Partner Diversity in a Multi-Partner R&D Alliance Context

Nils Doorten

23 June 2014

Word count: 13.662

Partner Diversity in a Multi-Partner R&D

Alliance Context

Author: Nils Doorten s2400871

N.doorten@student.rug.nl

First supervisor: Dr. I. Estrada Vaquero Second supervisor: Prof. Dr. W.A. Dolfsma

University of Groningen Faculty of Economics and Business

Master of Business Administration

3

Partner diversity within a multi-partner R&D alliance context

Faculty of Economics and Business, University of Groningen

This paper offers insights in the, so far, unexplored phenomena of multi-partner alliances. The focus lies on how partners diversity influences firm performance within multi-partner alliances. An analysis has been conducted within 96 firms (from 65 alliances) to find out if and how partner diversity dimensions influence a focal firms’ performance within multi-partner alliances. Prior research identifies three different types of partner diversity: (1) industry, (2) organizational, and (3) national diversity within dyadic relations (or a combination of dyadic relations) and found diverse results on their relation with performance. My findings reveal mixed results, as industry diversity seems to have a U-shaped relation. This indicates that with low levels of diversity, the coordination and management costs for the alliance will rise but offer too little novelty in terms of resources and information. With higher levels of industry diversity the benefits will outweigh the costs. For the other dimensions, no significant relations have been found. Therefore, managers should be advised to balance the different dimensions of diversity, upon which, I argue that managers should not be afraid to face some diversity as it can offer the opportunities needed to benefit from the multi-partner alliances.

Keywords: Multi-partner alliances, industry diversity, organizational diversity, national diversity

1. Introduction

4 such as research, development, sourcing, production, or marketing of technologies, products or services”(Estrada Vaquero, 2012, p.39).

With this shift from dyadic to MPAs, firms potentially gain access to an increasing amount of resources, knowledge and capabilities. A key factor that determines management and performance in these alliances is partner diversity. Diversity refers to the degree of variance in partners’ resources, capabilities, knowledge, and technological bases (Goerzen & Beamish, 2005; Jiang et al., 2010). Parkhe (1991) identified the problem of diversity by arguing that dissimilarities between the social actors in an alliance can impede interactions. Furthermore diversity can also be the source of reduced social integration, greater communication problems and more conflicts (Zenger & Lawrence, 1989). Researchers identify all kinds of diversity, but in the core they can be reduced to three different kinds of diversity. First diversity between partners in terms of resources, knowledge, information, nationality, industry and organizational, in this research referred to as partner diversity. Second diversity related to the governance or management of an alliance. Third diversity related to purpose of the alliance (Jiang et al, 2010; Parkhe, 1991; Goerzen & Beamish, 2005; Cui & O’conner, 2012). This research focuses on partner diversity that has already received some attention in the context of portfolio alliances and alliances network. Within prior research in the context of portfolio and network alliances a U-shaped relation between partner diversity and performance has been found (Jiang et al, 2010; Goerzen & Beamish, 2005). Whereas in the past researchers thought that in order to be able to cooperate with each other, firms needed to be similar to each other and not have too much diversity. Nowadays scholars argue that some level of diversity is desirable as too little diversity will offer no novelty and render little new and useful resources, information and knowledge. They argue there is a trade-off between on the one hand the need for novelty and the other hand increasing diversity which makes an alliance difficult to manage.

5 of this research is to provide insights on how different levels of partner diversity influence the degree to which firms are able to benefit from participating in MPAs.

I argue that to some extend this same U-shaped relation holds for partner diversity within a MPA. However I propose that this relation is not the same within MPAs, compared to other previously examined alliance forms, because there are additional complicating circumstances. Due to increasing complexity it will become harder to communicate, focus and integrate the exchanged information, knowledge or resources. Therefore, I expect that firms will earlier experience these downsides of diversity due to the increasing complexity. This current research answers the following research question: How does partner diversity within a multi-partner R&D

alliance influence a focal firm’s performance? Within this research partner diversity will be

divided into 3 dimensions. The first labeled ‘industry diversity’ referring to the industrial background of firms. The second ‘organizational diversity’ focusing on diversity between firms in terms of organizational structure, organizational governance and ownership. The third ‘national diversity’ refers to differences in firms country of origin and related cultural characteristics.

6 This paper is structured as follows. First a review is given of the current literature within partner diversity and MPAs resulting in the construction of three different hypotheses. Second the methodology of this research is explained. Followed by the result section showing the analysis of the data. Finally the discussion and conclusion address the findings, limitations and directions for future research.

2. Theoretical background and hypotheses

The theory section will treat the topics of diversity and MPAs by describing the developments of these topics in current literature. Sequentially hypotheses will be constructed combining these two constructs.

2.1 From dyadic to multi-partner alliances

Different types of alliances have received attention within literature. The type of alliance, which received most attention within literature, is the dyadic alliance. Dussauge & Garette (1999) argue that alliance are “links formed between two independent companies which choose to carry out a project or specific activity jointly by coordinating the necessary skills and resources rather than pursuing the project or activity on their own, taking on all the risks and confronting competition alone, merging their operations or acquiring and divesting entire business units” (p. 23). More recently other types of alliances start receiving attention too. Whereas dyadic alliances are characterized by a one-to-one exchange lately also other configurations of alliance are being examined. Researchers nowadays no longer view alliances as a single exchange relation but put this relation into a bigger context. According to Lavie and Singh (2012) alliance portfolios are the collection of all the immediate alliance partners a firm has. By adopting an alliance portfolio (view) complexity will increase for the focal firm. Hoffman (2005) argues that for alliance portfolios to be managed effectively they need to be observed and coordinated, and at the same time management practices need to be developed and institutionalized. The focal firm will be exposed to an increasing amount of ‘incoming’ information accessed through alliances, and will have to be able to recode and filter all this information. The different alliance partners within the portfolio will have a diversity of backgrounds and characteristics and all of them need specific collaborative approaches in order to efficiently manage them.

7 network as the unit of analysis. By doing so it recognizes that not only individual alliances (or the portfolio of them) can generate value, but also the network as a whole can do so. Davis and Greve (1997) argued that networks could also offer platforms for discussion, direct attention to new practices and facilitate the transmission of information. On top of this Gulati (1999) argued that alliance networks can be seen as imitable and non-substitutable resources as well as means to access unique capabilities.

In contrast, Das and Teng (2002) do not view the alliance network as a new stream of literature and describe it as simply a collection of several alliances. Authors that do research in the multi-partner alliance phenomenon often refer to it as constellations (Das & Teng, 2002), Multi-lateral alliances (Doz & Hamel, 1998; Li et al., 2012) or Multifirm alliances (Hwang & Burgers, 1997). MPAs, according to Lavie et al. (2007), are not simply a collection of dyadic alliances or a network of partners with direct ties with a focal firm. MPA’s essentially differ from the dyadic alliances, as they are formed by 3 or more firms (Das & Teng, 2002.) Lavie et al. (2007) argue that MPA settings ”rather entails multilateral interaction among partners and this generates unique dynamics” (p. 580). They argue that MPAs involve multilateral relations and interaction patterns between all of the participants, have far more complex governance structure and unique ways of collaborating. Moving from dyadic alliances to alliances with multiple partners’ changes the firm’s interpersonal exchanges. Das and Teng (2002) argue that constellations will result in a more generalized rather than restricted exchange. This has to do with the fact that firms no longer interact with firms individually but with a group of firms. Cooperating in this form has a number of implications as it gets harder to evaluate each firm’s contribution to the alliance. Furthermore reciprocity gets increasingly complex as firms no longer only exchange with one partner and show their appreciation by paying back their ‘debt’ but owe allegiance to a group of firms. This current research will focus on MPAs as little research has still been conducted in this field.

2.2 Partner diversity

8 resources which they would otherwise not be able to acquire or which would be extremely expensive. When firms start exchanging their resources, tangible or intangible, there are several reasons why firms fail to maximize benefits from these interactions. One of these reasons is the difference between firms, which hampers them to work together in the most efficient way, diversity.

Diversity first received attention in the context of a dyadic relationship between two partners. Parkhe (1991) identifies two different types of diversity: “Type І diversity includes the familiar inter-firm differences (interdependencies) that Global Strategic Alliances are specifically created to exploit. These differences form the underlying strategic motivations for entering into alliances” They are specifically concerned with differences that facilitate the formation and maintenance of the alliance. “Type ІІ diversity refers to differences in partner characteristics that often negatively affect the longevity and effective functioning of an alliance” (p.580). This current research will focus on the, by Parkhe (1991) labeled Type ІІ diversities.

More recently diversity has also received attention within alliance portfolio’s and alliance networks. Goerzen and Beamish (2005) researched network, product, geographic, network size, industry profitability and other types of diversity within alliance networks. Researchers so far have found different types of diversity but in concept they can all be reduced to three main categories. Jiang et al. (2010) labeled these three categories as: (1) partner diversity which deals with diversity related to the differences in characteristics of both partners, (2) functional diversity deals with the functional purposes of firms, as some involve exploring activities (R&D alliances) and others more exploiting activities (marketing, manufacturing and others) and (3) governance

diversity which has to do with diversity within the governance mode of the alliance, like for

9 country of origin and related differences in cultural norms.

2.3 Partner diversity within multi-partner alliances

The selection of partners for collaboration is an important activity as new skills can be learned from this partner. Prior studies have examined the impact of partner diversity on firm performance within different alliance types and contexts. Goerzen and Beamish (2005) found that partner diversity within Japanese firms’ foreign subsidiary network, has a U-shaped relation with performance when looking at industry and country background. In line with Goerzen and Beamish (2005) findings Jiang et al. (2010) also found a U-shaped relation between industry diversity and performance. Their other dimensions of partner diversity influenced performance in other ways, as organizational diversity had a more J-shaped relation.

All of the previous relations were found for partner diversity within dyadic alliance, whether portfolio or network. The focus of this research will shift from these dyadic alliances towards MPAs with their more generalized exchanges. Taking the RBV perspective, more partners would imply more possible complementary resources that can be accessed and integrated. Besides the positive effects, adding partners also can also have a number of downsides. Medcof (1997) argues that having more prospective partners within an alliance makes managing this alliance more complex. A higher number of partners often increase uncertainty due to the increasing divergence and diversity in goals and the potential for contingencies (Gonget al., 2007). Medcof (2007) in the same vein argues that there is a higher

need for a strategic fit between the several partners in terms of strategic added value within the alliance and the strategic purpose for entering the alliance. Having more partners makes it more difficult to ensure you receive and secure the right information about all partners’ intentions,

capabilities, behavior and commitments (Gong et et al., 2007). Shaw (1971) furthermore argued

10

(Gong et al., 2007). Overall MPAs could result in higher coordination costs, higher communication costs, greater conflict propensity, more opportunistic behavior and have higher

levels of diversity (Hennart & Zeng, 2002; Garcia-Canal et al., 2003).

Firms will face trade-offs as the diversity within their alliances increases. Jiang et al. (2010) state that on the one hand more can result in broadened search options, access to enriched pools and hence added value creations, capability development and opportunities. Whereas at the other hand too much diversity can also increase complexity, as it increases the contingent for conflicts and therefor will increase the costs related to coordination and management of the alliance.

2.4 Hypotheses

Figure 1 shows how the assumed relations of the three dimension of partner diversity (industry, organizational and national) influence financial performance of a focal firm. Including the control variables that further influence the performance of the focal firms to complete the model. Further explanation on why and how these relations influence performance is to be found on the next pages. Finally it is important to know that for all the hypotheses the focal firm’s performance is measured. Performance shows what the influence of the alliance has been on the performance of the firm and not the performance of the whole alliance.

11

2.4.1 Industry diversity

In this research partner industry diversity is defined as the level to which industrial backgrounds differ between the focal firm and those of the rest within a MPA. “Every industry has its own ‘industry recipe’ for success, the conventional wisdom on how a firm’s resources, knowledge, and processes should be combined and utilized to cope with the environment” (Lane & Lubatkin, 1998, p. 462). Jiang et al. (2010) argue that partners that are from the same industry and are competitors may bring the greatest learning through imitation and absorptive capacity because they have an overlap in their knowledge bases, backgrounds, experiences and technological bases. On the other hand they are also prone to fall into conflicts and learning races. Misfits can also occur in terms of resources or the lack of synergies. Industry diversity may provide access to resources, knowledge, capabilities and learning which would not be present without these differences. At the same time this diversity will be hard to manage and it’s more difficult to reach synergies.

Industry diversity has received a big deal of interest within the alliance portfolio perspective where several researchers found a curvilinear relation between industry diversity and performance. Goerzen and Beamish, (2005) found a U-shaped relation between performance and partner diversity within network alliances, including measures such as industry and national diversity. This same U-shaped relation between performance and industry diversity was found within portfolio alliances (Jiang et al., 2010). Jiang et al. (2010) argue that this relation exists because firms at first have to immediately face the level of diversity with one another. When these firms get more adept to the costs of cooperating and learning, performance will start increasing again. Firms will, from that point, be better able to manage the diversity resulting in an increasing performance.

12 opportunities, but the costs do rise. Therefore performance will first decrease and after a certain minimum, performance starts increasing after the firms adept to the circumstances. Industry diversity within MPAs is assumed to have a U-shaped relation with firm performance.

Hypothesis 1: Industry diversity is associated to a focal firm’s performance in a multi-partner alliance in such a way that low levels of diversity will result in decreasing firm performance up to a minimum point after which higher levels of diversity will result in an increasing performance (U-shape).

2.4.2 Organizational diversity

In this study organizational diversity will be viewed as the difference between firms in terms of organizational structure, organizational governance and ownership. Authors classically look at differences in organizational culture (Catwright & Cooper, 1993; de Man & Duysters, 2002) or the strategic fit in terms of complementary resources, compatible businesses and strategic objectives and goals (Das & Teng, 1997; Parkhe, 1991), when thinking of organizational diversity. Lavie and Singh (2012) more recently also offer attention to differences in organizational routines. This research will focus on that last view, where organizational diversity rather represents the more organizational centered differences.

13 On the contrary divergence in strategic goals, organizational structures and interests might also potentially be the reason for conflicts as it complicates alliance contract negotiations, alignment of goals, communication and information transfer (Piva & Rossi-Lamastra, 2013). Which in turn raises the management costs of communication and coordination in the alliance.

Jiang et al. (2010) concluded that the benefits of adding different types of partners to the alliance outweighs the costs of managing such a different partner resulting in a J-shape relation with performance. I assume this J-shape relation will not hold in the case of MPAs. When diversity levels rise at first the costs will outweigh the benefits, as it will become harder to manage the relation. Within MPAs a focal firm does not gradually get exposed to dyadic relations with new partners one by one. Instead a firm, just like with industry diversity, starts facing this increasing complexity as soon as it occurs. The increased complexity raises managing costs that at first outweigh the benefits resulting in a decreasing performance. Until the minimum performance has been reached, from this point on the additional diversity will start resulting in enough novelty to increase the performance. Therefore I assume there will be a U-shaped relation between organizational diversity and performance.

Hypothesis 2: Organizational diversity is associated to a focal firm’s performance in a multi-partner alliance in such a way that low levels of diversity will result in decreasing firm performance up to a minimum point after which higher levels of diversity will result in an increasing performance (U-shape).

2.4.3 National diversity

14 technical capabilities, and intangible assets. The other way around firms from developed countries want to ally for unique competencies, market knowledge/access, previous alliance experience and many others (Hitt et al., 2000). Jiang et al. (2010) point out those differences in political economical systems, societal and cultural institutions, government policies and national industry structure complicate allying for partners. These circumstances make managing the alliance harder and coordination costs will rise and conflicts are more likely to appear. High levels of national diversity are expected to increase the costs for coordination and integration and it will become even more complex to manage the alliance (Goerzen & Beamish, 2005). On the other hand, it is expected that allying with partners from different nations will bring along some advantages as the firms will bring in novelty in terms of resources and market access. Jiang et al. (2010) in their research for national diversity within a portfolio of alliances found that there was no U-shaped relation between national diversity and performance. Pothukuch et al. (2002) argued that national culture differences would negatively influence performance, but had to reject this hypothesis. Other authors however did find that too much diversity would decrease perfomance, Sivakumar et al. (2011) found that diversity would increase performance up to a certain point, after which the relationship becomes negative. In the same vein Richard et al. (2004) found that with increasing levels of cultural heterogeneity would continue to affect performance positively up to maximum levels of heterogeneity.

So national diversity offers firms opportunities to access new resources and novel information that they can benefit from. At some point however the increasing diversity will lead to increasing managerial and communicational costs resulting in a decreasing performance. Within MPAs especially communicating and coordinating will get increasingly difficult with higher levels of diversity impeding firms to efficiently work together. In contrast to the previous two hypotheses, the relation between performance and national diversity will first increase up till a maximum point, before decreasing again. Unlike the previous two hypothesis the national diversity will be analyzed relative to the focal firm, whereas the other two hypothesis focus on the group diversity.

Hypothesis 3: National diversity is associated to a focal firm’s performance in a multi-partner

15

2.4.4 Interaction effects

Together these three dimensions of partner diversity will give a clear view of how diversity influences firm performance in a multi-partner alliance. As the three dimenions are all part of the overarching construct ‘partner diversity’ it is likely that firms will not only have diversity related to one of the dimensions, but in two or even all three of them. Within institutional theory isomorphism is a widely accepted phenomenon. Isomorphism is "constraining process that forces one unit in a population to resemble other units that face the same set of environmental conditions" (DiMaggio & Powell, 1983, p.149). Therefore it is relatively likely that firms from a certain industry are also more likely to have adopted a particular organizational structure. The same applies to some degree for firms from a certain national background. Parkhe (1991) identified two different types of diversities from which diversity in partner characteristics is one. The different aspects of partner characteristics (industry, organizational and national in this research) all have some similarities in common making it sometimes hard to distinct between each other. Having levels of diversity within the different partner diversity dimensions might have an influence on performance. It is however not clear how these dimensions interact with each other as existing research does not allow theorizing clearly about the interaction between the dimensions. I will run an additional model on interaction effects, but the phenomenon are not hypothesized as it is not clear how the dimensions will react on one another.

3. Methodology

This section starts with describing how the data is gathered within this research, followed by an explanation of how the sample is formed. Finally I also describe and explain in detail how the variables are measured.

3. 1 Data collection

16 argued that these industries are specifically interesting for research on R&D because their survival and profitability is dependent upon the ability to develop and commercialize innovations in a quick and efficient manner. To identify these industries within the SDC database Standard Industrial Classification (SIC) codes have been used. For the range of SIC codes related to high-tech industries I used the Area broad high-high-tech categories used by Li, et al. (2008): high-high-tech manufacturing (SIC codes starting with 357,365, 366, 367, 381, 382, 384, and 386) and high-tech services (SIC codes 481, 482, 484 and 489). The range of SIC codes for biomedical and chemical consists of SIC codes 282, 283, 336, 338, 509, 737 and the complete codes 8062 and 8711. Data has been extracted from the SDC- database on MPAs, which have been announced between 2005 and 2009. I chose these years because the additional financial data (from Orbis) goes back in time for approximately 10 years. Alliances from before this cannot be linked with the necessary additional data. Newer data on the other hand might not be updated yet. Additional data had to be gathered from sources other than the SDC-platinum database, as this does not supply sufficient financial data. Therefore additional financial data of the individual firm performance is extracted from the Orbis (Company information across the globe) database and combined with the existing data. Other firm information, such as firm age also is gathered from the Orbis database. A final method for obtaining this information, in case previous sources were not sufficient, was to look up the firm’s age on the corporate firm’s website.

3.2 Sample

The initial sample consisted of 720 firms. This number dramatically declined as each firm was linked to financial data (net profit and profit margin), which was subtracted from the Orbis database. The sample further reduced as for a number of firms, the necessary R&D expenses could not be found in the Orbis database. All of these firms had to be excluded from the database. The final sample includes 96 firms of which 49 (so 51%) are high- tech and 47 (so 49%) biomedical/chemical. To conclude, when firms had participated in more than one alliance over the years, only the alliance that was announced first is included in the sample.

3.3 Study

17 years after the announcement of the alliance. To check for the hypotheses this research consists of several stages. In the first stage the correlations between the different variables (explained in the next subsector) are calculated. The next stage consists of checking both linear and curvilinear relations between the control and independent variables and the dependent variable. Finally the last stage shows possible interaction effects between the three independent variables industry, organizational and national diversity. The longitudinal research provides insights on how increasing levels of partner diversity influence performance over time.

3.4 Dependent variable

The dependent variable, which is assumed to be influenced by the other variables, is a focal firms’ performance. Net profit margin (Jiang et al., 2010; Faems et al., 2010) is a financial measure used for tracking performance in this research. Net profit margin (NPM) is a useful way of comparing corporate firm performance with that of others. Using the profit as a percentage of the whole turnover makes it possible to account for the size of firms, as bigger firms will naturally have bigger turnovers. In order to obtain a useful figure I first took the NPM average of one year before alliance announcement and the first year of the alliance. After this the average of the NPM of three years after announcement of the alliance is calculated. By subtracting the average of average initial net profit margin by the three after the announcement from, the financial performance is calculated (see equation 1).

18 Table 1 measures and data sources

3.5 Independent variables

The focus of this research lies on the three different dimensions of partner diversity: industry, organizational and national diversity.

Industry diversity. Industry diversity in this research is measured by using firms SIC-codes

19 this firm and the others is calculated. Combining all these figures results in figure for the diversity within the group . Compared to previous research this measure is more adequate for this research, because I examine the diversity of the alliance, whereas previous research focused on dyadic relations within a portfolio.

Organizational diversity. The age of the firm at the time of alliance announcement is used as a

measure for organizational diversity, which refers to differences in structures, goals and decision-making processes. Sørensen and Stuart (2000) argued that older firm would have perfected routines, structures, incentives programs and infrastructures with which they have been working. As firms age they become more dependent upon there established routines and are less likely to change strategic directions (Lavie & Rosenkopf, 2006). I therefor use the age of the firm, at the time of alliance announcement, to measure differences in organizational diversity. In order to calculate the differences in diversity I categorized the firms age into 5 groups (Lindow, 2013) which are labeled ‘1’ if they are between 0-10 years, 2’ if they are between 11-20 years, ‘3’ if they are between 21-50 years, 4’ if they are between 50-100 years and finally 5’ if they are over 100 years. Like industry diversity also organizational diversity is an absolute diversity measure which measures the diversity within the alliance.

For both industry and organizational diversity (separately) I obtained the level of multi-partner diversity by using the Blau Index of variability, which has been used widely within literature for calculating diversity (Jiang et al., 2010; Richard et al, 2004; Li et al, 2008; Koka et al., 2002). The Blau Index is used by, applying the next formula:

Equation 1 Blau index

20 calculated by P-1/P, where P stands for the number of categories. In contrast to previous research, which examined the dyadic relationships diversity, this research calculates diversity within MPAs (group diversity). When doing research after dyadic relations (or the collection of dyadic relations, like a portfolio of alliances) a specific figure for that focal firm can be obtained. However, for MPAs (where a number of firms participate) this is slightly different, as not only the focal firms have diversity with the other participants, but so do all of the other participants. In order to calculate (p) the proportion of belongings to a given category (so e.g. in case of industry diversity the 0,1,2,3and 4’s) all the relations between the different participants with the same level of diversity within the alliance are pooled. These ‘pools’, representing a certain degree of diversity, are divided by the total amount of relations within the alliance (i). This way one diversity figure is calculated for all the participants in the alliance (group diversity). So the method of calculating is the same except it is applied to all of the partners within the alliance to raise a general group diversity figure.

National diversity. All the firms within the SDC-platinum database contain information about

their country of origin. I use this information about the background of the focal firm to distract the number of different nationalities within an alliance. The amount of foreign partners is calculated from the focal firm’s perspective, whose origin is seen as the ‘home-country’. Like Jiang et al. (2010) I categorised the outcome into 5 different categories starting with ‘0’ meaning there are no foreign partners, ‘1’ there is one partner from a foreign country, ‘2’ there are two firms fom foreign countries and so on up till category ‘4’. In contrast to the other independent variables the national diversity figure is relative. I take the focal firms perspective and compare whether the firms are from the same country or from a different country. The reason for this difference in measuring has to do with the relative subjective characteristic of nation diversity. It is hard to say that one firm is relatively more diverse from the focal firm than the other. Therefore the number of different countries are used for measuring diversity.

3. 6 Control variables

21

Firm size. The size of the firm may have an impact on performance. As Sarkar et al. (2001)

suggest that smaller firms are disadvantaged against larger partners as they get disproportionally exploited and suffer on the long term. Firm size is measured its number of employees (Goerzen & Beamisch, 2005). Firm size is categorized according to the official European size definitions (European Commission, 2003) also used by Thorgren et al. (2012). Small firms consist of firms with less than 50 employees, medium firm employ between 50 and 249 employees, large firms employ between 250 and 999 employees and finally enterprises employ 1000 or more employees.

Alliance size. Furthermore performance may also be influenced by thee alliance size.

Gomes-Casseres (2003) argues that group size in constellations is a key design criterion. “For some firm

goals, the more the merrier; for example, more partners help build momentum and may share risk or reach markets better than fewer partners” (Gomes-Casseres, 2003, p.2). The same way Jiang et al. (2010) argued that alliance portfolio size might be influential I argue alliance size is.

Like argued mentioned by Jiang et al. (2010) and Gomes-Casseres (2003) number of partners is

an effective way to measure the alliance size. The actual size of the alliance, which is noted in

the SDC-platinum database, will be used as a number ranging from 3 till 6 representing the number of particpants.

Prior Experience. A number of researchers have argued that experience with alliances influences

firm performance. Duysters et al. (2012) conclude that previous alliance experience positively influences the relation between diversity and firm performance. Therefore prior alliance experience is also included as a control variable in this research like Sampson (2007) did when researching technological diversity. The number measures prior experience or previous alliances firms have formed before participating in their current alliance. This data is gathered from the SDC-platinum database. For this measure I used alliances in general, so not only MPAs. Data has been tracked back 5 years (like Li, et al. 2008) before the announcement of the alliance. Based on Draulans et al. (2003) their findings I created a dummy where ‘0’ represent the firms that have formed from 0 up to 6 previous alliances and ‘1’ the firms that have formed over 6 alliances in the past.

R&D intensity. The propensity of firms to invest in their R&D departments is likely to influence

22 & O’conner, 2012; Lavie & Rosenkopf, 2006). I gathered each firms 4 years of R&D expenditures since the year of announcement of the alliance. The average expenditure of these 4 years I use by dividing this number by the average net income (before taxes) of the same 4 years (Goerzen & Beamish, 2005). This way I calculated the R&D relative figure (see equation 3).

Equation 3 R&D intensity

Alliance year. I assume that firm performance is partially dependent upon the specific year the

alliance has started. Certain years might be more or less ‘fruitful’ than other years. One reason for this assumption is the financial crisis especially Europe and America faced around 2008. This crisis is likely to be influential for firm’s performance during those years. Thus the different year in which the alliance is announced and running will be controlled for like Sampson (2007) did in her research. After comparing the means of the different years related to firm performance I concluded that the means dramatically dropped in 2007. Therefore I control for this crisis with a dummy. The ‘0’ means alliance started before 2007 and ‘1’ started in the year 2007 or later.

Strategic importance. How important the alliance is to the focal firm is also an important issue.

Mothe and Quelin (2001) argue that involvement in an alliance will be higher if they attach higher strategic importance to that alliance. They assume that a higher importance will result in greater benefits to that partner. For that I also control for the influence of strategic importance to the focal partner. In the same way Mothe and Quelin (2001) use financial expenses as an indication for strategic importance, I use the SIC-codes of the focal firm and compare that to the SIC-code of the alliance. When three digits of the focal firm and those of the SIC of the alliance match, the alliance will be considered of high strategic importance to the firm. These firms are labeled ‘1’ whereas the firms without this match are labeled ‘0’.

Joint venture. Alliances can take different forms, some more ’hierarchical’ like joint ventures

23

4. Results

The analyses have been conducted in multiple stages. The first three models are used for analyzing my proposed relations whereas the other models are used for examining the interaction effects between my variables.

24 Table 2 Descriptive statistics

This study examines whether partner diversity influences a focal firm’s performance. In order to analyze the effect of diversity on a focal firm’s financial performance I conducted least squares regression analyses (GLS) with SPSS resulting in three models. Model one includes only the control variables and shows how these variables impact the focal firm’s performance. Model two adds the independent variables of multi-partner diversity. Finally model three contains the quadratic and root terms of these diversity variables to check for curvilinear relations with performance. Due to the relative small sample size N=96 not only the probability value of p< 0. 05 but also the probability value of P< 0.1 is included. Even though between P < 0.05 and P < 0.1 cannot be considered significant, these values do offer some suspect for the relation being more than completely random and shows a tendency towards significance (Fisher, 1992).

Model 1 shows that, in line with previous research findings ((Jiang et al., 2010), the control variables firm size and alliance size have no significant effect on performance. The other control variables prior experience, R&D, year (crisis), strategic importance and joint venture also don not have a significant relation with performance. All of the control variables show consistency across the three models. The variable inflation factors (VIFs) present an acceptable figure between minimum 1.059 and maximum 1.273. These VIF score measure to which extent collinearity occurs among the variables. In general VIF scores lower than 10 are regarded acceptable (Hair et al., 1998).

25 and national diversity shows a significant relation with performance. All three hypotheses propose a curvilinear relation of their independent variables with performance. The first method to check these relations was to divide each of the diversity dimensions into 2 groups, low and high. Based on these distinctive groups I would measure whether there would be an increasing or decreasing slope indicating a curvilinear shape. When analyzing, some of these groups turned out to be too small for conducting analyses. Therefore the same approach as Jiang et al. (2010) was chosen for checking the curvilinear relations. Adding a square and square root of the firm’s partner diversity scores to the model. A remarkable finding however was that when observing national diversity at model 2 it was still not significant but was close as it scored a 0.11 significant level. The control variables still show no significant relation with performance and their Beta values stay consistent. Despite the earlier correlation between the independent variables their low VIF scores, ranging from 1.273 till 1.592, suggest that the independent variables do not suffer from multicollinearity issues. The same can be concluded for the control variables.

26 Especially this coordination gets increasingly difficult and has to be structured differently when the number of employees increases. Number of employees is also used as a variable to measure organizational diversity. Comparable to firm’s age, employee diversity has been calculated using the Blau index. The 5 different groups in which employees have been divided are explained at employees control variable section. This new variable (also squared and square rooted) however did not show a significant relation with performance either. Hypothesis 2 is henceforth rejected. In order to still be able to check for hypothesis 3 its variable is also added to model 3. The assumed inverted U-shape relation would in this model have to show the opposite scores of industry diversity (which has a U-shape). The model however shows that there seems not to be any relation between national diversity and performance. Based on this information, I reject hypothesis 3. It’s noteworthy that relation between national diversity and performance rather seems linear (sign. level 0.11, β= 0.19) as it scores much higher significance compared to the significance levels of this score rooted (sign. level 0.79) and square rooted (sign. level 0, 85) for the curvilinear relation. This figure gets even more significant when dividing national diversity by the number of participants within the alliance, calculating a relative national diversity figure. If this relative national diversity figure is included in model 2, instead of the previously used national diversity figure, it shows a low significance (β= 0.22, P <0.1). This model is included in the appendix as model 11, relative national diversity. When this relative national diversity figure is squared and squared rooted to look for a curvilinear relation it is no longer significance anymore.

Finally the R square (or R²) in the second model shows a score of 7% and in model three a score of 15%. This tells us that 15% of the variation on performance is explained by my model. This could be interpreted as a relatively low explanatory value. It however makes sense that firm's overall profit margin is influenced by more than the just the variables in this model as it is a relatively ‘broad’ measure.

27 Table 3 OLS regression results

1. Regression coefficients (β) are shown 2. *p<0.05, **p<0,01

3. Number of firms (N)=96

28 present. The other dimensions do not moderate this mean that the relationship of each of the three dimensions of partner diversity with performance.

Finally the squared terms of the mean centered variables are introduced to the model 7 for an extra check on hypothesis 1,2 and 3. This model also shows a significant relation between the squared value of industry diversity and performance, thereby offering some support for hypothesis 1. Models 4,5,6 and 7 show that the VIF scores remain acceptable throughout the models as they do not get higher than 4.176 indicating that there is reason to assume issues of multicollinearity.

Table 4 interaction effects

1. Regression coefficients standardized (β) are shown 2. *p<0.05, **p<0,01

29 Post hoc analysis on interaction effects

In order to see whether there are any interaction effects when looking at specific groups, I

30 organizational and low national diversity have significant interaction effect(s), whereas the smaller sample sizes do not show these effects.

Table 5 concludes the results showing that hypothesis 1 is supported whereas in the case of hypotheses 2 and 3 I could not find support for the hypotheses, so these are rejected. Furthermore I did not find any clear significant interaction effects between the three dimensions of partner diversity. An additional analysis nonetheless showed that, with sufficient sample size, there might be such interaction effects. The limited sample size of this research however does not allow me to make any conclusion based on this analysis.

31 1. Regression coefficients Standard Beta (Stand Beta) are shown

2. *p<0.05, **p<0,01 3. Number of firms (N)=96

32

5. Discussion

A great deal of attention has been paid to partner diversity within dyadic alliances. For both individual alliances, as the collection of alliances (portfolio alliances) and alliance networks, the phenomenon has extensively been examined on its relation with performance. Authors have identified different types of diversity starting with Parkhe (1991) who found Type І, including the familiar inter-firm differences that are specifically created to exploit, and Type ІІ that refer to differences in partner characteristics. Elaborating on this Jiang et. al. (2010) divided diversity in partner characteristics, functional and governance diversity. This research focuses on partner diversity and its relation with performance for MPAs. Despite the fact that MPAs are

increasingly important collaborative phenomenon in competitive landscape (Estrada Vaquero, 2012) research on the phenomena remains scarce. This research will enter the unexplored field of partner diversity within multi-partner R&D alliances and provide insights on the partner diversity by answering the following research question: “How does partner diversity within a

multi-partner R&D alliance influence a focal firms performance?” Research has been conducted

including 96 firms from 65 alliances from both the high-tech and biomedical industries.

5.1 Partner diversity

33 decrease of performance like different cultures, experiences, the challenges of knowledge appropriation and the need for synergy and integration. Especially in the case of MPAs with its increasing number of partners and complexity the costs associated to managing distant knowledge and information, that derives from different industrial backgrounds, seems to be too high for firms at low levels of diversity to benefit from the alliance. Higher levels of industry diversity on the contrary seem to offer enough opportunities to compensate for these costs and start benefiting from this alliance.

Unlike the findings of a J-shape relation between organizational diversity and performance within portfolio alliances in prior research (Jiang, et al., 2010), this research found no relation between organizational diversity and performance within MPAs. Jiang et al. (2010) examines the phenomena in a collection of dyadic relations (portfolio) and this current research in a MPA setting. The latter setting gives birth to a more generalized social exchange rather than the restricted more interpersonal exchanges (Das & Teng, 2002). Instead of interpersonal interaction, one ‘platform’ for exchange is created to facilitate interaction between all participants. Hence, by exchanging in a more generalized manner the characteristics of the firms, like their structures as one determinant for communicating and coordinating, potentially loses its importance within firms’ interaction. The configuration of the type of alliance might be a reason for the non-significant findings on organizational diversity. Another possible explanation might have to do with the relative low proportion of diversity in organizational structures within this research. Compared to profit driven organizations, much fewer financial data could be found on non-profit organizations like universities or research centers resulting in a less diverse sample. This in turn might have led to not finding significant results for organizational diversity.

34 their cultural variation, enriches the focal form and offers enough opportunities to overcome the coordination and communication challenges. Also noted is that this research did not take into account that the cultural differences between some nations, for instance those from different continents, can be considered much higher than the differences between other countries (e.g. neighboring countries).

Finally it has been examined whether the three dimensions of partner diversity interact with one another. Studies in institutional theory have been researching the phenomena of isomorphism (DiMaggio & Powell, 1983) resulting in firms coming up with the same structures and strategies as other firms within their environment. It is possible that firms from certain nations or industries are more likely to have similar organizational structures. By checking for interaction effect between the dimensions of partner diversity, it has been examined whether the relation between each of the dimensions and performance, is influenced by the degree of diversity firms score on the other dimensions. Even though the descriptive matrix shows a correlation between industry and organizational, and industry and national diversity, the other models show no signs of multicollinearity. When checking for these interaction effects in the model itself it is confirmed that there are no linear interaction effects to be found between the three dimensions of partner diversity within MPAs. Not finding any support for these interaction effects somewhat contradicts the findings of Cui and O’Conner (2012) who found that for portfolio alliances interaction effects do exists between diversity dimensions.

All together the industry, organizational and national diversity dimensions represent partner diversity within a MPA. All three of them show different relations towards performance in a MPA. A one-sided conclusion about partner diversity within MPAs cannot be drawn based on the findings. The findings indicate that partner diversity definitely influences performance. Low levels of partner diversity seemingly don not offer enough opportunities for firms to benefit from their differences as communication and coordination costs increase more drastically compared to the performance. Higher levels of partner diversity insinuate they do offer these opportunities.

5.2 Control variables

35 expenses, year, strategic importance for the focal firm and joint venture were included in the models, but consistently across all models showed no significant relations with performance.

6. Implications and conclusion

In this final section the implications for both managers and scholars are addresses and a concise conclusion is provided.

6.1 Managerial and theoretical implications

The findings in this study support the notion that managers should balance the different dimensions of partner diversity as they influence performance in different ways. They also indicate that managers should not be afraid of cooperating with partners that would be considered relatively diverse from them. As the results of this study show that performance starts to increase after a certain degree of diversity. So for managers to benefit from MPAs, they will have to find partner with a relative high degree of diversity. Furthermore this research shows that there are no clear linear interaction effects between the different dimensions of partner diversity. This means firms can manage the different dimensions independently from one another. By finding some support on the influence of partner diversity on performance this current research shows that the partner diversity phenomena, proven to be present in dyadic relations, also play a role in MPA. Thereby contributing to the still scarcely examined field of MPA performance. More specifically industry diversity shows results, that are in line with prior research on portfolio alliances (Jiang et. al, 2010; Goerzen & Beamish, 2005), implicating that lower levels of diversity do not offer enough opportunities to overcome the rising managerial costs of communication and coordination. The other dimensions, organizational and national diversity, show no significant interaction effects showing, indicating that for MPAs these dimensions do not influence performance. In the case of organizational diversity this is contrasting to the J-shaped relation Jiang et al. (2010) found with performance. Furthermore this study is the first to research interaction effects between the partner diversity dimensions within MPAs and shows there are no linear interaction effects present.

6.2 Limitations and future research

36 obtain an in depth analysis and information about the differences between groups. For future research I therefore suggest to conduct this research with a larger sample size that can be used to divide the database into different groups. That would make the results of the analysis more accurate and could provide more insights into the different levels of diversity and their effect on performance. Finally this could also be useful for checking interaction effects, as my post hoc analysis already revealed that with sufficient sample sizes there might actually be interaction effects for different levels of diversity.

A second limitation is the database itself, which is the source for information. Especially Orbis, which was used to gather financial information on firms, chiefly for profit firms (listed). Information about non-profit firms like universities or research institutes is rarely supplied by the databases. This could have resulted in having a somewhat biased sample, in which the non-for-profit firms are underrepresented. Because of their differences in strategic goals (one aims for efficiency and stakeholder satisfaction and the other aims for scientific novelty and appreciation) these parties will influence performance in different ways. Moreover the increasing number of alliances formations by non-for-profit organizations (Guo & Acar, 2005) makes it important for research to have this group be sufficiently represented within the sample. Future research should make sure they draw financial performance from a database that offers sufficient information on, not only stock exchange listed but also, non-for-profit companies. Another way to achieve a better representation of non-for-profit firms could be by changing the performance indicator into a less financially oriented measure like patenting.

Thirdly, profit margin was a single (variable) measure used for examining performance. A corporate firm’s profit margin however can be influenced by a number of other things apart from those mentioned in this study. The R² shows that the model explains 15% of the variance influencing the dependent variable. For future research it might be interesting to choose different measures or cross-check performance by combining some measures, such as market share or return on equity, as these could have represented a better and more accurate measure for performance.

37 alliances and alliance networks) report alliances with numbers ranging from 2 till 354 dyadic relations (Jiang et al., 2010) whereas, my study for MPAs, the range is minimum 3 up to maximum 5. As the Blau-index results in a figure between 0 and 1 this seems not to be a problem, I observed however, that due to this small number of participants the possible outcomes are limited. This in turn results in a centration around certain scores instead of a nicely scattered range of outcomes. Future research should try to find different ways to measure diversity within MPAs.

6.3 Conclusion

This paper is one of the first to research the impact of partner diversity on firm performance within a MPA context. It provides empirical evidence that phenomenon which is at play for dyadic alliances also play a role within MPA diversity. The results indicate that for low levels of industry diversity, costs associated to managing this diversity (e.g. communication and coordination) outweigh the benefits of being exposed, to novel information and resources deriving from this diversity. At higher levels of industry diversity, the opportunities deriving from novel information and resources outweigh the costs and are positively related to performance. For the other dimensions of partner diversity no strong significant relations were found with performance. Although national diversity also seems to be more beneficial with increasing levels of diversity, as it shows a trend towards a linear relation between national diversity and performance.

Acknowledgement

38

References

Aldrich, J. (1995). Correlations genuine and spurious in Pearson and Yule. Statistical Science, 364-376.

Altenburg, T. (2005). The private sector and development agencies: How to successful alliances.

German Development Institute. Business and MDGs

Baer, M., & Oldham, G. R. (2006). The curvilinear relation between experienced creative time pressure and creativity: moderating effects of openness to experience and support for creativity. Journal of Applied Psychology, 91(4), 963.

Barney, J. (1991). Firm resources and sustained competitive advantage. Journal of

management, 17(1), 99-120.

Cartwright, S., & Cooper, C. L. (1993). The role of culture compatibility in successful organizational marriage. The Academy of Management Executive,7(2), 57-70.

Cui, A. S., & O'Connor, G. (2012). Alliance portfolio resource diversity and firm innovation. Journal of Marketing, 76(4), 24-43.

Commission, E. (2003). The new SME definition. User guide and model declaration. Official journal of the European Union. Enterprise and Industry Publications.

Das, T. K., & Teng, B. S. (1997). Sustaining strategic alliances: options and guidelines. Journal

of General Management, 22, 49-64.

Das, T. K., & Teng, B. S. (1998). Resource and risk management in the strategic alliance making process. Journal of management, 24(1), 21-42.

Das, T. K., & Teng, B. S. (2002). Alliance constellations: A social exchange perspective. Academy of management review, 27(3), 445-456.

Davis, G. F., & Greve, H. R. (1997). Corporate elite networks and governance changes in the 1980s. American Journal of Sociology, 103(1), 1-37.

De Man, A. P., & Duysters, G. (2002). The state of alliance management: Special ASAP report of research into alliance capability: 1-15. Association of Strategic alliance Professionals.

39

Doz, Y. L., & Hamel, G. (1998). Alliance advantage: The art of creating value through

partnering. Harvard Business Press.

Draulans, J., DeMan, A. P., & Volberda, H. W. (2003). Building alliance capability::

Management techniques for superior alliance performance. Long Range Planning, 36(2), 151-166.

Dussauge, P., & Garette, B. (1999). Competing succesfully through strategic alliances. Chichester: John Wiley & Sons, LTD.

Duysters, G., Heimeriks, K. H., Lokshin, B., Meijer, E., & Sabidussi, A. (2012). Do Firms Learn to Manage Alliance Portfolio Diversity? The Diversity‐Performance Relationship and the Moderating Effects of Experience and Capability. European Management Review, 9(3), 139-152.

Estrada Vaquero, I. (2012). Collaboration and value creation in multi-partner alliances: a longtitudinal case-study on the acuisost consortium. Universidad de Valladolid.

Etzkowitz, H. (2003). Research groups as ‘quasi-firms’: the invention of the entrepreneurial university. Research policy, 32(1), 109-121.

Faems, D., De Visser, M., Andries, P., & Van Looy, B. (2010). Technology Alliance Portfolios and Financial Performance: Value‐Enhancing and Cost‐Increasing Effects of Open Innovation*. Journal of Product Innovation Management, 27(6), 785-796.

Fisher, R. A. (1992). Statistical methods for resear ch workers. In Breakthroughs in

Statistics (pp. 66-70). Springer New York.

García-Canal, E., Valdés-Llaneza, A., & Ariño, A. (2003). Effectiveness of dyadic and multi-party joint ventures. Organization Studies, 24(5), 743-770.

Goerzen, A., & Beamish, P. W. (2005). The effect of alliance network diversity on multinational enterprise performance. Strategic Management Journal, 26(4), 333-354.

Gomes-Casseres, B. (2003). Constellation strategy: Managing alliance groups.Ivey Business

Journal, 1-6.

40

Gulati, R. (1999). Network location and learning: The influence of network resources and firm capabilities on alliance formation. Strategic management journal, 20(5), 397-420.

Guo, C., & Acar, M. (2005). Understanding collaboration among nonprofit organizations: Combining resource dependency, institutional, and network perspectives. Nonprofit and

Voluntary Sector Quarterly, 34(3), 340-361.

Hair, J., Anderson, R., Tatham, R., & Black, W. 1998. Multivariate data analysis (5th ed.). Upper Saddle River, NJ: Prentice Hall.

Harrison, J. S., Hitt, M. A., Hoskisson, R. E., & Ireland, R. D. (2001). Resource complementarity in business combinations: Extending the logic to organizational alliances. Journal of

Management, 27(6), 679-690.

Hennart, J. F., & Zeng, M. (2002). Cross-cultural differences and joint venture longevity. Journal of International Business Studies, 699-716.

Hitt, M., Dacin, M., Levitas, E., Arregle, J., & Borza, A. (2000). Partner selection in emerging and develeped market contexts: a resource-based and organizational learning perspective.

Academy of Management Journal, 43(3), 449-467.

Hitt, M. A., Hoskisson, R. E., & Kim, H. (1997). International diversification: Effects on innovation and firm performance in product-diversified firms. Academy of Management

journal, 40(4), 767-798.

Hoffmann, W. H. (2005). How to manage a portfolio of alliances. Long Range Planning, 38(2), 121-143.

Hwang, P., & Burgers, W. P. (1997). The Many Faces of Multi-Firm Alliances: LESSONS FOR MANAGERS. California Management Review, 39(3).

Jiang, R. J., Tao, Q. T., & Santoro, M. D. (2010). Alliance portfolio diversity and firm performance. Strategic Management Journal, 31(10), 1136-1144.

Kogut, B., & Singh, H. (1988). The effect of national culture on the choice of entry mode. Journal of international business studies, 19(3), 411-432.

41

Lane, P. J., & Lubatkin, M. (1998). Relative absorptive capacity and interorganizational learning. Strategic management journal, 19(5), 461-477.

Lavie, D., & Rosenkopf, L. (2006). Balancing exploration and exploitation in alliance formation. Academy of Management Journal, 49(4), 797-818.

Lavie, D., Lechner, C., & Singh, H. (2007). The performance implications of timing of entry and involvement in multipartner alliances. Academy of Management Journal, 50(3), 578-604.

Lavie, D., & Singh, H. (2012). The evolution of alliance portfolios: the case of Unisys. Industrial

and Corporate Change, 21(3), 763-809.

Li, D., Eden, L., Hitt, M. A., & Ireland, R. D. (2008). Friends, acquaintances, or strangers? Partner selection in R&D alliances. Academy of Management Journal, 51(2), 315-334.

Li, D., Eden, L., Hitt, M. A., Ireland, R. D., & Garrett, R. P. (2011). Mays Business School Texas A&M University.

Li, D., & Ferreira, M. P. (2008). Partner selection for international strategic alliances in emerging economies. Scandinavian Journal of Management, 24(4), 308-319.

Lin, Z., Yang, H., & Demirkan, I. (2007). The performance consequences of ambidexterity in strategic alliance formations: Empirical investigation and computational

theorizing. Management Science, 53(10), 1645-1658.

Lindow, C. M. (2013). A Strategic Fit Perspective on Family Firm Performance(pp. 49-93). Springer Fachmedien Wiesbaden.

Lonner, W. J., Berry, J. W., & Hofstede, G. H. (1980). Culture's Consequences: International Differences in Work-Related Values. University of Illinois at Urbana-Champaign's

Academy for Entrepreneurial Leadership Historical Research Reference in Entrepreneurship.

McGill, J. P., & Santoro, M. D. (2009). Alliance portfolios and patent output: the case of biotechnology alliances. Engineering Management, IEEE Transactions on, 56(3), 388-401.

42 Mintzberg, H. (1979). The structuring of organizations: A synthesis of the research. University of

Illinois at Urbana-Champaign's Academy for Entrepreneurial Leadership Historical Research Reference in Entrepreneurship.

Mothe, C., & Quelin, B. V. (2001). Resource creation and partnership in R&D consortia. The

Journal of High Technology Management Research, 12(1), 113-138.

Parkhe, A. (1991). Interfirm diversity, organizational learning, and longevity in global strategic alliances. Journal of international business studies, 579-601.

Piva, E., & Rossi-Lamastra, C. (2013). Systems of indicators to evaluate the performance of university-industry alliances: a review of the literature and directions for future research. Measuring Business Excellence, 17(3), 40-54.

Pothukuchi, V., Damanpour, F., Choi, J., Chen, C. C., & Park, S. H. (2002). National and

organizational culture differences and international joint venture performance. Journal of

International Business Studies, 243-265.

Richard, O. C., Barnett, T., Dwyer, S., & Chadwick, K. (2004). Cultural diversity in

management, firm performance, and the moderating role of entrepreneurial orientation dimensions. Academy of Management Journal, 47(2), 255-266.

Sampson, C.H. (2007). R&D alliances and firm performance: the impact of technological diversity and alliance organization on innovation. Academy of Management Journal, 50

(2), 364–386.

Sarkar, M. B., Echambadi, R., Cavusgil, S. T., & Aulakh, P. S. (2001). The influence of

complementarity, compatibility, and relationship capital on alliance performance. Journal

of the academy of marketing science, 29(4), 358-373.

Sarkar, M. B., Echambadi, R. A. J., & Harrison, J. S. (2001). Alliance entrepreneurship and firm market performance. Strategic management journal,22(6‐7), 701-711.

Shaw, M. E. (1971). Group dynamics: The psychology of small group behavior.

Sivakumar, K., Roy, S., Zhu, J., & Hanvanich, S. (2011). Global innovation generation and financial performance in business-to-business relationships: the case of cross-border alliances in the pharmaceutical industry. Journal of the Academy of Marketing

43 Sørensen, J. B., & Stuart, T. E. (2000). Aging, obsolescence, and organizational

innovation. Administrative science quarterly, 45(1), 81-112.

Thorgren, S., Wincent, J., & Boter, H. (2012). Small firms in multipartner R&D alliances: Gaining benefits by acquiescing. Journal of Engineering and Technology

Management, 29(4), 453-467.

Zahra, S. A., Ireland, R. D., & Hitt, M. A. (2000). International expansion by new venture firms: International diversity, mode of market entry, technological learning, and

performance. Academy of Management journal, 43(5), 925-950.

Zenger, T. R., & Lawrence, B. S. (1989). Organizational demography: The differential effects of age and tenure distributions on technical communication. Academy of Management

44

Appendix 1