• No results found

Startup Acquisitions on Incumbents’ Innovation Performance Escaping the Incumbent Curse? The Effect of Master Thesis International Business & Management

N/A
N/A
Protected

Academic year: 2021

Share "Startup Acquisitions on Incumbents’ Innovation Performance Escaping the Incumbent Curse? The Effect of Master Thesis International Business & Management"

Copied!
63
0
0

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

Hele tekst

(1)

Master Thesis International Business & Management

Escaping the Incumbent Curse?

The Effect of Startup Acquisitions on Incumbents’ Innovation

Performance

University of Groningen Faculty of Economic and Business

Supervisor: Dr. H. ul Haq Co-Assessor: Dr. A. Kuiken

Date of Submission: Monday, 18th of January 2021

(2)

Abstract

Even though the amount of research in the field of M&As is tremendous, it is not free of gaps. An innovation-seeking motive of M&As is a comparably fragmentary investigated topic that only offers mixed findings. Yet, gaining external innovation can be an important motive for firms that are lacking internal innovation capabilities but want to secure their market power, i.e. incumbent firms. The following study addresses this gap and examines the effect of startup acquisitions on incumbents’ innovation performance. It is hypothesized that the acquisition of startup firms is positively affecting incumbents’ innovation performance as startups are the epitome of innovation. Further, it is argued that this effect is moderated by incumbents’ cultural willingness to collaborate. Following a quantitative approach, a panel of 38 incumbents acquiring firms between 2010 and 2017 is investigated and a random-effects model is applied. However, no support is found for the derived hypotheses. Instead, the findings mark the importance of incumbents’ internal innovation creation. Generating structures that encourage and nurture internal innovation is therefore imperative.

(3)

Acknowledgments

First of all, I would like to express my gratitude to my supervisor Dr. Hammad ul Haq. I want to thank him for his detailed feedback and constant and reliable support during this whole process. Despite the COVID-19 situation, I could always rely on his full assistance and supervision. I highly appreciate all his effort and time that has flown into this thesis.

Further, I want to thank my parents and my sister for their constant support and encouragement not only during the process of my master thesis but during my whole time as a student. Thanks for always being there and cheering me up.

(4)

Table of Contents

List of Figures, Tables, and Appendices ... v

1. Introduction ... 1

2. Literature Review and Hypotheses Development ... 4

2.1 Resource-Based View ... 4

2.2 M&As and Innovation ... 6

2.3Hypothesis 1: Fit of Incumbents and Startups for Innovation Creation ... 10

2.4 Hypothesis 2: Moderating Effect of Incumbents’ Degree of Individualism ... 13

3. Methodology ... 15

3.1 The Energy Industry ... 15

3.2 Sample and Data Collection ... 16

3.3 Measurements ... 17 3.4 Statistical Model ... 21 4. Empirical Results ... 23 4.1 Descriptive Statistics ... 23 4.2 Regression Results ... 27 4.3 Robustness Tests ... 30

5. Discussion and Conclusion ... 32

5.1 Implications for Theory ... 32

5.2 Practical Implications ... 34

5.3 Limitations ... 35

5.4 Future Research ... 36

References ... 37

(5)

List of Figures, Tables, and Appendices

Figure 1: Conceptual Model………...14

Table 1: Pearson Correlation; VIF Results and Descriptive Statistics………...26

Table 2: Regression Results based on Random Effects Model……...………...29

Appendix 1: Energy Industry NACE Rev. 2 Codes…...………...………51

Appendix 2: Modification of Hofstede’s Individualism Dimension……….52

Appendix 3: Hausman Specification Test Results………..……..52

Appendix 4: Share of Startup Acquisitions………...52

Appendix 5: Share of Domestic and Foreign Acquisitions………...53

Appendix 6: Absolute Frequency Distribution of Countries……….53

Appendix 7: Histogram Main Industry and Main Industry Codes………54

Appendix 8: Spearman Test for Categorical Variables Acquisition Time and Industry.……...55

Appendix 9: Pearson Correlation with Centered Means....……….………...……..55

Appendix 10: Regression Analysis without † Acquisition Experience and † Firm Age…....…56

(6)

1. Introduction

Incumbent firms often appear to be unassailable and firmly anchored in their markets. Nevertheless, history has provided evidence that even incumbent firms can fail and be replaced. A well-known example of such failure is the photography company Kodak that missed out investing in innovation and therefore omitted the digital trend (Lucas & Goh, 2009). Another more recent example is based on the car manufacturer Tesla that seemed to appear out of nowhere and has meanwhile reached a higher firm value than previous incumbent firms such as Toyota and Daimler (Klebnikov, 2020). Derived from these observations, maintaining a constant innovation process is crucial for firms’ success regardless of their market power and size. It proves once again that innovation is a highly critical factor of firm performance (Franko, 1989).

However, referring to the “incumbent curse” that will be discussed in more detail at a later point, incumbent firms often lack innovation capabilities (Chandy & Tellis, 2000). Because of those lacking internal innovation capabilities, they have to increasingly seek external solutions to gain innovation. While the formation of strategic alliances is considered rather unsuitable (Lavie, 2006), M&As for innovation reasons have been indicated to be a more successful means (e.g., Andersson & Xiao, 2016; Sevilir & Tian, 2012; Zhao, 2009). However, other authors limit this assumption (e.g., Steigenberger, 2017; Ahuja & Katila, 2001; Bloningen & Taylor, 2000) or deny it entirely (e.g., Haucap et al., 2019; Cassiman et al., 2005). In sum, there is no consensus about the suitability of M&As for innovation reasons. Furthermore, only little research is done in this field (Zhao, 2009). Instead, the vast majority of M&A literature either focuses on the creation of market power (e.g., Renneboog & Vansteenkiste, 2019; Caiazza & Volpe, 2015) or economies of scale and scope, therefore synergies (e.g., Kiymaz & Baker, 2008; King et al., 2004; Larsson & Finkelstein, 1999). Nevertheless, retrieving clearer results concerning the innovation success of M&As for incumbents is of great importance, especially in present times characterized by change.

(7)

market and gain market power (Flauger, 2020). To defend their position in the market, incumbent firms are required to adapt and re-innovate their existing business models and products (Costa-Campi et al., 2014).

Derived from the overall question if M&As are a suitable source for innovation, it is further of high interest which M&A targets are most promising to increase incumbents’ innovation. The limited amounts of research investigating the link between M&As and innovation have, to a great extent, focused on M&As between large and publicly traded firms (Veugelers, 2006). Acquisitions of startups have been, with a few exceptions (e.g., Andersson & Xiao, 2016), neglected in previous research. Yet, startups seem to be an auspicious source for external innovation and can also be found in practice (e.g., Korosec, 2019; Eckert, 2019).

First of all, startups can provide access to new technologies (Andersson & Xiao, 2016). Incumbent firms can add these somewhat tangible resources to their pool of existing ones and therefore increase it (Ahuja & Katila, 2001). This, in turn, offers more possibilities for resource combinations, which ultimately might support the firm in creating new assets that help improving its competitiveness (Grimpe, 2007; Barney, 1991).

Moreover, startups provide intangible resources such as patents and human capital (Kwon et al., 2018). Also, startups are in particular characterized by their organizational structures and their “out of the box” creativity, which are assumed to be responsible for their high innovation output (Freeman & Engel, 2007). Through the acquisition, incumbent firms might benefit from these characteristics if they can successfully integrate certain processes and innovation triggering characteristics in their business. Following this thought, startup acquisitions could potentially support the incumbent itself in becoming more innovative and developing a more innovative internal environment.

(8)

In an attempt to connect the previous thoughts and address the identified gaps, this research aims to investigate the following question:

Do startup acquisitions improve the innovation performance of incumbent firms? Based on the resource-based view (RBV), this thesis proposes that a focus on startup acquisitions is positively affecting incumbents’ innovation performance. It argues that incumbent firms acquire startups to gain tangible as well as intangible resources that support their innovation performance, and thus help maintain or create a competitive advantage. Furthermore, the thesis claims that this relationship is moderated by incumbents’ cultural willingness to collaborate. To test these hypotheses, a quantitative approach is used. Secondary data from 38 incumbent energy firms acquiring between 2010 and 2017 are evaluated, and a random-effects model is applied for the panel. While no support is found for the postulated hypotheses, the results still provide relevant theoretical and practical insights that are of further use.

(9)

2. Literature Review and Hypotheses Development

2.1 Resource-Based View

The RBV represents one of the main theories in strategic management literature used to explain a firm’s success (e.g., Boyd et al., 2010; Galbreath, 2005). While theorists previously tended to focus on a firm’s external environment and, therefore, used the market-based view (Porter, 1980), the RBV is shifting the center of attention towards a firm’s internal resources (e.g., Barney, 1991; Wernerfelt, 1984).

Looking back in time, the first approaches of RBV can be traced back to several authors and their respective theoretical constructs, consisting of competitive advantage, capabilities, and resources (Hart, 1995). However, it is Wernerfelt (1984), Barney (1986, 1991), and Peteraf (1993) who mainly shaped the term RBV according to our current understanding (Makadok, 2001). Wernerfelt shifted the attention towards a firm’s internal matters. Following Penrose’s (1959) roots, he emphasized looking at a firm’s resources rather than its products. By doing so, Wernerfelt (1984) also acknowledged the value of intangible assets such as “knowledge of technologies”. In 1986, Barney picked up the idiosyncratic view of a firm’s internal resources and enriched it by focusing on creating a competitive advantage through the firm’s internal resources. For the context of this research, a competitive advantage is the “fundamental basis of above-average performance” (Porter, 1985, p. 11). At a later stage, Barney (1991) defined a framework of certain resource characteristics necessary to sustain a competitive advantage. Before, Barney (1991) had outlined resource heterogeneity and immobility as premises for gaining a competitive advantage. In the same course, he defined firm resources as physical assets and organizational processes, as well as knowledge capabilities that the firm must either own or control.

Furthermore, he clarified that not only singular resources are suitable to create a competitive advantage. Instead, one should also consider the different composition of multiple resources, which he labels as “resource bundles”. In this sense, whenever the term resources is used hereafter, this also includes resource bundles.

(10)

improves a firm’s effectiveness and efficiency and, therefore, becomes valuable. To fulfill the second criterion, being “rare”, the resource must not be freely accessible on the market (Barney, 1991). Else, it could be purchased by competitors via a market transaction. As a rule of thumb, Barney (1991) defines a resource as rare if the number of firms that demand that resource exceeds the number of firms that possess that resource. Yet, the fulfillment of only the first two criteria could also result from a first-mover advantage (Barney, 1991). For instance, it is conceivable that a new and multiple demanded technology is valuable and supplied by one singular firm. Consequently, the supplying firm appears to have a competitive advantage. However, to become a sustained competitive advantage and not only be branded as a first-mover advantage, this technology would also have to be “inimitable” (Lippman & Rumelt, 1982; Barney 1986). This implies that it must be impossible for competitors to obtain the technology by copying it. Ultimately, the resource has to be “non-substitutable”. There must be no other equivalent resources that allow a competitor to follow the same strategy (Barney, 1991). Illustrating this by the previous example, there must be no firm that can use a different technology that meets precisely the same purpose as the primary introduced technology. In sum, only the adherence of all four criteria will transform a resource into a sustainable competitive advantage.

Peteraf (1993), in turn, uses formerly established theories and frameworks and condenses these into a new coherent framework she calls ‘resource-based model’. Like Barney’s framework, it consists of four ‘cornerstones’ that must be met to sustain a competitive advantage (Peteraf, 1993). These cornerstones are heterogeneity, ex-post limits to competition, imperfect mobility, and ex-ante limits to competition. However, while there is a resemblance between both frameworks, differences between the authors’ work remain. Peteraf (1993) focuses on creating rents and therefore uses price-theory as the starting point of her theories (Foss & Knudsen, 2003). The practical application of the model is outlined for both the “single-business strategy” and the “multi-business corporate strategy” (Peteraf, 1993).

(11)

light of this. The two authors shift the reader’s attention from the focus on present success towards success in the future. In line with Barney, they argue that change is inevitable (Prahalad & Hamel, 1994). Instead of reacting to change, they encourage managers to act proactively (Prahalad & Hamel, 1994). As a result, Prahalad and Hamel (1994, 1990) outline that a firm must concentrate on its core competencies (Prahalad & Hamel, 1990). Thus, they portray core competencies as the firm’s real source of a sustained advantage. By doing so, they go one step further than Barney (1991, 1986) and Wernerfelt (1984), who acknowledge intangible assets next to tangible assets but do not specifically focus on these.

Furthermore, the importance of other firms for the creation of a competitive advantage has been expressed. Lavie (2006) builds upon Barney (1991) and Wernerfelt (1984) but focuses on competitive advantages that are created through the joined efforts of multiple firms. With this, he shifts the unit of analysis from singular firms to alliances between firms. Furthermore, Lavie (2006) argues that Mergers and Acquisitions (M&As) might be more effective than alliances if a firm seeks access to complementary resources. This is because within alliances both partners keep their independent status and thus might not uncover their business entirely to the other party (Lavie, 2006). Yet, the general idea of considering other firms for creating a competitive advantage is not new and has been expressed before “Mergers and acquisitions provide an opportunity to trade otherwise non-marketable resources and to buy or sell resources in bundles.” (Wernerfelt, 1984, p. 175).

2.2 M&As and Innovation

(12)

extension (e.g., Kiymaz & Baker, 2008; King et al. 2004; Larsson & Finkelstein, 1999; Sirower, 1997; Hall, 1988). The link between M&As and innovation, on the other hand, is often mentioned, but only little empirical research has been conducted within this field (Zhao, 2009; Cassiman et al., 2005; Chesbrough, 2003; Ahuja & Katila, 2001). This impression is further reinforced by Schulz (2007), claiming that the two motives are generally investigated independently from one another. Considering that innovation and the associated innovation process are often viewed as key resources to create and maintain a competitive advantage (e.g., Schulz, 2007; Dezi et al., 2018), retrieving meaningful results within this topic would, therefore, be extremely interesting. Hence, a sole focus on the suitability of M&As as an external knowledge source is more than reasonable.

Innovation is defined as the “practical application of an invention or a discovery to a process, product or service that ensures better results for the company, having a good impact on its competitiveness and long-term success.” (Dezi et al., 2018, p. 717). According to Ozcan (2016), M&As can either be conducted for complementing reasons or to substitute internal innovation. Hussinger (2010) supports this view by stating that acquisitions can be seen as an enhancement of existing capabilities, therefore as a complement, or means for technological diversification, which corresponds to the substitutional view. Grimpe (2007), for instance, focuses on the complementing view and argues that newly acquired knowledge can complement the acquirer’s existing resources. Furthermore, by merging two companies into one, synergies might be obtained (Rhodes-Kropf & Robinson, 2008). Ultimately, binding the complementary assets within the newly merged entity can not only lead to the improvement of already existing ones but also to the creation of new assets (Rhodes-Kropf & Robinson, 2008; Grimpe, 2007). This is because the amount of existing assets, including knowledge capabilities is more significant for two firms. Assuming a successful integration process, there are more possibilities for recombination (Ahuja & Katila, 2001). In turn, this might yield a new competitive advantage (see Barney, 1991).

(13)

and process innovation performance (Chesbrough, 2003). Moreover, M&As are not only claimed to be a crucial part of an acquirer’s corporate strategy but also of the target’s strategy (Andersson & Xiao, 2016). Referring to startup targets, Andersson and Xiao (2016) argue that startups initially aim to be acquired as they face, e.g. financial shortages.

However, while the two motives of complementing and substituting have been introduced distinctly from each other in the first place, one should bear in mind that M&As are a “multifaceted phenomenon” (Larsson & Finkelstein, 1999), and their motives are often interconnected (e.g., Andersson & Xiao, 2016). For this reason, the distinction is no longer maintained for the subsequent overview of previous findings.

First of all, the results concerning innovation success through M&A activities are relatively mixed. Bloningen and Taylor (2000) find that firms lacking research and development (R&D) intensity are generally more likely to conduct M&As for innovation reasons. Building upon that, Zhao (2009), and Sevilir and Tian (2012), confirm that M&As are a successful external source to improve a lack of innovation capabilities. Sevilir and Tian (2012) consider targets in addition to acquirers and find that the value creation through M&As holds most pronounced true for those firms that target a comparably more R&D intensive firm. Acquiring firms might benefit from adopting existing technologies or the usage of established patents (Sevilir and Tian, 2012; Hagedoorn & Duysters, 2002). Furthermore, Sevilir and Tian (2012) investigate the impact of the number of conducted M&As and, therefore, add a quantitative dimension. Here, they find a positive link between the number of conducted M&As and innovation, measured by new patents. A potential reasoning for this finding could be the establishment of absorptive capacity that allows the acquirer to routines that might apply to M&As in general (Cohen & Levinthal, 1990).

(14)
(15)

2.3 Hypothesis 1: Fit of Incumbents and Startups for Innovation Creation

Several industries, such as the energy industry, are characterized by the occurrence of incumbent firms. Incumbent firms have developed an established position in the market, which can further be reinforced by high barriers for new entrants. Such high barriers can, for instance, be cost-intensive assets (Costa-Campi et al., 2014). Not considering regulated markets and drawing on RBV instead, incumbents must possess a competitive advantage in the first place to gain such a favorable position (Barney, 1991). However, considering incumbents’ secured position within the market, their incentive to further improve and develop new resources is rather low compared to companies that must continually prove themselves to remain in the competition (Cho et al., 2016). Further, incumbent firms usually concentrate on their existing products if they still provide profits (Jiang et al., 2011). On the contrary, investments in future demands are connected with substantial opportunity costs (Jiang et al., 2011) and, therefore, often disregarded (Cho et al., 2016). Consequently, incumbents’ innovation levels tend to be low (Nanda et al., 2000). Moreover, this is reinforced by the size of incumbents. As most incumbent firms are comparably large (Chandy & Tellis, 2000), they are less agile due to a higher number of hierarchical structures (Shimizu & Hitt, 2005). These are likely to lead to organizational inertia and slower decision-making processes, ultimately hindering and decelerating innovation (Shimizu & Hitt, 2005). Consequently, changes coupled with innovation are likely to require much time and effort within an incumbent firm (Greve, 2011; Baregeh et al., 2009). Overall, the phenomenon of entrenched processes, routines, and low levels of innovation is often observed within incumbent structures and is therefore referred to as the “incumbent curse” (Chandy & Tellis, 2000). Yet, seeking further innovation and maintaining a constant innovation process is also pivotal for incumbents.

(16)

a little time to react. In sum, even though incumbents’ resources might be valuable, rare, inimitable, as well as non-substitutable, and provide, according to RBV, a sustained competitive advantage, this status is not guaranteed for an unlimited time. In this regard, the term sustainable must not be equated with indefinite or permanent.

Connecting the thoughts of incumbents’ lacking innovation capabilities and their constant need for innovation, M&As as a potential source of external innovation are of high interest. Focusing on suitable M&A targets in the next step, it must be noted that, each firm regardless of its size, theoretically possesses the ability to serve as an innovation source (Feldman, 1994). Notwithstanding, M&A activities between large firms tend to be complicated (Bruner, 2005). On the other hand, small firms seem to have a greater potential of being a suitable external innovation source. To begin with, the power distribution in acquisitions is more defined in comparison to mergers. As the incumbent firm acquires the smaller entity, it should automatically subordinate and adapt to the incumbent’s strategic vision (Epstein, 2005). Consequently, more resources can immediately be used for innovation purposes instead of integration endeavors (Ahuja & Katila, 2001). Besides, this implies fewer managerial resources (Hitt et al., 2009). In sum, the integration of a small firm in an established business construct decreases the risk of failure and tends to be easier than a merger between two equal firms (Kwon et al., 2018).

(17)

Furthermore, through the acquisition incumbents can also take over startups’ intangible assets. Those do include patents and specific knowledge of human resources. Picking up on the previous example from the energy sector, incumbents can make use of acquired knowledge in the field of hydrogen and combine it with existing resources, such as depleted gas fields where hydrogen can be stored (Tarkowski, 2019). In this case, the acquired resources would correspond to a rather complementary character, and existing resources would be enriched with newly acquired ones. Hence, new resource bundles would be created.

Finally, incumbents can also use startups’ intangible resources that promote their general innovation capabilities and processes. In general, startups set themselves apart from other companies because of their entrepreneurial spirit and organizational capabilities to innovate (Kwon et al., 2018; Freeman & Engel, 2007). They are characterized by loose structures and minimal levels of bureaucracy (Gulati, 2019; Freeman & Engel, 2007). Employees' vague ideas are encouraged, and novel ideas are rewarded (Freeman & Engel, 2007). This triggers startups’ creativity and “out of the box” innovation (Gulati, 2019; Freeman & Engel, 2007). Through the acquisition of such firms, incumbents might be able to adapt and integrate the targets’ innovation promoting characteristics. Ultimately, they might be able to create an innovation promoting environment and to “escape” the incumbent curse.

Startups’ suitability as acquisition targets is supplemented by the fact that they ordinarily face difficulties with the upscaling process of their innovations, possess limited financial resources (Krishna et al., 2016), and are lacking an awareness level within the market (Freeman & Engel 2007). Being aware of these shortages, they often aim to be acquired (Andersson & Xiao, 2016). On the contrary, incumbent firms possess the financial resources to acquire startups, to implement the innovation (-resources) in their business, and scale them up (Andersson & Xiao, 2016). As the firms theoretically complement each other’s resources, one can speak of a “David-Goliath symbiosis” (Baumol, 2002).

To sum up, startup acquisitions can provide crucial resources for incumbent firms that are lacking innovation capabilities. Through their acquisition, incumbents can access tangible as well as intangible assets that can be redeployed and might promote their innovation capabilities. Derived from RBV, these newly acquired resources can ultimately support incumbent firms to sustain an existing competitive advantage or even create a new one. Therefore, this thesis posits:

(18)

2.4 Hypothesis 2: Effect of Incumbents’ Cultural Willingness to Collaborate

Generally, M&As pose the opportunity to be successful external sources for innovation (e.g., Sevilir & Tian, 2012; Zhao, 2009) and might help firms prone to the incumbent curse. The suitability of startups as acquisition targets, including deploying their resources in combination with incumbents’ resources, has been discussed. However, except for startups’ small size, factors that influence a successful integration have not been focused on to this point. Yet, a successful integration has proven to be crucial for the outcome of M&As (Steigenberger, 2017). A vast amount of literature recognizes culture to be among the main drivers of M&A success (e.g., Datta & Puia, 1995; Birkinshaw et al., 2000; Brouthers & Brouthers, 2000; Angwin, 2001; Gomes et al., 2011). Drawing on the thought that companies’ cultural traits most likely affect the acquisition outcome, this study suggests that the incumbent’s firm culture is pivotal for the target’s successful integration.

To begin with, the acquiring incumbent firm is larger than the startup (Kwon et al., 2018). As previously elaborated, power distribution is therefore pre-determined (Epstein, 2005). Since the power comes from the incumbent firm, the respective one will most likely set the integration process frame. As the acquired knowledge base is further overall smaller than the incumbent’s one, most of the incumbent’s processes and structures are most likely maintained (Ahuja & Katila, 2001). Ultimately, this thesis refers to the fact that startups aim to be acquired (Andersson & Xiao). Consequently, they are likely to possess an intrinsic willingness to collaborate and to align with the acquirer to achieve a successful integration. For these reasons, the focus is set on the remarkably “bigger share”, i.e. the incumbent firm.

Before elaborating further on the incumbent’s cultural traits, an overview of the term collectivism is provided as it is important for the subsequent argumentation. Collectivism refers to an environment where members perceive themselves as being interdependent. Collectivists emphasize the well-being of the group and subordinate their self-interests. Further, collectivistic members have a strong sense of belonging and aim to share and cooperate to preserve harmony (Morosini & Singh, 1994; Triandis, 1993).

(19)

(Morris et al., 1994). This might also decrease potential frictions. Besides, collectivistic tendencies promote information and knowledge sharing (Michailova & Hutchings, 2006). Within collectivistic groups, individuals are treated equally (Morris et al., 1994). If the “new members” perceive to be part of the group, they might be more eager to share their information and tacit knowledge with the acquirer. In turn, this enhances the understanding of the resources, which is extremely important to deploy and recombine them to increase the incumbent’s innovation output ultimately. The importance of effective information flows is further emphasized because the targets likely stem from sectors that are not directly related to the incumbent’s one (Kline & Rosenberg, 1986).

While characteristics of collectivism are used within the former explanation, the term “cultural willingness to collaborate” relates to the general characteristics of collectivism. As it seems to capture the actual content of that explanation more intuitively, cultural willingness to collaborate is used hereafter.

This study concludes that incumbents’ cultural willingness to collaborate increases the likelihood of successfully integrating acquisition targets. It allows for greater collaboration between the involved parties and can lead to the full exhaustion of incumbents’ increased resource base. Hence, cultural willingness to collaborate can support incumbents to increase their innovation performance. The resulting hypothesis postulates:

Hypothesis 2: Incumbent firms’ cultural willingness to collaborate positively affects startup acquisitions and, ultimately, the resulting innovation performance.

The complete research model is depicted in figure 1.

(20)

3. Methodology

3.1 The Energy Industry

The previously developed hypotheses are based on the theory of RBV. The respective research investigates a specific question and follows a deductive approach. Therefore, a quantitative approach is applied (Silver et al., 2012). Using a quantitative method further allows processing larger amounts of data, which might increase the model’s overall power. Also, this approach is considered to be more objective (Basias & Pollalis, 2018).

This study investigates a sample of incumbent firms that stem from the energy industry for several reasons. For most of the time energy has been produced by using coal, petroleum, and natural gas, which are allocated to the group of fossil fuels (Zou et al., 2016; Kumar, 2013). Fossil fuels have the advantage of not only being cheap but also abundant (Johnsson et al., 2019; Kerr & Service, 2005). Firms within the energy industry have made high investments in assets that align with the usage of fossil fuels (Costa-Campi et al., 2014). Hence, their interest in further deploying these assets is considerably high, whereas innovation rates tend to be low. Besides, the high investments accounted for a relatively small number of actors in the market (Costa-Campi et al., 2014). Consequently, firms within the energy industry can be categorized as incumbent firms to which the incumbent curse of low innovation capabilities applies. However, fossil fuels have been proven to account for greenhouse gases and are causing environmental pollution and climate change (IPCC, 2018). In the course of a “sustainable evolution”, the energy industry is therefore forced by, inter alia, newly established laws (Ruggiero et al., 2015) and scrutinizing consumer perception (Johnsson et al., 2019) to develop more carbon-neutral approaches.

(21)

3.2 Sample and Data Collection

This study is using secondary quantitative data retrieved from several databases to test the previously elaborated hypotheses. Zephyr database is used to collect information about acquisition activities within the energy industry. Further, firm data, including financials, are collected from Orbis and WRDS database. Zephyr and Orbis have been used for similar research questions before (e.g., Bauer & Matzler, 2014; Moretti & Biancardi, 2020) and are both operated by Bureau van Dijk. As they provide numerous amounts of data (Bureau van Dijk, 2020a, 2020b), both are considered to be sufficiently comprehensive. Additionally, annual reports cover missing information about employees and R&D expenses of the respective firms. Further, each acquiring firm’s cultural score is collected from the internet source Hofstede (2020). Following, the collected data are processed with the statistical program Stata/SE 16.0. The overall investigated time frame for answering the research question is from 2010 until 2019. Growing numbers of M&As can be observed in most recent years (Institute for Mergers, Acquisitions, and Alliances, 2020). Furthermore, the development of sustainable solutions has been increasing, and firms are engaging more strongly in reducing greenhouse gas emissions and becoming more sustainable (e.g. Ruggiero et al., 2015). Initially, this movement has been triggered by previous events such as the Rio Conference in 1991 and the Kyoto Protocol in 2005 (Torney, 2014). However, the development has been further reinforced by the German “Energiewende” in 2010, the Paris Climate Agreement in 2015, and a shifting focus of the population towards sustainable solutions (Johnsson et al., 2019). Therefore, the latest M&A activities are considered to be most suitable for answering the research question.

In the first step, industries that can be assigned to the energy sector are classified. To do so, the industry categorization NACE Rev. 21 is used to identify suitable acquirers. This approach is in line with other authors (e.g., Costa-Campi et al., 2014) and is chosen for two reasons. First, compared to the NAICS industry categorization, more observations can be derived. Second, the majority of incumbent firms within the sample stem from Europe, which makes it reasonable to choose the European classification codes. The list of energy classified industries can be retrieved from Appendix 1. Further, the focus is on acquisition activities only as incumbent firms dominate and acquire startups. Besides, only fully completed acquisitions are

1 NACE is the statistical classification system for economic activities. The term is derived from “Nomenclature

(22)

investigated. From a geographical perspective, all activities are included to get a comprehensive overview. Applying these steps, the initially derived dataset includes 22,984 observations. Hereafter, the acquiring firms are assigned to their parent firm by using Orbis database. Next, the criterion for incumbents is included. Hereafter, four time frames are created: 2010-2011, 2012-2013, 2014-2015, 2016-2017. That is because the majority of incumbent firms do not acquire constant over each year. However, in line with other authors (e.g. Ward, 2016) it is aimed for a balanced panel data set. Hence, an incumbent must acquire at least one company within each time frame to remain in the dataset. Creating time frames and summarizing the acquisitions for two years therefore leads to a higher number of observations that can be kept in the model. The dataset derived from this includes 3.796 observations. However, multiple firms’ financial data cannot be found in the databases or published annual reports. Hence, these corresponding observations also have to be excluded. This leads to the strongly reduced number of 152 observations, including 38 incumbent firms that continuously acquire over the four time frames.

3.3 Measurements Dependent Variable

(23)

Besides, it must be mentioned that the perceptible effects of M&As often occur delayed (de Man & Duysters, 2005). In line with previous research (e.g., Greve, 2003; Hagedoorn & Cloodt, 2003), the dependent variable is, therefore, lagged by one year.

Independent Variable

To measure the independent variable, i.e. startup acquisitions, a ratio of startup acquisitions divided by a firm’s overall acquisitions is created. Using a ratio provides the benefit to decrease the distortion that might be provoked by singular firms that differ strongly in their absolute numbers of acquisitions. Else, in line with Delgado et al. (2020), incumbent firms are defined as companies older than eleven years at the point of acquisition. However, it must be noted that there is no universal definition for incumbent firms and researchers strongly vary in their definitions for incumbent firms (e.g., Santarelli & Tran, 2012). On the other hand, firms that are younger than six years are classified as startups (Sauermann, 2017; von Gelderen et al., 2000). As startups are further often characterized by their small size (Cockayne, 2019), this criterion is added. Regardless of their age, firms that have less than 100 employees when being acquired are, therefore, be categorized as startups as well (Sauermann, 2017).

Moderator Variable

The moderating variable cultural willingness to collaborate is captured by the Hofstede’s dimension of individualism versus collectivism. While one might argue that Hofstede’s scores are somewhat applicable to investigations on a national level, several authors have introduced the idea of national culture affecting firm culture (e.g., Chen et al., 2017).

Even though the characteristics of cultural willingness to collaborate can also be found within cultural dimensions of other authors (e.g., Schwartz, 2006), drawing on Hofstede is reasonable. His cultural dimensions are among the most commonly used in international business research (Tung & Verbeke, 2010). Also, they have repeatedly been verified and proven to be contemporary (Beugelsdijk et al., 2015). Besides, Hofstede’s scores have been determined for the broad majority of countries (Hofstede, 2020). Yet, it must generally be noted that Hofstede’s cultural scores are derived from an aggregated approach, meaning that individuals are characterized by a uniform cultural score (Hofstede, 2011). This can lead to an “ecological fallacy” (Hofstede, 2001, p. 16).

(24)

(Hofstede, 2021). Bearing in mind that it is a metric scale, the scale is turned around as this allows an easier interpretation of the results regarding collectivism. An example to illustrate this approach is provided in Appendix 2.

Control Variables

As innovation performance might be affected by other aspects, several control variables are added to control for firm and industry-specific effects. However, it must be pointed out that there is no limit for control variables, and this selection is not exhaustive (Entezarkhei & Moshiri, 2018).

First, the model entails the control variable acquisition experience. Multiple authors have investigated acquisition experience as it is likely to influence the acquisition outcome positively (King et al., 2004; Kusewitt, 1985). That is because firms are expected to develop certain organizational routines that simplify the integration of a foreign firm (Haspeslagh & Jemison, 1991). Through organizational learning, elements of a standardized integration procedure might be established. These, in turn, might make it easier for the acquiring company to familiarize with the target’s resources, including its human resources, to redeploy them and hence to increase innovation performance eventually (Chao, 2018). Hayward (2002) is measuring a firm’s acquisition experience as the total number of recent acquisitions undertaken by a firm. Yet, considering that the incumbent firms within the sample were founded many years ago, this measurement seems unsuitable. Firms are likely to develop skills and capabilities constantly with each acquisition. As Zephyr database does not provide M&A information before 1997, acquisition experience that has been gained before would be neglected. For this reason, acquisition experience is measured by an incumbent’s total number of firms within the group. This is reasonable because the management of multiple companies within a group is likely to require the development of similar skills and capabilities as for the conduction of M&As (Chao, 2018; Gupta & Govindarajan, 1991). Further, incumbents have likely developed problem-solving abilities, which might also be applicable to work aligned with the target’s employees. As the number of companies in a group strongly varies among incumbent firms, the variable is log-transformed to reduce its skewness (Hussinger, 2010).

(25)

Audretsch (1987) point out that only firms of a certain size invest in innovation to gain or maintain market power. In line with this, Kim et al. (2012) indicate that the insufficient financial resources of small firms are a barrier to innovation investment. In contrast, larger firms might be privileged as they tend to possess high financial resources (Andersson & Xiao, 2016) and therefore, can invest in innovation. In line with acquisition experience, this variable is log-transformed as well.

Also, the effects of firm age are investigated. Shimizu and Hitt (2005) find that a firm’s age is likely to entail information about its ability to successfully conduct business and, therefore, its ability to innovate. Besides, firms’ financial resources are related to their age (Hamilton. 2012). Hence, firm age would positively affect innovation performance. As the acquisition is measured in time frames, a firm’s average age per time frame is chosen. To account for skewness, the firm age is log-transformed as well.

Further, the acquisition time is used as a control variable. The investigated time includes the years 2010 until 2019. During this period, various events, e.g. in the economy or in politics, such as the oil spill of “Deepwater Horizon” have happened (Smith et al., 2011). As these events are likely to lead to incumbents’ reducing their R&D expenditures and therefore their innovation performance (Archibugi et al., 2013), the model must be controlled for the time of acquisition. The control variable is measured by one of the four time frames to which the respective acquisition can be assigned to.

(26)

3.4 Statistical Model

A panel data model is used to test the effect of startup acquisitions on an incumbent’s innovation performance. Panel data, which can also be referred to as longitudinal data (Grill, 2017), are used when observing a group of individuals repeatedly over time. Therefore, it includes a cross-sectional and time-series dimension (Frees, 2004; Hsiao, 1985). Implying large amounts of data, panel data models provide the benefit of increasing the econometric estimates’ explanatory power (Hsiao, 1985). Besides, less collinearity can be found between the variables, and there is more variation (Elhorst, 2003). Furthermore, predictions within panel data models are more accurate as the large pools of data lead to additional degrees of freedom, which increases the power of the test (Brooks, 2019, p. 527; Hsiao, 1985). The panel of this study is created based on the various firms (cross-sectional dimension) and the four time frames (time-series dimension). It is declared to be balanced because there are no missing values (Yaffee, 2003). Investigations for 38 firms over four time frames lead to 152 observations in total. Compared to other panel data models (e.g., Malen & Vaaler, 2017), it has to be acknowledged that this is relatively small, which might reduce the model’s power (Yaffee, 2003).

Within panel data models, one can distinguish between different analytical models (Yaffee, 2003). First, one has to differentiate between homogeneous and heterogeneous models. Homogeneous models imply no dependence within entity groups, whereas heterogeneous models allow parameters to vary across entities (Hurlin, 2018). As it is very likely that the sample includes unobservable characteristics that vary among the entities, it is decided in favor of a heterogeneous model. Among the heterogeneous models, either a model with fixed effects or random effects can be used (Clark & Linzer, 2015). Fixed effects refer to all regressors allowing endogeneity and individual effects (Baltagi et al., 2003). This means that fixed-effects models can allow the intercept within the regression model to either differ cross-sectionally, to differ over time or even to differ over both cross-section and time (Yaffee, 2003). Within random-effect models, on the other hand, unobserved heterogeneity evokes a correlation (Gardiner et al., 2009). In other words, the “variation across the entities is assumed to be random and uncorrelated with the predictor or independent variables included in the model” (Anna et al., 2014, p. 234).

(27)

should be used. Else, one should opt for a random-effects model (Baltagi et al., 2003). The Hausman test results for this sample can be retrieved in Appendix 3. Deriving from the test results, the null hypothesis cannot be rejected as the respective p-value is 0.310. Thus, a random-effects model is applied. Using a random-effects model poses the significant benefit that time-invariant variables such as industry can be kept as a regressor in contrast to fixed- effects models where time-invariant variables have to be omitted (Yaffee, 2003).

After deciding in favor of a random-effects model, four models are applied hierarchically to test the hypotheses. The first regression model solely includes the dependent variable and the chosen control variables. In model 2, the independent variable, startup acquisitions, is further added. Model 3 includes the moderator variable, a firm’s cultural willingness to collaborate for its own. To test if the moderator variable has the suggested impact on startup acquisitions, model 4 entails the interaction term that is determined as startup acquisitions multiplied with cultural willingness to collaborate. Furthermore, each model uses standard robust errors to account for potential correlations between the residuals, i.e. heteroscedasticity issues (Millo, 2014).

Besides, dummy variables have to be created for the categorical variables industry and acquisition time in order to interpret the results within the regression analysis. As firms within the sample are operating in ten industries (Appendix 7), ten binary dummy variables are produced. Industry 10, i.e. manufacturing and production of goods, serves as the “base scenario” as most of the investigated incumbents operate in that respective industry. On the other hand, four dummy variables are created for the four acquisition times. Here, the acquisition time 2016-2017 is chosen as the base scenario. Yet, each acquisition times could have served as the base scenario as they have the same frequency.

(28)

4. Empirical Results

4.1 Descriptive Statistics

For the subsequent results section, this thesis begins by describing the main characteristics of the sample, including its descriptive characteristics and correlation analysis. Afterward, the regression analysis results are presented. Ultimately, several robustness tests are conducted to verify the derived results.

The sample consists of 38 energy incumbents who have acquired 765 firms from 2010 until 2017. Of these 765 acquisitions, 371 can be classified as a startup acquisition, which equals roughly 50% (Appendix 4). Furthermore, 430 acquisitions are categorized as domestic ones, meaning that the acquirer’s home country equals the target’s home country (Appendix 5). In reverse, 335 acquisitions are foreign. An incumbent’s share of startup acquisitions per time frame is roughly 50% (Table 1). Yet, it shows that the share can vary strongly as the respective standard deviation is 33%.

Furthermore, the sample represents acquiring companies from 15 different countries that are placed all over the world. However, the majority of incumbent firms are located within Europe (Appendix 6). Besides, it is apparent that several countries are relatively often represented within this sample: Eight incumbent firms are based in France, seven are located in the US. Further, Japan and Great Britain are both represented four times. On the other hand, eight countries are only represented once within the sample (Appendix 6). Besides, the incumbent firms are operating in ten industries, with the majority being active within the manufacturing and production of goods sector (Appendix 7). Finally, it must be noted that the descriptive characteristics of the control variables acquisition experience, firm size, and firm age are log-transformed. Therefore, they are not further considered.

(29)

and several moderate significant correlations can be found within this sample. They are significant on a 5% level. A strong positive correlation (0.749) appears between the incumbent’s firm age and its respective acquisition experience. Bearing in mind that the acquisitions experience is measured as firms within an incumbent’s group, it can be assumed that older firms possess higher numbers of firms within their portfolio. Incumbents have established their position over many years and therefore, must have secured their market power. Potential ways to do so is by, e.g. diversifying into different business segments or entering new geographical markets (Montgomery, 1985; Tallman & Fladmoe-Lindquist, 2002). In both cases, this is likely to include foundations of new companies and acquisitions of existing ones, ultimately enlarging the corporate group. However, this explanation appears limited, considering that there is no significant correlation between firm size and firm age.

Further, firm size also positively correlates with a firm’s acquisition experience. This relationship does not seem astonishing as it is likely that the number of the incumbent’s employees increases with the number of firms within a corporate group. Surprisingly, this correlation is rather moderate (0.409).

Besides, there is another moderate but negative correlation between firm size and an incumbent’s cultural willingness to collaborate (-0.406). Bearing in mind that cultural scores are relatively stable (Kostis et al., 2018), a potential reasoning should be derived from the respective one. Further, one has to remember that collectivism and individualism are opposing (Hofstede, 2011). Therefore, the negative score for collaboration equals a positive score for individualism. Referring to Ezcurra (2020) individualism has a positive impact on political stability. In turn, firms that stem from countries with stable and strong political institutions might be more incentivized to invest in firm growth, which could lead to larger firm size (Boubakri et al., 2015).

(30)

Besides, not applying centered means for multicollinearity issues provoked by interaction terms is further supported by various authors (e.g., Allison, 2012; Echambadi & Hess, 2007). Referring back to the analysis results, the high correlation between startup acquisitions and firm age also raises concern for multicollinearity. For this reason, an additional VIF test is conducted to examine if certain control variables should be omitted for the subsequent regression analysis. The results derived from the VIF test are depicted in Table 1. While all scores are below the widely applied threshold of 10 (e.g., Midi et al., 2010; Chatterjee & Hadi, 2012, p. 250), it should be noted that there is no general agreement on a standardized score to rule out multicollinearity (Thompson et al., 2017). Jager et al. (2018), for instance, use the lower threshold of 2.5 to rule out multicollinearity. Applying the more conservative threshold of 2.5, the variables acquisition experience (3.11) and firm age (2.63) raise concern. For this reason, several regression models are run to ensure that their potential multicollinearity does not distort the subsequent regression analysis. One time, all variables are included (Table 2), one time acquisition experience is omitted and, lastly both acquisition experience and firm age are omitted (Appendix 10). However, it shows that the differences in excluding the variables acquisition experience and firm age are infinitesimal. Therefore, both control variables are kept in the model.

(31)

Table 1: Pearson Correlation; VIF Results and Descriptive Statistics.

Variables (1) (2) (3) (4) (5) (6) (7)

(1) Innovation Performance 1.000

(2) Startup Acquisition -0.091 1.000

(3) Cult. Willingness to Collab. 0.243* -0.042 1.000

(32)

4.2 Regression Results

To test the derived hypotheses, a regression analysis is conducted. Four models are hierarchically applied on each other. The individual results are depicted in Table 2.

Model 1 entails the dependent variable innovation performance and the chosen control variables acquisition experience, firm size, firm age, acquisition time, and industry. Within that model, no coefficient is significant. However, model 1 has an overall r-squared score of 0.391. This means that the chosen control variables can explain 39.1% of the variance within the model. Hence, the control variables still have an impact on the dependent variable innovation performance.

Model 2 shows the results for hypothesis 1, in which a high number of startup acquisitions was claimed to affect a firm’s innovation performance positively. To test this hypothesis, the independent variable startup acquisitions is added to the control variables. Table 2 shows that startup acquisitions have a positive but very low coefficient. However, the effect is not significant and can be provoked randomly. Therefore, hypothesis 1 cannot be supported. Furthermore, the overall r-squared value only increases slightly from 39.1% to 39.4% when adding the independent variable. This shows that the independent variable startup acquisitions does not strongly influence the model’s explanatory power.

(33)

bundled with startup acquisitions. Hence, the moderating effect cannot be supported. Furthermore, the r-squared value decreases slightly to 50.8%, which indicates that the model’s fit is negatively affected by the interaction term. Incumbent’s cultural willingness to collaborate for its own is again significant and negative. Nonetheless, the results should be treated tentatively for the previously elaborated reason of multicollinearity.

Considering the control variables once again, the acquisition time 2010-2011 and the electricity and gas industry become significant on a 10% level in Models 3 and Model 4. For illustrative purposes, the meaning of a categorical variable’s coefficient is explained. The electricity and gas industry has a coefficient of -3.6%. This is to be interpreted so that, on average and holding all other variables constant, firms operating in the electricity and gas sector have an innovation performance 3.6% lower than companies operating in the manufacturing industry (base scenario). However, it is striking that these control variables became significant in Models 3 and 4, which entail the moderator variable cultural willingness to collaborate. Referring back to the multicollinearity issues of the respective variable, it is most likely that these issues cause the controls’ significance. The same holds true for the intercept that also becomes significant on a 1% level. This assumption is also reinforced by its comparably large robust standard error (2.68).

(34)

Table 2: Regression Results based on Random Effects Model.

Variable Model 1 Model 2  Model 3    Model 4   

Startup Acquisitions 0.001 0.001  0.000 (0.46) (0.45) (-0.15) Cult. Willingness to Collab.   -0.001**  -0.001** 

(-2.17) (-2.22) Interaction Term       0.000 (0.85) † Acquisition Experience 0.002 0.002 -0.006  -0.006 (0.21) (0.21) (-0.49) (-0.45) † Firm Size -0.011 -0.010 -0.017 -0.017 (-0.71 (-0.65) (-1.53) (-1.56) † Firm Age -0.002 -0.002 0.000 0.000 (-0.32) (-0.37) (-0.03) (0.06) Acquisition Time 2010-2011 -0.004 -0.004 -0.005*  -0.005*  (-1.46) (-1.42) (-1.78) (-1.77) Acquisition Time 2012-2013 -0.002 -0.002 -0.003  -0.003 (-0.95) (-0.90) (-1.35) (-1.33) Acquisition Time 2014-2015 0.000 0.000 0.000 0.000 (0.14) (0.16) (-0.16) (-0.15) Acquisition Time 2016-2017 (b.s.) omitted omitted omitted omitted Industry I: Mining -0.003 -0.002  -0.020  -0.020

(-0.17) (-0.12) (-0.81) (-0.78) Industry II: Service -0.010 -0.009 -0.019  -0.019 (-0.88) (-0.84) (-1.21) (-1.22) Industry III: Electricity & Gas -0.012 -0.013 -0.036*   -0.036* (-0.84) (-0.94) (-1.66) (-1.65) Industry IV: Water Supply -0.012 -0.011 0.000  -0.001 (-0.81) (-0.78) (-0.02) (-0.05) Industry V: Construction 0.178 0.179 0.168  0.167

(1.28) (1.28) (1.59) (1.59) Industry VI: Wholesale 0.029 0.030 0.022  0.022 (1.62) (1.66) (1.08) (1.07) Industry VII: Transport & Stor. 0.001 0.001 -0.006  -0.006 

(0.05) (0.12) (-0.41) (-0.40) Industry VIII: Information -0.011 -0.010 -0.020  -0.020  (-0.95) (-0.96) (-0.86) (-0.87) Industry IX: Tech. Activities -0.020 -0.018 -0.033  -0.034 (-0.76) (-0.68) (-1.52) (-1.56) Industry X: Manuf. & Product. (b.s.) omitted omitted omitted omitted Constant 0.063 0.061 0.173*** 0.176***

(1.01) (1.00) (2.67) (2.68)

N 152 152  152    152   

r-squared 0.391 0.394  0.509 0.508   

(35)

4.3 Robustness Tests

In the following section, four robustness tests are conducted to confirm the previously derived results. To do so, the measurements of several variables are changed (Lu & White, 2014). The results of the robustness tests are presented in Models 5 to Model 16 within Appendix 11. For the first robustness test, the measurement of the dependent variable innovation performance is changed. Innovation performance has been determined as the ratio of R&D expenses divided by the total revenues. To control if a different measurement changes this study’s outcome, it is referred to Aase (2018) who determines innovation performance as the ratio of sales divided by R&D expenses. Energy incumbents that acquire startups might be able to conduct R&D more effectively, which could ultimately strengthen their resources. Following this thought, a firm that is innovating more effectively should generate more revenue output for the same amount of R&D input. As in the original model, four regressions are run (Model 5 to Model 8). The results show that the new model has an overall lower explanatory power of 27.2%. In line with the original model, startup acquisitions and the interaction term further remain insignificant. Hence, hypothesis 1 and hypothesis 2 find no support.

For the next robustness test, the moderator variable cultural willingness to collaborate is changed. While the previous focus used to be on the acquirer's cultural willingness to collaborate only, the target’s willingness is now also considered. Thus, the underlying assumption that startups aim to be acquired (Andersson and Xiao, 2016) and hence do their best to align and share their resources is challenged. To see if differences in acquirer’s and target’s cultural willingness to collaborate affect the overall outcome, this study follows the approach of Kogut and Singh (1988). The Euclidean Distance of collectivism between the acquirer and each target per time frame is calculated. Afterward, the mean is determined for each time frame. As the overall procedure remains, only the last two models, including the moderator variable and the interaction term, have to be recreated. The results are illustrated in Model 9 and Model 10. According to the r-squared score (0.398), the explanatory power of the model is lower than the explanatory power of the original model. It also shows that the interaction term remains insignificant. Thus, firms’ combined cultural willingness to collaborate has no moderating effect on an incumbent’s innovation performance.

(36)

collectivism dimension, similar characteristics can be found in other authors’ cultural dimensions. As Schwartz's dimension of embeddedness shows similarities to the characteristics of Hofstede’s collectivism dimension (Schwartz, 2006; Gouveia & Ros, 2000), it is used to proxy cultural willingness to collaborate. The results can be seen in Model 11 and Model 12. The interaction term remains insignificant. The model’s overall r-squared value is 0.426. In the last conducted robustness test, the dependent variable innovation performance is changed. R&D as well as incumbent’s revenue are lagged by two years. Within the original model, both R&D and sales were lagged by one year. However, several authors have addressed that acquisitions’ effects might show after a more extended period (e.g., Cloodt et al., 2006). Hence, the largest span possible is investigated, i.e. 2 years. Again, four models (Model 13 to Model 16) are run. However, the results show no noteworthy differences compared to the original model.

(37)

5. Discussion and Conclusion

5.1 Implications for Theory

This study contributes to existing research as it bridges the gap concerning the effects of startup acquisitions on incumbents’ innovation performance. Several authors propose that acquisitions can increase the acquirer’s overall resource pool and, therefore the number of resource combinations (e.g., Sevilir & Tian, 2012; Ahuja & Katila, 2001). According to RBV, this is likely to support the acquiring firms in either sustaining an existing competitive advantage or in creating a new one (e.g., Grimpe, 2007; Barney, 1991). In this context, startups were considered to be an ideal target for incumbent firms suffering from the incumbent curse. However, the derived results of this study lead to the tentative conclusion that a focus on startup acquisitions does not increase incumbents’ innovation performance. This result is in line with the research of several authors (e.g., Haucap et al., 2019; Cassiman et al., 2005; Bloningen & Taylor, 2000) who find that acquisitions do not create value, respectively innovation. Several reasons might explain the outcome of this study.

First, it might be that the acquiring firm cannot effectively use the acquired resources. Ozcan (2016) states that M&As can have a complementing or substituting character. In the case of acquiring complementing resources, it is of even greater importance for the incumbent firm to fully understand the new resources. Else, the acquirer will not be able to recombine them with existing resources to ultimately create innovation (Hitt et al., 2009). However, incumbent firms acquire startups to gain access to foreign resources they do not possess (Zhao, 2009). The acquired startups often have very specific and unique resources and knowledge (Kwon et al., 2018). Hence, the acquired resource base is likely to be extremely different from the resource and knowledge base of the incumbent. As a consequence, the acquirer is probably not capable to understand and deploy it successfully (Steigenberger, 2017). In that case, the larger resource base is not adding value as the incumbent cannot create new resource bundles and respectively innovation (Zhao, 2009; Cloodt et al., 2006; Hagedoorn & Duysters, 2002).

On the contrary, if the acquired resources are strongly related to the incumbent’s existing ones, hence substitutive, they do not provide the potential for innovation (Steigenberger, 2017). In this sense, it is challenging for the incumbent to acquire the right amount of “foreignness”.

(38)

innovation in the long-term. Startups’ intangible resources are also pivotal for their own innovation capabilities and include, for instance, knowledge, creativity, flexibility, and organizational processes (Freeman & Engel, 2007). However, when a startup is acquired, it is dominated by the incumbent firm (Epstein, 2002). Hence, the relatively small target is embedded in the incumbent’s existing structures. Therefore, the startup is likely to adopt the acquirer’s working methods including its structures. Bearing in mind that incumbent firms are characterized by high bureaucracy levels and other innovation harming characteristics (Chandy & Tellis, 2000), startups might not be able to preserve their flexible structures. Instead, they probably diminish or lose their creativity and ultimately their innovativeness (Gulati, 2019). Hence, the startup’s elution from its original environment and the subsequent acculturation negatively affect its innovation capabilities. As a result, the startup might not be able to further create innovation. Depending on how fast the startup is fully integrated and affected by the incumbent curse, the acquirer might not be able to adopt any innovative characteristic and implement it in the overall business.

A contrary potential explanation for the derived findings is also related to the startup’s integration time. Referring to the startup’s small size (Cockayne, 2019) and the incumbent’s large size (Chandy & Tellis, 2000), the integration process of the startup’s resources might be more time-demanding. The startup, has to integrate its innovative characteristics into the incumbent’s existing structures to ultimately increase innovation performance. The overall process might be further decelerated by the fact that change is likely to provoke resistance (Dent & Goldberg, 1999). Therefore, it might take up to several years until innovation effects finally occur and can be observed (de Man & Duysters, 2005).

Furthermore, contrary to the expectations of this thesis, incumbents’ cultural willingness to collaborate seems to have no moderating effect on acquisitions.

(39)

However, cultural willingness to collaborate might impact innovation performance negatively but independently from the acquisition context. Bearing in mind that collectivism and individualism are opposing sides (Hofstede, 2011), this finding is in line with the suggestions of Tian et al. (2018) as well as Nakata & Sivakumar (1996). A potential reason for this is that individualistic tendencies increase the freedom to initiate innovation (Kaasa & Vadi, 2010). Hence, the incumbent firm might be able to create new resources and to increase its innovation performance.

5.2 Practical Implications

The results derived in this study give essential insights especially for managers in industries that are characterized by the occurrence of incumbent firms.

To begin with, incumbents will likely face the need to react to a fluctuating environment. However, the results of this study indicate that startups are no exceptionally promising target to acquire for incumbent firms who seek external innovation. They appear to neither support the recombination of existing resources nor the creation of new ones. With regards to potential power struggles within merger activities (Epstein, 2005), it is questionable if M&As of other firm types can increase incumbents’ innovation performance remarkably. Instead, managers of incumbent firms have to find different ways to adapt to external changes and demands. Therefore, the improvement of internal innovation behavior is imperative. This implies that managers have to build up and further strengthen routines and structures that are improving internal innovation creation. In line with this, they might consider increasing the freedom to initiate innovation within the firm. Also, they might examine allocating their financials differently and increasing their investments in internal R&D activities and projects.

Referenties

GERELATEERDE DOCUMENTEN

The results show that CEO turnover is significantly related to stock price performance, while board of management turnover (excluding CEO) is related to accounting

Master Thesis – MSc BA Small Business & Entrepreneurship.. University

In this research paper, three hypotheses were tested by examining the relationship between the use of big data and firm performance, and the interaction effect that

The organizational learning perspective is used to examine how accumulated prior experience of internal acquisitions, acquisition programs and experience of other firms may

A case study found that an overall decline in innovativeness and creativity was felt under a psychopathic CEO (Boddy, 2017), and the literature review illustrates

Future research, exploring the moderating influence of the target firm’s alliance management capabilities on the relationship between acquisitions with or without

The authors argue that technological and non-technological innovations should not be viewed as substitutes, but rather as complimentary to each other, suggesting

On the other hand, I found that the acquiring firm’s firm size had a positive moderating effect on this relationship, insinuating that the positive effect of alliance experience