• No results found

MASTER THESIS MSc Business Administration Specialization: Strategic Innovation Management

N/A
N/A
Protected

Academic year: 2021

Share "MASTER THESIS MSc Business Administration Specialization: Strategic Innovation Management"

Copied!
44
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

1

University of Groningen | Faculty of Economics and Business Department of Innovation Management & Strategy

MASTER THESIS

MSc Business Administration

Specialization: Strategic Innovation Management

Extending Alliance Portfolio:

The Impact of Acquisitions on the Relationship between

Alliance Portfolio and Firm Performance

Aneta Anna Oleksiak

Student No: 2375346

First supervisor: Dr. Pedro de Faria Second supervisor: Prof. Dr. Dries Faems

Word count: 14075

(2)

2

ABSTRACT

Previous research has given a lot of attention to alliance portfolio configuration and its impact on firm performance. The purpose of this thesis is to deepen the understanding of this relationship by examining if the share of the acquired (indirect) alliance portfolio moderates the relationship between alliance portfolio and firm performance. Through merging data from Thomson SDC Platinum Alliances and SDC Platinum Mergers & Acquisitions databases and triangulating information on financial data from Compustat, Orbis, Amadeus databases and annual financial statements, a panel dataset was created for a sample of 31 firms from biotechnological and pharmaceutical industries. The results show that the share of acquired alliance portfolio moderates both the relationship between industrial diversity and firm performance and between functional diversity and financial performance. Namely, firms that have a low share of the indirect portfolios benefit from increased industrial and functional diversity of the alliance portfolio. However, in the presence of a high share of the acquired portfolios, increasing industrial and functional diversity has a negative influence on firm performance. The advantages of having a portfolio seem to dissipate in this case. Jointly, this evidence demonstrates that when choosing a potential acquiree, the acquirer has to consider the shape of the own alliance portfolio and the portfolio of the acquiree. This is essential because if the indirect alliances constitute a big part of the alliance portfolio, this would most probably reduce the potential advantages of having a portfolio. For this reason, the acquisitions are not necessarily the best tool for enlarging alliance portfolios.

Key words: alliance portfolio dynamics, acquisition, alliance portfolio size, industrial

(3)

3

TABLE OF CONTENT

1 INTRODUCTION ... 4

2 THEORETICAL BACKGROUND ... 7

2.1 ALLIANCES AND THEIR GROWING STRATEGIC MEANING ... 7

2.2 ALLIANCE PORTFOLIO DIVERSITY ... 9

2.2.1 Industrial Diversity ... 10

2.2.2 Functional Diversity ... 12

2.3 ALLIANCE PORTFOLIO SIZE ... 14

2.4 ACQUISITION CHANGES THE IMPACT OF ALLIANCE PORTFOLIO ON FIRM PERFORMANCE ... 16

2.4.1 Acquisition, Acquired Alliance Portfolio and its Different Nature ... 16

3 METHODOLOGY ... 19

3.1 DATA COLLECTION AND ALLIANCE PORTFOLIO CONSTRUCTION ... 19

3.2 SAMPLE ... 20 3.3 ANALYTICAL METHOD ... 21 3.4 MEASURES ... 21 4 RESULTS ... 23 4.1 DESCRIPTIVE STATISTICS ... 23 4.2 REGRESSION RESULTS ... 25

4.2.1 Step One – Analysis of Panel A ... 25

4.2.2 Step Two – Analysis of Panel B ... 27

4.2.3 Step Three – Analysis of Interaction Terms with Data from Panel B ... 29

5 DISCUSSION ... 32

6 CONCLUSION ... 37

7 LIMITATIONS AND FUTURE RESEARCH ... 39

ACKNOWLEDGEMENTS ... 40

(4)

4

1 INTRODUCTION

In the last two decades strategic alliances have become a crucial part of business strategy (Duysters et al., 2012; Gulati and Singh, 1998; Jiang et al., 2010). By means of this kind of cooperative agreement, firms are able to gain access inter alia to unique and scarce resources and expert knowledge which they do not have at their disposal in-house. Owing to the fact that nowadays the majority of firms is unable to develop all needed resources on their own (Van Beers and Zand, 2014), strategic alliances became a useful strategic tool to enhance firm’s internal activities (Faems et al., 2010). In order to ensure the feasibility of those activities by having access to essential resources, firms engage not only in one but in multiple alliances at the same time and establish alliance portfolios. For this reason, a shift in research from investigating individual alliances to analyzing alliance portfolios as a cumulative representation of single alliances can be observed.

The existing literature on portfolios highlights the importance of alliance portfolio configuration. Scholars put a lot of emphasis on size and its implications (Deeds and Hill, 1996; Lahiri and Narayanan, 2013; Rothaermel and Deeds, 2006; White and Lui, 2005) but more importantly, on diversity of partners and its impact on firm performance (Cui and O’Connor, 2012; Duysters et al., 2012; Faems et al., 2010; Jiang et al., 2010; Noseleit and de Faria, 2013; Sampson, 2007). Authors try to explain how particular portfolio composition influences firm performance, however, the results are still not fully conclusive. In order to shed new light on existing research, in this thesis the study of Jiang et al. (2010) will be partially replicated. The approach will be used in order to investigate the impact of direct alliance portfolio configuration, i.e. of portfolio which has been formed by the company directly, on firm’s financial outcomes.

(5)

5 decrease, instead of enhance the potential of a portfolio. Thus, the main objective of this study is to contribute to alliance portfolio research by revealing whether the share of the indirect alliance portfolio has an impact on the relationship between alliance portfolio characteristics, in particular industrial diversity, functional diversity and size, and firm performance. It is important to fill this gap, since an understanding of portfolio dynamics as well as obtaining a complete picture of portfolios in terms of direct and indirect alliances is essential for determining a profitable alliance portfolio management strategy. If companies gain a better understanding of which portfolio composition is the most beneficial and how acquiring an indirect alliance portfolio may influence the relationship between alliance portfolio and the firm performance, they will be able to make better decisions in the choice of acquirees and their alliance portfolio partners. Moreover, filling this gap will provide a novel way of constructing variables related to alliance portfolio composition, i.e. considering both direct and indirect alliance portfolios when looking at alliance portfolio as a whole.

Throughout the study, it will be distinguished between industrial and functional diversity of partners. Further, the quantitative aspect of portfolios, namely, their evolving size, will be explored. To test the hypotheses, a panel dataset was created for 31 companies in biotechnological and pharmaceutical industries for a five-year time period (2002 – 2006). For this purpose, the data from Thomson SDC Platinum Alliances and Thomson SDC Platinum Mergers & Acquisitions databases was merged. Financial data was obtained from databases such as Orbis, Amadeus, Compustat and firm’s financial statements, i.e. annual reports and SEC forms. The fixed-effects model was applied in the empirical analysis.

(6)

6 acquired portfolios is high, then the advantages from having a portfolio seem to diminish. In this case increasing both industrial and functional diversity has a negative impact on firm performance. With regards to portfolio size, no significant relationship with firm performance was found. Similarly, the interaction term of portfolio size and share of indirect portfolios did not demonstrate significant results meaning that there is no moderation effect of the share of acquired alliance portfolio on the relationship between portfolio size and financial performance of the firm.

In such a way, this study sheds new light on alliance portfolio dynamics. It is shown that the share of the indirect alliance portfolio plays a meaningful role in influencing the relationship between alliance portfolio diversity measures and firm performance. A valuable contribution of this study is that the acquisitions are not necessarily the most appropriate tool for enlarging portfolios. This is because having a high share of the acquired alliances in the portfolio requires a lot of commitment, which in turn highly increases the costs of having a portfolio. In that way, having a high share of indirect alliances may reduce all potential advantages of having a portfolio. Further, this thesis demonstrates that it is necessary to consider partner diversity in terms of different dimensions because different elements of diversity can exhibit varying implications for firm performance. Hence, the partner diversity should not be treated as a single dimension. Moreover, another meaningful contribution for the research methodology is provided, namely variable construction. The approach applied in this thesis shows that when investigating alliance portfolios, researchers ought to make sure that they check whether firm’s alliance portfolio includes indirect alliances. Since the Thomson SDC Platinum Alliances database links only to the direct alliances, it is important to go beyond that and merge data from different sources. Only combining the information about the direct and indirect alliances makes it possible to obtain a clear and complete overview of company’s alliance portfolio.

(7)

7

2 THEORETICAL BACKGROUND

In this part an extensive literature review is provided. The first section (2.1) describes the phenomenon of the strategic alliances and the growing relevance of alliance portfolios. The second (2.2) and third (2.3) subsections touch upon the implications of alliance portfolio diversity and size respectively. Here, the hypotheses that attempts to replicate the results of previous research are presented. The last section (2.4) leads to the literature gap and provides reasoning why the share of the acquired alliance portfolio may moderate the relationship between alliance portfolio and firm performance.

2.1 ALLIANCES AND THEIR GROWING STRATEGIC MEANING

The majority of firms do not have at their disposal all resources and capabilities to both develop and bring an innovation to the market (Van Beers and Zand, 2014). In the last twenty years, especially companies operating in very dynamic industries such as biotechnological and pharmaceutical, have made an increasing use of strategic alliances in order to enhance and secure the firm’s resource pool as well as win the battle for the competitive advantage (Eisenhardt and Schoonoven, 1996; Hoffmann, 2007).

(8)

8 this trend greater development risks and rising costs. Through sharing or exchanging resources between or among alliance partners companies can achieve more, since often without such a resource combination a development of certain capabilities, knowledge or even new innovative products would be impossible (Das and Teng, 2000; Deeds and Hill, 1996; Tsang, 2000). This implies a synergistic effect of resource combination (Tsang, 2000; Wang and Zajac, 2007; Van Beers and Zand, 2014). St John and Harrison (1999) argue that combining assets makes the resource utilization more efficient compared to a situation where each of the companies would operate on its own. Van Beers and Zand (2014) go further and state that in order to develop and successfully commercialize a completely novel product companies have to make use of synergistic effects of resource pooling. For example, a company that is strong in R&D, may be limited in its commercialization capabilities. Therefore, in order to remain ahead of the competition with their innovation, it is wise to combine complementary forces and take advantage of partners marketing competence, and bring the product faster onto the market. Cooperation, therefore, may result in significantly greater benefits. Following this line of arguments, the knowledge-based view (Spender, 2007; Van Beers and Zand, 2014) considers strategic alliances as a source of expertise, know-how and external novel knowledge (Das and Teng, 2000; Ireland et al., 2002) as well as information about markets (Ireland et al., 2002), all of which can help filling the knowledge gaps and positively influence firm performance (Das and Teng, 2000; Ireland et al., 2002).

(9)

9 similarities and dissimilarities of partners and their unique compilation determine the success of an alliance portfolio (Brouthers et al., 1995; Darr and Kurtzberg, 2000). Thus, it is essential to investigate its composition, namely the diversity but also the size which corroborate to have an influence on firm performance.

2.2 ALLIANCE PORTFOLIO DIVERSITY

The partner diversity, which is defined as the extent to which partners differ from each other, and its effects on firm performance has received more attention in the last decade. However, the current evidence is still somewhat ambiguous. On the one hand, there are findings which are in favor of heterogeneous alliance portfolios. For example, Baum et al. (2000) in their study on biotechnology companies have found a positive impact of diverse partners on firm’s financial performance. Authors argue that such partners are a source of various information, often capabilities and other resources that the focal firm needs to develop and market an innovative product. Due to a rich and diversified resource pool, firms are able to shorten the time-to-market, outstrip rivals, sell products before others do and enjoy revenues. Further, diversity gives an opportunity to explore and exploit firms’ resources and ideas that when translated into products are a source of short-term revenues from exploitation and long-term revenues from exploration (Jiang, et al., 2010). Interestingly, being in a network of diversified partners teaches companies to deal with the surrounding complexity of the business they are in. This in turn makes it possible to establish routines, best practices and embed in the network which ease a firm’s survival and ensure future earnings (Ozcan and Eisenhardt, 2009).

(10)

10 distant to be absorbed and utilized. The more dispersed the knowledge between partners, the more difficult to manage and the higher the transaction costs (McGill and Santoro, 2009). In turn, the focal firm’s initial assumptions about the cooperation cannot be achieved. High alliance portfolio management costs and problems with knowledge transfer bear negative consequences for firm performance and outweigh initial benefits arising from greater heterogeneity in portfolio.

As it has been shown above, the dispute about the impact of alliance portfolio partner diversity is still very vivid and ambiguous. This can be explained by the fact that the partner diversity is differently assessed by scholars. Whereas some consider partner diversity as a single dimension (Duysters et al., 2012; Faems et al., 2010), others divide it into two or even more dimensions (Duysters and Lokshin, 2011; Goerzen and Beamish, 2005; Jiang et al., 2010). In this study, the partner diversity will be considered in terms of two constructs: the industrial and the functional diversity, in order to give a better understanding of dependencies between these two diversity types and firm performance.

2.2.1 Industrial Diversity

One of the most relevant characteristics that influences the shape and outcomes of the alliance portfolio is its industrial heterogeneity, i.e. the extent to which alliance partners belong to different industries. Although previous research investigated the impact of industrial diversity on firm performance, the outcomes of those studies are still not fully conclusive. Some authors praise more the positive effects (Jiang et al., 2010; Luo and Deng, 2009) whereas others stress that negative implications may arise together with higher levels of industrial diversification and outweigh the positive ones (Goerzen and Beamish, 2005; Noseleit and de Faria, 2013; Sampson, 2007).

(11)

11 complementary resources, industrial diversity teaches and forces companies to become flexible to environmental changes and learn how to deal with the business and technological uncertainty and react to them. Such an acquired ability increases firms’ chances for survival (Hoffmann, 2007; Jiang et al., 2010; Ozcan and Eisenhardt, 2009). In addition, Deeds and Rothaermel (2006) in their investigation of biotechnology firms, highlight the importance of alliances with non-profit organizations such as universities, which represent completely unrelated industry to biotechnology or pharmaceuticals. Universities have proven to be an excellent source of high-quality knowledge and human capital capable to combine their expertise with firms’ current innovations. This is being translated into new products and later sales. Next to the resource complementarities, they argue that industrially diversified portfolios offer a faster way of achieving economies of scale which positively affect firm performance.

(12)

12 internal R&D unit, which in turn can contribute to lower performance. Although partners with different industrial backgrounds are a potential source of distinct and inspiring knowledge, collaboration between unrelated partners is often not fruitful due to problems with knowledge assimilation since the know-how is too distant to firm’s own knowledge base. What is more, Noseleit and de Faria (2013) argue that the most beneficial composition for the internal R&D activities consists of partners from the related and the same industries. This is due to the fact that with such partners a firm is able to create new knowledge, however this knowledge will not be that distinct, thus it will be possible to recognize, transfer and assimilate it. This is in line with one of the first studies which argued for the inverted U-shaped relationship between partners industrial diversity and firm performance (Sampson, 2007). This research indicated that a company can obtain the most benefits from moderate levels of industrial diversity among its partners in the alliance portfolio. Too homogenous partners are limited in their ability to develop completely novel ideas and yield additional value since their understanding has similar limitations (Sampson, 2007), whereas too highly heterogeneous portfolio will create knowledge transfer, attention (Deeds and Rothaermel, 2006; Hoang and Rothaermel, 2005) and scarce resources allocation problems (Hoang and Rothaermel, 2005). Just as important, industrially diversified partners may have different strategic goals which hinder achieving set aims and cooperation intentions by the focal company (Hoffmann, 2005).

Although increasing industrial diversity can be beneficial, turning the portfolio into a structure consisting of partners with totally distinct backgrounds will lead to a decrease in firm performance due to management and understanding problems of industrially unrelated firms. Hence, it is expected that the industrial diversity among alliance portfolio partners will have positive effects up to a certain point from which it will impact firm performance negatively. Therefore the following hypothesis is proposed:

Hypothesis 1: The relationship between industrial diversity of the alliance portfolio and firm

performance follows an inverse U-shaped relationship.

2.2.2 Functional Diversity

(13)

13 there is evidence (Cui and O’Connor, 2012) that highly diversified partners, with regards to their function, may negatively affect firm performance due to higher coordination costs and demand for greater effort to share information and resources, the majority of the research body declares itself in favor of positive implications of functional diversity (Baum et al., 2000; Jiang et al., 2010; Lavie and Rosenkopf, 2006; Van Beers and Zand, 2014). Even though the aspects of the experience in alliance portfolio management and the alliance portfolio management capability go beyond the scope of this thesis, it is worth mentioning that with rich experience (Cui and O’Connor, 2012; Duysters et al., 2012) established portfolio management capability (Duysters et al., 2012) and an adequate alliance portfolio management system (Kale and Singh, 2009; Neyens and Faems, 2013), firms are able to overcome or at least diminish the negative effects of inter alia functional heterogeneity of the alliance portfolio. Therefore, scholars are inclined towards the positive aspects of functional diversification.

(14)

14 knowledge and widen thinking horizons not only concerning technology but also information about the market (Baum et al., 2000; Van Beers and Zand, 2014). Having access to multidisciplinary and complementary knowledge boosts firm’s creativity and understanding of the market factors. Further, it enhances a firm’s technological competences. This in turn accelerates new product development and the outcome can be translated into sales numbers much faster than it would be without access to partners’ functional resources, knowledge and expertise (Van Beers and Zand, 2014).

Based on this evidence, it can be expected that the positive effects of increased functional diversity among the alliance portfolio partners will fully outweigh the coordination costs as well as greater effort in resource and information sharing. Thus, it can be hypothesized that:

Hypothesis 2: Functional diversity of the alliance portfolio has a positive impact on firm

performance.

2.3 ALLIANCE PORTFOLIO SIZE

(15)

15 terms of time, effort and resources at their disposal. Another reason why alliance portfolio size should not be too large is given by White and Lui (2005). Namely, the bigger the alliance portfolio, the higher the cooperation costs are. At some point a firm may not have enough resources in order to participate actively in the alliances and thus will become an unattractive partner. Participation in alliances requires certain investments and the maintenance is costly. If a firm is unable to face the cooperation costs, an alliance will probably fail. An interesting reasoning behind the number of alliances in a portfolio is given by Rothaermel and Deeds (2006) who also revealed an inverted U-shaped relationship between alliance portfolio size and firm performance. They argue that both too few as well as too many alliances are unfavorable for the firm performance. The authors state that a very small number of alliances puts a company in a situation of competitive disadvantage. This is because companies are not able to develop new products as fast as companies that have more alliances on their disposal. Regarding too many alliances, Rothaermel and Deeds (2006), point out three major issues that arise: mismanagement, expropriation and opportunistic behavior. For those reasons, engaging in too many alliances should be avoided. In line with the prior evidence, Lahiri and Narayanan (2013) argue that collaboration in numerous alliances positively affects both innovation and financial performance of the firm, however, in highly innovative firms too large alliance portfolio will result in a lesser financial performance. In particular, high-technology companies, which are also the objective of this thesis, should find the right balance between the size and potential benefits, otherwise the huge number of partners will lead to lower firm performance. Lahiri and Narayanan (2013) suggest that high-tech innovative companies in the first place, share and communicate the knowledge internally. Therefore, too many external partners to collaborate with will disperse their attention, and as a result decrease performance. Laursen and Salter (2006) support this by arguing that although having many sources of external knowledge or resources can provide a company with a variety of novel ideas, too many sources of external input may lead to an ‘over-search’, a dissipation of search efforts and an attention allocation dilemma in which employees while concentrating on exploiting external sources forget about caring for the already existing ones. This can lead to lower performance of the firm. Hence, this study hypothesizes:

Hypothesis 3: The relationship between the alliance portfolio size and firm performance

(16)

16

2.4 ACQUISITION CHANGES THE IMPACT OF ALLIANCE PORTFOLIO ON FIRM PERFORMANCE

The evidence from the previous sections gives us the irrefutable impression that in order to be successful, a company has to ensure the right shape of the alliance portfolio. Interestingly, scholars argue that the alliance portfolio composition is dynamic in its nature, thus the shape of the portfolio is not static and changes over time. According to them, these dynamics are shaped by the acts of new alliance formation (Ahuja, 2000) and current alliance termination (Makino et al., 2007). The underlying reasons for changing alliance portfolio configuration are to influence the competitive environment by enhancing own competitive position (Wassmer, 2010), control business uncertainty and simply fulfill company’s alliance strategy (Hoffmann, 2007). Even though it has been shown that alliance portfolio is not static and its (evolving) shape has a strategic meaning, there has been very little research that has investigated the changes in alliance portfolio configuration and its implications (Wassmer, 2010). Most interestingly, the event of an acquisition and its implications for alliance portfolio composition and firm performance has been omitted. It is important to give this query attention because the acquirer takes over not only the particular target but also its partnerships, for instance, strategic alliances which are rich in resources, potential knowledge and expertise. However, acquirers have to face the challenge of integration and appropriation of technologies, social capital and capabilities that are enclosed in indirect alliance portfolios, i.e. in the alliance portfolios that belonged to the acquiree and are now owned by the acquirers. In turn, the inclusion of the acquired alliance portfolio may affect the relationship between the alliance portfolio and firm performance.

2.4.1 Acquisition, Acquired Alliance Portfolio and its Different Nature

(17)

17 firm’s network of relationships should be considered as a unique and inimitable structure which not only is a valuable resource itself but it also gives access to potentially precious resources such as information, joint action, knowledge, etc. As a result, a firm can enhance its competitive advantage by extending current alliance portfolio through acquiring a firm with a valuable alliance portfolio. Nowadays, companies compete on access to external resources which may give them the prospect of greater performance. Therefore, in order to reduce competition and strengthen own position while exploiting rivals resources, some firms decide to acquire other companies to take advantage of their resource portfolio (Wang and Zajac, 2007). In this vein, an acquisition and strategic alliance portfolio can be seen as complementary rather than excluding cooperation strategies. This is because an acquisition can be a strategy to obtain access to other firm’s network, which could be an alliance portfolio (Wang and Zajac, 2007). By doing so, firms are able to profit from the resources embedded in the alliances. Relevant partners, who have at their disposal information, know-how and physical resources interesting for the acquirer, may bring a lot of benefits when implemented in the acquirer’s projects, for instance, better product solutions and shorter development time. Further, by incorporating partners which are highly diverse among each other in terms of their functions might be helpful in achieving the right balance between exploration and exploitation, thus developing current activities while inventing new propositions (Lavie and Rosenkopf, 2006). Ensuring the right proportion allows firms to retain financial health in both short and long term which is essential for firm’s sustainable growth.

(18)

18 acquiring company has to face the challenge of dealing with additional relationships that have already been developed between present partners and manage them as well as gain benefits from this inclusion. Thus, it implies the necessity for development of organizational routines to establish communication and a fundament of joint cooperation with new members of the alliance portfolio. This, however, can generate enormous costs of the management of a portfolio consisting of both direct and indirect alliances. Dealing with alliances that obviously demand special managerial care may cause attention and resource allocation problems (Hoang and Rothaermel, 2005). What is more, firms that have a relatively high share of acquired alliances, in proportion to their portfolio, may especially experience problems with dealing with high diversity of knowledge-bases (Ahuja and Katila, 2001). The authors argue that the acquirer ought to be able to recognize, assimilate and most importantly implement the knowledge that it obtains. Dealing with high levels of diversity within the direct portfolio is very difficult in most cases (Sampson, 2007). Thus, increasing the diversity, while having many acquired alliance partners, may become an even more serious challenge. The knowledge exchange and assimilation processes that need to be created for acquirer and acquired portfolio will most probably involve very high costs. Thus, firms that have to cope with a large share of the acquired alliances in their portfolio may experience difficulties in exerting advantages of having diversified alliance portfolio due to the demanding commitment they already have to make to the management of indirect alliances. All this, i.e. a lot of indirect alliances together with the high diversity of portfolio partners, may make it difficult to increase performance.

Thus, considering the conditions of the inclusion of acquiree’s alliance partners into acquirer’s alliance portfolio and the special nature of the indirect alliance portfolio, it can be predicted that the share of indirect alliance portfolio will influence the relationship between alliance portfolio characteristics (industrial diversity, functional diversity and size) and firm performance. Hence, it can be expected that:

Hypothesis 4: The share of acquired alliance portfolio will moderate the relationship between

alliance portfolio industrial diversity and firm performance.

Hypothesis 5: The share of acquired alliance portfolio will moderate the relationship between

alliance portfolio functional diversity and firm performance.

Hypothesis 6: The share of acquired alliance portfolio will moderate the relationship between

(19)

19

3 METHODOLOGY

For this research, a longitudinal study approach has been chosen. By using this method, it is possible to investigate the impact of a changing degree of industrial and functional diversity, and alliance portfolio size on firm performance over time. Based on the collected information, a unique panel dataset was developed in which the focus was on biotechnological and pharmaceutical industries. Consistently with previous research (Sampson, 2007), the assumption that each alliance lasts five years was used. The data was collected for a period from 1998 to 2006, whereas five points in time (tn;n∈{1 (2002), 2 (2003), 3 (2004), 4 (2005), 5 (2006)}) constitute the main observation period of this study.

The chosen industries are representative both in terms of strategic alliances and acquisitions. Cooperative agreements are essential for firm performance in these industries (Deeds and Rothaermel, 2006). Biotechnology and pharmaceutics are very closely related in terms of the domain they operate in. Both are characterized by similar dynamics, competitive market structure, uncertainty and the needs for the same type of, or similar resources. Such a business environment is best suited for answering the research question posed in this paper.

In the following subsections, each methodological step will be described in depth.

3.1 DATA COLLECTION AND ALLIANCE PORTFOLIO CONSTRUCTION

In order to identify companies in biotechnological and pharmaceutical industries, the primary Standard Industrial Classification (SIC) codes were used, where for biotechnological it corresponds to 2836 and for pharmaceutical to 2834. Using Thomson Securities Data Corporation (SDC) Platinum database, in particular its part on Mergers & Acquisitions, companies in biotechnological and pharmaceutical industries that took over a firm(s) in a period from 2002 to 2006 and their targets were identified. Initially, 2760 deals were considered but after restricting the sample to only full takeovers the number dropped to 1362 deals1. In the next step, Thomson SDC Platinum database on Alliances was used to construct the alliance portfolios. Based on the information on the acquisition deals, in SDC Alliances acquisition targets and acquirers were searched by name. Acquisitions wherein targets did not have any alliances were excluded. However, the number of acquisitions per year per firm was retained in order to control for it in the statistical analysis. For each of the acquirers as well as for each of the targets that were engaged in alliances, the announcements for period 1998-2006 were downloaded from the database in a text file. In the next step, these files were converted into Excel files. The approach of this study required a construction of firms’ direct

1

(20)

20 alliance portfolios (Panel A) as well as portfolios including both direct and indirect alliances (Panel B). In case of direct alliances, these were the alliances that were directly formed by the focal companies and were found by searching in SDC Alliances by focal firms’ names. In order to create an alliance portfolio that encompasses both direct and indirect alliances, the targets of the acquisition were searched by name in the SDC Alliances. Their alliance partners were then added to the direct alliance portfolio of the acquirer. Here the assumption of five-year alliance duration (Sampson, 2007) was also strictly followed. For example, if a firm acquired a company in 2003 but the alliance between acquiree and other company was formed in 2000, then this alliance was taken into consideration only for years 2003 and 2004. Having done this, it was possible to calculate the alliance portfolio size, industrial and functional diversity for each acquirer and for both direct alliance portfolios and these including acquired alliance portfolios. A firm-level dataset was constructed in which the information for each of these three measures was aggregated as the sum in the focal year and four previous years (respectively, for each Panel A and Panel B). All these measures and exact explanation of how they were computed will be described in more detailed in the Section 3.4.

The financial data as well as the firm size were collected from such databases as Orbis, Amadeus and Compustat. Due to some discrepancy in the numbers among these databases, it was controlled for their correctness. To do so, the annual reports and comprehensive summary reports of firm performance, i.e. the Securities and Exchange Commission (SEC) forms: 10-K and 20-F, were obtained by conducting an extensive Internet recherché and contacting two companies. As a final step, a unique Excel-format dataset was created in which complete firm-level data for each of the firms was put together. Later, it was exported to the software program STATA12® in which the entire statistical analysis was performed.

3.2 SAMPLE

(21)

21

3.3 ANALYTICAL METHOD

In this thesis panel data is used. In the unique dataset created for the purpose of this study, the same firms are observed year after year and the goal is to investigate the relationship between predictor and dependent variables within and not between companies. The fixed-effects model will help to avoid the bias caused by the unobserved heterogeneity between firms (Noseleit and de Faria, 2013). Therefore, in all analysis conducted in this thesis the conservative fixed-effects model is applied.

3.4 MEASURES

Dependent variable. Net profit margin as a measure of firm performance has been applied in

the previous research (Faems et al., 2010; Jiang et al., 2010). It expresses the percentage of net profit (loss) relative to revenue. In order to maximize the chance for reliable and valid outcomes while minimizing the chance of bogus results caused by an exceptional financial performance in a particular year, a corporate level three-year average net profit margin is used in this study. For time point tn=1 the performance measure will include the average net profit margin from years 2003-2005. This rule was implemented analogically till the last tn=5 (year 2006), where the financial performance corresponds to the three-year average net profit margin between 2007-2009.

Independent variables. In order to test the hypotheses of this study, seven independent

variables were built: industrial diversity of the alliance portfolio with and without acquired alliance portfolio, functional diversity of the alliance portfolio with and without acquired alliance portfolio, alliance portfolio size with and without acquired alliance portfolio, and

share of acquired alliance portfolios.

(22)

22 Thanks to the categorical character of the scale it is possible to measure the degree of diversity in terms of variety. In one of the recent studies (Jiang et al., 2010) Blau Index of Variability (BIV) was chosen to calculate the diversity. Its formal notation is known as:

where pi is equal to the fraction of category i in the alliance portfolio and k is the total number of categories. The Blau Index diversity score ranges from 0 (perfectly homogenous group) to (k-1)/k, where the maximum score is a function of the total number of categories. The higher the number of categories, the higher the possible degree of diversity within a certain group (Blau, 1997). Based on the categorization in this study, the maximum diversity score for the industrial variation is 0.8 and for the functional one it equals to 0.75. In case of unequal number of categories across variables and from this resulting varying maximum scores, Agresti and Agresti (1978) advise to adjust the Blau Index by using the Index of Qualitative Variation (IQV) which multiplies Blau Index by k/(k-1); IQVID=BIV*1.25 for industrial diversity and IQVFD=BIV*1.333 for functional diversity. In such a way, a standardized range from 0 to 1 was obtained. This procedure was repeated for alliance portfolios with and without acquired alliances.

In line with previous studies (Deeds and Hill, 1996; Lahiri and Narayanan, 2013), the measure of the next independent variable, i.e. alliance portfolio size, is determined by a cumulative number of alliances formed, in case of this study, between years 2002 and 2006; observation period: 1998-2006. Firstly, the aggregated number of direct alliances was included (for Panel A). In the second step, the acquired alliances were added (for Panel B).

The last independent variable, namely the share of acquired alliance portfolios, expresses the portion that the indirect alliance portfolios constitutes in the complete alliance portfolio which includes both direct and indirect alliance portfolios.

Interaction terms. In order to test hypotheses 4-6, three interaction terms were built

and tested with the data of Panel B, i.e. including direct and indirect portfolios. In each of the interaction terms, the share of acquired alliance portfolios has the role of moderator. The interaction terms were created by multiplying the moderator with consecutively industrial diversity, functional diversity and alliance portfolio size.

Control variables. Lahiri and Narayanan (2013) argue that the firm performance is

highly dependent on scale effects. Therefore, they state that it is important to control for the

(23)

23 more resources at their disposal and a greater absorptive capacity to manage alliance portfolios (Duysters and Lokshin, 2011). To measure, the authors used the total number of employees. This data has been collected for each of the five points in time. Due to its skewness, the logarithm of the total number of employees has been applied. The second control variable is the number of acquisitions. Owing to the fact that the study is considering the events of acquisition(s) and previous research provides evidence that they have a negative influence on firm performance (Gregory, 1997; Longhran and Vijh, 1997; Riviezzo, 2013), it is important to control whether the intensity of firm’s acquisition activities influences the firm performance.

4 RESULTS

In this section the results of the statistical analysis are reported. It is divided into two main parts, where Section 4.1 is dedicated to the descriptive statistics of the sample and Section 4.2 provides the outcomes of each of the three steps of the empirical analysis.

4.1 DESCRIPTIVE STATISTICS

Two tables with descriptive statistics and correlations between predictor and control variables, and firm performance are presented (p. 24). Table 1a shows the summary for Panel A, i.e. the direct alliance portfolio which does not include acquired alliances. Correspondingly, Table 1b depicts the descriptives for Panel B in which the acquired partnerships are added.

(24)

24

Table 1a. Descriptive statistics and correlations (for Panel A, excluding acquired alliance portfolios) 2,3

Variable Mean S.D. 1 2 3 4 5

Net Profit Margin (three-year av.) 11.30 22.85

Portfolio Size 15.98 17.81 0.23*** Industrial Diversity 0.59 0.31 0.29*** 0.51*** Functional Diversity 0.67 0.30 0.05 0.46*** 0.66*** Number of Acquisitions 1.02 1.17 0.02 0.23*** 0.15* 0.20** Firm Size 3.98 0.84 0.35*** 0.72*** 0.52*** 0.37*** 0.22** 2 Number of observations = 138; *** p<0.01, ** p<0.05, * p<0.1. 3

For the sake of descriptives, all variables have been pooled.

Table 1b. Descriptive statistics and correlations (for Panel B, including acquired alliance portfolios) 4,5

Variable Mean S.D. 1 2 3 4 5 6

Net Profit Margin (three-year av.) 11.30 22.85

Portfolio Size 18.70 19.92 0.22***

Industrial Diversity 0.64 0.28 0.28*** 0.48***

Functional Diversity 0.70 0.28 0.07 0.41*** 0.48*** Share of Acquired Portfolios 0.16 0.21 -0.08 -0.05 0.01 -0.05

Number of Acquisitions 1.02 1.17 0.02 0.28*** 0.10 0.25*** 0.21**

Firm size 3.98 0.84 0.35*** 0.71*** 0.41*** 0.37*** -0.21** 0.22**

4

Number of observations = 138; *** p<0.01, ** p<0.05.

5

For the sake of descriptives, all variables have been pooled.

This evidence demonstrates two important observations. Firstly, it shows that acquiring a target which has alliances in its resource portfolio changes the configuration of the acquirer’s alliance portfolio. Not only in terms of size but also in terms of diversity structure. Secondly, it provides true numbers which indicate that when constructing variables related to alliance portfolio composition, it is important to include both direct as well as indirect (acquired) alliances since not considering them together may cause spurious results. Investigation of only direct alliances does not express the true shape of the portfolio.

Furthermore, the sample encompassed large firms where on average 33219 employees were employed. The average number of acquisitions equaled to 1, where the lowest value for this variable in a particular year was 0 and the highest 5.

(25)

25

4.2 REGRESSION RESULTS

With regards to regressions, the analysis was divided into three steps. In the first step (Section

4.2.1), Panel A including only direct alliance portfolios is analyzed, and thereby the aim of

this part is to replicate the outcomes of the previous studies, and confirm the first three hypotheses. In the second step (Section 4.2.2), Panel B, in which complete portfolios are taken, is analyzed. Specifically, it is examined whether the results alter in comparison to the outcomes of regressions of the Panel A. In the final step (Section 4.2.3), data from the Panel B is investigated, however here the interaction terms were added. In this place an attempt to reveal whether the share of acquired alliance portfolios moderates the relationship between alliance portfolio characteristics and firm performance is made.

An extensive analysis of the results of each step will be provided in the Section 5.

4.2.1 Step One – Analysis of Panel A

The results of the fixed-effects regressions which consider only the direct alliance portfolios are provided in Table 2a (p. 26). In Model 1 only control variables were included. The coefficients of firm size and number of acquisitions are not significant. In Models 2, 4 and 6 the linear terms of alliance portfolio size, industrial and functional diversity were augmented respectively. In the following step, in Model 3 the squared term of alliance portfolio size and in Model 5 the quadratic term of industrial diversity were added in order to investigate whether a curvilinear relationship exists between these constructs and firm performance. In Model 7 all linear terms were examined jointly. Finally, in the Model 8 all variables were considered together with the quadratic terms of alliance portfolio size and industrial diversity.

As expected in the first hypothesis, the relationship between the industrial diversity of the alliance portfolio and the firm performance has a shape of an inverted U. In the joint analysis presented in Model 8 significant results are revealed. It can be seen that the linear

term is positive (β = 0.395, p-value = 0.030), whereas the quadratic term is negative (β = -0.290, p-value = 0.069). This implies a curvilinear relationship (inverted U-shape).

Therefore, Hypothesis 1 is supported.

(26)

26 6

Standardized coefficients are presented. Standard errors are provided in the parentheses.

7

The indication of the significance level is done by: *** p < 0.01, ** p < 0.05, * p < 0.1.

Table 2a. Fixed-effects regression outcomes for Panel A (excluding acquired alliance portfolios) 6,7

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8

Number of Acquisitions -0.018 -0.018 -0.018 -0.018 -0.018 -0.022 -0.023 -0.030

(0.031) (0.031) (0.031) (0.031) (0.031) (0.030) (0.031) (0.031)

Firm Size (log) -0.398 -0.401 -0.396 -0.398 -0.397 -0.259 -0.217 -0.162

(0.240) (0.241) (0.243) (0.241) (0.242) (0.244) (0.247) (0.247)

Portfolio Size 0.013 -0.040 0.030 0.154

(0.082) (0.236) (0.081) (0.242)

Portfolio Size (squared) 0.039 -0.078

(27)

27 In the third hypothesis it was expected that the relationship between the alliance portfolio size and firm performance will take the shape of an inverted U. Although in Model 8 it is shown that the coefficient of the root term is positive, whereas the squared term takes a negative value, which would indeed imply an inverted U-shaped relationship, the coefficients are non-significant. Therefore, Hypothesis 3 cannot be supported.

Lastly, both control variables do not show significant relationship with firm performance.

4.2.2 Step Two – Analysis of Panel B

In Table 2b (p. 28) the results of the fixed-effects regressions, in which indirect alliance portfolios were added, are presented. Here, the same approach as for Panel A was applied, i.e. Model 9 (the first model of the Panel B) includes only control variables; Models 10, 12 and 14 the linear terms, while 11 and 13 roots and quadratic terms. Consecutively, two joint analyses are performed, where in Model 16 all linear terms are investigated and in Model 17 roots and relevant quadratic terms are included. In this panel an independent variable was added, namely the share of acquired alliance portfolio, which is considered in separate analysis in Model 15 as well as in two joint analyses. The goal of the analyses of Panel B is to see whether after the inclusion of indirect alliance portfolios, the results alter in comparison with the outcomes of analyses of Panel A where only the direct alliance portfolios were considered.

With regards to Panel B, it is noticeable that the results alter when compared with the outcomes of the Panel A. Significant changes are observable with respect to industrial diversity and functional diversity. Additionally, the analysis reveals that the share of acquired alliance portfolios has a positive influence on firm performance (β = 0.066, p-value = 0.094). Therefore, it is possible to state, that the indirect alliance portfolios are not without importance.

(28)

28 8

Standardized coefficients are presented. Standard errors are provided in the parentheses.

9

The indication of the significance level is done by: *** p < 0.01, ** p < 0.05, * p < 0.1.

Table 2b. Fixed-effects regression outcomes for Panel B (including acquired alliance portfolios) 8,9

Model 9 Model 10 Model 11 Model 12 Model 13 Model 14 Model 15 Model 16 Model 17

Number of Acquisitions -0.018 -0.019 -0.019 -0.017 -0.016 -0.021 -0.032 -0.032 -0.031

(0.031) (0.031) (0.032) (0.030) (0.030) (0.030) (0.031) (0.029) (0.030)

Firm Size (log) -0.398* -0.412* -0.412 -0.483** -0.435* -0.234 -0.612** -0.395 -0.377

(0.240) (0.244) (0.254) (0.234) (0.238) (0.247) (0.255) (0.245) (0.249)

Portfolio Size 0.024 0.023 -0.019 -0.093

(0.070) (0.210) (0.068) (0.215)

Portfolio Size (squared) 0.000 0.052

(0.149) (0.147) Industrial Diversity 0.131*** 0.011 0.158*** 0.101 (0.047) (0.121) (0.049) (0.126) Industrial Diversity (sq.) 0.141 0.064 (0.130) (0.130) Functional Diversity -0.097** -0.155*** -0.147*** (0.044) (0.043) (0.046)

Share of Acquired Portfolios 0.076** 0.060* 0.066*

(29)

29

4.2.3 Step Three – Analysis of Interaction Terms with Data from Panel B

The results of the fixed-effects regressions, which include the interaction terms, are presented in the Table 2c (p. 31). Models 18, 20 and 22 show the joint analysis where the linear terms of the variables were included. With regards to interaction terms, in Model 18 the linear term of industrial diversity interacts with the share of acquired alliance portfolio (moderator), in Model 20 the interaction takes place between linear term of functional diversity and moderator, and in a similar vein, in Model 22 the linear term of alliance portfolio size interacts with the moderator. In Models 19, 21 and 23 the joint analyses with additionally relevant quadratic terms of the independent variables are reported. Next to the linear terms, the squared terms of relevant independent variables are applied in the interaction terms. Namely, in Models 19 and 23 the interaction terms encompassing quadratic values of the industrial diversity and portfolio size were added respectively.

The fourth hypothesis predicted that the share of acquired alliance portfolios moderates the relationship between the industrial diversity and firm performance. The analysis provides insightful and significant results with respect to this interaction. Although in Model 18, where only linear terms were included, the interaction term between industrial diversity and share of acquired alliance portfolios shows a positive and significant interaction, a change in the results is remarkable when the quadratic term is added. In Model 19, the coefficient of the interaction term of the moderator and the linear term of industrial diversity equals to β = -0.917 (p-value = 0.001), whereas the coefficient of the interaction term of the moderator and the quadratic term of industrial diversity shows value of β = 0.464 (p-value = 0.000), indicating a nonlinear effect. In order to better understand these results as well as be able to interpret them, Figure 1 depicts this interaction effect.

-2 -1 0 1 2 3 4 5 6 F irm P er form ance

Low Share of Acquired Portfolios

High Share of Acquired Portfolios

Figure 1. The interaction effect between industrial diversity and share of acquired portfolios and its effect on

firm performance.

(30)

30 2 2,2 2,4 2,6 2,8 3 3,2 3,4 3,6 3,8 4

Low Functional Diversity High Functional Diversity

F irm P er form ance

Low Share of Acquired Portfolios

High Share of Acquired Portfolios

Figure 1 (p. 29) shows that in case of firms with a low share of the acquired alliance

portfolios, increasing industrial diversity has a positive impact on firm performance. On the other hand, if a firm has a high share of acquired alliance portfolios, increasing industrial diversity will have a negative effect on financial outcomes. These results provide evidence that the share of the acquired alliance portfolios moderates the relationship between the industrial diversity and firm performance. Therefore, Hypothesis 4 is supported.

In the fifth hypothesis it was expected that the share of acquired alliance portfolios will moderate the relationship between functional diversity and firm performance. Also with respect to this hypothesis, significant results are found. Consistently, both in Models 20 and 21 the coefficients of the examined interaction term between functional diversity and moderator is negative and significant (in Model 20: β = -0.255, p-value = 0.001; in Model 21: β = 0.257, p-value = 0.001). In order to be able to state explicitly what this interaction effect means for the financial performance, a figure is presented.

Figure 2 demonstrates that with regards to firms that have a low share of acquired

alliance portfolios, increasing functional diversity has a positive effect on firm performance. Opposite to that, in case of companies with a high share of acquired alliance portfolios, increasing functional diversity has a negative impact on firm performance. This outcome outlines that the share of acquired alliance portfolios indeed moderates the relationship between the functional diversity and firm performance. Thus, Hypothesis 5 is accepted.

Figure 2. The interaction effect between functional diversity and share of acquired portfolios and its effect on

(31)

31

Table 2c. Fixed-effects regression outcomes for Panel B with the interaction terms 10, 11

Model 18 Model 19 Model 20 Model 21 Model 22 Model 23

Number of Acquisitions -0.030 -0.017 -0.028 -0.028 -0.037 -0.038 (0.029) (0.027) (0.028) (0.028) (0.029) (0.030)

Firm Size (log) -0.416* -0.222 -0.318 -0.306 -0.415* -0.366

(0.242) (0.230) (0.234) (0.238) (0.245) (0.249)

Portfolio Size -0.055 -0.220 0.031 0.027 0.005 -0.045

(0.069) (0.199) (0.066) (0.207) (0.071) (0.218)

Portfolio Size (squared) 0.092 0.001 0.065

(0.135) (0.141) (0.148) Industrial Diversity 0.125** 0.469*** 0.090* 0.017 0.145*** 0.068 (0.051) (0.143) (0.051) (0.122) (0.050) (0.127) Industrial Diversity (sq.) 0.007 0.082 0.072 (0.119) (0.123) (0.129) Functional Diversity -0.130*** -0.115*** -0.095** -0.088* -0.162*** -0.158*** (0.044) (0.043) (0.045) (0.047) (0.043) (0.046) Share of Acquired Portfolios -0.093 0.950*** 0.297*** 0.302*** 0.051 0.049 (0.083) (0.254) (0.077) (0.078) (0.038) (0.040) Share * Industrial Div. 0.175** -0.917***

(0.086) (0.268) Share * Industrial Div. (sq.) 0.464*** (0.108)

Share * Functional Div. -0.255*** -0.257***

(0.074) (0.075)

Share * Portfolio Size -0.046 -0.224

(0.039) (0.138)

Share * Portfolio Size (sq.) 0.178

(0.135) _constant 0.301*** 0.303*** 0.293*** 0.290*** 0.300*** 0.293*** (0.038) (0.036) (0.037) (0.037) (0.038) (0.039) Number of Observations 138 138 138 138 138 138 Number of Firms 31 31 31 31 31 31 F - test 4.71*** 5.72*** 6.13*** 4.74*** 4.21*** 3.15*** R2 within 0.248 0.371 0.300 0.303 0.223 0.245 R2 between 0.452 0.328 0.435 0.422 0.449 0.447 10

Standardized coefficients are presented. Standard errors are provided in the parentheses.

11

(32)

32 Lastly, this study hypothesized that the share of acquired alliance portfolios will moderate the relationship between alliance portfolio size and firm performance. In model 22 only the linear roots of the independent variables are considered as well as a linear term of the alliance portfolio size is used in the interaction term. Model 23 includes relevant quadratic terms of independent variables and of the alliance portfolio size applied in the interaction term. Contrary to the expectations, both Models 22 and 23 do not show any significant results for this interaction, neither the interaction term with the linear nor quadratic term of alliance portfolio size. For this reason, Hypothesis 6 cannot be supported.

In the light of these results, it can be stated that if a company has a low share of acquired alliance portfolios, it is beneficial to bring industrial and functional diversity onto higher levels which in turn enables a firm to achieve higher performance. However, if a firm has a high share of the acquired alliance portfolios, the potential advantages from having a portfolio seem to vanish. The arguments behind this statements will be discussed in the section below.

5 DISCUSSION

A lot of attention has been given to alliance portfolio configuration and its impact on firm performance. However, the changes in the portfolio composition over time and their influence on companies’ outcomes remained under-researched. For instance, the impact of alliance portfolio enlargement with acquisitions has not been investigated yet. Therefore, the main goal of this study was to reveal whether the share of acquired alliance portfolio moderates the relationship between certain alliance portfolio characteristics and firm performance. For this purpose, a sample of 31 firms from biotechnological and pharmaceutical industries was investigated. The results demonstrate that the indirect alliance portfolios influence the relationship between alliance portfolio characteristics, especially industrial and functional diversity, and firm performance. Throughout all the analyses, no impact of alliance portfolio size on firm performance was found. Further, no moderation effect of share of acquired alliance portfolio on the relationship between portfolio size and financial performance was revealed. This implies that it is not necessarily the pure quantitative aspect of portfolios that matters, but the qualitative one, i.e. the diversity, as well as the presence of the indirect alliances in the portfolio.

(33)

33 much by the portfolio’s size but by the characteristics of the firms that a focal organization is connected to.” This is also in consonance with Hagedoorn and Schakenraad (1994) who claimed that it is not necessarily the number of alliances a firm participates in but the qualities of partners. Thus, the success does not lie in the number of alliances but in the right selection of alliance portfolio partners. In a similar vein, the results of this thesis show that the share of indirect alliance portfolio does not impact the relationship between alliance portfolio size and firm performance. Therefore, the further discussion will concentrate on the effects regarding industrial and functional diversity discovered in this study.

The analysis of the Panel A, in which only direct alliance portfolios were included, attempted to replicate the results from previous research. With regards to industrial diversity this study confirms the inverted U-shaped relationship between industrial diversity and firm performance. This outcome is in line with the evidence of previous literature (Goerzen and Beamish, 2005; Noseleit and de Faria, 2013; Sampson, 2007) and shows that increasing industrial diversity positively affects firm performance, however if it is brought onto very high level, the operational and managerial costs arising from the management of industrially heterogeneous portfolio outweigh the benefits that were visible at a somewhat lower level of industrial diversity. Further, opposite to the predictions, a negative relationship between functional diversity and firm performance was found. Cui and O’Connor (2012) argue that studies which praise the positive effects of functional diversity of portfolio, which could be, for instance, a better balance between explorative and exploitative activities, underestimate the difficulties with information as well as resource exchange. Authors highlight that costs of coordination and cooperation with functionally heterogeneous portfolio partners often exceed the potential benefits. For this reason, such negative relationship between functional diversity of partners and firm performance ought to be not surprising. However, this study also shows that these results are not the ultimate representations of the true relationships between investigated constructs. This is because under such a study and sample design as the one of Panel A, only the impact of direct alliance portfolio characteristics on firm performance is measured. This, however, does not depict the true shape of the portfolio since it omits the impact of the indirect alliance portfolios. Thus, by studying only the direct alliance portfolios researchers are not able to discover the true relationships and dependencies.

(34)

34 the shape of the portfolio and the diversity structure shift when the indirect alliance portfolios are added. Further, the empirical analysis and the results of the regressions demonstrated significant changes in the relationships between independent variables and firm performance. Additionally, the share of acquired alliance portfolios appeared to have a significant impact on firm performance. These results clearly highlight the need of constructing complete alliance portfolios since considering only the direct portfolios generates spurious results. In order to examine in more depth the hypothesized impact of the indirect alliance portfolios on the relationship between alliance portfolio characteristics and firm performance, a third analysis, in which the interaction terms were included, was conducted.

With regards to the relationship between industrial diversity and firm performance as well as between functional diversity and financial outcomes it was found that the share of the acquired alliance portfolios indeed moderates the just-mentioned relationships. This evidence shows that it is not enough to look at the direct relationship between industrial diversity and firm financial results as well as the direct effect of functional diversity on firm performance because these are dependent on the moderating effect of the indirect alliance portfolios.

(35)
(36)

36 knowledge. Additionally, when looking at the Figure 1 (p. 29), another intriguing observation can be made. It can be seen that dependent on the share of acquired alliance portfolios, the low industrial diversity can provide either low or high firm performance. The difference is tremendous. One could ask why it is so big if in both cases it concerns low industrial diversity. A possible explanation could be that companies with a low share of acquired alliance portfolios have only their own deeply ingrained practices for managing low industrial diversity which appear to be insufficient to be able to benefit from very low level of industrial heterogeneity. On the other hand, companies with a high share of indirect alliance portfolios are able to exert high financial performance at low level of industrial diversity. This can be possibly explained by the fact that the acquired company may have very good means of managing its portfolio and better ideas for recombination of the same or very similar resources. Since the acquisition target is incorporated in the acquirer’s ownership, the acquirer can benefit from these practices. Next to the own way of managing portfolios, it gains the acquired knowledge on how to deal with a low industrial diversity. However, in order to be sure of this explanation, further research should explore this issue.

(37)

37 alliances. Thus, having many indirect alliances implies that a firm has to give a lot of attention to the acquired alliances since their management is complex. At the same time, the direct alliances may suffer from less managerial interest (Laursen and Salter, 2006). In such a way, the advantages of having a portfolio will be destroyed and the firm performance will decrease. All things considered, the share of acquired alliance portfolio has a significant influence on the relationship between alliance portfolio and firm performance. In case of a large share of the indirect alliances the challenge of managing industrially diversified portfolios becomes even greater. Already in the investigation of the direct alliances, scholars (e.g. Noseleit and de Faria, 2013; Sampson, 2007) highlighted that high industrial diversity often poses many problems with exerting benefits since the complexity of knowledge and information sharing becomes too high to be able to derive benefits from it. This study adds by showing that if a firm has a high share of acquired alliance portfolios, it has major problems with managing even slightly increased industrial diversity which results in lessened firm performance. This also applies to the functional diversity. The indirect alliance portfolios are different in their nature and their management demands a lot of effort and commitment which are not easy to reconcile with increased industrial nor functional diversity. The cost effect of portfolios with a high share of indirect alliances may even emphasize the negative effects of having a diversified portfolio.

6 CONCLUSION

The main objectives of this study were to show that it is important and relevant to construct alliance portfolios in terms of direct and indirect portfolios as well as answer the research question whether and how the share of acquired alliance portfolio influences the relationship between alliance portfolio and firm performance.

The findings of this study contribute to the literature stream encompassing changes in the alliance portfolio composition, which is clearly under-researched (Wassmer, 2010). The empirical analysis shows that it is not enough to look at the direct alliance portfolio and its characteristics. An alliance portfolio is a dynamic structure that changes throughout the years not only due to formation of new alliances and termination of others. It is essential to include alliances that a firm takes over in an acquisition since having new partners may significantly change the impact of the portfolio characteristics on firm performance.

Referenties

GERELATEERDE DOCUMENTEN

The cross-case analysis shows that all eight managers mention that macro-economic elements have a significant influence on SOE internationalization. As mentioned by the managers,

The second table refer to Table.5 is the results of regression between the Mediator variables which is gender diversity and financial performances (ROA &amp; ROE). As can

This study extends the literature on home region influence on MNC strategy by exploring the impact of culture, legal- and business environment on the percentage of exports

This research was designed to test a possible relationship between unethical organizational behavior and reputational risk as perceived by the customer, as well as the

To some extent, nationality diversity helps the firm to improve the success rate of the R&amp;D activities because nationality diverse managers can offer more ideas

That is in line with the second hypothesis, where we predicted that the higher the proximity is, or the closer the firms are, the more positively it will influence the

Furthermore, this study suggests that the level of investor protection negatively moderates the relationship between the firm’s inclination to be dependent on

A significant study that demonstrates the barriers that small and medium sized firms face when they try to achieve external linkages for innovation, is that of Nietto and