Trust and innovation in asymmetric strategic
alliances, what is the effect of partner fit?
Caitlin A.D. Deuling S2702703
Combined Thesis
Msc Strategic Innovation Management Msc Small Business & Entrepreneurship Date of Submission: 3rd of August 2020 Word count: 11881
First Supervisor
Florian Noseleit
Co-assessor
Abstract
Table of Contents
Abstract 1 Table of Contents 2 1. Introduction 3 2. Literature Framework 4 2.1 Trust 4 2.2. Partner fit 53. Conceptual Model & Hypotheses 7
3.1. Signals of Trust 7 3.2. Partner fit 8 3.3. Conceptual Model 10 4. Research Design 10 4.1. Data Collection 10 4.2. Measures 11 4.3. Data Analysis 14 5. Results 15 5.1 Descriptive statistics 15 5.2 Main Results 18 6. Discussion 21 7. Conclusion 22 7.1 Managerial Implications 23
8. Limitations & Directions for Future Research 23
1. Introduction
Since the end of the 20th century there has been a growing body of research on how firms can benefit from collaborations with other firms. Collaborations such as strategic alliances can help firms innovate and respond to massive market changes (Doz & Hamel, 1998). Strategic alliances are enduring, yet temporary, interfirm exchanges that member firms join to jointly accomplish predetermined goals (Rahman, 2006). Being able to make the most of a firm's strategic alliances continues to remain an important skill, as this allows firms to continue to learn and develop new capabilities and firm specific advantages. Even now as markets continue to increase in its turmoil, there remains discrepancy on what factors influence the success of an alliance. Furthermore, research also indicates that strategic alliances are often characterised by high failure rates and high instability (Das & Teng, 2000) and often involve high levels of risk and uncertainty. Due to high risks and uncertainty, firms are often careful with who they can trust within collaborations, and often put many protective measures in place, to substitute for leaving themselves vulnerable to competitors. Yet, trust has many advantages as it facilitates cooperation, lowers agency and transaction costs, benefits the promotion of smooth and efficient market exchanges, and can help a firm adapt to complexity and change (Molina-Morales et al., 2011). Mutual trust provides the basis of exchanging tacit know-how and learning, crucial in alliance performance, especially when that performance is focused on the innovation (Molina-Morales et al., 2011). At the same time trust curbs opportunistic behaviour, allowing firms to pursue common goals (Kale et al., 2000). Trust is especially important for collaborations between partners that are asymmetric in size, managerial resources, finances, technical resources, organizational culture and tolerance for losses and risk (Blomqvist, 2002). But as more and more firms engage in alliances with partners that are distinctly different from themselves, knowing how to cultivate trust between partners depending on different structural characteristics could be of the upmost importance.
As whether or not partners can be perceived as trustworthy has not been explored thoroughly, this paper will contribute to the innovation and alliance literature by exploring whether or not signals of trust by partners can actually positively benefit the innovation performance of the firm. Furthermore, interaction effects will be explored specifically focusing on how partner fit can help create an environment in which trust can become more or less effective in alliances. This research could contribute to management by exploring how signals of trust can be interpreted, and how firms and alliances should invest in trust between partners, to make the most of all alliances in which innovation is crucial.
2. Literature Framework
2.1 Trust
Within the innovation and alliance literature the many benefits of trust have been described and empirically tested. When high levels of trust are maintained, information, knowledge and ideas can be transferred more easily, enhancing the innovative capabilities of the firm (Wang et al., 2011). Similarly, Kale et al. (2000) found that trust deters opportunistic behaviour of alliance partners, but also prevents leakage of critical know-how to alliance partners. Additionally, trust within a partnership has the ability to reduce transaction costs, and prevent opportunistic behaviour (Laaksonen et al., 2009). Through trust firms can spend more time focusing on how each firm can absorb and utilize the knowledge they acquire through the partnership (Lane et al., 2001), use resources and time and spend it on collaborative innovation activities (Zaheer et al., 1998). These same resources and energy would otherwise be spent governing the partnership and negotiating contracts (De Jong & Woolthuis, 2008). In the end, firms strive for value creation instead of striving to make sure the alliance does not fail. Trust is essential in this process of alliance formation and management, working towards greater innovation.
According to Ganesan (1994), trust can operationally be defined along two dimensions; (1) benevolence (2) competence. Benevolence based trust refers to the motives and intentions of the alliance partner and refers to the goodwill of a partner and the avoidance of opportunism. When a firm believes the partner is exploiting them for their benefit, the alliance will start to fall apart or stronger control mechanisms will be put in place (Judge & Dooley, 2006). Competence based trust refers to the consistent display of traits such as credibility and expertise. Competence based trust reflects that willingness of the partner to trust on each other’s expertise, capabilities and judgements. Both these forms of trust are needed to continue to support the alliance, it provides stability as the partner trusts the relationship will continue to do well in the future (Ganesan, 1994).
Wang, Yeung & Zhang (2011) found a linear relationship between trust and innovation performance. They found that due to the fact that there are large information asymmetries when innovating, making trust more important. In contrast, Molina-Morales et al. (2011) found an inverted u-shape for the relationship between trust and innovation and describes there are optimal levels of trust for innovation performance. This level of diminishing returns is determined by trusting too much, overinvesting in trust, or investing in trusting relationships that have no, or little value. Gaining information on where to invest a firm's trust and how much might be crucial for the performance of the alliance, and the firm.
In order to deal with the many information asymmetries the focal firm is faced with, it helps to understand what drives trust, and what signals trust to potential partners. Mutual trust has many drivers. These drivers reflect both the trustor and the trustee within the relationship. Examples of such are the ability, or propensity of the focal firm to trust a partner, this can be enforced through trusting cultures and a firm's past experience with trust (Bierly & Gallagher, 2007). Additionally, partners, the trustee, gives signals of trust to the trustor, these can be through previous experience with that partner, or a partner's reputation and past alliance experience (Bierly & Gallagher, 2007).
Within this paper I focus specifically on how signals of trust can be beneficial for innovation performance. By looking at the effect of partner specific experience, and partner alliance experience, I will determine how signals of benevolence and competence-based trust by the partner are affecting the focal firm.
2.2. Partner fit
Before an alliance, firms often spend considerable efforts in evaluating and determining risks and uncertainties of the partnership (Cullen et al., 2000). Especially risks and uncertainties associated with collaborations with different partners and the dynamic and interrelationship of elements of fit is crucial to knowledge creation, creating synergies and developing new innovation (Nielsen, 2005). Li et al. (2008) even argues that partner selection is an alternative mechanism by which firms in an alliance can control the threat of knowledge leakage and retain their core proprietary knowledge. Partner fit is one of the first things considered when choosing a partner for an alliance as this can determine who possesses the necessary resources as well as with whom strategic and economic incentives can be achieved (Sarkar et al., 2001). Partner fit can be explained by the fact that firms both need different resources and capabilities yet need to share a similar social structure in order to enhance performance and create value through an alliance (Sarkar et al., 2001). Effective inter-organisational alliances are associated with the selection of appropriate partners, as it is critical to find those partners who possess the necessary resources and those with whom economic and strategic incentives can be aligned (Sarkar et al., 2001).
Partner fit describes the relationship between a firm and its environment or among its strategy, structure and processes (Nielsen & Gudergan, 2012). Partner fit can encompass several different structural and behavioural characteristics that would determine the fit between partners. Within literature, the fit of compatibility and complementarity has been identified as most critical to alliance performance (Shakeri & Radjer, 2017; Harrigan, 1988; Tucchi, 1996). The following sections will discuss both elements further.
2.2.1. Partner Compatibility
Partner compatibility refers to the compromise of the differences between alliance partners. Partner compatibility allows for optimal communication, collaboration and interaction between the partners (Shakeri & Radjer, 2017). Partner compatibility can refer to broad historical and, philosophical elements, strategic grounds, values and principles, expectations for the future (Kanter, 1994) as well as cultural and organisational elements (Shamdasani & Seth, 1995).
When partnering firms can be considered more compatible, a firm, and its top management team will have more trust and confidence in another firm that is or feels more familiar, or less threatened by the way that other firm conducts business (Bierly & Gallagher, 2007). This will help firms to actively create new capabilities, competitive advantages and new routines for shared knowledge creation and allow firms to develop new innovative solutions. Madhok & Tallman (1998) mentions that partner compatibility influences the extent to which partners are able to realize the synergistic potential of an alliance, where the actualization of this collaborative potential is created through the dynamic process of integration and interaction (Koza & Lewin, 1998).
obtaining synergies and cultivating rapid learning and growth (Hoffmann & Schlosser, 2001). Furthermore, partnerships with larger firms also allows SMEs to create a buffer against their liability of smallness and enhance their chances of survival (Baum et al., 2000; Stuart, 2000).
Smaller firms can benefit greatly from the collaboration with a large established firm. A study conducted by Koh & Venkatraman (1991) suggests even that the returns on average the smaller partner gains substantially more than that of the larger partner. Yet, SMEs often face risks of appropriation and risks of leaking their core proprietary assets to their partner, and eroding their competitive position (Oxley, 1997). Partner firms strive to internalize valuable knowledge and gain expertise from one another (Hamel, 1991). This internalisation of knowledge has also been referred to as the learning race. Knowledge appropriation by a partner firm can have a catastrophic impact on the competitive position of the focal firm.
On the contrary the larger organisations are often faced with information asymmetries with respect to the quality of smaller firms (Shane et al., 2006). Large organisations are often characterised as bureaucratic and suffering from inertia (Rumelt, 1995). In order to survive in the current turbulent market, they need SMEs to develop new innovative solutions and create new competitive advantages. These large organisations invest significant large sums of money and resources in the relationship. Large organisations therefore need to trust the competence of the smaller firms to be able to deliver on the set goals of the alliance.
2.2.2. Partner Complementarity
Partner complementarity refers to the lack of similarity or overlap between the core business or capabilities between alliance partners (Mowery et al., 1996). Complementarity ensures that both partners bring different but valuable capabilities to the relationship and creates potential for the other firm to learn from its partner (Kale et al., 2000). Research has confirmed that resource complementarity provides greater potential for firms to create synergies from alliances, leading to higher levels of firm performance (Harrison et al., 2001). It provides opportunities for enhanced learning and development of new capabilities. But despite a keen desire to learn, partner firms often fail to achieve effective learning and knowledge transfer in alliances (Lam, 1997).
3. Conceptual Model & Hypotheses
I have developed several hypotheses that discuss the relationship between trust and innovation performance. Using signaling theory, theories of familiarity, alliance learning and theories of absorptive capacity as underlying reasoning, each hypothesis is explained below.
3.1. Signals of Trust
3.1.1. Partner Specific ExperiencePrior research has used partner specific experience between both partners as a proxy for trust (Gulati, 1995).
According to Gulati (1995), familiarity between partners will help build trust. Additionally, in a study by Li et al. (2008) it was found that firms are more likely to work with partners they know and worked with before when pursuing risky and radical innovation goals. When firms already have had prior alliance(s) with a partner firm, it is likely there is mutual trust between both partners. The more a firm works with a prior partner, the more likely the familiarity increases between both firms, and thus the more trust will be developed. Furthermore, they are more likely to trust the partner firm compared to other firms with whom they have no prior experience (Ring & Van de Ven, 1989).
Furthermore, when deciding to work with a partner a firm has worked with before shows commitment. Both partners are invested in the relationship and are committed to growing the relationship even further. Commitment and trust are often mentioned as the crucial dimensions of relationship capital. These two aspects are the essential threads in the social fabric of the alliance relationship (Cullen et al. 2000) and are mutually reinforcing in alliances. Commitment specifically concerns the intention of the partner to continue in a relationship. Commitment is driven by the expectations of rewards in some shape or form, that can be obtained by the firm when forming and managing this alliance, while also wanting to go beyond the contractual obligations to achieve the set goals of the alliance. The commitment by both firms made when choosing to partner again after previous experiences, will create pressure to conform to expectations of the alliance, and are thus less likely to act opportunistically (Molina-Morales et al., 2011).
I believe that the partner specific experience will have a positive direct effect on innovation performance based on the premise of familiarity and commitment.
H1: Partner specific experience has a direct positive effect on innovation performance.
3.1.2. Partner Alliance Experience
Prior ties to partners might not be the only way in which firms will be able to reduce information asymmetries and determine how much a partner can be trusted. Firms might also look at partners of partners as viable collaborators (Baum et al., 2005). Knowing who a potential partner has partnered with might tell the focal firm a lot about the partner's past behaviour. It gives signals of benevolence and competence to the focal firm, as past experience can engender trust among partners (Granovetter, 1985, 1992; Marsden, 1981). Many alliances could also indicate that the firm is considered legitimate and a good partner. When a firm is considered legitimate it is perceived as an organization that is held in high esteem or regard by its key constituents on the basis of its past actions and future appeal (Fombrun & Shanley 1990).
When the partner has greater alliance experience, they will have learned from their past experience and developed some form of collaborative processes, while also signaling to potential partners that the partner can be considered as a desired partner. This makes me believe that partner alliance experience will have a positive effect on the innovation performance of the focal firm.
H2: Partner Alliance Experience has a direct positive effect on innovation performance.
3.2. Partner fit
3.2.1. Partner Compatibility
In order to effectively build new innovations, tacit information needs to be transferred from one firm to another. In a world in which agents of the firms do not share exactly the same experience, and the same understanding of the world around them, a collective knowledge base is required for coordination (Cremer, 1990). Tacit understanding is facilitated by similarities in operational capabilities and trust. Two firms which are similar in size, and thus operationally similar are more likely to understand one another as their operations will feel more familiar to their own. According to Bierly & Gallagher (2007) this familiarity will help promote trust.
Incompatibility can represent itself in different ways. When comparing large and small organisations, a way to determine the compatibility is looking at firm size. When compatibility is positive it means that the focal firm is larger than its partner firm. When compatibility is negative the focal firm is smaller than its counterpart. Firms can be considered most compatible when the compatibility value is approaching zero, whereas a high absolute number of compatibility actually indicates a very low compatibility.
When two firms are not compatible, they need trust to bridge the gap and provide mutual understanding for the exchange of tacit information. Although high partner compatibility will allow firms to better exchange between firms, many organisations form alliances with firms very different in size. According to Brouthers et al. (1995) strategic alliances work better when there is symmetry in size, and Geringer (1988) actually found that asymmetry in partner's size has negative effects on the stability of joint ventures. The authors argue that a collaboration between small and large organisations suffers from an imbalance and mismatch in strategic missions, corporate culture and level of bureaucracy. When firms are more compatible, they will be facing similar risks and uncertainty. This will make the partner's behaviour more predictable as both are striving towards the same goal.
The familiarity of actually knowing the partner will be increased by having similar operational routines, as both have worked within an alliance before. They will have developed their own routines and path dependencies in the alliance which will be facilitated in similarities in their own organisational routines. Partner specific experience will benefit from partner compatibility such that these firms will be able to deepen the learning from each other's knowledge base, as well have already developed deep connections, enforcing trust signals between one another.
H3a: Partner compatibility has a positive interaction effect on the relationship between partner specific experience and innovation performance.
within a study by Lane & Lubatkin (1998) that a greater similarity in organisational properties and knowledge base between pharmaceutical and biotech companies enhanced alliance success. Furthermore, according to the alliance learning literature, when firms are more similar, they will have better ability to learn from each other as each other's knowledge base will share some common ground to create understanding.
Routines and superior management capabilities that result from experience can constitute intangible resources that are likely to improve performance in future alliances (Barney, 1991). These inferences from past experiences, might be encoded in routines (Nelson & Winter, 1982), partner compatibility can stimulate the activation of these routines that the encoding thereof, and create a tacit understanding of the partner's knowledge base. Additionally, signals of trust can be better interpreted when there are similar routines present in both firms, creating an understanding of the accumulated capabilities of the partner. Firms can therefore benefit more from the alliance experience of the partner when there is higher compatibility.
H3b: Partner compatibility has a strengthening effect on the relationship between partner alliance experience and innovation performance.
3.2.2. Partner Complementarity
Partner complementarity makes sure that both partner firms bring differences, skills, knowledge and expertise to the alliance that will allow firms to learn from another (Kale et al., 2000). Morina-Morales et al. (2011) argues that firms run the risk of investing their trust in alliances that yield no value for the firm.
When knowledge is not sufficiently different from their own knowledge base, there is little risk, and little uncertainty. Additionally, there are little learning opportunities for firms, and therefore will have little influence on how firms are innovating. When complementarity is high however, there is much to gain in an alliance, but firms might not always have the skills to assimilate that knowledge. This uncertainty and risk will increase the opportunities for opportunism and mismatch in competence-based expectations, yet this also allows firms to create an environment where trust could be more effective, by the simple notion that there is more to gain. In a study by Wang et al. (2011) environmental uncertainty was found to increase the effectiveness of trust-based governance of alliances, where it did not increase the effectiveness of contracts. I hypothesize that partner complementarity will moderate the relationship between signals of trust and innovation performance in a similar way.
With regards to partner specific experience, a study by Li et al. (2008) illustrates that when there are high risks and radical innovation goals are set for the alliance, firms prefer to work with partners they have collaborated with before. These high risk and uncertainty situations call for higher trust. When goals for the alliance are less ambitious, firms are more likely to seek out a partner which they have not worked with before instead of firms they had casual exchanges with. These firms would be able to provide new complementarities, and will allow firms to learn, regardless if that means that the alliance is riskier. Competence based trust, is in the case of low ambition more important, whereas for ambitious goals there is a need for both benevolence trust and competence-based trust in the alliance. When firms are choosing to collaborate with partners they have worked with before, and have high complementarity between the two, it will allow the two firms to deepen the relationship and allow firms to learn more tacit information. Partner complementarity will strengthen this signal of trust, improving the innovation performance of the firm.
Theories of absorptive capacity suggests that firms need capabilities to properly value, assimilate, and use that new information to commercial ends, in order to effectively incorporate new knowledge to create innovation and other possible new competitive advantages (Cohen & Levinthal, 1990). This knowledge can be accumulated over different alliances, as firms can create superior alliance management capabilities. Baum, Calabrese & Silverman (2000) found that biotech start-ups were more innovative when they had many alliances, suggesting that alliances contribute to a firm’s knowledge base. Partner complementarity will enhance the positive effect of partner alliance experience, as firms will be able to put their experience of managing and absorbing new knowledge into practice within the alliance. Additionally, signals of trust will be strengthened as more complementary partners will signal strong learning capabilities and a strong and diverse knowledge base to partners.
H4b: Partner Complementarity has a positive interaction effect on the relationship between partner alliance experience and innovation performance.
3.3. Conceptual Model
Figure 1 Conceptual Model
4. Research Design
4.1. Data Collection
For the purpose of this research data has been collected from the pharmaceutical and biotechnology industry. These industries are known for using many alliances between both large and small companies,
Innovation Performance
Partner
Compatibility
Partner
Complementarity
with each firm having specific areas of expertise. I have collected a sample of the top 100 pharmaceutical and biotechnology firms based on their revenue in 2018 (Statista, 2020). This list consists of both large and small firms from different countries. By looking at the press releases of each individual firm, alliances between several firms are revealed. Within this research there is a specific focus on alliances between 2010 and 2015, as I would like to measure innovation performance 4 years after the announcement of this alliance, as indicated by Sampson (2007). This sample of firms can be classified as a convenience sample and consists of dyadic relationships between two firms which indicate that they will collaborate on some level during this partnership. To collect the different alliances, I looked at press releases of the different firms indicated by Statista. Within the record I detailed the date of the announcement of the alliance, as well as the company with whom the alliance was initiated. All alliances were recorded from both sides, where both firms were registered as the focal firm and the partner firm. Furthermore, in order to test my hypothesis, within each alliance one of the two firms are indicated as the focal firm, from which we will measure the innovation performance.
Within this research I am only looking at alliances between competitors, also known as horizontal alliances. Alliances with research institutions, universities or non-profit organisations were left outside the scope of this research. Within this research I am specifically interested in the effects of trust and partner fit in situations where competitors choose or are forced to collaborate in order to further their competitive position.
After alliances have been recorded, data to support the analysis were collected from the ORBIS database, and press releases on company websites.
4.1.1 Sample
I started with a sample of alliances which consisted of 361 unique firms and 942 registered alliances. But as not all data were available for all of these firms, I ended up with a final sample of 60 unique focal firms and 229 registered alliances. The sample is characterised by alliances whose focal firm headquarter in several different countries. Additionally, these alliances have firms with different sizes. From the 229 alliances, 9 focal firms can be considered a small or medium sized company according to the European Union (European Commission, 2020).
4.2. Measures
4.2.1. Dependent Variable
The dependent variable, innovation performance will be measured by the as a count variable, by the number
of patents given out to the focal firm after the announcement of the alliance. Patents are strongly correlated with new product development (Comanor & Scherer, 1969), counts of inventions based on literature (Basberg, 1982) and non-patentable innovations (Patel & Pavitt, 1997).
Patents are reliable indicators of innovation performance (Sampson, 2007). They can be considered better indicators of innovation performance as for example R&D spending, which measures the input into innovation, and not the output (Comanor & Scherer, 1969; Griliches, 1990), which is reflected by a patent count. The patent data will be collected from the ORBIS database, which can provide patent counts between 2010 and 2019.
4.2.2. Independent Variables
As independent variables I measure two different signals of trust. Partner Specific Experience, and Partner
alliance experience.
In order to measure partner specific experience, I looked at Gulati (1995) for measuring the partner specific experience. I will look at the number of prior alliances with the specific partner between 2010 until the moment of the announcement of the alliance. The more two partners work with one another within a period, the stronger the partner specific experience. When partners have never worked with each other before, the variable will register as zero. For every new alliance between the two partners within the time period, will increase the number of partner specific experience.
Partner alliance experience indicates the alliance experience of the alliance partner will be measured by
accumulation of the number of alliances, between 2010 until 2016. This measure used by Kalaignanam et al. (2007) does not distinguish between successful and unsuccessful alliances, as firms can learn from both, and as the length of each alliance is taken into account, the effect of both represented. As I would like to capture possible signals of trust to the market by the partner firm, the count of firms does not fully encapsulate what signals are important in measuring alliance experience. Rothermel & Deeds (2006), measure alliance experience by the aggregated number of years in which the firm is involved in alliances, as firm-level alliance experience, such as forming, managing and exiting alliances, is accumulated over time by learning by doing, and signals a firm as an desirable and trust-worthy partner. An average alliance is formed for 5 years. When an alliance is ended prematurely this measure will indicate this as such, by decreasing the number of years. When an alliance is mentioned to last longer than 5 years within the announcement, this will also be reflected. This accumulation does not include the alliances that the focal firm has had with the partner firm within the time frame. This is to avoid collinearity between partner alliance experience and partner specific experience. Additionally, this measure does not include the experience accumulated within the focal alliance.
4.2.3. Moderating Variables
In order to measure partner compatibility, I will look at the difference in operational compatibility between both alliance partners, measured by the difference in the number of employees. A smaller firm will operate differently compared to larger organisations as both have different resources (e.g. human resources, financial resources, etc.) at their disposal. This will be reflected by measuring firm size. According to the European Union firm size can be measured by number of employees (European Commission, 2020).
!"#$%&# ()*+"$,-,.,$/ = −1 ∗ 456 (8)(". 9,#* :,;& − !"#$%&# 9,#* :,;&)
When subtracting the partner firm size from the focal firm size, the number could be negative, when the focal firm is smaller compared to the partner firm. By taking the negative absolute value for partner compatibility I can clearly indicate when a firm is more or less compatible. The higher the number, the closer this value is to zero, the more compatible both firms will be. The larger the size asymmetry between both firms, the smaller the number will be.
4.2.4. Control Variables
Alongside the variables determined in the conceptual model, we will also measure the following control variables.
Firm age has been indicated to affect the results of this study (Baum et al., 2000), and alters the
organisational context in which the innovation is created (Lin et al., 2012). Therefore, firm age is included as a control variable in this research. Firm age will be measured in the number of years the firm has existed at the moment of the announcement of the alliance. Data to determine firm age will be collected from Orbis and will encapsulate the year the firm was incorporated until the moment of the alliance announcement. Second, I will also collect data on the size of the firm's R&D Investment. Investments in R&D will allow firms to acquire, build, assimilate and apply it to commercial ends. As investment in R&D will not only be specific to the alliance but also affect individual organisational learning as based on path-based accumulation of knowledge, overall R&D investment will have an impact on the innovative performance of the firm. This data will be collected from the ORBIS database and will encapsulate the average of percentage of the R&D expenses as part of the operating income between 2010 and 2019. I have taken the average between this time period as investments in R&D in alliances do not happen on a single occasion but instead require multiple investments over a period of time.
Equity based alliances, equity is another governance mechanism used to control the partners willingness to
cooperate and deter opportunistic actions by the both firms as it allows interests of both firms to be aligned (Kale et al., 2000). Additionally, markets may perceive equity exchanges between small and large organisations as a start of an acquisition process. Equity based alliances are measured as a simple dummy variable where 1 stands for an equity exchange which was mentioned in the press release of the alliance. When no equity exchange was mentioned, a zero was denoted for this variable.
Fourth, the control variables that will be included is a dummy variable indicating whether the collaboration in question was registered as a joint venture. This would drastically change the structure of the alliance, as it will be registered as a separate entity to both firms. This will impact innovation performance. When the alliance was registered as a joint venture, the variable will indicate a 1, when no joint venture was registered it means the variable will denote a 0.
As the fifth control variable is based on a firm’s own alliance experience. The firm’s alliance experience will be measured as by Rothearmal & Deeds (2006) in the same time period as the measure of partner alliance experience of and will accumulate the number of years of each separate alliance between 2010 and 2016. An alliance is formed for an average of 5 years. When the duration of the alliance was otherwise indicated in the press release the number will be edited. This could be the case when alliances are ended prematurely, or alliances are announced for longer periods such as a period of 10 years. Additionally, the focal alliances and alliances between the partner firm and the focal firm are removed from this measure. This is to avoid collinearity with partner specific experience and partner alliance experience.
Finally, both partner complementarity & partner compatibility have been mentioned in other studies (Kale et al, 2000) to have a direct effect on innovation performance, and the performance of an alliance. To account for this effect, and not solely look at the effect of the interaction between trust and partner fit. We will include both measures as a control variable in the model.
Studies such as Kalaignaman et al. (2007) used fixed effects within their models to account for unregistered effects that could have an effect on the dependent variable. The authors within their study adjusted for the alliance formation year to account for year specific factors and firm, which could be interesting for my study as well. By using control variables I have tried to account for as many factors that could influence the model but still some elements could be missing. First when looking at alliance formation year, there could be certain events in specific years that influence the number of patents registered. Yet, after preliminary testing, I found that the year the alliance was formed did not increase the explaining power of my model, nor were the dummy variables created found significant. Which is why I have chosen to leave this fixed effect out.
Additionally, there are no clear reasons why firm fixed effects should be added, other than the ones that have been accounted for in the other variables such as partner compatibility, alliance experience and country of origin. There is no reason why firm A would be able to perform better, simply because of being firm A, there are underlying factors that determine this difference, which is something that this study is trying to uncover. For that reason, I have decided against adding any fixed effects in my models as this will not help coming to conclusions with the study.
4.3. Data Analysis
In order to test my hypothesis, I will use a form of analysis that will be able to deal with a discrete count variable. The test I will perform will be a form of a Poisson regression analysis. By the use of Poisson distribution, I can find the probability of a certain event happening within a period of time. In this case the regression analysis will measure the probability that higher trust influences the innovation performance of the focal firm in an alliance. A Poisson regression analysis has several key advantages as it is appropriate for the use of integer data, as well as being able to aggregate data over longer periods of time. The Poisson distribution method is appropriate as this study uses patent data from a 4-year period, which account for positive integer values in a time-series panel (Katila, 2000).
The test that I performed could either have been a Poisson regression or the negative binomial regression analysis. In order to determine which is most appropriate I determined the overdispersion of the dependent variable, which I refer to in the results section. In the case of overdispersion, when the conditional variance is greater than the conditional mean, the negative binomial regression will be more appropriate. That is because the negative binomial regression takes into account the overdispersion as an additional parameter. The Poisson regression analysis assumes that the mean is equal to the variance and will therefore keep it constant.
5. Results
5.1 Descriptive statistics
In table 1, below you can find the descriptive statistics for all variables. Here one can find the number of observations, the mean, the standard deviation and the minimum and maximum are presented. Alongside the descriptive statistics I also performed a Pairwise Correlation test where I included all variables including control and moderator variables. These results can be found in table 2.
Table 1 Descriptive statistics
Variables N Mean Standard
deviation Minimum Maximum
Innovation Performance
(1) Innovation Performance 229 5343.402 8413.263 3 71237
Trust
(2) Partner Specific Experience 229 0.245 0.615 0 3
(3) Partner Alliance Experience 229 60.943 59.391 0 200
Partner Fit (4) Partner Compatibility 229 -43948.85 48532.06 -422480 -314 (5) Partner Complementarity 229 0.489 0.501 0 1 Control Variables (6) Firm Age 229 39.445 33.284 1 151 (7) R&D investment 229 23.212 17.580 0.0305 88.135 (8) Equity Exchanged 229 0.416 0.493 0 1 (9) Joint Venture 229 0.192 0.394 0 1 (10) Alliance Experience 229 92.576 74.455 0 285
Table 2 Correlation Matrix
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)
(1) Innovation Performance 1.000
(2) Partner Specific Experience -0.038 1.000
(3) Partner Alliance Experience 0.034 0.261*** 1.000
(4) Partner Compatibility -0.656*** -0.060 0.016 1.000 (5) Partner Complementarity 0.082 -0.070 -0.130* -0.064 1.000 (6) Firm Age 0.485*** -0.003 -0.100 -0.169* -0.028 1.000 (7) R&D Investment -0.202** -0.011 0.213** 0.079 0.068 -0.238*** 1.000 (8) Equity Exchanged 0.029 -0.045 -0.112 0.132* 0.036 -0.098 -0.066 1.000 (9) Joint Venture 0.046 -0.052 -0.067 0.125 -0.003 -0.023 -0.074 -0.015 1.000 (10) Alliance Experience 0.211** 0.136* -0.166 * 0.285*** -0.115 0.073 -0.149* 0.004 -0.011 1.000
(11) Same country of origin -0.131* 0.220*** 0.077 -0.129 -0.071 0.052 0.122 0.007 -0.089 0.094 1.000
My variable for innovation performance is an integer count variable that ranges between 3 and 71237, and has a mean of 5343.402, while the standard deviation is 8413.263. This shows that this data might be over dispersed. To take a closer look at the data I also computed the variance of this variable along with the skewness and kurtosis level. The variance for the innovation performance is 7.08e+7. This indicates that the conditional variance is larger than the conditional mean, indicating overdispersion. This means that I will use a negative binomial regression as this test is similar to a Poisson regression but additionally accounts for overdispersion of the data. Furthermore, the correlation matrix in table 2 shows that innovation performance is positively correlated with firm age and alliance experience, and negatively correlated with partner compatibility, R&D investment, and same country of origin. Especially partner compatibility and firm age show high correlations, this would suggest that a firm is more likely to have a high number of patents when the firm is older. Furthermore, the correlation between innovation performance and partner compatibility shows that an increase in partner compatibility is correlated with an increase in innovation performance.
Partner Specific Experience is a positive count variable where the highest number of partner specific experience is 3, only 4 alliances were registered with a 3 for partner specific experience, 10 alliances obtained a 2 for this variable, and 24 alliances had an alliance with the same partner in the predetermined time period before, 191 alliances had no prior experience with their partner. When looking at table 2, the correlation matrix showed that partner alliance experience and partner specific experience are correlated. This can be explained due to the fact that more experience with alliances, can coincide with experience with the same partner. Furthermore, partner specific experience is positively correlated with alliance experience, and same country of origin.
Partner Alliance Experience is also a positive count variable where the lowest number of alliance years was 0, and the highest was registered at 200. When within an alliance the partner had announced no further alliances except for the focal alliance, the partner alliance experience was registered as zero. When press releases registered additional alliances this number was increased by the number of years for each alliance. The average firm registered a partner alliance experience of 60.9 years. Furthermore, when looking at table 2, one can see that partner alliance experience is negatively correlated with partner complementarity and alliance experience, while being positively correlated with R&D investment.
Partner compatibility is a continuous negative variable which has 229 observations and has an average of -43948.85. The minimum is -442480, where the maximum is -314. This variable is the negative absolute value of the difference between the number of employees between both firms. The closer this value gets to zero the more compatible both firms are. When looking at the correlation matrix, it showed that when partner compatibility increases, alliance experience also increases.
Finally, the last dummy variable indicates both partner firms have their HQ in the same country. Of all alliances, 70 indicated they had the same country of origin and 159 had a different country of origins.
5.2 Main Results
5.2.1 Negative Binomial Regression Analysis
I started my analysis by regressing my two independent variables on my dependent variable, innovation performance. This model shows no significant results. As a negative binomial regression is a log model, the R2 cannot be interpreted in the same way, instead I will look at the log likelihood of the model and the Akaike information criteria (AIC) of each individual model. The Akaike information criteria (AIC) is an estimator of out-of-sample prediction errors. The lower the AIC, the better one can predict/estimate future events. This first model had a log likelihood of -2185.95 and an AIC of 4379.900. For my second model I included my control variables alongside my independent variables. Model 2 showed a log likelihood of -2111.220 and an AIC of 4246.441. Finally, the third model includes the interaction effects between the independent and moderator variables alongside my control variables. This model showed a log likelihood of -2106.351 and an AIC 4244.701. These results would indicate that model (3) is statistically the best model. Several of the variables and interaction effects showed significant results within both model (2) and model (3).
Looking at my first hypothesis, I hypothesized the following: H1: Partner specific experience will have a positive direct effect on innovation performance. In my test results I would have expected a positive and
significant coefficient, yet as P > |z| = 0.535, which falls outside my confidence interval of 95%. I cannot with certainty say that the coefficient is different from 0, which is why I reject hypothesis H1.
My second hypothesis: H2: Partner Alliance Experience has a direct positive effect on innovation
performance, indicates an expected positive result for the coefficient of partner alliance experience. The
results indicated that with P > |z| = 0.002, that that partner alliance experience has a positive effect on innovation performance. When there is one unit increase in partner alliance experience, the log value of innovation performance will be increased by 0.002. I can confirm hypothesis H2.
Apart from looking at the direct effects of trust on innovation performance, I also specifically look at the interaction effect between trust and partner fit by testing 4 hypotheses. The hypothesis 3a considers partner compatibility as a moderating variable and goes as follows: H3a: Partner compatibility has a diminishing effect on the relationship between partner specific experience and innovation performance. The expected
results should show a negative coefficient, yet the results from the analysis indicated that the results are insignificant. The results fell outside my confidence interval of 95% where P > |z| = 0.342 and had a coefficient of 0.000. Hypothesis 3a is rejected.
The second hypothesis testing the effect of partner compatibility is hypothesis 3b, which looks at the interaction effect between partner compatibility and partner alliance experience. The hypothesis is as follows. H3b: Partner compatibility has a diminishing effect on the relationship between partner alliance
experience and innovation performance. The test showed that there are interaction effects between partner
Table 3 Negative Binomial Regression Analysis Variables Model (1)
Model(2) Model (3) Expected Sign
Partner Specific Experience -0.168
(0.223) -0.027 (0.808) 0.116 (0.535) + (H1)
Partner Alliance Experience 0.001
(0.292) 0.002 (0.115) 0.006** (0.002) + (H2)
Partner Specific Experience x Partner Compatibility
0.000 (0.342)
+ (H3a)
Partner Specific Experience x Partner Complementarity
0.038 (0.875)
+ (H4a)
Partner Alliance Experience x
Partner Compatibility 0.000* (0.040) + (H3b)
Partner Alliance Experience x
Partner Complementarity -0.002 (0.450) + (H4b) Firm Age 0.011*** (0.000) 0.010*** (0.000) R&D Investment -0.018*** (0.000) -0.018*** (0.000) Equity Exchanged 0.458 (0.397) 0.151 (0.783) Joint Venture 0.456 (0.329) 0.226 (0.626) Alliance Experience 0.003** (0.002) 0.003** (0.002)
Same country of origin -0.138
(0.321) -0.071 (0.605) Partner Compatibility 0.000*** (0.000) -0.000*** (0.000) Partner Complementarity 0.119 (0.347) 0.282 (0.110) Intercept 8.529*** (0.000) 7.404*** (0.000) 7.111*** (0.000) N 229 229 229 Pseudo R2 0.0004 0.0346 0.0368 AIC 4379.900 4246.441 4244.701
Log Likelihood (null) -2168.86 -2186.86 -2186.86
The final two hypotheses test the interaction effect of partner complementarity on the two proxies for trust, partner specific experience and partner alliance experience. The first hypothesis goes as follows: H4a: Partner Complementarity will strengthen the relationship between partner specific experience and innovation performance. The results show that there is no significant interaction effect between partner
specific experience and partner complementarity. The results showed that P > |z| = 0.875 falls outside the 90% confidence interval and thus I cannot say that the coefficient 0.038 is or is not different from zero. I reject hypothesis 4b.
The final hypothesis that I tested is the hypothesis that describes the interaction effect between partner complementarity and partner alliance experience. The hypothesis goes as follows: H4b: Partner
Complementarity will strengthen the relationship between partner alliance experience and innovation performance. The expected result would show a positive coefficient. The results from the test show however
that that I cannot infer any conclusion from the test as P > |z| = 0.918, and it falls outside my confidence interval of 90%, I cannot interpret these results, showing a negative coefficient. I reject hypothesis 4b.
The only control variables that showed significant results were firm age (0.010***, P > |z| = 0.000), R&D investment (-0.018***, P > |z| = 0.000) and alliance experience (0.003***, P > |z| = 0000).
Finally partner compatibility also showed significant results as a control variable (-0.000***, P > |z| = 0000). The results show however that with a 99% confidence interval that partner compatibility does not influence innovation performance, as this coefficient is zero.
Within the discussion I will further explain the results, in combination with literature and give some indication of what these results mean for the alliance and innovation literature. First, I give a short overview of the robustness checks for the model.
5.2.2. Robustness Check
As a Poisson regression will not accurately be able to check for the robustness of my model, I have performed several other tests. To check the robustness of my variables, I have performed two different analyses. First, I checked the robustness of my partner compatibility measure and did so by comparing the results from both negative binomial regressions within one another, which can be found in appendix A. Here I split the dataset into negative and positive compatibility, the difference being that a part of the dataset the focal firm was smaller as the partner firm, resulting in a negative compatibility, the other part shows the alliances in which the focal firm was larger as the focal firm.
Furthermore, I have checked the robustness of partner complementarity by creating several more dummy variables based on the last digit or more of the SIC code being different (Comple_3), and the last 2 digits or more of the SIC code being different (Comple_2). Referring to the results in appendix E, I can conclude that the results show little difference in terms of coefficients, significance and AIC. This indicates that the measure for partner complementarity is quite robust.
6. Discussion
Within the discussion, I will further discuss the results from the negative binomial regression analysis. It was found that partner specific experience did not affect the innovation performance of the focal firm, whereas partner alliance experience did positively influence the innovation performance. There was only one negative interaction effect found, between partner alliance experience and partner compatibility. The results showed however, that the effect was so small that it approached zero. The explanations for these results supported by literature will be discussed below.
First, I would like to discuss the signals of trust and how partner specific experience and partner alliance experience were found to have no or little effect on innovation performance. Within this research I specifically look at signals of trust and how each could inform the focal firm on how trustworthy their partner is. The signals that are sent out into the alliance, are not always interpreted in the same way by the focal firm some firms might have developed a better propensity to trust than other firms and might influence when a signal is interpreted as positive or not.
A firm's propensity to trust is strongly influenced by past experience of collaborations (Bierly & Gallagher, 2007). A firm's previous experience with alliances could have impacted a firm's ability to trust greatly. Especially when firms have been taken advantage of before, partners that acted against the common purpose of the alliance, or partners that were not able to perform as promised could leave a bad taste in the mouth of the firm, and affect future alliances. According to the learning perspectives, people and by extension firms make decisions based on historical experiences from which they learned specific lessons (Reiter, 1993). Organisational learning is an iterative process where firms engage repeatedly in similar activities and draw lessons from these engagements, these lessons are stored and used for future engagements (Levit & March, 1988). Thus, when the alliance experience was negative for the firm, the after effects could lead to reduced trust in future endeavors. Learning from experience in this case does not only relate to how well the firm has learned to manage and form alliances, but also relates to how well a firm is able to learn to trust a firm in alliance.
Furthermore, the conditions under which the alliance was formed might play a big role in how much trust a firm places in its partner. When alliances with partners were formed out of necessity, and not out of choice, it might play a big role on how a firm is willing to trust the other. It was previously discussed that commitment also plays a role in reasoning why former partners might be more successful for innovation performance. Yet, prior alliance can be about other partner selection issues such as power, search costs, inertia and path dependence (Bierly & Gallagher, 2007). Additionally, the commitment by one firm might not be reciprocated as the partner firm might not be faced with the same issues as the focal firm. This makes the focal firm increasingly vulnerable to the opportunistic actions of the partner firm. Small firms especially might be vulnerable to the aforementioned risks. They face risks of appropriation, but additionally might be forced to work with specific partners, as they lack the resources to overcome their liability of smallness (Baum et al., 2000). They are very committed to the outcome of the alliance, as the firm might not survive the failure of the alliance.
differently than does the larger firm. In order to investigate this further I performed a post-hoc analysis by creating a dummy variable indicating when the partner compatibility was negative or positive.
When looking at the descriptive statistics (Appendix F) and the model created. I can see that there are 82 firms that registered as negative compatibility and 147 firms with positive partner compatibility. The model (Appendix G) shows that there is no significant interaction effect for the interaction effect with partner specific experience, and partner alliance experience, but it did show that partner compatibility as a control variable was significant and showed that there is a negative direct effect of the dummy variable. This means that compared to positive compatibility, innovation performance is lower for firms with negative compatibility. This reinforces that partner compatibility behaves differently for smaller and larger firms in the alliances.
Returning to results regarding the interaction effects, none of my hypotheses were confirmed. The results did show that partner compatibility had extremely small positive effects on partner alliance experience, so small that it can be considered zero. It shows that in the case of operational compatibility, the similar working patterns, routines and habits, allowing new tacit information to be easily assimilated, did not strengthen the signals of trust. Yet as has been mentioned before, partner fit has many more dimensions including cultural compatibility. Cultural compatibility could provide a social construct for trust instead of an operational framework of understanding. According to Bierly & Gallagher (2007) tacit understanding is created through similar cultures. Organisational culture provides a sense of control, as it unifies the way members process information and react to the environment (Das & Teng, 1998). When members of partnering firms have similar cultures, they will be able share the same values and share a common understanding and trust might be easier to develop in the course of the alliance. Organisational culture also changes how working patterns, routines, and habits are understood within the organisation, and could possibly be a better determinant on how trust and signals of trust can be improved within an alliance. Finally, the results also showed that partner complementarity has no impact on innovation performance, not as an interaction effect nor as a direct effect. This discrepancy with the hypotheses could be explained by the fact that the measure of complementarity did not take into account that complementarity is not set in stone. Firms can continuously learn from other simultaneous alliances which knowledge they can bring along to new alliances. The firm's complementarity is not necessarily is not set, and can change as the firm develops. This cannot be illustrated by the SIC codes. Furthermore, the difference in corporate portfolio might not be sufficient to explain the diversification of different firms. The pharmaceutical and biotechnology industry can be considered as very diversified in itself, as many different pharmaceutical companies focuses on different areas of medicine and product development, especially large pharmaceutical companies diversified themselves in several different areas (Suzuki & Kodama, 2004). Sampson (2007) measures the technological diversity between firms. Technological diversity resembles functional diversity, except for the fact that the diversity focuses on the technological background rather than on aspects such as sales experience or other functional diversity elements.
This perspective could give a different view on the complementarity between firms in the pharmaceutical and biotechnology industry.
7. Conclusion
Furthermore, the analysis showed that partner compatibility showed to have a direct effect on innovation performance and showed significant interaction effects with partner alliance experience. Within the post-hoc analysis it was found that partner compatibility has different effects for the larger and the smaller firm, where the smaller firm showed lower innovation performance compared to the larger organisation. This study contributes to exploring what gaps of alliance literature still remain unexplored, emphasizing how trust can be interpreted differently by firms, and how trust might have different effects for small and large organisations.
7.1 Managerial Implications
Managers should be aware when selecting and managing collaborations with possible competitors. Trust has been shown to be a strong determinant of innovation performance in past literature. Yet firms should not solely look at signals of trust in the partner but should instead organize their own organisational characteristics to create strong trust relationships. When firms choose their next partner, they should consider broadening their alliance portfolio in order to create more experience, broadening their ability to deal with uncertain situations.
Partnering with the same partner did not show to improve the innovation performance of the firm, whereas partner alliance experience did show improvement in innovation performance. Firms can better focus on broadening their alliance portfolio with different sources of knowledge. This will also help to develop new alliance know-how and improve the trust propensity of the firm.
8. Limitations & Directions for Future Research
In this section I will comment on the limitations of my research as well as give directions to future research that could be interesting to explore to shed further light on the dynamics between partner fit and trust mechanisms.
First, I would like to discuss how this study would benefit from a larger sample size. First the period in which alliances are recorded could be expanded. This will give more specific insights into the partner specific experience, as a longer time period will allow firms to register more alliances with the same partner. Furthermore, my sample only contained 9 SMEs compared to 200 large organisations. Directions of future research could focus on comparing the relationships between small and large organisations in terms of partner fit and trust dynamics. In order to do so, a larger sample of small and medium sized firms should be collected to come to conclusions. Additionally, different signals of trust could be explored to see how small firms specifically encourage trust between partners, and whether or not these will be interpreted differently by large and small organisations.
9. Bibliography
Aulakh, P. S., Kotabe, M., & Sahay, A. (1996). Trust and performance in cross-border marketing partnerships: A behavioral approach. Journal of international business studies, 27(5), 1005-1032.
Baum, J. A., Calabrese, T., & Silverman, B. S. (2000). Don't go it alone: Alliance network composition and startups' performance in Canadian biotechnology. Strategic management journal, 21(3), 267-294.
Bierly III, P. E., & Gallagher, S. (2007). Explaining alliance partner selection: fit, trust and strategic expediency. Long range planning, 40(2), 134-153.
Bleeke, J., & Ernst, D. (1995). Is your strategic alliance really a sale?. Harvard business review, 73(1), 97-105.
Blomqvist, K. (2002). Partnering in the dynamic environment: The role of trust in asymmetric technology partnership formation. Lappeenranta University of Technology.
Brouthers, K. D., Brouthers, L. E., & Wilkinson, T. J. (1995). Strategic alliances: Choose your partners.
Long range planning, 28(3), 2-25.
Chen, M. J., & Hambrick, D. C. (1995). Speed, stealth, and selective attack: How small firms differ from large firms in competitive behavior. Academy of management journal, 38(2), 453-482.
Cohen, W. M., & Levinthal, D. A. (1990). Absorptive capacity: A new perspective on learning and innovation. Administrative science quarterly, 128-152.
Comanor, W. S., & Scherer, F. M. (1969). Patent statistics as a measure of technical change. Journal of
political economy, 77(3), 392-398.
Cullen, J. B., Johnson, J. L., & Sakano, T. (2000). Success through commitment and trust: The soft side of strategic alliance management. Journal of World Business, 35(3), 223-240.
Das, T. K., & Teng, B. S. (2000). Instabilities of strategic alliances: An internal tensions perspective.
Organization science, 11(1), 77-101.
De Jong, G., & Woolthuis, R. K. (2008). The institutional arrangements of innovation: antecedents and performance effects of trust in high‐tech alliances. Industry and Innovation, 15(1), 45-67.
Deutsch, M. (1960). Trust, trustworthiness, and the F Scale. The journal of abnormal and social psychology, 61(1), 138.
Douma, M. U., Bilderbeek, J., Idenburg, P. J., & Looise, J. K. (2000). Strategic alliances: managing the dynamics of fit. Long range planning, 33(4), 579-598.
Fombrun, C., & Shanley, M. (1990). What's in a name? Reputation building and corporate strategy.
Academy of management Journal, 33(2), 233-258.
Ganesan, S. (1994). Determinants of long-term orientation in buyer-seller relationships. Journal of
marketing, 58(2), 1-19.
Granovetter, M. (1985). Economic action and social structure: The problem of embeddedness. American
journal of sociology, 91(3), 481-510.
Griliches, Z. (1990). Patent statistics as economic indicators: A survey. Journal of Economic Literature, 27, 1661–1707.
Gulati, R. (1995). Does familiarity breed trust? The implications of repeated ties for contractual choice in alliances. Academy of management journal, 38(1), 85-112.
Harrigan, K. R. (1988b). ‘Strategic alliances and partner asymmetries’, Management International Review, 28, pp. 53–72
Hausman, J. A., Hall, B. H., & Griliches, Z. (1984). Econometric models for count data with an application to the patents-R&D relationship.
Hoffmann, W. H., & Schlosser, R. (2001). Success factors of strategic alliances in small and medium-sized enterprises—An empirical survey. Long range planning, 34(3), 357-381.
Hoskisson, R. E., Hitt, M. A., Johnson, R. A., & Moesel, D. D. (1993). Construct validity of an objective (entropy) categorical measure of diversification strategy. Strategic management journal, 14(3), 215-235. Hosmer, L. T. (1995). Trust: The connecting link between organizational theory and philosophical ethics. Academy of management Review, 20(2), 379-403.
Houston, M. B. (2003). Alliance partner reputation as a signal to the market: Evidence from bank loan alliances. Corporate Reputation Review, 5(4), 330-342.
Judge, W. Q., & Dooley, R. (2006). Strategic alliance outcomes: a transaction‐cost economics perspective.
British Journal of Management, 17(1), 23-37.
Kalaignanam, K., Shankar, V., & Varadarajan, R. (2007). Asymmetric new product development alliances: win-win or win-lose partnerships?. Management Science, 53(3), 357-374.
Kale, P., Singh, H., & Perlmutter, H. (2000). Learning and protection of proprietary assets in strategic alliances: Building relational capital. Strategic management journal, 21(3), 217-237.
Kanter, R. M. (1994). Collaborative advantage. Harvard business review, 72(4), 96-108.
Katila, R. (2000). Using patent data to measure innovation performance. International Journal of Business
Performance Management, 2(1/2/3), 180-193.
Koza, M. P., & Lewin, A. Y. (1998). The co-evolution of strategic alliances. Organization science, 9(3), 255-264.
Krishnan, R., Geyskens, I., & Steenkamp, J. B. E. (2016). The effectiveness of contractual and trust‐based governance in strategic alliances under behavioral and environmental uncertainty. Strategic Management
Journal, 37(12), 2521-2542.
Krishnan, R., Martin, X., & Noorderhaven, N. G. (2006). When does trust matter to alliance performance?.
Academy of Management journal, 49(5), 894-917.
Lam, A. (1997). Embedded firms, embedded knowledge: Problems of collaboration and knowledge transfer in global cooperative ventures. Organization studies, 18(6), 973-996.
Lane, P. J., & Lubatkin, M. (1998). Relative absorptive capacity and interorganizational learning. Strategic
management journal, 19(5), 461-477.
Lane, P. J., Salk, J. E., & Lyles, M. A. (2001). Absorptive capacity, learning, and performance in international joint ventures. Strategic management journal, 22(12), 1139-1161.
Laaksonen, T., Jarimo, T., & Kulmala, H. I. (2009). Cooperative strategies in customer–supplier relationships: The role of interfirm trust. International Journal of Production Economics, 120(1), 79-87. Levitt, B., & March, J. G. (1988). Organizational learning. Annual review of sociology, 14(1), 319-338. 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.
Lin, Z., Yang, H., & Arya, B. (2009). Alliance partners and firm performance: resource complementarity and status association. Strategic management journal, 30(9), 921-940.
Lui, S. S., & Ngo, H. Y. (2004). The role of trust and contractual safeguards on cooperation in non-equity alliances. Journal of management, 30(4), 471-485.
Lu, X., & White, H. (2014). Robustness checks and robustness tests in applied economics. Journal of
econometrics, 178, 194-206.
Madhok, A., & Tallman, S. B. (1998). Resources, transactions and rents: Managing value through interfirm collaborative relationships. Organization science, 9(3), 326-339.
Molina-Morales, F. X., Martínez-Fernández, M. T., & Torlo, V. J. (2011). The dark side of trust: The benefits, costs and optimal levels of trust for innovation performance. Long Range Planning, 44(2), 118-133.
Morrow, J. D. (1991). Alliances and asymmetry: An alternative to the capability aggregation model of alliances. American journal of political science, 904-933.
Nielsen, B. B. (2005). The role of knowledge embeddedness in the creation of synergies in strategic alliances. Journal of Business Research, 58(9), 1194-1204.
