AEH are mentioned in Table A2 (Appendix A). The Cronbach’s (Reference Cronbach1951) α for the composite construct AEH is 0.78.
Control variables
Prior literature suggests various factors that may influence focal firm’s alliance portfolio characteristics. Firm size has been identified as a predictor of alliance portfolio characteristics such as size and diversity (Duysters & Lokshin, Reference Duysters and Lokshin2011; Jacob, Belderbos, & Gilsing, Reference Jacob, Belderbos and Gilsing2013). Hence, firm size is included as a control variable. Following prior studies, such as Leiblein and Madsen (Reference Leiblein and Madsen2009), firm size is operationalized as the natural logarithm of focal firm’s revenue in the focal year. Following prior studies Ahuja (Reference Ahuja2000a), firm age is included as a control variable. Firm age is measured as the natural logarithm of the count of years since incorporation of the firm to the start of the focal year. Furthermore, researchers have viewed prior firm performance as an important factor that may influence firm behavior, and hence controlled for it while predicting alliance behavior (Gulati, Reference Gulati1999). Prior firm performance is operationalized as return on assets of the focal firm in the year prior to the focal year. Additionally, Collins (Reference Collins2013) found significant positive relationship between focal firm’s capital intensity and APD. Hence, following his study, capital intensity is included as a control variable. Capital intensity is measured as the ratio of capital expenditure to revenue in the focal year (Mayer, Stadler, & Hautz, Reference Mayer, Stadler and Hautz2014).
Furthermore, the sample firms were spread across 27 two-digit SIC industries. To control for the industry-level factors, 26 dummies were included. Similarly, since the panel data set is prepared for 11 years (2004–2014), 10 dummies were included to control for the year effects. Table 1 presents descriptive statistics of the sample, along with Pearson’s correlation coefficient between key variables used in the study.
Table 1 Descriptive statistics
Notes. n (firm-year observation)=776; N (firms)=90.
APD, alliance portfolio diversity; AEH, alliance experience heterogeneity.
*p<.05; **p<.01.
Model Specification and Estimation
This study took panel data analysis approach to examine the hypothesized relationships. The Hausman (Reference Hausman1978) specification test suggested using fixed effect panel analysis approach. The DV (APD) is a fraction, which is bounded between 0 and 1. The nature of the DV poses challenges to linear regression model, as using a linear model may yield predictions outside the unit interval (Baum, Reference Baum2008). Papke and Wooldridge (Reference Papke and Wooldridge1996) proposed fractional regression model, based on quasi-maximum likelihood estimation (QMLE) method, to handle bounded DVs in cross-sectional data. However, one of the limitations of their model was its inadequacy to control for unobserved heterogeneity, which make it less useful for panel analysis (Wagner, Reference Wagner2003; Gallani, Krishnan, & Wooldridge, Reference Gallani, Krishnan and Wooldridge2015). Later, Papke and Wooldridge (Reference Papke and Wooldridge2008) extended their earlier fractional response models for cross-section data to panel data. They proposed fractional probit model (FPM), based on quasi-maximum likelihood estimation method, which allow controlling for the unobserved firm fixed effects. Following a few prior studies (Wagner, Reference Wagner2008; Kölling, Reference Kölling2012; Pericoli, Pierucci, & Ventura, Reference Pericoli, Pierucci and Ventura2013), this study considered the FPM as the primary methodological reference (Papke & Wooldridge, Reference Papke and Wooldridge2008). Additionally, for robustness checks, estimations of FPM were compared with fixed effect panel estimates, withFootnote 6 and without log-odds transformed DV (Baum, Reference Baum2008; Phelps, Reference Phelps2010). In all the estimation methods, experience-related predictor variables were lagged by a year with respect to the DV to ensure the temporal precedence of the predictor variables (Phelps, Reference Phelps2010).
Results
Table 2 presents the results of FPM based on quasi-maximum likelihood estimation. For models 1 and 2, the DV is composite construct APD. Moreover, the separate effects of predictor variables on each dimension of APD are examined in models 3–6.
Table 2 Results of fractional probit model, quasi-maximum likelihood estimatesFootnote a
Notes.For models 1 and 2, DV is the composite construct APD; For models 3 and 4, DV is partner type diversity; and, for models 5 and 6, DV is partner geographic diversity.
All models include individual time averages of explanatory variables as suggested by Papke and Wooldridge (Reference Papke and Wooldridge2008).
a Unstandardized coefficients are reported, with robust standard errors in parentheses.
DV, dependent variable; APD, alliance portfolio diversity; AEH, alliance experience heterogeneity.
† p<.1; *p<.05; **p<.01; ***p<.001.
In model 1, only control variables are included. The coefficients of firm size and prior firm performance are not significant. The coefficients of capital intensity (β=0.06, p<.01) and firm age (β=0.79, p<.05) are positive and significant, suggesting that capital intensity and firm age are positively associated with APD. In model 2, the main predictor variables are included. The coefficient of product diversification experience (β=0.28, p<.05) is positive and significant, supporting Hypothesis 1 that product diversification experience has positive association with APD. The coefficient of international diversification experience (β=0.45, p<.01) is also positive and significant, supporting Hypothesis 2 that international diversification experience has positive association with APD. Similarly, findings also support the other two hypotheses. In line with Hypothesis 3a, the coefficient of alliance experience (β=0.043, p<.01) is positive and significant. Finally, in line with Hypothesis 3b, the coefficient of AEH (β=0.40, p<.05) is positive and significant. Among the control variables, the coefficients of prior firm performance (β=−0.30, p<.1) and capital intensity (β=0.05, p<.01) are significant. However, the sign of the coefficient of the prior firm performance is negative.
In the next four models (models 3–6), APD is decomposed to isolate the effects of explanatory variables on each dimension. This refined analysis is motivated by prior studies that consider the dimensions of APD separately (Jiang, Tao, & Santoro, Reference Jiang, Tao and Santoro2010; Rogbeer, Almahendra, & Ambos, Reference Rogbeer, Almahendra and Ambos2014). For these models, APD is decomposed into partner type diversity and partner geographic diversity. The construct AEH is also decomposed into AEH partner type and AEH partner geographic region. In models 3 and 4, the DV is partner type diversity. In these models, AEH partner type is used as measure of diversity of alliance experience. Model 3 examines the effects of control variables. As in the model 1, the coefficients of capital intensity and firm age are significant. The findings of the model 4 support all the four hypothesized relationships. The coefficients of product diversification experience (β=0.31, p<.05), international diversification experience (β=0.37, p<.05), alliance experience (β=0.05, p<.01), and AEH partner type (β=0.54, p<.05) are positive and significant. Thus, findings of models 3 and 4 support that diverse experiences positively affect a firm’s partner type dimension of APD. In models 5 and 6, DV is partner geographic diversity. Diversity of alliance experience is measured by AEH partner geographic region. Model 5 introduces the control variables. Unlike models 1 and 3, coefficients of none of the control variables are significant. Findings in model 6 support all the hypothesized relationships regarding effects of: product diversification experience (β=0.25, p-value<.1); international diversification experience (β=0.54, p-value<.01); alliance experience (β=0.053, p<.01); and AEH partner geographic region (β=0.74, p<.001). However, the coefficient of product diversification experience is marginally significant at p<.1. On the other hand, the coefficient of AEH partner geographic region is significant at p<.001.
Thus, overall, findings support that product and international diversification experience, alliance experience, AEH are positively associated not only with overall diversity of the portfolio (APD), but also separately with each dimension of APD. Similarly, regarding the AEH, the findings suggest that the overall AEH of the focal firm is positively associated with overall APD. Additionally, the study suggests that there is positive association between the corresponding dimensions of AEH and APD. In sum, findings support all the four hypotheses.
Robustness checks
For robustness checks, additional models were estimated using panel fixed effect methods, with and without log-odds transformed DV. Table 3 presents findings of these two estimation methods. For models 7–10, the DV is APD. Models 1A and 1B present the average partial effects (APE) calculated for the models 1 and 2, respectively (Papke & Wooldridge, Reference Papke and Wooldridge2008). APEs of a FPM are comparable with the coefficients of the panel fixed effects. However, this comparison cannot be made with the coefficients of the log-odds transformed panel fixed effects model (Wooldridge, Reference Wooldridge2010; Pericoli, Pierucci, & Ventura, Reference Pericoli, Pierucci and Ventura2013). Models 7 and 8 present the estimates of the panel fixed effects, and models 9 and 10 present the estimates of the log-odds transformed panel fixed effects. Models 7 and 9 include only control variables, and models 8 and 10 are full models. The coefficient of international diversification experience is not significant in model 8, whereas the coefficients of other predictor variables are as hypothesized. In model 10, coefficients of all the predictor variables are positive and significant. However, the coefficient of international diversification experience is marginally significant at p<.1. As log-odds transformed panel FE and FPM are considered better estimation methods than a simple panel FE (Baum, Reference Baum2008; Papke & Wooldridge, Reference Papke and Wooldridge2008), the findings, in general, support all the four hypothesized relationships (Hypotheses 1–3b). Moreover, it can be observed that findings of the FPM are robust to other estimation methods.
Table 3 Standard estimation methods, robustness checksFootnote a
Notes.FE, fixed effects; AEH, alliance experience heterogeneity; APD, alliance portfolio diversity.
Dependent variable: APD
a Unstandardized coefficients are reported, with robust standard errors in parentheses.
b Firm-specific dummies are included in the panel FE models (models 7–10) to control for unobserved firm-level heterogeneity.
c Since log-odds transformation is undefined when dependent variable is equal to 0 or 1, standard practice was followed to recode these values as 0.0001 and 0.9999, respectively (Phelps, Reference Phelps2010).
d Models 1A and 2A present average partial effects (APE) calculated for the models 1 and 2, respectively. Model 1A and 2A, with APE, are included to facilitate comparison with the panel FE model. APE can directly be compared with coefficients of panel FE, but it cannot be compared with coefficients of log-odd transformed panel FE model (Pericoli, Pierucci, & Ventura, Reference Pericoli, Pierucci and Ventura2013; Wooldridge, Reference Wooldridge2010).
e Time-invariant industry dummies are not included in the fixed-effect models.
† p<.1; *p<.05; **p<.01; ***p<.001.
Furthermore, an unbiased measure of diversity was used to address the concerns of biasednessFootnote 7 of APD (Biemann & Kearney, Reference Biemann and Kearney2010). FPM is re-estimated with the unbiased measure of APD. As expected, the coefficients of product diversification experience (β=0.38, p<.05), international diversification experience (β=0.74, p<.01), AEH (β=0.75, p<.01) are positive and significant; thus, supporting Hypotheses 1, 2, and 3b. However, the coefficient of alliance experience is no longer significant. Additionally, the log-odds transformed panel FE is also re-estimated with the unbiased measure of APD. The results are qualitatively similar to those of the FPM. Findings of this analysis are interesting, considering that the Blau’s (Reference Blau1977) index underestimates the variety of a group when group sizes are relatively small (Biemann & Kearney, Reference Biemann and Kearney2010, p. 585). A detailed explanation of this interesting finding is provided in the discussion section.
For additional robustness checksFootnote 8 , alternative measures of the product diversification experience, international diversification experience, and alliance experience are used. Product diversification experience is measured by counting the number of product segments in which a firm is present (Barkema & Vermeulen, Reference Barkema and Vermeulen1998). International diversification experience is measured as the ratio of the number of outside subsidiaries to the total number of subsidiaries (OSTS) (Sullivan, Reference Sullivan1994). The correlation between FSTS and outside subsidiaries to the total number of subsidiaries is 0.63, suggesting that any of them can be used as proxy for internationalization experience. However, considering that FSTS incorporates export sales also, it is a better indicator of a firm’s internationalization experience (Sullivan, Reference Sullivan1994). Results remained robust to the alternative measurements of product and international diversification experience. Additionally, following Rothaermel and Deeds (Reference Rothaermel and Deeds2006), cumulative alliance portfolio ageFootnote 9 was used as an alternative measure of alliance experience. It is interesting to note that the coefficient of alliance experience became insignificant. Some alternative explanation for the same is presented in the Discussion and Conclusion section.
In sum, findings of the main model (model 2) are robust to the alternative measures of predictor variables, with an exception of alliance experience variable. One explanation for this exception is that the alternative measure may be inferior to the measure used in the main FPM (i.e., model 2). In general, the findings provide support for the hypotheses related to the learning from diverse experiences perspective (Barkema & Vermeulen, Reference Barkema and Vermeulen1998).
Discussion and Conclusion
Drawing primarily on organizational learning perspective, this study examines the experiential antecedents of APD. A longitudinal investigation on a panel data set of 90 Indian firm for the period 2000–2014 provide insight into how a firm’s diverse experiences influences its APD. Findings of this study are robust to multiple estimation methods and have implications for theory as well as practice.
Theoretical implications
Literature suggests that experience is considered the most important source of learning (Levitt & March, Reference Levitt and March1988; Rothaermel & Deeds, Reference Rothaermel and Deeds2006). Diversity enriches experience and facilitates learning (Levitt & March, Reference Levitt and March1988; Barkema & Vermeulen, Reference Barkema and Vermeulen1998). Learning from diverse experiences saves firms from falling into competency trap (Levitt & March, Reference Levitt and March1988). Broadly, this study considered two experience and learning-related factors: first, focal firm’s learning from prior diversification experience, including both product-market and international market; second, focal firm’s learning from prior alliance experience. It is argued that learning attained from prior diversification and alliance experience enhances focal firm’s ability to manage a portfolio of diverse alliances. A diversified alliance portfolio indicates that focal firm has access to diverse set of resources and that firm is able to strike competitive balance (George, Zahra, Wheatley, & Khan, Reference George, Zahra, Wheatley and Khan2001). Moreover, a focal firm may be interested in balancing its exploration–exploitation intent by purposefully engaging with diverse partners (Jiang, Tao, & Santoro, Reference Jiang, Tao and Santoro2010). Based on these arguments, four hypotheses (Hypotheses 1-3b) are made.
In Hypotheses 1, it is hypothesized that there exists a positive relation between a firm’s product diversification experience and its APD. This hypothesis is based on the argument that a firm’s presence in multiple product markets may be an important source of learning (Barkema & Vermeulen, Reference Barkema and Vermeulen1998). Learnings from such experience improve focal firm’s capabilities to manage complexities (Doz & Prahalad, Reference Doz and Prahalad1991; Bartlett & Ghoshal, Reference Bartlett and Ghoshal1993). A capable firm may maintain and get the intended benefits of high level of APD. Results provide support for Hypotheses 1 that focal firm’s product diversification experience is positively associated with its APD. Additionally, APD was decomposed into its two dimensions to isolate the effects of product diversification experience. Findings suggest that product diversification experience has positive association with each of the APD dimensions. Prior studies have not examined the effect of product diversification experience on a firm’s APD. Thus, findings of this study make valuable contribution to the alliance portfolio literature by emphasizing the effect of learning from diversity (Barkema & Vermeulen, Reference Barkema and Vermeulen1998).
In Hypothesis 2, it is hypothesized that there exists a positive relation between international diversification experience and its APD. This hypothesis is based on the argument that an internationally diversified firm is exposed to customers, rivals, suppliers and partners from various country- and market-contexts (Barkema & Vermeulen, Reference Barkema and Vermeulen1998), and such exposure helps a firm enhance its capabilities to manage diversity (Bartlett & Ghoshal, Reference Bartlett and Ghoshal1999; Nadolska & Barkema, Reference Nadolska and Barkema2007). Findings support Hypothesis 2 that international diversification experience is positively related to its APD. Additionally, the effect of international diversification experience on the each dimension of APD is also examined. Findings suggest that international diversification experience has positive relationship with each dimension. These findings are in line with those of prior studies. For example, Duysters and Lokshin (Reference Duysters and Lokshin2011) argued that firms benefit from the multinational experiences of their parent firms, as the linkages of the parent multinational company facilitate focal firm in forming and maintaining alliances with diverse set of partners. Similarly, Jacob, Belderbos, and Gilsing (Reference Jacob, Belderbos and Gilsing2013) found that firms enhance geographic diversity of their alliance portfolio as they attain international alliance experience. As the effect of firm’s own overall international experience on APD has not been examined thoroughly in the prior literature, the findings of this study are valuable contribution to the alliance portfolio literature.
In Hypothesis 3a, it is hypothesized that a firm’s prior alliance experience is positively related to its APD. This hypothesis is based on the argument that prior alliance experience enhances focal firm’s abilities to form and manage alliances (Gulati, Reference Gulati1999), and such experienced firms may be able to mitigate the negative effects of high level of APD (Duysters et al., Reference Duysters, Heimeriks, Lokshin, Meijer and Sabidussi2012). Results of the FPM (model 2) provide support for this hypothesis. Additionally, the effect of alliance experience is examined separately on each dimension of the APD. It is found that alliance experience has positive association with each dimension. The findings related to the general alliance experience are in line with prior studies that have found significant positive association between alliance experience and APD (Jacob, Belderbos, & Gilsing, Reference Jacob, Belderbos and Gilsing2013; van Beers & Zand, Reference van Beers and Zand2014). However, robustness checks suggest that the positive relation between alliance experience and APD is not significant when APD is corrected for its bias (Biemann & Kearney, Reference Biemann and Kearney2010) or when alternative measure of the alliance experience is used. These results are discussed below.
Regarding the relationship between alliance experience and unbiased measure of APD, the panel data were reanalysed and it was observed that there were many firm-year observations for which portfolio sizes were between 2 and 5. It should be noted that Blau’s index underestimates the diversity of the portfolio when the group size is relatively small, compared with a reference case having groups size equal to 10 (Biemann & Kearney, Reference Biemann and Kearney2010, p. 587). Indeed, when APD is corrected for the bias, the mean of the variable is 0.47, which is higher than the uncorrected value of APD (0.33). The correlation between unbiased APD and alliance experience is 0.37, which is lower than the correlation between uncorrected value of APD and alliance experience (0.56). Furthermore, the coefficient of alliance experience becomes insignificant for the unbiased APD. Thus, it appears that when measure of APD is corrected for the bias, the alliance experience loses its explanatory power. From theoretical perspective, this result is insightful as it suggests that at higher level of portfolio diversity, i.e. for the unbiased measure of APD, variables related to the diverse experiences (i.e., product diversification experience, international diversification experience, and AEH) have more explanatory power than the general alliance experience. This supports the contention that, compared with general alliance experience, diverse experiences are more important when it comes to managing diversity. However, one should be cautious in accepting this claim, as more rigorous analysis is required to generalize it and to better understand the relationship between general alliance experience and the level of portfolio diversity.
Regarding the alternative measurement of the alliance experience, it appears that cumulative alliance experience pronounces the overestimation of overall alliance experience, as experience decay was not accounted for, despite that experience was measured since 1990 (Stettner & Lavie, Reference Stettner and Lavie2014). Although, this overestimation might be present in the simple count measure also, it would be more pronounced when the portfolio age is used as a proxy for alliance experience. For example, suppose one observes an alliance portfolio in 2013; a portfolio that has three alliances, all being formed in 1990s, will appear to have more experience in terms of portfolio age, compared with a portfolio having three alliances that were formed during last 5 years. However, when one accounts for experience decay, recent experiences might appear more helpful, as the focal firm may not be able to benefit much from the temporally distant experiences (Stettner & Lavie, Reference Stettner and Lavie2014). Thus, further examination with some other measure of alliance experience, which accounts for experience decay, may be helpful in understanding the relationship between alliance experience and APD. This may be an avenue of future research to examine how experience decay may influence alliance portfolio characteristics.
In Hypothesis 3b, the nature of alliance experience is considered. Drawing on learning from diversity literature, it is argued that firms that have prior experience of collaboration with diverse partners learn and develop alliance management capabilities (Gulati, Reference Gulati1999). Firms with such capabilities may maintain a diverse portfolio (George et al., Reference George, Zahra, Wheatley and Khan2001). Findings support the hypothesis that focal firm’s AEH is positively associated with the APD. Moreover, AEH was decomposed into its dimensions – AEH partner type; AEH partner geographic region – to examine how they affect the corresponding dimensions of APD (i.e., partner type diversity and partner geographic diversity). Findings provide support that each dimension of AEH has positive association with the corresponding dimension of APD. Prior studies have not examined the influence of AEH on APD. In this regard, findings of this study make valuable contribution to the literature.
Additionally, like the findings of Collins (Reference Collins2013), findings of this study show that the coefficient of firm capital intensity is positive and significant in models 2, 8, and 10, suggesting that this finding is robust to all the three estimation methods. Although, prior studies have not provided any theoretical insight into the nature of relationship between firm’s capital intensity and alliance behavior, it appears that capital intensive firms may be keeping diverse alliance portfolio to manage their resource dependencies (Pfeffer & Salancik, Reference Pfeffer and Salancik1978; Pangarkar & Wu, Reference Pangarkar and Wu2012; Cui, Reference Cui2013). However, a further analysis with strong theoretical background is required to better understand the relationship between firm capital intensity and APD.
In sum, the study makes some valuable contributions to the alliance portfolio and organizational learning research. First, findings of this study contribute to alliance portfolio literature by providing empirical support to the predictions of organizational learning theory about APD. More importantly, this study integrates learning from similar experiences with learning from diversity perspectives with regard to antecedents of APD (Barkema & Vermeulen, Reference Barkema and Vermeulen1998; Jacob, Belderbos, & Gilsing, Reference Jacob, Belderbos and Gilsing2013; van Beers & Zand, Reference van Beers and Zand2014). Second, drawing on the insights from Barkema and Vermeulen (Reference Barkema and Vermeulen1998), this study considered firm diversification an important source of experience. Prior studies have not given much attention to the question that how the learning accrued through product or international diversification may affect a firm’s alliance behavior, in particular a firm’s APD. Thus, findings of this study adds to this less researched stream of learning from diversity literature (Barkema & Vermeulen, Reference Barkema and Vermeulen1998; Schulz, Reference Schulz2002). Third, findings of this study adds to the less researched area of antecedents of APD by examining experiential antecedents of APD. Fourth, although prior studies have examined the influence of prior alliance experience on APD (Jacob, Belderbos, & Gilsing, Reference Jacob, Belderbos and Gilsing2013; van Beers & Zand, Reference van Beers and Zand2014), the effect of the nature of alliance experience has not been investigated thoroughly. Findings of this study suggest that leaning from partners of different types and from diverse geographies positively influences a firm’s APD. In this regard, this study provides deeper insight into the relationship between nature of alliance experience (AEH) and APD. Fifth, to the best of our knowledge, this is the first study in the Indian context which examines antecedents of APD. Additionally, it is one of the few studies in emerging markets context that deal with alliance portfolio characteristics (Golonka, Reference Golonka2015). Alliances have become significant strategic choice for firms from emerging markets, as well (Hitt et al., Reference Hitt, Ahlstrom, Dacin, Levitas and Svobodina2004). Thus, this study complements the earlier studies, which are mostly undertaken in the developed markets context.
Managerial implications
The study has implications for managers as well. Strategic alliances have become significant drivers for firm growth (Wassmer, Reference Wassmer2010; Faems, Janssens, & Neyens, Reference Faems, Janssens and Neyens2012). However, the extent to which a firm may create and capture value through its alliances depend on its ability to simultaneously manage its multiple and diverse alliances (Gulati, Reference Gulati1999; Rothaermel & Deeds, Reference Rothaermel and Deeds2006; Duysters et al., Reference Duysters, Heimeriks, Lokshin, Meijer and Sabidussi2012). In this regard, findings of this study inform managers that a firm’s diverse experiences are transferrable to its alliance portfolio. Thus, the learning from prior diverse experiences may help a firm better manage its APD, which may lead to superior firm performance (Duysters et al., Reference Duysters, Heimeriks, Lokshin, Meijer and Sabidussi2012). Furthermore, as this study considers alliance portfolio at the level of corporate, it brings out how a firm’s learning from prior venturing into different product and international markets may make it capable of managing diversity. Such enhanced capabilities would be helpful in getting the intended benefits of high level of APD. Moreover, the study has emphasized the central role of learning and how it may influence a firm’s future alliance behavior. Thus, implicitly, this study highlights the managerial role to create routines and processes to maximize learning from experiences (Levitt & March, Reference Levitt and March1988).
Limitations and directions for future research
There are a few limitations of the present study. First, this study instantiates alliance portfolio in terms of JVs only. There are two primary reasons behind inclusion of only JVs in the portfolio. First, authors could find accurate information regarding termination of JVs by scanning annual reports. This information is not present for contractual alliances. Second, authors could find information regarding formation of additional JVs by scanning annual reports and searching IBID. These additional JVs were not reported in the SDC. Such information is not easily available for contractual alliances. Thus, in order to preserve accuracy of the data set, this study does not include contractual alliances. However, considering that broader definition of alliance includes both contractual and equity JVs (Culpan, Reference Culpan2009), it was suspected that the instantiation of portfolio in terms of only JVs may limit the generalizability of the findings of this study. However, additional analysis of the data set (see Appendix B) suggests that exclusion of the contractual alliances may not have major implications for generalizability of the findings of this study. Second, since the study analyses alliance portfolio at the level of corporate, it assumes that each business unit (or business segment) shares same level of importance in the eyes of the corporate executives, and, hence, the alliance strategy is homogeneous across all business units. However, such assumption may be strong if a firm’s alliance strategy varies across different business units (Hoffmann, Reference Hoffmann2007; Wassmer, Reference Wassmer2010). A survey-based study may help overcome this limitation.
Furthermore, business groups are known phenomenon in emerging markets (Khanna & Palepu, Reference Khanna and Palepu2000). It will be interesting to examine whether a firm’s alliance portfolio characteristics vary for business group affiliated and non-affiliated firms. Since majority of the firms in the sample (>80%) were affiliates of business groups, the influence of business group affiliation on APD has not been examined in this study. Future studies may compare business group affiliated and non-affiliated firms, by taking a balanced sample. Additionally, during the study, we observed that in some cases it is the ultimate group holding company which enters into the alliance. This observation suggests that there exists a higher level at which alliance portfolio can be conceptualized in the context of emerging markets (Wassmer, Reference Wassmer2010). Future studies may explore antecedents and consequences of APD by conceptualizing it at the level of business group.
Acknowledgments
The authors highly appreciate the feedback received from the Associate Editor Felix Arndt and two anonymous Reviewers. Additionally, the authors are grateful to some of their colleagues and friends at the IIM Lucknow for their valuable comments.
Financial Support
This research received no specific grant from any funding agency, commercial, or not-for-profit sectors.
Originality
The authors assure that this is an original work. The authors also declare that this paper has not been submitted elsewhere for publication.
Declaration
Both the authors of the article have read and approved the paper and meet the authorship criteria mentioned by the journal.
Contribution
A few influential articles concerning the phenomenon of alliance portfolio have been recently published in this esteemed journal, encouraging authors to consider this journal a potential outlet for publication of their work. The authors believe the article fits to the journal, as it extends the extant literature on alliance portfolio and makes valuable contributions by empirically validating the predictions of organizational learning theory about alliance portfolio diversity (APD). Findings of the study suggest how experiential learning is valuable in managing diversity. Findings inform managers that the learning accumulated through diverse experiences are transferrable toward the management of alliance portfolio.
Appendix A: DATA SET, VARIABLES, AND MEASUREMENTS
Table A1 The data set
Table A2 Variables and measurements
Appendix B: ASSESSMENT OF THE IMPACT OF EXCLUSION OF CONTRACTUAL ALLIANCES ON THE GENERALIZABILITY OF THE FINDINGS
An analysis of the SDC database revealed that as many as 32 of the sample firms (around 35% of the firms in sample) had no reported contractual alliance in the SDC during the study period 2004–2014. Furthermore, it was observed that firms from the biopharmaceutical and information technology (IT) and enabled services (IT/ITES) were involved in more number of contractual alliances, compared with the firms from other industries. This observation has also been highlighted in prior studies (Lee, Kirkpatrick-Husk, & Madhavan, Reference Lee, Kirkpatrick-Husk and Madhavan2014; Wassmer, Reference Wassmer2010). There were 23 firms in our sample that belonged to biopharmaceutical and IT/ITES industries. These firms were separated and APD (both Blau’s index and unbiased measure) for these firms is calculated for the period 2004–2014 by merging the JV data with the contractual alliance dataFootnote 10 , reported in the SDC. A t-test revealed that there was no significant difference between the means of the APD (both Blau’s index and unbiased measure) calculated with and without inclusion of contractual alliances. Furthermore, high correlationFootnote 11 is found between the APD, calculated with and without inclusion of contractual alliances. Thus, this additional analysis provides some confidence that exclusion of the contractual alliances should not have any major implications with regard to the generalizability of the findings of this study. Nonetheless, it is suggested that future studies should include both types of alliances by following the approach suggested by Lavie and Rosenkopf (Reference Lavie and Rosenkopf2006), who gathered information related to alliance formation and termination from corporate announcements and press releases.
