Innovation and technological standards : the effect of patent strength on the relation between dominant designs and innovation in European industries

(1)

Thesis

Innovation and Technological Standards: The Effect of Patent

Strength on the Relation between Dominant Designs and

Innovation in European Industries.

MSc. Business Administration – Strategy Track Amsterdam Business School, University of Amsterdam

Name: Renske Modderman Student ID: 11909420 Date of Submission: 22th of June, 2018 Supervisor: Andreas Alexiou

(2)

2

Statement of Originality

Statement of Originality

This document is written by Student Renske Modderman who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are original and that no

sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents

(3)

3

1. Abstract

In this thesis the ambiguous effect of a dominant design on the innovative performance of an industry is studied. To understand whether this relationship is influenced by one of the characteristics of a dominant design, it examines the moderating effect of the strength of the patent that protects the dominant design. The results of the study confirm expectations made in previous research and establish that a dominant design has a negative effect on the subsequent innovative performance in European industries. Moreover, the results show that the negative effect of a dominant design is enhanced when it is protected by a stronger patent. Hence, this study contributes to previous research by empirically establishing the relation between a dominant design and innovation performance within an industry, and suggests that this effect is contingent of the strength of the patent that protects the dominant design. The effect was tested by a composite indicator that was created for this research. On that account, this thesis additionally contributes to the literature by introducing the new measure. Whereas this thesis aimed to understand effects of the measure by a sensitivity analysis it might still be improved by further research. Future research might commence by validating the measure in other studies, adding other dimensions of patent strength or play with the aggregation method of the measure.

Keywords: Dominant Design, Industry Standards, Innovative Performance, Intellectual Property Rights, Patent Strength, Patent Scope

(6)

6

2. Introduction

The emergence of new technologies resulting from innovation lead to changing dynamics in industries (Karlsson, Johansson, & Stough, 2012). Whereas a large percentage of innovations fail or are not profitable (Castellion & Markham, 2013), innovations require market acceptance to be successful. Subsequently, the topic of standardization has not only gained interest in the academic literature but also on the level of regulation of European industries. This is shown by a report created by the European Commission which is called “Towards an increased contribution from standardization to innovation in Europe”. In this report they attract focus towards a greater contribution from standardization to innovation and competitiveness (European Commission, 2008). Along the side of this, questions addressed in the literature on standardization discuss standardization should be managed without obstructing innovation and how innovation should be managed whilst introducing and maintaining effective standards (Viardot, Sherif, & Chen, 2016).

Traditionally, the effect of a standard on a markets’ innovative performance is considered negative (Suárez & Utterback, 1995). This is resulting from the idea that innovation is about creativity while standardization is about uniformity. As a result standardization would constrain entrepreneurs in seeking opportunities (Swann & Lambert, 2010). However, where standardization conventionally has been identified as an obstacle to innovation, there is a growing amount of academic literature that perceive standardization as an ‘enabler’ or ‘catalyst’ for innovation at a national or a company level (Choi, Lee, & Sung, 2011). It is argued that standardization facilitates access to markets and enables interoperability between new and existing technologies, products, services, and processes (Blind, 2011; Farrell & Saloner, 1985; Galvin & Rice, 2008; Swann, 2010). Hence, mixed results describe the relationship between a standardization and innovative performance (Blind, 2013; Brem, Nylund, & Schuster, 2016). Despite the gain in attention on standards, it is argued that the

(7)

7 concept of dominant design and its influence on innovation still has been investigated to a limited extent only (Blind, 2013; Brem et al., 2016). Moreover, only few have strived to empirically research the relationship between a dominant design and innovative performance (Swann, 2010).

Similarly as standards, it is argued that intellectual property rights (IPR) and especially one of its component ‘patents’ influence innovative performance on both firm and industry level (Acemoglu & Akcigit, 2012; Bekkers, Duysters, & Verspagen, 2002; Delgado-Verde, Martín-de Castro, & Amores-Salvadó, 2016; Kanwar & Evenson, 2003; Murray & Stern, 2007; Papageorgiadis & Sharma, 2016; Sweet & Eterovic Maggio, 2015a; Tamura, 2016).This is as patents guarantee an innovator it can appropriate its invention and subsequently protect against imitation from competitors. This changes the opportunities for all players in the market whereas innovators can no longer base their invention on the new technology. Hence, IPR affects the competitive nature of innovation. Subsequently, it is argued that this effect is influenced by the strength of a patent as this increases the scope of the technology that is protected (Chen, Pan, & Zhang, 2014; Lippoldt & Park, 2005; Papageorgiadis & Sharma, 2016; Trerise, 2010). Once a technology is better protected it becomes more difficult for competitors to work around the invention. Where dominant designs are based on patents it is conceivable that when it is based on a stronger patent this affects its initial effect on innovation of an industry. This is as it not only prevails its effect on innovation as the main technology but also protects this technology more comprehensively. Therefore, while this thesis aims to understand the effect of a dominant design on the innovative performance of an industry, it is regarded important to take the strength of the relevant patent into account as it affects subsequent innovation as well. Although, the literature provides research on the relation between standards and innovation, IPR and innovation, and IPR and standards, literature on the combined effect of patent strength and

(8)

8 standards on innovation is limited. Whereas some researches have researched the topic from the framework of formal standardization there is not yet empirical research that aims to understand the strength of a patent in combination with dominant designs. Hence, this thesis aims to understand its moderating role empirically. On that account, the main focus of this thesis is to test if and how the contradictory relationship between a dominant design and innovative performance of an industry is dependent on the effect of a patents’ strength.

By investigating this topic this thesis contributes to the literature in several ways. It will contribute to the gap that currently exists around dominant designs and innovation, in comparison to literature on formal standardization processes. This is done by extending our understanding of dominant designs in two ways. First of all, it sheds light on the contradictory relationship between a dominant design and innovation by testing it empirically on the industry level. Second of all, it looks whether patent strength has a contingent effect on this relationship, thus showing whether a dominant design can better be understood in relation with patent strength. Whereas literature is mixed on the operationalization of patent strength a final contribution of this thesis is the investigation into a new measure of patent strength.

(9)

9

3. Literature Review

This chapter provides a review of the existing research on technological innovation, standardization, and intellectual property rights i.e. patent scope. First of all, it discusses how standards are related to innovation in the literature. After that, it is explained how standards are derived. It continues by discussing the implications of intellectual property rights in both standards and on innovations. It finishes with the possible effects of patent scope on innovation. Subsequently a literature gap is identified, and a research question is posed. The chapter concludes with the expected results formulated in hypotheses.

3.1 Innovation vs. Standardization

As posed by the Schumpeterian view innovation can lead to creative destruction which can be the source of rent for firms (Pisano & Teece, 2007; Teece, 1988). Besides that, innovation drives long-term economic growth (David, 1975; Phillips, 1968; Rosenberg, 2004). Following these notions, technological innovation is a widely discussed and researched topic from both the strategic as economic perspective. In the field of strategy, the aim of research is to understand the role and strategic approach of firms to take advantage of technological and market opportunities (Viardot et al., 2016). The economic perspective focuses on optimizing incentives for innovation to drive economic growth. Whereas innovation is highly complex and uncertain of nature (Featherston, Ho, Brévignon-Dodin, & O ’Sullivan, 2016), the dynamics of innovation are researched from various angles. Lately, there is a growing understanding that the role of standardization is important in technological innovation, especially as technical standards offer the potential as a source of competitive advantage (Featherston et al., 2016; Swann, 2010; Tassey, 2000). However, the field of innovation and technological standards has originally has received little attention (Choi et al., 2011; Viardot et al., 2016).

(10)

10 This might be the result of the traditional view that sees standards as an obstacle of innovation due to their conflicting characters (Choi et al., 2011). Whereas innovation is about creativity and using new ideas, standardization is about uniformity which subsequently constrains entrepreneurs in seeking opportunities (Swann & Lambert, 2010; Viardot et al., 2016). However, more recently research has come to understand that standardization plays an important role in technological innovation and can promote and stimulate it both national or firm level (Choi et al., 2011; Farrell & Saloner, 1985; Featherston et al., 2016; Galvin & Rice, 2008; Gamber, Friedrich-Nishio, & Grupp, 2008; Swann, 2010; Viardot et al., 2016). This is as standards facilitate interoperability. This means that companies that are part of a technical community are aided to work together by interface standards. Hence, it facilitates knowledge diffusion (Featherston et al., 2016) that leads companies to work innovative solutions together and thereby encourages new technologies, products, services and processes (Blind, Gauch, Blind, & Gauch, 2009; Choi et al., 2011; Farrell & Saloner, 1985; Featherston et al., 2016; Galvin & Rice, 2008; Swann & Lambert, 2010; Viardot et al., 2016). Subsequently, a growing stream of literature aims to understand the implications standards have on innovation. This stream identifies how standards can be used as a tool for diffusion, integration and regulation and this affects innovation (Choi et al., 2011).

Similarly, it has become acknowledged that standardization is strategically important for firms (Balakrishnan & Wernerfelt, 1986; Bekkers et al., 2002; Srinivasan, Lilien, & Rangaswamt, 2006; Suárez & Utterback, 1995; F. Wang, Chen, Wang, Lutao, & Vanhaverbeke, 2014). When firms are able to transform their in-house R&D outcomes into innovations that become standards, it may lead a firm into a favorable quasi-monopoly status in market competition (Blind & Mangelsdorf, 2016; Choi et al., 2011; Henderson & Clark, 1990; Suárez & Utterback, 1995; Viardot et al., 2016). Besides that, when a standard is emerged this may cause market concentration. This subsequently raises cost for rivals,

(11)

11 reduces choice for customers, and causes lock- in on old technologies (Blind, 2011). All in all, the establishment of a new standard can be a major advantage for a firm to reach dominant market share and survive on markets with rapid technological change (Brem et al., 2016; Suárez & Utterback, 1995). Hence, a second stream of research from a strategic perspective on standards and innovation assesses the relationship between standards and innovation.

All things considered, research into standards and innovation has gained attention, and is considered a relevant topic to research when concerning technological change and innovation. Hence, this will be the focus of this research. In order to have a better understanding on standards and their effects, one should be aware of the different standardization processes that lead to standards.

3.2 Standardization processes

Standards can be the result of the process of formal standardization or the process of de facto standardization, which are addressed separately in the literature. (Allen & Sriram, 2000; Belleflamme, 2002; Featherston et al., 2016; Tassey, 2000; J. Wang & Kim, 2007). Whereas Formal standards are established by voluntary open and transparent, consensus based on standardization processes organized by Standards Development Organizations (SDO’s) (Belleflamme, 2002; Blind, Petersen, & Riillo, 2017; Manders, de Vries, & Blind, 2016; Wakke, Blind, & De Vries, 2015). De facto standards emerge naturally through market processes (Belleflamme, 2002). However, gross of the current literature focuses on formal standardization.

3.2.1 Formal Standardization

Formal standards are the result of a process of development and application of standards; and additionally ‘standards’ are rules, guidelines or characteristics, established by consensus and

(12)

12 approved by a recognized body (Choi et al., 2011). An example is ISO 9001 which is an international standard that specifies requirements for a quality management system and has been implemented in more than one million organizations (ISO/IEC, 2004; Manders et al., 2016). It is argued that Formal standards are preferred over De facto standards because they are developed through agreed, open and transparent procedures based on a consensus of all interested parties (Belleflamme, 2002). On the other hand, it is argued formal standards suffer a considerable drawback as Belleflamme (2002) argues: “The pace of reaching them is often too slow in a context of rapid technological progress compared to the rather quick emergence of de facto standards.” The literature on formal standardization aims to understand the strategies firms can undertake to engage in or establish standardization (Blind et al., 2017; Ehrhardt, 2004; Soh, 2009; Tamura, 2016) and looks into the effects of policies (Blind, 2011; Fleming & Sorenson, 2001; Gans & Stern, 2003; J. M. Utterback & Afuah, 1998), such as the influence of regulations on innovation like the ISO 9001 standard (Blind et al., 2017; Manders et al., 2016; Wakke, Blind, & Ramel, 2016). Research on de facto standardization focuses on processes such as market competition.

3.2.2 De Facto standards

De facto standards, are the result of market processes, which are either “sponsored” or “unsponsored”. This means that it depends on whether there exists an identified originator with a proprietary interest or not (Belleflamme, 2002). De facto standardization processes are often aggressive battles between firms to position their technology as the dominant one (Srinivasan et al., 2006). When market forces lead to acceptance of a product's design as the leading design in an industry or product category they result in a standard called the dominant design (Abernathy & Utterback, 1978; Soh, 2009; Srinivasan et al., 2006; J. Utterback, 1994). Whereas a dominant design emerges naturally, it is not always the best innovation that becomes the standard. Correspondingly, it is argued that in some product categories a

(13)

13 dominant design not always emerges (de Vries, de Ruijter, & Argam, 2011; Srinivasan et al., 2006).

Characteristics of a Dominant Design

Scholars have used different definitions for dominant design (Murmann & Frenken, 2006) and the term has been used interrelated with standardization in studies over the years (Gallagher, 2007). However, key characteristics of the concept come from the seminal work of Abernathy and Utterback (1978). The first characteristic is that a dominant design must be widely adopted. Secondly, the emergence of a dominant design apparently changes the nature of competition within the corresponding industry (Abernathy & Utterback, 1978; Murmann & Frenken, 2006). A product category evolves due to the processes of variation, selection, and retention (Anderson & Tushman, 1990; Murmann & Tushman, 1998). In the early stages of evolution of a market, technical and market uncertainty lead to a diversity in product designs (Abernathy & Utterback, 1978; Dosi, 1982; J. Utterback, 1994). Technological breakthroughs create rivalry among alternative designs, resulting in a period of design variation. The emergence of a dominant design is the transition point between the periods of variation and selection (Cecere, Corrocher, & Battaglia, 2015). A dominant design can be defined to exist if the market accepts a particular product’s design, or single architecture as the standard for the whole industry or product category (Abernathy & Utterback, 1978; J. Utterback, 1994). Although scholars differ on whether a dominant design is the cause or consequence of changing competitive dynamics, it is clear that the emergence of a dominant design is a turning point for the competition in an industry (Fujimoto, 2014). It directly affects the technology life cycle and indirectly affects the strategies and performance of firms in that industry (Brem et al., 2016; Srinivasan et al., 2006). This change implies that studying dominant designs is interesting because the change of nature of the game, leads to winners and losers (Suarez, 2004).

(14)

14 Therefore, to better understand the dynamics around a dominant design, gross of the literature on dominant designs looked into factors influencing the emergence of a dominant design, and the strategic participation of firms in this process (Featherston et al., 2016). However, factors influencing the emergence of a dominant design are found on both firm and environmental level. On the firm level factors such as technological superiority, complementary assets, and strategic maneuvering have been investigated (Suarez, 2004). On the environmental level the effect of network patterns has received a lot of attention (Ehrhardt, 2004; Soh, 2009; Srinivasan et al., 2006). Besides network patterns, Srinivasan et al. (2006) have looked into the effect of characteristics of the product market such as appropriability, value net, and radicalness of a technology on the probability and time of emergence of a standard. One of the aspects they took into account was the appropriability of a technology. They found that a dominant design is more likely to emerge with weak appropriability. Although one could expect during in innovation firms are looking to secure their rents from innovations (Teece, 1986; 2006) the research shows that the emergence of a dominant design has opposing effects on this. Srinivasan et al. (2006, p. 4) give as an explanation that: “On the one hand weak appropriability can result in a loss of rents because other firms can imitate the innovator's design. On the other hand, weak appropriability is associated with a greater probability of emergence and the earlier emergence of a dominant design, increasing rents for all firms in the market”. Appropriability embodies the aspect of patents protection the innovation from duplication.

Although research on standards has grown and the interrelation of standards or dominant designs and innovation seems to be a major contributor to a firm’s competitiveness, research suggests that effects on innovation after the emergence of a standard has been investigated to a limited extent only (Blind, 2013; Brem et al., 2016; Farrell & Saloner, 1985; Galvin & Rice, 2008).

(15)

15 Besides that, the evidence found on the relationship between dominant design and innovative performance is considered contradictory. Traditionally it has been argued that standardization leads to a lower innovative performance. This is explained by the idea that innovation is about creativity and generating new ideas, where standardization is about uniformity, thus inhibiting entrepreneurs in their search for new opportunities (Swann, 2010; Viardot et al., 2016). Subsequently, research has been done to establish the relationship of standardization leading to lower innovation (Suárez & Utterback, 1995). However, more recent research indicated that this traditional relationship does not hold in several conditions and counterevidence has been given (Blind, 2013; Brem et al., 2016; Soh, 2009; Swann, 2010). For example, Allen and Sriram (2000) argue that in general the benefits of standards on innovation in design and manufacturing outweigh the possible limitations on creativity imposed by such standards. In addition, on the firm level empirical, evidence of the relationship between a dominant design and a firm’s innovative performance is given by Soh (2010). He argues that tight collaboration with other partners can play a role. Soh concludes that firms with high proximity to other firms in industry alliances, an extensive information flow between firms in the collaboration network, and the strategic intent to open up the innovation process and share knowledge with partners, achieve better innovative performance. Swann and Lambert (2010) support this conclusion as they find that respondents in the Community Innovation Survey are more successful in innovation activities when they perceive standards as an instrument to receive information instead of perceiving standards as an obstacle to innovation. On the macro level, Swann (2010) also argues that the perception of standards as a framework or infrastructure condition combines both positive and negative impacts. This is because any type of infrastructure generates opportunities for its users but also limits the user options. Consequently, the question remains whether dominant

(16)

16 design and subsequent standardization generally constrains or enables innovation, which will be the main focus of this thesis. Hence this will be the focus of this study

To better understand the effects of dominant designs and its contradictory relationship with innovation performance of firms, one could look for factors that might influence the nature of a dominant design. Dominant designs may lead to a firm’s favorable market position (Cusumano, Kahl, & Suarez, 2008; Porter & Linde, 1995; Soh, 2009; Suarez, 2004; Suárez & Utterback, 1995; Tassey, 2000). Thus, as an effect of strategic maneuvering firms often protect their innovations with intellectual property rights (IPR), especially patents. The integration of IPR into standards is possible during the overlapping phase of patent application and standardizations, and generates a series of benefits for both the owner and those interested in implementing these standards, but also some challenges (Berger, Blind, & Thumm, 2012; Blind, 2013). Firms may benefit from a temporary monopoly position, however it may lead to inefficiencies of the market when this monopoly power is maintained. These negative effects are especially stressed when a dominant design is protected by strong intellectual property rights (Woo, Jang, & Kim, 2015). However, a system of intellectual property rights is necessary to ensure that firms will carry out innovative activities (Bekkers et al., 2002). Without the protection of new inventions, imitation will erode the inventor’s profit rate, and hence lower the incentive for innovative activities. Patents grant the inventor of an innovation the legal monopoly to appropriate their rents from the invention. Subsequent research has been done to understand the effect the strength of a patent.

3.3 Patent Strength

Patent strength can be viewed from different angles (Chen et al., 2014). For example it can be positively related to the probability that its owner will prevail if challenged in a litigation. It

(17)

17 can also be seen as the degree of enforcement to prevent infringement by imitators. Yet, another possibility is to view a higher patent strength as indicating a broader patent scope.

That is why research considering the characteristics and strength of a patent can be divided in a few streams. One of the streams research is conducted from an economic perspective. It looks at the optimal breadth and scope of patents to incentivize innovation, and thus increase social welfare (Chen, Jang, & Wen, 2010; Chen et al., 2014; Encaoua, Guellec, & Martínez, 2006). The main argument within this stream of literature is that there is market failure in the production of technological knowledge by innovation (Encaoua et al., 2006). Patent systems are thus needed to incentivize innovation (Acemoglu & Akcigit, 2012; Sweet & Eterovic Maggio, 2015b). Alongside the literature on patent systems, the strength of intellectual property rights has been at the attention of scholars on the level of countries. Ginarte and Park (1997) constructed a patent rights index that shows the strength of a national patent right protection which is used to understand the national and international effect on innovation. For example Allred and Park (2007) find there is a difference between countries with developed and developing economies. In developed countries patent strength positively affects R&D and domestic patent filings. In line with this research, it is found that patent protection promotes the valuation of innovation efficiency of multinational firms, as they are mainly diversified into developed markets or markets with better patent protection (Gao & Chou, 2015). However, where this stream of research focuses on the effects on a national, to understand the effect of patent strength on the firms and industry level a final stream of literature concerning patent scope is evaluated.

3.3.1 Patent scope

The scope of patents on inventions strongly influences the incentives of inventors and its potential competitors (Merges & Nelson, 1994). This follows a general rule, stating that if the allowed patent scope is broad, today’s inventors may proceed into the next stage of inventing

(18)

18 without fear of infringement by outsiders. The outsiders are deterred from participating because of the likelihood that their invention will be held infringing. In contrast, if allowed scope is narrow, outsiders are less deterred from competing in the next round of inventing (Scotchmer, 1991).

Whereas Gilbert and Shapiro(1990) and Klemperer (1990) considered the optimum length of a patent, Merges an Nelson (1994) argue that the amount of claims of a patent determines the ability of competitors to produce substitutes without fear of infringement suits, and hence the real ‘monopoly power’ of the patent holder (Merges & Nelson, 1994). These claims are one of the two components that are considered during the application of a patent, in order to protect the invention from variations of the idea. The first component is a specification of the invention. This specification describes the techno-economic problem faced by the inventing firm and provides according to Merges and Nelson (1994, p.9) : “ a precise characterization of the ‘best mode’ of solving the problem’. The second is a set of claims. This set of claims specifies possible improvements or variations that could be made to the patented invention to adapt it for different uses (Merges & Nelson, 1994; Walker, 1995). Thus, the number of claims have been used in many studies to measure the patent scope, suggesting that it protects the idea in an additional area (Lanjouw & Lerner, 2000; Lanjouw & Schankerman, 2004; Merges & Nelson, 1994). Next to the amount of patent claims, the amount of patent classes in which a patent is filed is used to describe the scope of a patent. This is based on the idea that a patent with broader scope would include more distant applications (Lerner, 1994; Novelli, 2015; Shane, 2001). Studies from both streams argue that patents with broader scope should enjoy stronger protection against the risk of imitation.

Thus, where prior research in this area has used both patent claims and patent classes as alternative measures of patent scope, more recent literature suggests that they reflect different dimensions. While claims reflect the number of variations identified to an initial core

(19)

19 invention, patent classes would reflect the extent to which these variations are spread out in the technological space (Novelli, 2015). Novelli (2015) builds on these two operationalizations and argues that the strength of protection provided by the patent varies, depending on both the number of variations identified, i.e. included in the patent claims, and their positioning in the inventive space, i.e. their spread over patent classes. Where most of the literature on patent scope has been theoretical, Novelli (2015) and Lerner (1994) provide empirical evidence of the effect of patent scope. The bigger the scope of a patent the more difficult it is for competitors to work around the invention. Hence, Lerner (1994) has validated that a broader scope leads into a higher number of external forward citations explaining that it is an important element of market power for patent applicants (Berger, Blind, & Thumm, 2012; Lerner, 1994).

3.2 Literature Gap & Research Question

To conclude, empirical research into the relation of a dominant design and its effect on innovative performance of a firm has provided contradictory results. Hence, scholars suggest further research is needed to understand the aspects that influence this relationship, and thus the effects of a dominant design. Whereas it is argued that the topic of dominant design and its effect on the innovative performance is considered understudied (Blind, 2013; Brem et al., 2016; Choi et al., 2011; Murmann & Frenken, 2006; Viardot et al., 2016), only few attempts have been made to test moderating factors of this relationship(Blind, 2001; Brem et al., 2016; Swann & Lambert, 2010). Hence, it is not surprising that patent strength has not yet been used to explain the effect of a dominant design on innovation performance in an industry. However, since the strength of a patent has been researched to affect innovation by firms, and dominant designs are regarded pivotal in the changing dynamics of a market competition during innovation, it seems logical that there is a possible effect between these two concepts, and its subsequent effect on innovation. Hence, to contribute to the literature on the

(20)

20 contradictory relationship between dominant design and innovative performance, and to fill a part in the literature gap on dominant designs, in this thesis I will research the moderating effect of patent strength on the relationship between dominant design and innovative performance of an industry. Therefore, the research question of this thesis will therefore be as follows:

“What is the effect of patent strength of a dominant design on the innovative

performance in an industry?

3.3 Hypotheses Development

3.3.1 Dominant designs and innovation performance

Research acknowledges that standardization can have both positive and negative effects on innovation (Blind, 2013; Brem et al., 2016; Featherston et al., 2016; Swann, 2010; Tassey, 2000). Despite the traditional view that standards inhibit innovation by constraining certain innovation activities (Blind, 2013; Featherston et al., 2016) it is also suggested that the benefits of standards on innovation in design and manufacturing outweigh the possible limitations on creativity imposed by such standards (Allen & Sriram, 2000).

Especially when standards are the result of a formal standardization process and provide compatibility and interoperability, they might have a positive effect on innovation on both firm and industry level. This is as it might have a ‘mediating’ role in knowledge diffusion among innovation actors. (Allen & Sriram, 2000; Blind et al., 2009; Tassey, 2000). Nonetheless, this positive effect is dependent on several conditions. For example it has been proven by empirical research that on the firm level, innovation performance can increase by being close to network alliances. In the race of standards to become a dominant design, firms can shift the technological trajectory towards their preferred standard by positioning themselves central in network alliances. In this ways they attract suppliers of complementary

(21)

21 products, in order to turn market resources away from competing standards. As a result, centrality in a dense network and the strategic intent to acquire and share knowledge within the technological community leads to a better innovation performance of the firms (Soh, 2009). Similarly, close collaboration of the industry with Standards Development Organizations might foster innovation by solving difficult technical and commercial problems in standards development (Leech & Scott, 2011). Furthermore, when appropriate interfaces between old and new technologies are formed instead of leading to a lock-in, leading standards might enable innovation. This is because they allow simultaneous use of new and old technologies and ensures their compatibility. Hence, a requirement is open standardization processes and shared interface as they enable a competition between and within technologies and contribute therefore to innovation-led growth (Blind, 2013). However, it should be noted that these conditions are the result of formal standardization and not de facto standardization.

De facto standards-setting processes are characterized by aggressive market competition without intervention of standards bodies (Srinivasan et al., 2006). Technological battles pertinent to this aggressive character might not allow the conditions of formal standardization in which innovation flourishes. Besides that, the result of technological battles may inhibit innovation. Once technological battles are settled one firm might benefit from a monopoly power of an emerged dominant design. The dominant design shapes future generations of products in the category. Resulting in an ‘architectural franchise’ for the firm with a dominant design and potentially locking out competitors (Schilling, 1998). Despite the opportunity to create complementary technology and products, it inhibits the innovation in the relevant industry temporarily. Especially when the dominant design benefits from strong network externalities innovation might be inhibited (Katz & Shapiro, 1992). Moreover, strong network externalities might consequently result into the lock-in into old technologies.

(22)

22 This has especially a negative effect on radical innovations (Arthur, 1989; Katz & Shapiro, 1992). A negative consequence is that an obsolete technology can hinder the formation of potentially superior new standards (Farrell & Saloner, 1985; Tamura, 2016). Whereas these negative effects result from a lack in interoperability, reduction of variety also has an effect on innovation.

Innovation is a process of variation, selection and retention. As a consequence of an established standard there is focus on solely one innovation and variation will be reduced. Although this brings about negative effects, this does provide opportunities in terms of innovation too. For example, it allows opportunities for economies of scale. As a result of this, profits can be generated and incentives to innovate can rise to re-appropriate investments into innovation. Besides that a focus on specific technologies promotes the development of critical masses. Consequently, it increases the credibility especially in new technologies attracting further investments and the development of complementary technologies (Blind, 2013). However, it is argued that in the end the negative effects of variety reduction might overpower the positive effects (Blind, 2013; Brem et al., 2016).

As a result of the self-regulating nature of the patent citation mechanism markets concentrate on the selected technologies with the potential to breakthrough (Suárez & Utterback, 1995). Subsequently, market concentration reduces the variety of technologies. This results in a lower innovative performance in that industry overall. Also the proposed development of complementary technologies is limited by the reduced choice of available technologies. Besides that, it is shown that if companies enter a market after the rise of a dominant design, they are subject to fail more often than companies that entered the market in the pre-dominant design phase (Suárez & Utterback, 1995). As a result of this, fewer players remain on the market and less innovative activities will be conducted (Brem et al., 2016). Thereupon it is

(23)

23 argued that there is a negatively influences the overall innovative performance in an industry (Blind, 2013; Swann & Lambert, 2010).

Hence, in the end it can be argued that dominant designs have the potential of both positive and negative effects on innovation. However, in this thesis the innovative performance is based on the cumulative output of patents on the level of an industry. Positive effects such as economies of scale stimulate innovation on the level of the firm. The negative effects resulting from monopoly power, lock-ins and variety reduction occur on the level of the industry. Even though the stimulation of innovation on the firm level in the end might contribute to innovative output on the industry level, there is no guarantee this will lead to effective output on the level of the industry. Moreover it might be that this behavior is already overpowered by the industry level effects. Besides that the framework conditions, such as appropriate interfaces and collaboration with SDO’s, which are considered to stimulate innovative performance are especially likely to occur in the process of formal standardization and not in de facto standardization that gives rise to dominant designs. This supports that when a technology becomes the dominant one, conditions for positive effects are lacking and the negative effects might prevail.

All in all, arguing within a framework on the industry level and de facto standardization it is expected the negative effects prevail as conditions for the positive effects are less supported. Hence it is expected that the aggressive nature of the market process leading to a dominant design and the subsequent monopoly power and market concentration and variety reduction lead to reduced innovation in an industry. Therefore it is hypothesized that:

Hypothesis 1 (H1): The existence of a dominant design in an industry has a negative effect on the innovative performance in that industry.

(24)

24

3.3.2 The moderating effect of patent strength

It is the nature of intellectual property rights (IPR), and especially patents that it protects the invention of a firm. This follows the costly investments companies have to make to generate innovation. If a firm wants to benefit from the monopoly position resulting from their invention, it is useful to protect it with a patent. This is as companies take advantage of the “free ride” of an innovation (Soininen, 2007; Viardot et al., 2016). Costs of innovation is one of the factors that influences the degree firms protect their patent. It is argued that high marginal innovation cost reduces the need for strong patent strength, whereas high fixed innovation cost tends to require greater patent protection (Chen et al., 2014). Especially, the development of a standard is costly, and subsequent protection of the standards increases cost for the competitors (Blind, 2013). It is shown that driven by strategic motives to patent research results of firms are increasingly protected by IPR (Blind, 2011).

To understand the effect of a stronger patent specifically in the presence of a dominant design the specific effects of patent strength should be understood. Almost the total of the literature on the effects of patent strength debate from an economical perspective, to find the optimal balance to stimulate innovation. Both stimulating as inhibiting effects are found, thus a distinct effect remains ambiguous (Gallini & Scotchmer, 2002). First of all, it is shown by previous research that strong IPR stimulates technological innovation by incentivizing the inventors who drive economic growth (Allred & Park, 2007; Woo et al., 2015). On the other hand, inhibiting effects can be explained by strategic patenting and monopoly power. Strategic patenting is done to block competitors entering the market. Strong IPR negatively affects technology transfer, diffusion, and commercialization due to the monopoly power an inventor receives from the protection of its innovation (Allred & Park, 2007; Encaoua et al., 2006; Gallini & Scotchmer, 2002; Hall & Ziedonis, 1979). Hence, it may impede further innovation.

(25)

25 However, this is seen in the static framework of a single invention, while innovation is cumulative (Chen et al., 2014; Gallini & Scotchmer, 2002; Gilbert & Shapiro, 1990; Klemperer, 1990). Innovation is especially the basis of future improvement when two fields of technology are combined (Scotchmer, 2004). Hence, to understand the effect of patent strength one should also argue from a two-stage innovation framework, where a second innovation builds upon the first (Chen et al., 2014; Scotchmer, 2004).

In the context of continuous innovation there are several ways to understand the effect of patent strength. First of all, several scholars argue that stronger patents further innovation by delaying the next patentable discovery (Bessen, 2008; Encaoua et al., 2006; Gallini & Scotchmer, 2002; Hunt, 2004; O’Donoghue, Scotchmer, & Thisse, 1998; Segal & Whinston, 2007). This is explained as follows. If inventions follow straight from previous ones, the exclusive rights may cut off access to the knowledge embedded in the development of previous inventions. Thus, it slows down technological progress. As current patent holders attempt to hold up the future innovation of a rival through a strong IPR regime, the inventor will also be delayed by previous patent holders (Woo et al., 2015). Hence, stronger patents delay further innovation, when build on knowledge of previous technologies.

In addition, the effect of stronger patent can be understood by the effect of profit division between innovators. This profit division is the result of the interplay between incumbents and other inventors (Chen et al., 2014). A character of continuous innovation is that an ‘incumbent’ firm who has produced a product of a certain quality previously, is followed by a potential entrant that enters the industry. If the entrant successfully discovers a higher-quality product it will replace the current incumbent, and become the new incumbent. Patents both provide the division between the sequential innovators and expand profits from innovation by deterring imitation (Chen et al., 2014). Consequently, it is argued that high patent strength reduces continual innovation due to a “front-loading” effect. This effect beholds that an

(26)

26 innovator benefits from the innovation immediately as an entrant but with a discount as the future incumbent. The incentive to innovate is reduced as stronger patent protection shifts innovation profit from the entrant to the incumbent (Chen et al., 2014; Segal & Whinston, 2007; Vickers, 2010).

In order to understand the implications of these effects in the light of our research question, they should be seen in relation to the effect of standards on an industry. This research is focused on the effects of a dominant design which are derived from a market process and pose the leading technology in the industry. Therefore, it is expected that a dominant design negatively influences innovation by the creation of monopoly power. Stonger patents enhance monopoly power by negatively effecting technology transfer, diffusion, and commercialization of the technology. Thus it hinders stimulating conditions for innovators. Accordingly, it is argued that the effect of monopoly power is especially pronounced when a dominant design is protected by strong intellectual property rights (Woo et al., 2015).

Secondly, as discussed above it is argued that stronger patents delay further innovation based on previous technologies. A dominant design is the most preferred technology of an industry, firms might be inclined to invent around this technology. However, innovation based on this technology will be delayed further by stronger patents.

This is also argued by Blind (2013) who finds that as a result of the network externalities of a standard the combination with IPR might lead to lasting longer than the maximum length of patent protection (Blind, 2013).

Finally, whereas the operationalization of patent strength is based upon measures of patent scope the effects of patent scope are considered. The number of claims represents technological specifications of the patented technology. More claims limit other innovators to

(27)

27 build variations of the technology. Besides that it is argued that the higher the number of patent classes, the higher the market value of the patent. A higher number of patent classes, thus market value might reinforce the the dominant position of a technology. As a result of this it is argued that innovation subsequent on the technology a of dominant design is increasingly inhibited by stronger patents as the field of technology that is protected is bigger.

Taking all of these considerations into account, it is expected that if the negative effects of a dominant design are confirmed, stronger patent protection will even enhance this effect. Hence, this study predicts:

Hypothesis 2 (H2): The negative effect of a dominant design on an industries innovative performance is enhanced with the increase of a patent’s strength.

(28)

28

4. Methodology

This chapter explains the steps undertaken to perform this research. The chapter starts with a description of the data and an explanation of the data collection and sampling strategy. The second part provides an overview of the independent, dependent and control variables. The last section explains the statistical models used to analyze the data.

4.1 Sample and Data Collection

This research used secondary data retrieved from OECD Patent Datasets. These datasets have been produced using primary data of the latest edition (Autumn 2017) Worldwide Statistical Patent Database (PATSTAT). The specific datasets used for this research are Patent Citations, Patent Quality and International Patent Classes (IPC) that were combined using the surrogate key of the datasets. The datasets contained information about patents filed at the European Patent Office (EPO) from the period of 1979 until 2017. The IPC dataset contain all patents granted at the EPO, sorted by patent class. This database was used to extract the dependent variable ‘Innovative Performance’. The Patent Citations dataset listed a total of 11,795,845 patent citations which were used to construct the independent variable ‘Dominant Design’. The dataset on Patent Quality was used to compute the moderating variable ‘Patent Strength’, which will be elaborated upon in section 4.2.3.

Patent classes based on the IPC classification were used to structure the dataset. This classification system uses an “application-oriented” approach to distinguish patent by technological field. As using the classification permits the assessment of technological development in various areas, it is the basis for industrial property statistics (Wipo, 2018). Hence, structuring the dataset on patent classes provides the analysis on different industries. A total of 631 patent classes where identified, which were all included in the sample to be able to study the effect across industries. Whereas the dependent variable is based on a lagged

(29)

29 effect the final sample consisted of patent class years up to 2016, whereas the lagged value of 2017 is not yet known. The remaining 37 years were taken to be able to control for industry life cycles (Klepper, 1997). Since some patent classes have been established after 1979 not all patent classes provided information on a range of 37 years. Hence, the sample contained data on 21,093 patent class years from the period of 1979 until 2016.

The citations database was used to extract the patents relevant to this study. The PATSTAT dataset contained information citations of all granted patent applications filed at the EPO. Patent citations are considered to be a good indicator of the value of a patent (OECD, 2009; Van Zeebroeck & Van Pottelsberghe De La Potterie, 2011). Based on this qualification the most important patents within each patent class year were identified. These patents were used as an input for the variable dominant design. The citations in the PATSTAT dataset are the result of “European searches”, meaning that they indicate the value of European patents within European industries.

4.2 Variables

4.2.1 Independent variable

Dominant Design: In this research a dominant design is defined as follows. A dominant design exist if the market accepts a particular product’s design, or single architecture as the standard for the whole industry or product category (Abernathy & Utterback, 1978; J. Utterback, 1994). In order to operationalize this variable previous research was followed and is based on patent citations. Patent citations are considered to be a good indicator of technologically important patents (Carpenter, Narin, & Woolf, 1981), and the value of a patent (Bessen, 2008; OECD, 2009; Van Zeebroeck & Van Pottelsberghe De La Potterie, 2011). On that account, the importance, and value of a patent to others in an industry is measured as the number of citations a patent received in a patent class in a specific year. To

(30)

30 compare a the citations of a patent relative to other patents in a patent class, the percentage of all citations in a patent class year were calculated for every single patent. A design can be considered dominant if a majority of other innovations in a patent class includes the same design (Brem et al., 2014). Hence, a dominant design exists in a patent class year, if the percentage of patents that cite the same patent is above a threshold value of 50 percent as it represents a majority of citations. This leads to a binary operationalization of the variable dominant design, whether it exists in a patent class year or not.

4.2.2 Dependent variable

Innovative performance: Innovative performance is in this study regarded as the cumulative output of an industry. As mentioned in the previous section industries are proxied by patent classes. Innovation in an industry is proxied as the patenting frequency of each patent class, calculated as the number of patent applications in a given year. This proxy for innovative performance is consistent with prior studies that use patent applications for the innovation output of respectively firms, and industries (Ahuja & Katila, 2001; Austin, 1993; Beneito, 2006; Brem et al., 2016; George, Zahra, & Wood, 2002; Hall & Ziedonis, 1979). While dominant design is a phenomenon related to technological classes, the dependent variable innovative performance is measured for each patent class. Whereas this study is interested of the effect of a dominant design the year after the establishment of it, the variable is lagged.

4.2.3 Moderating variable

Patent Strength: In this research patent strength is defined as the degree of enforcement to prevent infringement by imitators. In order to operationalize the concept a composite indicator was computed, based on two operationalizations of patent scope. According to Saisana & Tarantola (p. 6, 2002) a composite indicator is: “The mathematical combination of individual indicators that represent different dimensions of a concept whose description is the objective of the analysis”. In the literature there are two traditional ways to operationalize

(31)

31 patent scope. The first measure uses the number of claims in a patent (Kitch, 1977; Lanjouw & Schankerman, 2004; Merges & Nelson, 1994; Walker, 1995) i.e. the number of variations to an invention. The second measure looks towards the positioning of these claims i.e. variations across multiple technological patent classes (Lerner, 1994). Recent research presumes that both constructs reflect different dimensions of patent scope and that both operationalizations should be taken into account (Novelli, 2015). Moreover, it is argued that using a single parameter of patent scope to denote for patent strength is highly restrictive (Chen et al., 2014), and that combining the two measures of patent scope provides better info on the strength of protection is provided (Novelli, 2015). Hence, to study the effect of patent strength for this study a new measure was created by combining both operationalizations of patent scope into a composite indicator. The measures of patent scope were both obtained from the PATSTAT dataset on Patent Quality.

In order to create the variable, research was done into other composite measures. For example Choi et al. (2007) base the amount of knowledge creation on the balance of resource allocation towards exploration and exploitation. More closely towards patent strength, Ginarte and Park base the strength of national patent protection on the unweighted sum of patent laws. Pursuing these studies and following the OECD handbook on patent composite indicators (2008) the measure was based on a linear aggregation of the measures. This is as the construct is based on measures that reflect different dimensions of the same concept and allows to include weighting of the different variables. Following the steps of the OECD handbook the measures were normalized by standardization. Normalization is required prior to data aggregation as both indicators are based on different measurement units. (OECD, 2008; Saisana & Tarantola, 2002). Subsequently, the measures were combined in the

following formula:

(32)

32 When there is insufficient knowledge of causal relationships or a lack of consensus on the alternative composite measures are usually based on equal weights (OECD, 2008). Current literature does not mention relative values of the number of claims and number of patent classes. Thus, so far it lacks a statistical or an empirical basis that could function as a base for the weighting of the measures. In order to prevent that the indicator is an element of double counting the correlations of the measures were checked. Table 1 shows that the correlation between the two measures is significant but very low. Hence, the measure is initially based on equal weights. Since the weighting of the variable can have significant effect on the composite indicator a sensitivity analysis can be found in section 5.3 to understand the effect of different weightings on the robustness of the results. Besides improving the indication of a patent’s strength, one of the gains is that gain of combining the two dimensions is that flaws in traditional operationalization are diminished (Novelli, 2015).

Table 1: Descriptive Statistics and Pearson correlation of Patent Scope operationalizations

N Mean Std. Deviation 1) 2) 1) # Patent Classes 20,386 5.51 3.13 1

2) # Claims 20,386 35.85 36.86 0.08** 1

** Correlation is significant at the 0.01 level (2-tailed) * Correlation is significant at the 0.05 level (2-tailed)

4.2.4 Control Variables

It is expected that the hypothesized relationships are moderated by different characteristics of the dominant design (Brem et al. 2016). Hence, the model will be controlled by the variables: Time Lag and Age of the dominant design.

6.2.4.1 Time Lag

Time lag is constructed by calculating the difference between the year a patent was first published and the year it became a dominant design. It thus measures the number of years it

(33)

33 takes a certain patent to become dominant in a patent class. In other words, it shows the time it takes an innovation to become dominant in an industry. The measures were derived from the PATSTAT Patent Citations database. The measure was based on the earliest date of publication, instead of application or priority date, as it this leads to a better estimate of the citation lag (De Rassenfosse, Dernis, & Boedt, 2014; Webb, Dernis, & Harhoff, 2005).

6.2.4.2 Age

Age of the dominant design measures the duration a dominant design maintains it status. Hence, it shows how long it takes before new innovations become predominantly based on other technologies. Age of the dominant design is constructed by calculating the number of years a design has been dominant. The measures were derived from the PATSTAT Patent Citations database.

4.3 Descriptive Statistics about Sample

The total sample consisted of 21,093 patent class years spread over a period of 1979 until 2016. Figure 1 shows an overview of the numbers of granted patents in all patent classes, which illustrates changes in patenting behavior over the years.

(34)

34 The number of patents granted in different patent classes seem to increase in the first 10 years up to a maximum of 20,000 granted patents per year. While the number of granted patents fluctuate over the years, as of 2004 this number exceeds 20,000 granted patents per year in certain patent classes. The highest amount of granted patents occurs in 2010 in patent class C12N with 139,225 patents.

Figure 2: Frequency of Dominant Designs distributed over the years

A total of 2,709 dominant designs were identified. This represents 13 percent of the total sample. Figure 2 shows that the occurrence of a dominant design seems higher before 1990. This might be explained by the change in patenting behavior. While the number of patents filed seems to increase in some patent classes it might become harder for patents to become the dominant design. However, this should not be misinterpreted as the figures do not show the amount of citations on which the dominant design is based.

Table 2 Descriptive statistics of control variables in years

N Minimum Maximum Mean Std. Deviation

Age 21093 0 5 0.01 0.124

(35)

35 Concerning the control variable table 2 shows that the maximum length a patent was most influential in a patent class is 5 years. This means that inventions are not able to maintain their position as the most important one longer than this period. Whereas the average period a patent remains most influencing is far below one it is quite exceptional if a patent is able to maintain its status as the most influencing patent. The biggest lag found in the sample was 42 years. The average of 7 years shows that it might take some time for an innovation to get noticed by the industry.

4.4 Statistical Model

The data is analyzed using a within Fixed Effects model, which was chosen for several reasons. The dataset contains a longitudinal structure whereas the patent class years reflect 37 years of data. Hence, there was reason to believe the model needed to control for time-variant fixed effects. Besides that each patent class might be different in nature. Thus, it can be assumed there are individual effects of patent classes. To identify the individual effects the Breusch-Pagan Lagrange multiplier test was performed. The results were significant (Χ2 (4) =304.15 , p < 0.01). Accordingly, the null hypothesis for homoscedasticity must be rejected. This is because the variance of the residuals of the dependent variable increase as a function of independent variable. It is thus implied that there are significant differences across patent classes. In order to check the panel data for serial autocorrelation the Wooldridge test was conducted (Drukker, 2003). Similarly, the test was significant (F = 364.986, p <0.01). Hence, the null hypothesis of no serial autocorrelation should be rejected. Whereas the data contained both heteroscedasticity and serial autocorrelation using an OLS regression would lead to a wrong estimate of the standard errors. Hence, to figure out which estimator would estimate the standard errors correctly, a Hausman test was performed. This test identifies whether the heterogeneity bias would be solved by estimating the within or between effects. Whereas the test statistic is significant ((Χ2 (4) = 1469.91, p < 0.01) it is implied that the

(36)

36 estimation method used for this data should control for within effects, thus tested with a fixed effects model. To control for heteroscedasticity and autocorrelation, the estimation method was performed with robust standard errors.

In order to test our hypothesis the estimation model includes a lagged effect of the dependent variable as it is of interest to identify changes over time. Using a lagged effect in a fixed effects model gives rise to a dynamic panel and can be a potential source for endogeneity in a fixed effects panel estimation (Nickell, 1981). However, this problem diminishes when time period is relatively large. It is argued that if the time period is large, the bias arising from using fixed effects when strict exogeneity fails is minimal (Wooldridge, 2009). Similarly, it is argued that he Nickell bias may be large in short panels but it vanishes as the time period increases (Arellano, 2003). Whereas, the data in this research contains 37 years is based on a relatively large time period. Thus, the model was tested with a within fixed effects model.

In order to test the interaction effect between the dependent and moderating variable an interaction term was created. Where it is commonly argued that the interaction effect should be based on mean centered or standardized values in order to decrease the change of multicollinearity, this step is not necessary (Hayes, 2012). Whereas the composite indicator is based on standardized values, and the dependent variable is binary it seemed redundant to transform the measures before creating the interaction term. The interaction effect was added to the equation of the main effect which resulted into the following model:

𝐼𝑃𝑖𝑡−1= 𝛼𝑖+ 𝛽1𝑗∗ 𝐴𝑔𝑒𝑖,+ 𝛽2𝑗∗ 𝑇𝑖𝑚𝑒 𝐿𝑎𝑔 + 𝛽3𝑗∗ 𝐷𝐷 + 𝛽4𝑗∗ 𝑃𝑎𝑡𝑒𝑛𝑡 𝑆𝑡𝑟𝑒𝑛𝑔𝑡ℎ +𝛽_5𝑗∗ 𝐷𝐷 ∗ 𝑃𝑎𝑡𝑒𝑛𝑡 𝑆𝑡𝑟𝑒𝑛𝑔𝑡ℎ + 𝑢_𝑖𝑡

𝛼𝑖(i = 1 … . n)

(37)

37

5. Results

In this chapter the results from the statistical tests of the research will be explained. The chapter starts with descriptive statistics on all the variables tested in the model to provide an overview of the data. The section follows with an examination of the significant correlations between the variables. The concluding section provides clarification statistical significance of the findings from the regression analyses that have been conducted to test the hypotheses constructed in chapter 3.

The results are based on a dependent variable that was transformed using a log transformation. This transformation was done as it improved normal distribution of the data significantly as it scaled the skewness. Since this study contains database research, reliability and scaling analyses were not applicable. Thus, these tests were not performed.

5.1 Descriptive Statistics and Correlations

Table 3 provides descriptive statistics and the bivariate correlation analysis between the variables in the final sample. The relationships between the variables were measured by the Pearson product-moment correlation coefficient. First of all, the correlation between the dependent and independent variable is presented. Secondly, the correlations between the dependent and control variables are discussed. Finally the correlations between the dependent, independent and control variables with the moderating variable are outlined.

The final sample consisted of 21,093 observations within patents classes from 1979 to 2016. Considering the correlation between the dependent and independent variable, there is a significant, moderate, negative relationship (r = -0.52) between innovative performance and whether there is a dominant design within an industry or not.

Secondly, innovative performance is very weak but significantly and positively related to the variable time lag (r = 0.09). Besides that, there is a very weak but significant relationship

(38)

38 between innovative performance and the control variable age (r = -0.11). Additionally, both time lag and age are significantly related with the independent variable dominant design. Whereas time lag has a negative relationship with dominant design (r = -0.12), age has a positive relationship with dominant design (r = 0.21).

Looking at the relationships with the moderating variable patent strength both the independent and dependent variable have a significant correlation. Innovative performance has a moderate and positive correlation with patent strength (r = 0.62). On the contrary, dominant design has a weak and negative correlation (r = -0.31) with patent strength. Both control variables, time lag and age, have very weak correlations with the moderating variable. Time lag has a positive relationship (r = 0.08), whereas age has a negative relationship (r = -0.05). An overview of the full set of the correlations can be found in table 3 below. In order to check the moderate correlations dominant design and patent strength with innovation the variables were checked for multicollinearity. In order to confirm there is no multicollinearity between the variable the variance inflation factors (VIF) were obtained. Whereas none of the VIF’s exceeded the value of 4 (min = 1.00 max = 1.18), it was concluded there was no multicollinearity between the variables.

(39)

39

Table 3: Sample Size, Mean, Standard Deviations and Correlations Matrix of Variables

N Mean Std. Dev. (1) (2) (3) (4) (5) (1) Innovative Performance 21,093 1.89 0.91 1 (2) Dominant Design 21,093 0.13 0.33 -0.52** 1 (3) Time Lag 21,093 6.84 6.12 0.09** -0.12** 1 (4) Age 21,093 0.01 0.21 -0.11** 0.21** -0.01 1 (5) Patent Strength 20,386 0.00 1.47 0.63** -0.31** 0.08** -0.05** 1

** Correlation is significant at the 0.01 level (2-tailed) * Correlation is significant at the 0.05 level (2-tailed)

Innovation and technological standards : the effect of patent strength on the relation between dominant designs and innovation in European industries

Thesis