Advantage by Design: the effect of matrix structures on
post-acquisition innovation performance
___________________________________________________________________________ __
Sylvain Swaen S3523276
M.Sc. Strategic Innovation Management June 21, 2019 Supervisor Dr. K.J. McCarthy Co-assessor Dr. J. Dong Abstract
Mergers and acquisitions are still an often-used way to acquire knowledge to create innovations. However, 83% of all the mergers and acquisitions fail to deliver value. This research investigates the effect that organizational design has on the success of mergers and acquisition. This paper focuses on the effect that matrix organizational structures has on the post-acquisition innovation performance. It is believed that matrix structures have the necessary characteristics that make them better at mergers and acquisitions and therefore show higher post-acquisition innovation performance with regard to inventive quantity and exploitative inventions. The hypotheses are tested using a database consisting of 685 acquisitions. The results of this study show that matrix structures do indeed have an impact on the post-acquisition innovation performance. Matrix organizational structures have a higher inventive quantity, and a higher exploitative invention rate.
1. Introduction
According to the resource-based view, firms can obtain a competitive advantage by possessing resources that are valuable, rare, inimitable, and non-substitutable (Barney, 1991). Furthermore, besides the unique resources a firm possess, the way in which managers alters, acquire, and recombine resources is the determinant for capturing the potential of obtaining a sustainable competitive advantage (Eisenhardt and Martin, 2000). Here, I follow the definition of Barney (1991, p.101) of firms’ resources as ‘all assets, capabilities, organizational processes, firm
attributes, information, knowledge, etc. controlled by a firm that enable the firm to conceive of and implement strategies that improve its efficiency and effectiveness.” As can be seen in this
definition, the knowledge and capabilities can be seen as the resources that potentially lead to a competitive advantage. Such a competitive advantage could be attributed to a better innovative performance. Especially for innovation, knowledge is an extremely important resource.
However, when a firm wants to improve their innovative capacity, their existing resources, capabilities and knowledge might not be enough. New resources and capabilities need to be either internally developed or externally acquired through, i.e., mergers and acquisitions (Ahuja and Katila, 2001; Cassiman and Veugelers, 2006; De Man and Duysters, 2005; Puranam, Singh, Zollo, 2006). In a turbulent environment, it is necessary to introduce new innovations as fast as possible to be ahead of competition. Since the internal development of new knowledge is a long process, many firms use mergers and acquisitions as a means towards acquiring external knowledge. Besides, it has been proven that the most successful innovations were built upon a combination of existing internal knowledge and newly acquired external knowledge (Cassiman and Veugelers, 2006). Although the value of mergers and acquisitions is known for a long time, a recent study by KPMG shows that still 83% of all mergers and acquisitions fail (Forbes, 2015). They state that one of the reasons for the high percentage of failure is due to the absence of proper management capacity. Due to the high failure rate of mergers and acquisitions it is still required to investigate the determinants for success.
integration success of acquirers, such as the speed of integration. However, the underlying capabilities that facilitates the merger and acquisition process leading to superior post-acquisition performance, such as a faster integration, is still largely underexplored.
Recently, Sytch et al. (2018), have found that matrix organizations, those organizations that are characterized by dual reporting lines, have rare capabilities that lead them to have a higher probability to enter complex alliances and show a higher ability to manage those complex interorganizational alliances. Said differently, matrix structures have the management capacity to properly manage alliances. According to their research, matrix organizations have a management capability, within complex alliances, due to their complex internal structures which makes them more resilient to external complexity. In this study, I follow the example of Sytch et al. (2018) and use the resource-based view for arguing that matrix organizations are better than non-matrix organizations in mergers and acquisitions. Since the managers in a matrix organization are prone to simultaneously managing the needs of multiple managers, and working together with many other business units, they have developed the valuable and rare management capability for successfully managing mergers and acquisition and create alignment between the acquirer and the acquired firm, so said a ‘acquisition management capability’. Due to the management capability of managers in a matrix firms, scientists in the acquiring firm will be better able to use and combine the newly acquired knowledge with their own knowledge (Bauer et al., 2016), leading to better innovation performance relative of the innovation performance of non-matrix organizations. Additionally, previous research has indicated that firms can possess valuable resources that lead to acquisition capability and better post-acquisition financial performance. Thus, this paper will give an answer to the question ‘What is the influence of a matrix organization on the post-acquisition innovation
performance?’.
limitations of this research and future research directions, and to conclude on the research question respectively.
2. Background
2.1 Organizational structure
Organizations can perform tasks more effectively and efficient than individuals separately. This is the basic premise of why organizations exist. However, bundling forces together comes with certain challenges. Therefore, these organizations need to be structured to formally divide labor and coordinate this division through formalization and standardization. According to Mintzberg structure can be defined as ‘the sum total of ways in which it [an organization] divides its labor
into distinct tasks and then achieves coordination among them’ (1979: p. 2). These ways in
which it divides labor can be divided in six dimensions; (1) specialization, (2) standardization, (3) formalization, (4) centralization, (5) configuration, and (6) flexibility (Pugh et al. 1968). Together, these six dimensions form the basis on which an organizational structure can be built.
The ‘contingency theory’ literature stream argues that there is no singular ‘best’ way of structuring an organization, but that organizational design must be aligned to the contingency of the internal and external environment. The first factor that influences the structure of an organization is the characteristics of the organization. According to Child (1973) size is one main predictor of the structure of an organization, showing evidence that with a growing number of employees, organizations structure their tasks with higher specialization and standardization, and lower centralization. A second determinant according to Mintzberg is the Technology within the organization, which is described as the technologies that employees can use in their daily operations. The third factor is the environment in which the company operates. The environment influences the structure choice by its stability, complexity, market diversity, and hostility (Mintzberg, 1979). The last contingency factor is power, which can be explained as the power of internal and external stakeholders.
Other authors also considered the amount of uncertainty and information that needs to be processed as a determinant of the ‘best’ organizational structure chosen (Galbraith, 1973). In his 1980s paper, Mintzberg argues for ‘structure in 5’s’ arguing that it couldn’t be coincidental that the organization consists of five parts – strategic apex, middle line, technostructure,
support staff, and operating core – five kinds of the centralization/decentralization
configuration – vertical and horizontal centralization, selective horizontal decentralization,
The basic archetypes considered by Mintzberg, from which organizational designs can emerge, are described below.
2.2 Mintzberg’s’ Structure in fives
The Simple Structure, as described already by its name, is a non-complex organizational
design. It is often centralized in the hands of the chief executive officer and information flows informally throughout the structure. Most important in this structure is that it is organic. The structure is often adopted when the future is difficult to predict and therefore the organic characteristic of the simple structure is very important. The business units are often formed by a single person, which can handle the external environment by itself and therefore decision making is controlled by that single person. This structure is most often used by the smaller and younger organizations. Once organizations tend to grow in size and have routine operating tasks, which are specialized within business units and distinct across business units, the power become rather centralized.
In the Machine Bureaucracy the technostructure, in which the analysists are housed, becomes more important since they are responsible for the standardization of tasks. These organizational structures are most often found in environments that are relatively stable. The machine bureaucracy is most often found in mature firms, that have the size to exploit scale of operations. The bureaucracy form of organizing is often related with a high level of external control, which explains the highly centralized and formalized nature of the machine bureaucracy.
The Professional Bureaucracy, although the bureaucracy is characterized by the
formalization and standardization of tasks, is not centralized by definition. The existence of highly trained specialists – professionals – in the operating core of the organization, makes decentralization possible. The bureaucracy is effective in a stable environment, but due to the expertise of the operating core it is able to operate in complex environments.
Next, the Divisionalized Form can be described as a bundle of different organizations (divisions) that is overseen by the headquarters where each division is responsible for their own market. This organizational structure is preferred when the market is diverse in terms of products and services. The distinct characterization of the divisionalized form is that, within the divisionalized firm, multiple structures can exist, tailored to the variables contingent in the various business groups.
mixtures of line managers and staff and operating experts get high authority. As a result, the line-staff distinction gets blurred and it relies extensively on matrix structure. This configuration, which is often found in practice as the matrix structure, is unique as it is the only one that relies on dual reporting lines, combining functional and market bases. Thus, the matrix structure is characterized by a high level of decentralization. Matrix structures are mainly used in environments that require fast ‘ad-hoc’ decision making and that are dynamic and complex. As stated by Davis and Lawrence (1977), a matrix structure is only the solution when three necessary conditions are satisfied: (1) outside pressure necessitates dual reporting lines (due to the constraint that humans can’t be at two places at the same time and have mental limits), (2) pressures for high-information-processing capacity, and (3) pressure for shared resources, or said differently; a pressure for economies of scale. The purpose of the dual reporting line is eventually to create a power balance between the two managers. Most often, the two managers have conflicting goals; one’s goal could be the cost efficiency while the other strives for full utilization of resources. Thus, one can say that a matrix structure is purposively designed to create conflict. As a result, the search behavior of the manager, which reports to two supervisors, is to discover current information and to resolve the conflict (Galbraith, 1973). It is this structure that I will focus on in this paper, which will be emphasized more in chapter 3.
2.3 Structure and innovation performance
In innovation, the cost of not including essential information could be detrimental for firms pursuing exploratory innovation, while the cost of including the wrong information could be detrimental to exploitative innovation (Csaszar, 2013). Imagine a firm that wants to develop a radical new product. The radicalness of the product implies that there is currently no alternative for the new product, that is being totally new to the market or firm. If the omission error is relatively high this means that the probability of not including the essential information is high. This essential information could be decisive for the product becoming the dominant design in the new market (Abernathy and Utterback, 1978). While, on the other hand, in markets in which a dominant design is existing, innovation shifts towards an exploitative focus. Here, changes to a product or service based on information that is not correct could be very costly. A firm that invests a considerable amount of resources in the development of an improved product, will face negative returns when the innovation is based on faulty information (Csaszar, 2013). Firms with a high management capacity will be more likely to make the distinction between high- and low-quality knowledge, which lowers the omission and commission error and therefore will outperform firms without a high management capacity.
The ideal situation in which the firm has both a low omission and commission error, is known as being ambidextrous. Previous research shows that achieving ambidexterity, thus achieving both explorative and exploitative innovation, has a positive effect on firm performance (Gurtner and Reinhardt, 2016; He and Wong, 2004; Raisch et al., 2009). According to Csaszar (2012) ambidexterity is most easily achieved in a structure configurated between hierarchical and polyarchal and thus being a combination of decentralization and coordination. The combination between decentralization and coordination makes a firm capable of better utilizing external knowledge (Foss et al. 2013).
organizational structure determines to which level the authority is delegated and thus has an effect on the type of innovation pursued.
These reasons imply that organizational structure could have a considerable impact on the innovation performance of a firm. Throughout this section, it has become clear that a decentralized structure favors radical product innovations. Therefore, it is fair to believe that there could be a relationship between a matrix organizational design and the innovation performance of a firm, which will be emphasized more in chapter 3. Besides, when organized properly it could give the organization the necessary capacities to outperform their competitors.
2.4 Acquisitions and innovation performance
The results of prior research show that the effect of M&As on innovation performance is either neutral or negative (De Man and Duysters, 2005). Cloodt et al. (2006), for example, show that non-technological M&As have a negative effect on a firm’s post-acquisition innovation performance. Also, technological M&As have a, eventually, negative effect on the post-acquisition innovation performance. Besides the negative effects, firms that acquires other firms also get access to a larger pool of knowledge. The availability of this larger pool of knowledge will subsequently lead to a larger innovative output, measured as the total quantity of patents (Valentini, 2012), since the extent to which scientists create new inventions is partly dependent on the size of the knowledge base on which scientists can rely. Thus, mergers and acquisitions have a positive effect on the innovative quantity for the acquiring firm.
the exploitative focus of a firm. Therefore, it is expected, taking the combinative potential of the target and the bidder into regard, that M&A’s will have a negative impact on the exploratory innovation performance of a firm.
2.5 Structure and acquisition innovation performance
As organizational design has an impact on the performance of a firm, structure can also impact the acquisition innovation performance of an organization. In this section I’ll discuss two streams of research that investigate the impact of M&A on innovation performance. The first stream of research investigates the effect of unique resources on the post-acquisition innovation performance, where the structure functions as the basis for creating valuable resources. The second stream of research investigates the antecedents of structure and their reliance on internal or external knowledge, and thus the extent to which they are active in mergers and acquisitions. Based on the resource-based view, the research by Popli et al. (2017) show a relationship between business group affiliated organizations and acquisition performance, stating that BG affiliated firms possess unique resources and capabilities that can reduce ex-ante and ex-post acquisition challenges. Therefore, BG affiliated firms, which are connected through a structure with one parent company, show higher abnormal returns after acquisition relative to stand-alone firms. This research shows evidence that structure, in the sense of how people are connected within or between firms, could create valuable capabilities which improve the acquisition performance. Another aspect related to structure that could influence the post-acquisition performance, is the flexibility of the organization (Calore et al., 1994; Hitt et al., 1998). This shows that organizational structure forms the basis on which valuable, rare, inimitable and non-substitutable resources can be built. Together, these arguments give enough reason to believe that organizational structure can have a considerable impact on the post-acquisition innovation performance. Just like Popli et al. (2017) argues, different types of organizational structures could lead to management capacities, which can be seen as a valuable resource, that will lead to above average returns post-acquisition.
capability, the M&A process will require less attention and the post-acquisition innovation performance will be higher than their counterparts.
It can be concluded that structure, the way in which firms are organized and the extent to which they are centralized or decentralized, can have an impact on the post-acquisition innovation performance. Nevertheless, in the research by Arora et al. (2014), they state that prior literature has stressed structure as a ‘forgotten pillar’ of organization science. Therefore, it is necessary to look more into the function that organizational structure has with regard to the post-acquisition innovation performance.
3. Hypothesis
3.1 Matrix structure and post-acquisition innovation performance
The performance difference between firms, with regard to post-acquisition innovation output, is dependent on the degree to which firms share newly acquired knowledge throughout the organization. Firms that are good in organizational wide knowledge sharing are likely to profit more from the increased knowledge pool. However, acquisitions are often prone to conflict, due to the differences in the formal organizational structure and organizational culture. The conflict between the merging firms is one of the main reasons why many acquisitions fail (Weber and Camerer, 2003). Alignment between the acquired and acquiring firm is therefore imperative for the post-acquisition performance. As stated throughout this paper, a high management capacity will enable a firm to overcome the negative effects of mergers and acquisition.
Above all, matrix structures are characterized by having a coordinated decentralized organizational structure which is proven to result in a higher acceptance rate of projects, and more reliance on external knowledge for innovation. The matrix structure, which could be compared to the hybrid structure in Csaszar (2012), is a form that has the lowest overall error and accepts the most project. Next, matrix structures are characterized by having a high extent of organizational wide knowledge sharing, which benefits innovation as explained earlier. Therefore, it is expected that organizations with a matrix organizational design have a higher innovation output (Csaszar, 2012; Hobday, 2000).
Hence, there are two reasons to believe that matrix structures will outperform non-matrix structures with regard to the post-acquisition innovation quantity. First, non-matrix organizations have an acquisition management capability and thus are able to create a higher level of synergy between the acquirer and target. As a result, more knowledge will be shared between the focal firm and the target firm, and scientists in both companies will remain more productive which will lead to a higher innovative output. Second, matrix structures have the characteristics that favor innovation, and therefore their innovation output will be higher. Said differently, the characteristics of a matrix structure could create the management capabilities needed to effectively create alignment between the two firms (acquisition capability). Together with the structural characteristics that benefit innovation, such as the lower omission and commission error and the complexity of the organization (Csaszar 2012; Damanpour, 1996), the following hypothesis is formed:
Hypothesis 1: Matrix firms will have a higher post-acquisition innovative quantity than non-matrix firms.
3.2 Matrix structure and post-acquisition innovation type
Besides the inventive quantity, firms can differ with respect to which they experience the effects of M&A on the type of innovation as well. Certain organizational factors have an impact on the extent a firm focuses on exploratory or exploitative R&D. For example, It has been proven that firms with a high degree of centralization show lower levels of exploratory innovations (Foss et al. 2013; Janssen, van den Bosch and Volberda, 2006).
theory is stated in term of losses. The ultimate goal of a company, according to the decision theory, is to achieve the lowest error possible, given the external and internal pressures and conditions. An interesting insight by Csaszar is that the extent a firm tries to lower the omission and commission error differs by their design. Firms vary with regard to their decision-making structure, and in some firms’ decisions to proceed with a project are taken by a single person (polyarchy), where in other firms multiple people need to approve a project before it can proceed (hierarchy). Previously, it is explained that a combination between decentralization and coordination (hybrid structure) could favor ambidexterity. As Csaszar (2013) explains, a hybrid structure is ideal for an organization that aims to have a low commission error without the high costs of omission error. A matrix structure is an example that shows the combination between decentralization (high authority to managers) and coordination (duality of authority). Due to this structure, it is likely that a matrix structure focusses on exploitative innovation, while also keeping an eye for exploratory innovation.
Thereby, the chance that firms innovate in the current trajectory gets higher with the more managers involved in the decision-making process. Here, I use an example as used by Csaszar (2013). Imagine two firms with both two managers. The first firm is characterized by a polyarchy and thus a single manager can make decision to proceed with a project. When the first manager rejects the innovation project, the second manager can still accept the innovation project and the project will be continued for development. Therefore, the probability that an innovation project gets accepted is 50%. The second firm is characterized by a hierarchy and thus both managers need to accept a project in order to proceed. The chance that a project gets accepted in a hierarchy is smaller than in a polyarchy since both managers need to accept the project in order to continue. The chance a project will get accepted by one manager is 50% and thus the chance of two managers accepting a project is 25% (0.5*0.5). This equation gets reinforced when we involve risk as a factor that needs to be considered. Managers differ in the extent that they are risk-averse. Since in the polyarchy only one manager needs to accept a project in order to let it proceed, the chance that this type of organization pursues more exploratory innovations (which are characterized with a higher extent of risk) is higher. Therefore, it is expected that, due to the duality of authority, matrix organizations have a more exploitative focus than non-matrix structures.
structures likely to excel in exploitative innovation. Taken this characteristic together with the characteristics of a matrix structure that will lead to a more exploitative focus, matrix organizations will have a stronger effect on exploitative innovation post-acquisition than non-matrix firms. Therefore, I hypothesize the following:
Hypothesis 2: Matrix firms will have a higher post-acquisition exploitative innovation rate than non-matrix structures.
The hypothesis stated above are graphically shown in figure 1.
Figure 1 - Conceptual Model
4. Methodology 4.1 Empirical setting
test the hypothesis in this setting. Secondly, the novelty of technological production is the main determinant of the competitive landscape, making it suitable to test the innovative output. Following Cloodt et al. (2006) and McCarthy and Aalbers (2016) I define the high-tech industry as consisting of the aerospace and defense (SIC-codes 372 and 376), computers and office machinery (SIC-code 357), electronics and communications (SIC-code 36), and pharmaceutical (SIC-code 283) industries.
4.2 Sample
For the acquisition sample, I use the SDC Platinum Mergers and Acquisitions dataset. This dataset consists of all announced deals from 1979 to onwards and is produced by Thomson Reuters. The dataset is refined to the extent that it includes all acquisitions announced by large, publicly listed, high-tech acquirers, in the period from Jan 1990 – Dec 2015. The dataset is further restricted to those acquisitions that acquired a majority stake of the target but excluded those acquisitions in which the majority stake was only increased. The process resulted in a sample of 252 unique firms with acquisitions from the period 1990 - 2015. Each unique firm could occur multiple times in the sample if they announced multiple acquisitions in one year and therefore the total sample size resulted in 831 acquisitions.
Next, for the sake of the dependent variable, Data from the European Patent Office (EPO) patent data, extracted from OECD REGPAT, is used to gather the inventive outcomes of the firms in the sample. The use of EPO instead of the commonly used USPTO has three advantages. First, the filing fees from EPO are higher than for USPTO, which reduces the count of preemptive patenting (Ceccagnoli, 2009). Secondly, where other patent offices only provide information about granted patents, the EPO also share information about the patent applications (not granted yet) which gives a better overview of the actual innovative activities from a focal firm. At last, the EPO provides detailed information about the firm which will increase the reliability of measures.
4.3 Dependent variable
For the second hypothesis, I will look at the distribution of the type of patents filed. When assessing the patent, I look at the Cooperative Patent Classification (CPC) codes (McCarthy and Kok, In Production). The CPC code is used as an indication for the underlying technology used. In other words, when a company files a patent with CPC codes that they have used before, it can be assumed that they work on an existing technology and further develop or refine the technology. When a firm file a patent with a CPC code that has been used in the last ten years, it is labeled as an exploitative patent. The total amount of exploitative patents is divided by the total patent count, which will give a share of exploitative patents of the total patents.
4.4 Independent variable
The independent variable ‘matrix structure’ will take the form of a binary variable, consisting of the values of ‘0’ for non-matrix structures and ‘1’ for matrix structures. Since an organizational structure is difficult to measure, I have used two estimates, based on the paper from Sytch et al. (2018), to assign a firm in a particular year as a matrix-firm. First, I have downloaded the annual report of the company via www.annualreports.com, or the company’s website. If the annual report wasn’t available, I have searched for the 10-k or 20-f forms via ‘EDGAR’, the search engine of the United States Securities and Exchange Commission. Once collected, I have searched for words as ‘Manager’, ‘Executive’, ‘President’, and ‘Officer’ in the annual report, 10-k, or 20-f form, and looked for dual reporting lines. Matrix organizations often use geography as one of the dimensions to which authority is allocated. Therefore, I have looked for geographical managers and other non-geographical managers (such as product or division managers) as an indication of the dual reporting lines of a matrix organization. Once an organization had such an allocation of authority, I have checked if the vertical authority lines corresponds with meaningful hierarchies by checking the allocation of revenues by geographies and non-geographical groups. If a firm satisfied both criteria, I have assigned the organization with a ‘1’ for matrix structures for that specific year, otherwise it was assigned with a ‘0’. In table 1 an example of matrix firms and non-matrix firms is presented.
results. In the research from Sytch et al., matrix structures forms about 5% of the total database, which seems to be a bit low. In this research, we have been able to identify 135 matrix structures, which is 19.7% of the total (usable) database. However, the fact that the database of Sytch et al. outnumbers the database in this research based on size, is easily explainable. In Sytch et al. (2018) they investigate the effect of matrix structures on the probability to enter and ability to manage alliances. Now, this research investigates the effect on mergers and acquisitions, a decision that involves more risk and resources. Therefore, it is grounded to assume that M&A’s occur much less than alliances, which explains the size difference of the database.
Table 1 – Matrix firms and Non-Matrix firms Examples
1 (Matrix Structure) 0 (Non-Matrix Structure)
Abbott Laboratories Apple Inc.
General Electric Co. Applied Materials Inc.
Intel Corp. Hewlett Packard Co.
Koninklijke Philips Electronics NV. International Business Machines Corp.
Teva Pharmaceuticals Industries Ltd. Johnson & Johnson
Schneider Electric SA. Texas Instruments Inc.
ABB Ltd. Qualcomm Inc.
4.5 Control variables
I control for several variables that have been proven to affect the post-acquisition performance of a firm. Note that I purposively look for variables that influence ‘post-acquisition performance’ rather than solely ‘post-acquisition innovation performance’ since merely focusing on the innovation performance might exclude important control variables. For the control variables I follow Zollo and Meier (2008) which state that acquisition success can be measured on multiple levels, where I focus mainly on the firm and acquisition level.
To begin with the firm-level of analysis, the first variable I will control for in this paper is the revenue (Revenue) of a firm in the year of acquisition. Firms that have a higher revenue are more likely to reap the benefits from acquisitions since they have ‘deeper pockets’ (Ahuja and Katila, 2001). A variable related to the revenue of a firm is the R&D-intensity (R&D
Intensity). A firm that is more active in R&D is also likely to have a higher patenting frequency.
intensity are better able to combine the newly acquired knowledge to the existing knowledge and therefore have a higher inventive output (Cohen & Levinthal, 1990). The R&D intensity of a firm is measured by the total R&D investments of a firm in a particular year divided by the revenue of the firm.
Where R&D intensity is an estimate for the current year efforts, firms are likely to change from year to year regarding the intensity of R&D. However, the intensity of firms in previous years does have an effect for the success of post-acquisition innovation performance in subsequent years. Therefore, it is important to control for estimates that are related with the R&D intensity of previous years. The firm’s size of the knowledge base (Focal Firm
Knowledge Base Size) is such an estimate. Thereby, Ahuja and Katila (2001) have shown that
firms with a larger knowledge base are more productive. The size of the knowledge is measured by a simple count of the firm’s total patents. Besides the knowledge base size, I also control for the knowledge base’s age (Focal Firm knowledge base age), since older firms show a lower probability to innovate (Huergo and Jaumandreu, 2004).
Besides the necessity for controlling for variables related to the characteristics of the firm, there are several acquisition related characteristics where need to be controlled for. First, I control for the number of acquisitions (Number of Acquisitions) made in the focal year, since firms that have multiple acquisitions, and therefore suffer from multiple disruptions, are proven to have poorer performance (Makri et al. 2010). Next, with regard to the percentage of international acquisitions (International Acquisitions), I control for the difficulties experienced in acquisitions due to the differences in culture since these differences lowers the acquisition performance (McCarthy and Aalbers, 2016). Also, firms that are in the same industry are likely to be complementary in each other’s’ knowledge domains and therefore will be better able to recombine knowledge (Makri et al. 2010). Therefore, I control for the percentage of acquisitions in a particular year in which the firm performed inter-industry acquisitions
(Inter-Industry Acquisitions). In the same line of reasoning, I control also for the diversity of the target
(Target’s Knowledge Base Diversity) and focal firm’s knowledge base (Focal Firm Knowledge
Base Diversity) since it has been proven that knowledge diversity improves innovation
(Dell’Era and Verganti, 2010). Furthermore, I control for the complementary of the knowledge bases (Acquisition’s Knowledge Base Distance), since it has been proven that firms with knowledge bases more closely related are better able to combine their knowledge (Cloodt et al., 2006). At last, I control for the acquisitions that erupted from alliances (Transitional
each other due to previous alliances outperform non-transitional acquisitions (McCarthy and Kok, in progress).
4.6 Method of Analysis
To test hypothesis 1, I have used a negative binomial regression. For a regular regression, the data must be normally distributed. However, as one can see in figure 2 in appendix A, the distribution of inventive quantity is positively skewed. Also, since the data for inventive quantity is count data, an ordinary least square regression analysis is not suitable. Thus, when you want to analyze count data, the Poisson regression is an alternative. However, a basic assumption of the Poisson regression is that the mean and variance of the data are equally distributed. Though, as can be seen in the descriptive statistics, the data on inventive quantity is highly dispersed, and therefore I will use a negative binomial regression. The second hypothesis is tested using a tobit regression. The variable that measures the share of exploitative inventions is bounded to a range from 0 to 1. With a tobit regression you can create the boundaries to which the dependent variable has to be measured. Since all the observations for Exploitative Inventions ranges from 0 to 1, the tobit regression is a suitable method of analysis for hypothesis 2. Thereby, dummy variables for year and time are included in both regressions to control for variance in firm and year.
5. Results
5.1 Sample statistics and correlations
Table 2 – Correlation matrix *Significance in parenthesis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1. Matrix structure 1.00 2. Inventive Quantity 0.28 (0.00) 1.00 3. Exploitative Inventions 0.14 (0.00) -0.21 (0.00) 1.00 4. Revenue 0.03 (0.45) 0.29 (0.00) -0.02 (0.56) 1.00 5. R&D Intensity -0.13 (0.00) -0.08 (0.03) -0.22 (0.00) -0.04 (0.36) 1.00 6. Number of Acquisitions 0.13 (0.00) 0.15 (0.00) -0.02 (0.66) 0.04 (0.26) -0.12 (0.00) 1.00 7. Inter-industry acquisitions 0.08 (0.04) 0.14 (0.00) 0.08 (0.01) 0.04 (0.26) -0.25 (0.00) 0.22 (0.00) 1.00 8. International acquisitions 0.25 (0.00) 0.13 (0.00) -0.01 (0.73) -0.01 (0.81) -0.12 (0.00) -0.05 (0.21) -0.02 (0.68) 1.00
9. Focal firm knowledge
base diversity 0.17 (0.00) 0.25 (0.00) 0.08 (0.02) 0.05 (0.16) -0.22 (0.00) 0.03 (0.4) 0.13 (0.00) 0.11 (0.00) 1.00
10. Focal firm knowledge base size 0.31 (0.00) 0.80 (0.00) -0.20 (0.00) 0.26 (0.00) -0.11 (0.00) 0.19 (0.00) 0.17 (0.00) 0.18 (0.00) 0.30 (0.00) 1.00
5.2 Main results
In table 3, the main effects can be found. Here, model 1 and 2 represent the negative binomial regression, where I test the effect of a matrix structure on the inventive quantity. Model 3 and 4 represents the tobit regression where I test the effect of matrix structures on exploitative inventions.
Model 1 and 3 represents the baseline model including all control variables and their effect on, respectively, inventive quantity and exploitative inventions. Model 2 and 4 includes the
Table 3 – Regression results
* = P £ 0.1 ** = P £ 0.05 *** = P £ 0.01
Standard error in parenthesis
Model 1 Model 2 Model 3 Model 4 Negative Binomial Regression Negative Binomial Regression Tobit Regression Tobit Regression Control Variables 1. Transitional acquisitions 0.335** 0.148 0.08** 0.072** (0.144) (0.162) (0.031) (0.034) 2. Revenue 0.036 -1.449 -0.013 -0.083 (0.155) (2.203) (0.039) (0.506) 3. R&D intensity -0.118 -0.254 -0.101 -0.278** (0.450) (0.514) (0.105) (0.113) 4. Number of acquisitions -0.022* -0.001 0.003 0.005 (0.013) (0.014) (0.003) (0.003) 5. Inter-industry acquisitions -0.084 -0.077 -0.013 -0.024 (0.084) (0.093) (0.018) (0.019) 6. International acquisitions -0.104 -0.160* -0.009 0.003 (0.079) (0.089) (0.017) (0.018) 7. Focal firm knowledge base diversity 4.221*** 3.882*** -0.122 -0.599**
(1.148) (1.285) (0.229) (0.242) 8. Focal firm knowledge base size 0.000*** 0.000*** 0.000** 0.000
(0.000) (0.000) (0.000) (0.000) 9. Focal firm knowledge base age -0.346*** -0.318*** -0.002 -0.002
(0.028) (0.031) (0.006) (0.006) 10. Target’s knowledge base diversity 0.220** 0.225** -0.014 0.003
(0.101) (0.110) (0.022) (0.023) 11. Target’s knowledge base distance 0.218** 0.243** -0.031 0.001
(0.107) (0.117) (0.023) (0.024) Independent Variable 12. Matrix Structure 0.245** 0.038* (0.097) (0.021) Constant -0.530 -0.507 1.053*** 1.343*** (1.040) (1.137) (0.195) (0.204) Observations 831 685 831 685
Year FE YES YES YES YES
Firm FE YES YES YES YES
ll -3758 -3132 545.9 494.8
independent variable ‘Matrix Structure’. Surprisingly, in model 1, the size of the knowledge base is highly significant but has almost no effect on the inventive quantity. In contrast, the results for a focal firm’s knowledge base diversity and age both show a large impact on the inventive quantity of the firm while being highly significant (diversity b 4.221, p < 0.01, age b -0.346, p < 0.01), which is in line with previous research. Moreover, the positive relationship between the acquisition knowledge base distance and inventive quantity is surprising as well. In theory it is proven that post-acquisition innovation is dependent on the complementarity of the knowledge bases. However, here it shows that the more distant, the more firms invent.
Model 2 includes the independent variable ‘Matrix Structure’. The results show a significant and strong relationship between matrix structure and inventive quantity (b 0.245, p<0.05) in line with hypothesis 1. This figure shows that matrix structures have, on average, 24,5% more patent applications post-acquisitions than non-matrix firms. Therefore, it can be said that firms with a matrix organizational structure have a higher post-acquisition innovative quantity within this research. The underlying reason is that matrix structures have the characteristics that fosters innovation and therefore will have a higher innovative output. Combined with the purposively creation of conflict within the organization makes a matrix organization better in integration and managing the acquisition process. Together, these characteristics result in a higher post-acquisition innovation quantity than for non-matrix firms.
5.3 Robustness check
In order to check the robustness of the results, I have used a different measure for the dependent variables. First, as can be seen in Appendix A, the inventive quantity variable is non-normal with some extreme outliers. Most of the observations center between the values of 0 and 500 (patents), but some observations have an inventive quantity of 2500. Therefore, the results can be influenced by these extreme observations. To account for these outliers, I winsorize the inventive quantity variable so that the observations outside the 10th and 90th percentile are equal
to the 10th and 90th percentile (Ghosh and Vogt, 2012). Next, where I use the share of
exploitative patents for testing the exploitative inventions, I use the share of the exploratory patents for the robustness check. Since the share of exploratory patents is also bounded between 0 and 1, I use a tobit regression to test this alternative model. Furthermore, I will use a poisson regression as an alternative test for hypothesis 1 and a fractional probit regression to test hypothesis 2. The results remain consistent for the winsorized variables, different type of innovation measure and for the different estimation models. The results of the robustness analysis can be found in Appendix B.
6. Discussion
The results of this research indicate that matrix structures, indeed, have an impact on the acquisition innovation performance of a firm. It shows that matrix structures have a higher post-acquisition innovation quantity. This result is in line with the research from Popli et al (2017), which state that business group affiliated firms can leverage the resources and capabilities from partner firms and therefore their structure creates an acquisition capability, which has shown that structure can be a determinant of post-acquisition performance. As Popli et al. shows that business group affiliated firms have special capabilities, I show that matrix structures have an acquisition capability that enables them to strengthen their innovative quantity. Besides the effect on innovative quantity, this research also proves that matrix firms strengthen their exploitative innovation performance by means of mergers and acquisitions. This result is in line with Valentini (2012), which has shown that acquisitions lowers the inventive quality of firms, measured as the impact, generality, and originality. However, where Valentini argues that mergers and acquisitions have a negative impact on inventive quality, I show that matrix organizations have unique capabilities that makes them able to excel at exploitative innovations compared to other organizational structures.
structure could benefit the wanted post-acquisition performance. It is important to note that the ‘right’ organizational structure is very much contingent on the desired results. For firms that are very active in mergers and acquisitions and that want to improve their inventive quantity, it is worth considering reorganizing the company to a matrix organizational structure since this can improve the innovative performance of a firm with regard to the inventive quantity. Next to that, this research can give input to professionals into the decision making whether to acquire knowledge or to develop it internally. Developing knowledge internally is regarded as time consuming and expensive. On the other hand, mergers and acquisitions are regarded as a fast alternative to gaining new knowledge. However, as this study shows, for some firms this may have contradictory results. Moreover, this study contributes to the literature in two ways. First, as noted by Zollo and Meier (2008), relatively little research has been done on the post-acquisition innovation performance (about 5% of the total literature). Therefore, this study enriches this stream of research. Second, this research has indicated a new predictor for post-acquisition innovation performance. This creates the possibility to investigate new directions of post M&A performance with structure as independent variable. As Arora et al (2014) have stated in their paper, structure is a forgotten pillar of organizational science, and with this research I fill in this gap.
7. Limitations and future research
Although this research presents some interesting new insights for the mergers and acquisition literature field, just as any research it suffers from some limitations. First, the organizational structure in this research is assessed through manually checking the annual reports of the firms. As a result, the assigned matrix structures are selected based on the ideas of the author. Therefore, it might be that, although carefully approached, the structures are prone to mistakes.
Secondly, due to the lack of data, moderating analysis was not possible for this research. However, since there are a lot of factors that could influence the relationship between structure and post-acquisition innovation performance it might be very interesting to investigate the moderating effect of, i.e., relative size, culture distance, and technological distance.
subsequent effect on innovation performance could change the outcomes presented in this study.
At last, as Teece (1986) explains in his paper, firms differ in the way they protect their innovations. There are different mechanisms with which innovations can be protected, and patents are just one of them. Therefore, there is a chance that this research doesn’t include the full picture of the effect of M&A on post-acquisition innovation performance.
This study is the first to match matrix structures with acquisition capabilities. Therefore, there is still much to explore between matrix structured firms and their relation to acquisitions. As Zollo and Meier (2008) present in their paper, M&A performance could be measured on different levels. This paper has been limited to the firm-performance level. However, it could be very interesting to investigate the effect of a matrix structure on the acquisition-level performance, such as the survival rate of acquisitions, or the task-level performance, such as the acquisition process performance, as well.
Next, this study has focused on the long-term performance of innovation performance. However, the acquisition performance could also be measured on the short-term effect, such as the post-acquisition financial performance in an event study. At last, as mentioned earlier, the M&A literature is rich in providing examples on factors that could influence the post-acquisition performance, such as the relative size (Cloodt et al. 2006) and complementarity of knowledge (Makri et al. 2010). Therefore, the results of this research could be very dependent on the situation, which is not taken into account in this research. Therefore, it is very important for future research to investigate what the effects are of these characteristics.
8. Conclusion
innovation rate than non-matrix firms. With these results I can answer the formulated research question ‘What is the influence of a matrix organization on the post-acquisition performance?’. The answer to this question is that the matrix organizational design impacts the post-acquisition performance by stimulating the inventive quantity and exploitative innovation rate.
9. References
Abernathy, W. J., & Utterback, J. M. (1978). Patterns of industrial innovation. Technology
Review, 80(7), 40-47.
Ahuja, G., & Katila, R. (2001). Technological acquisitions and the innovation performance of acquiring firms: A longitudinal study. Strategic Management Journal, 22(3), 197-220.
Alegre, J., & Chiva, R. (2008). Assessing the impact of organizational learning capability on product innovation performance: An empirical test. Technovation, 28(6), 315-326.
Armour, H. O., & Teece, D. J. (1978). Organizational structure and economic performance: A test of the multidivisional hypothesis. Bell journal of Economics, 9(1), 106-122.
Arora, A., Belenzon, S., & Rios, L. A. (2014). Make, buy, organize: The interplay between research, external knowledge, and firm structure. Strategic Management Journal, 35(3), 317-337.
Barney, J. (1991). Firm resources and sustained competitive advantage. Journal of
Management, 17(1), 99-120.
Bauer, F., Matzler, K., & Wolf, S. (2016). M&A and innovation: The role of integration and cultural differences—A central European targets perspective. International Business
Review, 25(1), 76-86.
Birkinshaw, J., Bresman, H., & Håkanson, L. (2000). Managing the post‐acquisition integration process: How the human integration and task integration processes interact to foster value creation. Journal of Management Studies, 37(3), 395-425.
Bradt, G. (2015, January 27) 83% Of Mergers Fail -- Leverage A 100-Day Action Plan For Success Instead. Retrieved from: https://www.forbes.com/sites/georgebradt/2015/01/27/83-mergers-fail-leverage-a-100-day-value-acceleration-plan-for-success-instead/#1e9d19345b86
Cassiman, B., & Veugelers, R. (2006). In search of complementarity in innovation strategy: Internal R&D and external knowledge acquisition. Management Science, 52(1), 68-82.
Calori, R., Lubatkin, M., & Very, P. (1994). Control mechanisms in cross-border acquisitions: An international comparison. Organization studies, 15(3), 361-379.
Ceccagnoli, M. (2009). Appropriability, preemption, and firm performance. Strategic
Chiara Di Guardo, M., & Valentini, G. (2007). Explaining the effect of M&A on technological performance. Advances in Mergers and Acquisitions (pp. 107-125).
Child, J. (1973). Predicting and understanding organization structure. Administrative Science
Quarterly, 18(2).
Cloodt, M., Hagedoorn, J., & Van Kranenburg, H. (2006). Mergers and acquisitions: Their effect on the innovative performance of companies in high-tech industries. Research
policy, 35(5), 642-654.
Cohen, W. M., & Levinthal, D. A. (1990). Absorptive capacity: A new perspective on learning and innovation. Administrative Science Quarterly, 35(1), 128-152.
Csaszar, F. A. (2012). Organizational structure as a determinant of performance: Evidence from mutual funds. Strategic Management Journal, 33(6), 611-632.
Csaszar, F. A. (2013). An efficient frontier in organization design: Organizational structure as a determinant of exploration and exploitation. Organization Science, 24(4), 1083-1101. Damanpour, F. (1991). Organizational innovation: A meta-analysis of effects of determinants and moderators. Academy of Management Journal, 34(3), 555-590.
Damanpour, F. (1996). Organizational complexity and innovation: developing and testing multiple contingency models. Management Science, 42(5), 693-716.
Davis, S.M., Lawrence, P.R. (1977). Matrix. Reading, Massachusetts: Addison-Wesley Publishing Company.
De Man, A. P., & Duysters, G. (2005). Collaboration and innovation: a review of the effects of mergers, acquisitions and alliances on innovation. Technovation, 25(12), 1377-1387.
Dong, J., McCarthy, K. J., & Schoenmakers, W. W. (2017). How central is too central? Organizing interorganizational collaboration networks for breakthrough innovation. Journal
of Product Innovation Management, 34(4), 526-542.
Duncan, R. (1979). What is the right organization structure? Decision tree analysis provides the answer. Organizational Dynamics, 7(3), 59-80.
Eisenhardt, K. M., & Martin, J. A. (2000). Dynamic capabilities: what are they? Strategic
Management Journal, 21(10‐11), 1105-1121.
Foss, N. J., Lyngsie, J., & Zahra, S. A. (2013). The role of external knowledge sources and organizational design in the process of opportunity exploitation. Strategic Management
Journal, 34(12), 1453-1471.
Galbraith, J. (1973) Designing Complex Organizations. Reading, Massachusetts: Addison-Wesley Publishing Company.
Ghosh, D., & Vogt, A. (2012). Outliers: An evaluation of methodologies. In Joint Statistical
Gurtner, S., & Reinhardt, R. (2016). Ambidextrous idea generation—antecedents and outcomes. Journal of Product Innovation Management, 33, 34-54.
He, Z. L., & Wong, P. K. (2004). Exploration vs. exploitation: An empirical test of the ambidexterity hypothesis. Organization Science, 15(4), 481-494.
Hitt, M., Harrison, J., Ireland, R. D., & Best, A. (1998). Attributes of successful and unsuccessful acquisitions of US firms. British Journal of Management, 9(2), 91-114. Hobday, M. (2000). The project-based organisation: an ideal form for managing complex products and systems? Research policy, 29(7-8), 871-893.
Homburg, C., & Bucerius, M. (2006). Is speed of integration really a success factor of mergers and acquisitions? An analysis of the role of internal and external
relatedness. Strategic Management Journal, 27(4), 347-367.
Huergo, E., & Jaumandreu, J. (2004). How does probability of innovation change with firm age? Small Business Economics, 22(3-4), 193-207.
Jiménez-Jiménez, D., & Sanz-Valle, R. (2011). Innovation, organizational learning, and performance. Journal of business research, 64(4), 408-417.
Keats, B. W., & Hitt, M. A. (1988). A causal model of linkages among environmental
dimensions, macro organizational characteristics, and performance. Academy of Management
Journal, 31(3), 570-598.
Makri, M., Hitt, M. A., & Lane, P. J. (2010). Complementary technologies, knowledge relatedness, and invention outcomes in high technology mergers and acquisitions. Strategic
Management Journal, 31(6), 602-628.
McCarthy, K. J., & Aalbers, H. L. (2016). Technological acquisitions: The impact of geography on post-acquisition innovative performance. Research Policy, 45(9), 1818-1832. Mintzberg, H. (1979) The Structuring of Organizations. Englewood Cliffs: Prentice-Hall, Inc.
Mintzberg, H. (1980). Structure in 5’s: a synthesis of the research on organization design.
Management Science, 26 (3), 322-341.
Popli, M., Ladkani, R. M., & Gaur, A. S. (2017). Business group affiliation and post-acquisition performance: An extended resource-based view. Journal of Business
Research, 81, 21-30.
Pugh, D. S., Hickson, D. J., Hinings, C. R., & Turner, C. (1968). Dimensions of organization structure. Administrative Science Quarterly, 65-105.
Puranam, P., Singh, H., & Zollo, M. (2006). Organizing for innovation: Managing the coordination-autonomy dilemma in technology acquisitions. Academy of Management
Puranam, P., & Srikanth, K. (2007). What they know vs. what they do: How acquirers leverage technology acquisitions. Strategic Management Journal, 28(8), 805-825. Raisch, S., Birkinshaw, J., Probst, G., & Tushman, M. L. (2009). Organizational ambidexterity: Balancing exploitation and exploration for sustained
performance. Organization Science, 20(4), 685-695.
Sytch, M., Wohlgezogen, F., & Zajac, E. J. (2018). Collaborative by Design? How Matrix Organizations See/Do Alliances. Organization Science, 29(6), 1130-1148.
Teece, D. J. (1986). Profiting from technological innovation: Implications for integration, collaboration, licensing and public policy. Research Policy, 15(6), 285-305.
Valentini, G. (2012). Measuring the effect of M&A on patenting quantity and quality. Strategic Management Journal, 33(3), 336-346.
Weber, R. A., & Camerer, C. F. (2003). Cultural conflict and merger failure: An experimental approach. Management Science, 49(4), 400-415.
Zollo, M., & Meier, D. (2008). What is M&A performance? Academy of Management
Appendix A – Descriptive statistics
Figure 2 - Histogram Inventive Quantity
Figure 3 - Histogram Exploitative Inventions
Appendix B – Robustness Analysis
In model 1, I use the winsorized dependent variables for Inventive Quantity. Model 2 represents the alternative estimation test for testing hypothesis 1. In this test, I use a poisson regression instead of a negative binomial regression. Model 3 consists of the fractional probit regression with regard to hypothesis 2, instead of a tobit regression model. At last, in Model 4, I change the dependent variable ‘Exploitative Inventions’ to consisting of the rate of exploratory patents with regard to hypothesis 2. All the results remain consistent with the results in the main analysis.
Table 4 - Robustness Analysis
* =p<0.1 ** = p<0.05 *** = p<0.01
Standard deviation in parenthesis
Model 1 Model 2 Model 3 Model 4 Negative
Binomial Regression
Poisson
Regression Fractional Probit Regression Tobit Regression 1. Transitional acquisitions 0.144 0.177*** 0.361*** -0.072** (0.162) (0.028) (0.136) (0.034) 2. Revenue -1.49 -1.79*** -0.096 0.083 (2.194) (0.156) (0.709) (0.506) 3. R&D intensity -0.296 1.439*** -1.123*** 0.278** (0.511) (0.113) (0.421) (0.113) 4. Number of acquisitions -0.001 0.001 0.016 -0.005 (0.139) (0.002) (0.011) (0.003) 5. Inter-industry acquisitions -0.07 -0.099*** 0.087 0.024 (0.925) (0.017) (0.072) (0.019) 6. International acquisitions -0.166* -0.129*** 0.07 -0.003 (0.089) (0.013) (0.071) (0.018) 7. Focal firm knowledge base diversity 3.888*** 7.663*** -2.196* 0.599**
(1.279) (0.275) (1.143) (0.242) 8. Focal firm knowledge base size 0.000*** 0.000*** -0.000 -0.000
(0.000) (0.005) (0.000) (0.000) 9. Focal firm knowledge base age -0.315*** -0.357*** 0.011 0.002
(0.031) (0.006) (0.028) (0.006) 10. Target’s knowledge base diversity 0.219** 0.198*** 0.019 -0.003
(0.109) (0.016) (0.095) (0.023) 11. Target’s knowledge base distance 0.248** 0.005 -0.001 -0.001
(0.117) (0.016) (0.096) (0.024) 12. Matrix Structure 0.236** 0.263*** 0.129* -0.037* (0.097) (0.012) (0.073) (0.021) Constant -0.503 -3.348*** -2,872*** -1.343*** (1.132) (0.490) (0.945) (0.204) Observations 685 685 685 685
Year FE YES YES YES YES
Firm FE YES YES YES YES
ll -3128 -3758 -3132 545.9
