Analysis of convergence in paper and printing industry. Matti Karvonen* and Tuomo Kässi

Transcription

1 Int. J. Engineering Management and Economics, Vol. 1, No. 4, Analysis of convergence in paper and printing industry Matti Karvonen* and Tuomo Kässi Industrial Management, Technology Business Research Center, Lappeenranta University of Technology, P.O. Box 20, Lappeenranta, Finland *Corresponding author Abstract: Fusion of new technologies and converge has led to the emergence of new industries. This paper uses patent data of the paper and electronics companies to evaluate their overlapping technologies in discovering new businesses based on technology convergence. Patent data was collected from the 87 main players operating in the RFID value chain. Backward citations and technology cycle time have been used to get an idea of knowledge spillovers and rate of technological innovation between the players. The recognised trends of the growing overlaps of the technological fields show indications for knowledge spillovers and convergence between industries. This type of patent analysis helps to recognise trajectory changes early in the industry and helps companies to take strategic decisions accordingly. Keywords: evolution; patent analysis; knowledge spillovers; technology cycle time; TCT; technological trajectories; convergence. Reference to this paper should be made as follows: Karvonen, M. and Kässi, T. (2010) Analysis of convergence in paper and printing industry, Int. J. Engineering Management and Economics, Vol. 1, No. 4, pp Biographical notes: Matti Karvonen works as a Researcher at Technology Business Research Center. He received his Master s degree from Joensuu in 1999 and Lappeenranta in He is currently pursuing for his Doctoral thesis related to industry renewal in converging environments. His research interest focuses on technology management, industry evolution and industry renewal. Tuomo Kässi has an MSc in Technical Physics and in Economics. He wrote his Doctoral thesis in Industrial Engineering and Management. He has worked in research and development, and sales and marketing; and as the Director of machinery production and a Consultant in industry some 25 years before entering his academic career. His research interests are in technology management, innovation and strategy. Copyright 2010 Inderscience Enterprises Ltd.

2 270 M. Karvonen and T. Kässi 1 Introduction The fact that technological development leads to convergence or fusion of technologies (e.g., camera or TV in mobile devices) has become the focus of many studies (Bonnet and Yip, 2009; Lei, 2000; Palmberg and Martikainen, 2006; Stieglitz, 2003; Wirtz, 2001) and is commonly referred to by the concept of convergence. However, at least so far in the literature the convergence phenomenon has predominantly been analysed with a focus on information and communication technologies and there have been relatively few studies (Bröring et al., 2006; Bröring and Cloutier, 2008) from other industries. Convergence has been often used but rarely defined concept. Blurring, or even disappearance, of industry boundaries, overlapping technologies and markets are used to describe this phenomenon. In industries such as telecommunications, information technologies and electronics digital convergence have meant that formerly distinct industry boundaries have already largely faded (Duysteers and Hagedoorn, 1998; Suoranta, 2008; Curran and Leker, 2009). However, there exists conceptual confusion within the field of convergence, and as a result convergence often means different things to different authors. The literature also presents very contradictory conceptions of the effects of the convergence phenomenon, and it is evident that in order to make right conclusions about the phenomenon it is essential to understand the nature and stage of change in the converging environments (Hacklin, 2008; Yoffie, 1996). Regardless of the growing importance of convergence phenomenon from a theoretical perspective there seem to be little consensus in what convergence means for and how it affects to the different industry sectors. Similarly from a managerial perspective the effects of convergence on firm strategy and behaviour have not been addressed adequately. The need for deeper understanding of this phenomenon represents underlying rationale of this work. In this paper we give an overview of the current changes in the printing and electronics industries where important technological innovations are moving them closer together as more cross-scientific research can enable the printing sector to utilise technological developments in neighbouring disciplines. A general trend towards these smart products has started a production of new printed intelligence products in the intersection of paper, plastic and electronics industries. These new industry segments present incumbents opportunities for new field of business. On the other hand they are often quite challenging as firms have to utilise knowledge and technologies not within their traditional core fields of expertise. In most cases of convergence sourcing the essential knowledge from beyond their own industry is necessary and key to successful innovation management. Furthermore, they are potentially facing new competitors, who may have been the strong incumbents in their own segments prior the formation of the new segment. Anticipating convergence would enable firms to form strategic alliances or acquire new technologies at early stages in the process of convergence. Thus, it seems to be strategically important to anticipate fading technological areas and industry boundaries at the earliest possible moment (Curran and Leker, 2009). It is also important to understand the nature of convergence process and stage of convergence in order to help to do right strategic management decisions. This study utilises the major RFID industry players in order to analyse the intersection between the paper and electronics industries. Among other areas of printed intelligence, radio frequency identification, especially printed RFID offer new business opportunities for the paper industry.

3 Analysis of convergence in paper and printing industry 271 Patents are generally regarded as precursors of technological developments. While patents information is mainly intended for specialists, especially their relations to other patents can make it valuable source of knowledge also for non-specialists trying to identify general trends. In this paper we first briefly define the concept of convergence in the technological management context. After providing understanding to the concept of convergence we take an overview of the current changes in the paper and electronics industries. Overall this study has two main goals. First, it aims to discover the mechanism of new businesses and knowledge spillovers between sectors based on technology convergence. Second, the paper provides understanding of the stage of convergence and rate of technological innovation between the players. In addition, the paper provides insights to the future competitive environment between paper-based and electronics industries focusing especially on the paper and printing industry transformation. The paper is structured as follows: In Section 2, a concept of convergence will be discussed and previous literature about the stage of convergence will be reviewed. Section 3 describes the data and research methodology. Section 4 presents the results of empirical analysis and finally the study is discussed and concluded in Sections 5and 6. 2 Industry convergence Convergence has been widely used but rarely properly defined term. Its meaning has been tremendous as the vision of convergence justified a number of large corporate mergers and aggressive diversifications within the IT, telecom, media and consumer industries, of which many eventually failed (Lind, 2004). Because of the conceptual confusion we first define what is meant by convergence in our context and then specify the basic forms of convergence. 2.1 Definitions The term convergence has been first applied in mathematics and natural sciences, and it has been used as well in social sciences, architecture and economic sciences. In management literature technological convergence has been mostly associated with the developments in the ICT industry since the introduction of personal computers in the late 1970s and the implied digitalisation of media. (Lind, 2004; Hacklin, 2008) In a managerial context the first use of the term technological convergence is attributed to the historical work of Rosenberg (1963), who in his article of the machine tool industry observes the nature and the consequences of technological convergence. He employs the phrase to describe processes that are commonly used throughout the machinery and metal-using sectors by different (unrelated) industry sectors. Rosenberg termed this process technological convergence because different industries began increasingly to rely on the same set of technological skills in their production processes. Since then the term has been used in various industry settings and there have been a multitude of definitions of convergence (Table 1). Many of the definitions of convergence are based on the idea of growing overlaps in technologies and services, and boundaries between industrial sectors are blurring. Nyström (2008) argues that there should be a clear distinction between technological and industry convergence. According to Nyström (2008) the terms are strongly related, but they are still not interchangeable, because the forces driving technological convergence

4 272 M. Karvonen and T. Kässi may not be the same as the forces driving the convergence of industries of product markets. Furthermore Curran et al. (2010) see that there is clear difference between the terms convergence and industry fusion, which have been so far mainly used interchangeably. In this classification convergence describes a process, where objects move or stretch further from their prior and discrete spots, to a new and common place and fusion describes a process, where objects begin to merge with each other in the very same place of at least one of the objects. In this definition industries would be labelled converging if they began to merge with each other in a new field, which must not have been part of any of the hitherto distinct industries. In contrast in the fusing industries new and replacing segment of sub-segment would substitute parts of the prior segments. (Curran et al., 2010) Regardless of this significant distinction and its meaning when evaluating new industry segments in the intersection of industries, we will be using the both terms throughout the paper as in many cases it is not clear in advance whether industries are actually fusing or converging. Table 1 Definitions of convergence Source Definition Management scope Choi and Välikangas (2001, p.426) Lind (2004) Bally (2005, p.13) OECD (1992, p.93) European Commission (1997) Pennings and Puranam (2001) Wirtz (2001) Yoffie (1996, p.33) Bores et al. (2003, p.1) Thielmann and Dowling (1999).. blurring the boundaries between industries by converging value propositions, technologies and markets.. a market/industry definition generated by technological change.. the growing together of technologies, which fundamentally alters the boundaries of previously distinct industry or market sectors and merges them into new competitive environment the growing overlaps between the technologies, services and firms active in each sector telecommunications merging with media and information technology, i.e., converging with a presumed outcome of e.g., one unified market.... the erosion of boundaries that define and isolate industry-specific knowledge dynamic approach or partial integration of different communication and information based market applications the unification of functions the coming together of previously distinct products which employ digital technologies a process by which the telecommunication, broadcasting, information technologies and entertainment sectors may be converging to a unified market process of interaction between firm environment respectively competitive structure as well as corporate strategy, which leads to a structural connection between previously disparate markets Source: Adapted from Hacklin (2008) Industry and market boundaries Technological change Industry and market boundaries; competition Competition Innovation through recombination; market redefinition Knowledge Innovation through recombination Innovation through recombination Market redefinition Competition and strategy

5 Analysis of convergence in paper and printing industry 273 Table 1 Definitions of convergence (continued) Source Definition Management scope Gill (2008, p.47) Dowling et al. (1998) Fai and von Tunzelmann (2001) Greenstein and Khanna (1997) Lei (2000) Curran et al. (2010).. enabling the addition of disparate new functionalities to existing base products.. Convergence describes a process change in industry structures that combines markets through technological and economic dimensions to meet merging consumer needs. It occurs either through competitive substitution or through the complementary merging of products or services or both at once. The notion of general purpose technologies (GPT) draws upon the view of inter-industry spill-overs... all industries will come to be influenced sooner or later by these new technological paradigms (GPT), we can label this as a process of technological convergence between industries Industries that have been distinct historically.. at the boundaries where formerly separate industries come together in a new industry converging in substitutes or converge in complements convergence occurs when advances or innovations commercialised in one industry begin to significantly influence or change the nature of product development, competition, and value-creating processes in other industries Convergence [fusion] is a blurring of boundaries between at least two hitherto disjoint areas of science, technology, markets or industries. Through this convergence [fusion], a new (sub)segment is being created in a new spot [the same spot] as a merger of (parts) of the old segments. Source: Adapted from Hacklin (2008) 2.2 Evolution and types of convergence Innovation through recombination Technological change and market redefinition; competition Technological change, knowledge spillovers Industry and market boundaries Industry and market boundaries Science, technology, market and industry boundaries In the process of convergence technology bases of companies operating in different industry sectors are becoming increasingly similar, which eventually means that companies compete with the same technological competencies. The new industry segment will either replace the former segments or will complement them at their intersection (Dowling et al., 1998; Bröring et al., 2006). In the substitutive paradigm the new industry segment will replace the former segments (1 + 1 = 1) leading to the competitive convergence. In the cooperative paradigm a new market emerges (1 + 1 = 3) that requires combination of resources and competencies from previously separate industries (e.g., through strategic alliances or other forms of collaboration) leading to complementary convergence. In the competitive paradigm convergence may also imply a need to collaborate and compete at a same time. (Figure 1) Depending on the type of convergence, different forms of new entries may emerge. Firms with new technologies may seek to enter and take over existing markets. In competitive convergence the number of competitors in the relevant market increases

6 274 M. Karvonen and T. Kässi rivalry, while the complementary convergence tends to lessen rivalry. Convergence typically changes the basis for competitive advantage and firms must adapt their strategies depending on the nature and degree of convergence (Dowling et al., 1998). Theoretically one of fundamental questions is related to technological convergence versus path-dependency view where the Shumbeter s work brought the notion of pathdependency and diversification together as it progressed and developed over time (Fai and von Tunzelmann, 2001). The young Schumpeter view was highly path-dependent and supposes that the advent of new products and technologies leads to creative destruction of older products and hence that firms which produce them. The old Schumpeter view instead suggests the adaptability of large enterprises and their ability to initiate as well as passively absorb new technologies (see e.g., Nelson and Winter, 1982; Sood and Tellis, 2005). Figure 1 Basic forms of convergence Industries/Firms (Supply) Technology Needs (Demand) Converging Markets Complementary Coopetitive Competitive Figure 2 S-shaped paths for firms adopting industry-specific and spillover technologies Technological opportunities Spillover paradigm combining established technology, with extraindustry paradigm for new applications 1st trajectory followed by firm in industry s main technological paradigm 2nd trajectory followed by firm in industry s main technological paradigm Paradigm spillover from extra industry Time Source: Fai and von Tunzelmann (2001)

7 Analysis of convergence in paper and printing industry 275 Figure 3 Phases of convergence Distinct knowledge bases Science/knowledge convergence Distinct scientific disciplines beginning to cite each other and collaborate Possible indicators Scientific articles Unassociated technologies Technology convergence Distance between applied science and technology develoment decreases Patent data Intersected technologies Market convergence New product-market combinations General market information Integrated technologies Industry convergence Fusing or converging industry segments General market information Source: Curran and Leker (2008) and Hacklin (2008) We might expect the growing intra-industry diversity in patenting, if the technological convergence hypothesis holds, or alternatively growing cumulativeness in patenting, if the path-dependency view is valid. Cumulativeness or path dependency is a fundamental property of innovative activities while at the same spill-over paradigm can lead to technology diversification followed by product and market diversification (Figure 2). As the spillover paradigm dominates and technology from another industry diffuses, we would expect increased technological convergence between industries or at the minimum least a decrease in the degree of divergence between industries. Gambardella and Torrisi (1998) found that electrical firms performance was positively associated with its technological diversification. However, they found that performance was also positively correlated with greater focus on downstream business operations. Their study indicates that technological competencies are not sufficient to enter different market because other complementary assets are required.

8 276 M. Karvonen and T. Kässi Curran and Leker (2009) base their concept of anticipating convergence upon the assumption of an idealised time series of events (Figure 3) starting with scientific disciplines, where distinct scientific disciplines beginning to cite each other and eventually develop further into closer research collaborations. After the distance between basic science have been decreasing, applied science and technology development should follow (Meyer, 2000) leading to technology convergence. Then in the market convergence, new product-market combinations will emerge, and finally the fusion of firms and industry segments will lead to industry convergence. However, full industry convergence will only take place when technologies and markets converge (see Nyström, 2008). In this paper, we take an evolutionary perspective, as illustrated by the process of cascading transitions from knowledge and technology to application and industry (Hacklin, 2008; Hacklin et al., 2010). It is clear that the process of technological convergence does not necessarily lead to industries converging and complete industrial convergence. Sectors and technologies differ greatly in terms of the knowledge base and learning processes related to innovation. In some sectors, science is the force driving knowledge growth, while in others, learning by doing and the accumulation of advancements are the major forces. (Malerba, 2005) Technological opportunities in science-based firms in general emerge horizontally in related product markets and downstream in user sectors, whereas in scale-intensive and supplier-dominated firms opportunities tend to be upstream in related production technologies (Pavitt et al., 1989). Firms operating principally in the paper and printing industry have characteristics between the supplier-dominated and scale-intensive firms. Typically the process innovations in these industries rely on non-patent methods of protecting intellectual property rights, such as secrecy or tacit knowledge. High tech industries, such as electronics and RFID industries, have characteristics that are clearly science-based. In these industries product inventions are more typical and the use of non-patent options is much more difficult (e.g., when the product enters the market, it becomes vulnerable to reverse engineering.) Patents have been criticised for distinguishing both inter-industry and inter-firm differences in the propensity to patent and differing levels of significance of each patent in relation to technological advantage (e.g., Fai and von Tunzelmann, 2001) and this should be remembered in interpreting the findings. 3 Methods and data Patents have long been recognised as a rich source for the study of innovation and economics of technological change. Patent data include also citations to previous patents opening up the possibility to study spillovers between technologies and industrial sectors. 3.1 Technologies, industrial sectors and knowledge spillovers Economic indicators tend to be collected, classified, and made available by categories based on industry classification systems such as the Standard Industrial Classification (SIC) system. By contrast, the European Patent Office (EPO) relies on classification systems that are based on the International Patent Classification (IPC) system to classify their patent grants. These patent classification systems are not based on the associated industry of manufacture or sector of use. Regardless of the challenges related to

9 Analysis of convergence in paper and printing industry 277 associating technologies and industries in recent years IPC classifications based concordances have become more popular for associating patents with industry-based categories. Concordances between the IPC and industry-based classifications may be attractive, because they have the potential to enable a researcher to obtain patent data from many different patent offices and to aggregate those data by industry using a potentially uniform and consistent methodology (Verspagen et al., 1994; Johnson, 2002; Silverman, 2002; Smoch et al., 2003). Pennings and Puranam (2001) suggest that a growing overlap in the SIC codes from cross-sectional patent analysis could be used as a proxy for convergence. In the supply side convergence, by assuming the validity of the classification schemes, we can track changes in the relationships between different industries or patent classes. A growing overlap among the SIC codes and patent classes signals convergence. Such an overlap can be approximated through the increase in patent citations between classes of patents that are anchored in the specialised intangible assets of firms (Pennings and Puranam, 2001). In order to identify a firm s technological domains, the observed IPC codes in the firm s patent records were identified and classified into technology fields representing the respective firm s major business domains. Patent application in each field indicates an accumulation of knowledge and advancement of technological trajectory (Fai and von Tunzelmann, 2001; Suzuki and Kodama, 2004). IPC codes are a hierarchical way of assigning the category to which every patent belongs. There are eight sections, 120 classes, 628 sub-classes and about 70,000 groups. In our analysis we have utilised a higher level classification, by which the 628 sub-classes are aggregated into 35 technological fields, and these are further aggregated into five main categories: electrical engineering; instruments; chemistry; mechanical engineering; and others. (WIPO, 2008; see Appendix 1) We will call categories chemistry and mechanical engineering as traditional fields and electrical engineering and instruments as emerging fields Citation analysis Citations made are perhaps the most common empirical approach to serve as indications of knowledge flows and spillovers, i.e., the fact that patent B cites patent A may be indicative of knowledge flowing from A to B. (Hall et al., 2001; Jaffe et al., 2000; Deng, 2008). The study of patent citations has its own limitations. Advantages and disadvantages of using patent citation data are extensively discussed by Griliches (1990) and Jaffe et al. (2000). We approach the knowledge spillovers of patents by using patent citations collected from PATSTAT database. The backward citations (i.e., citations made) in the patent document position the new invention technologically with respect to previous patents and forward citations represents citations received to the patent (Figure 4). These citations to previous patents serve an important legal function, since they delimit the scope of property rights awarded by the patent and have also been linked to the originality of a patent and thus economic value. The forward citations (citations made by other patents) are considered to reflect the patent s technological significance, the applicability and the ability of the inventors to benefit from their inventions, i.e., appropriability. All in all, patent citations convey information on two major aspects of innovation. The first is knowledge transmission between inventions and assignees enabling quantitatively to study spillovers and knowledge flow along geographical and industrial dimensions (mainly backward citations). The second line of research counts

10 278 M. Karvonen and T. Kässi the number of time a patent is cited in subsequent citations ( forward citation count) to measure its importance or value. In this paper we concentrate mainly on the first aspect, i.e., knowledge spillovers between industries. (Appleyard, 1996; Jaffe et al., 1998; Lukach and Plasmans, 2002; Deng, 2008). One issue is to what extent citations to previous inventions are patented by the same assignee (self-citations) as presumably citations that belong to the same assignee represent transfers of knowledge that are mostly internalised and cannot be regarded as representing spillovers as they are commonly defined. Self-citations would suggest that the firm has a strong competitive position in that particular technology and is in a position to internalise some of the knowledge spillovers created by its own developments whereas citations to patents of others are closer to the pure notion of spillovers. In other words, a more appropriable technology does not transmit readily through external spillovers. In the intellectual asset management the building of protective patent fences by filing extensively around a valuable patent position have been a common strategy. (Hall et al., 2001; Hall et al., 2005; Graff, 2003). Figure 4 Patent citations Patent Patent Patent Patent Patent Patent Patent Patent Time Generation -1 Generation 0 Generation +1 Backward citations/ Citations made Forward citations/ Citations received In the empirical part we classified patent citations altogether into four groups: 1 citations occurring within the same firm (self-citations) and within the industry 2 self-citations beyond the industry 3 external citations within the industry 4 external citations beyond the industry. Such distinction is made because the meaning for the industry structures and economic value of each type of citations may be different. Self-citation would imply that the firm is gaining a more competitive position in that field while an external citation to another patent within the industry may suggest that the citing firm is entering a technological competition where the cited firm might have already built a strong competitive position.

11 Analysis of convergence in paper and printing industry 279 The knowledge spillovers from external citations and beyond industry patent may have much weaker implications on the technological competition because the cited patent is in a different technology field (see e.g., Deng, 2008). Thus making distinctions between these four types of citations may shed light on the intensity of technological competitions between the paper and printing and other electronics industry players Technology cycle time The technology cycle time (TCT) measures the rate of technological innovation and how quickly firms are absorbing new technologies. The TCT is defined as the median age of the patents cited on the front page of a patent document (Kayal, 1999). When the TCT was long, the technology was older, and when the TCT was short, the developed technology was younger. Typically emerging technologies have short cycle times, four years or less, whereas more mature technologies can display TCT that averages 15 years or more (Hirschey and Richardson, 2004). TCT uses the rate of substitution in relation to technological progress with the shorter cycle times reflecting faster substitution (fast progress) and longer cycle times reflecting slower substitutions (slow progress). Long TCT s could indicate the cumulativeness of innovative activities implying that the knowledge base that forms the foundation of the technology is large. Short TCT indicate that something new has been added to the knowledge base (Daim et al., 2008; Daim and Jordan, 2008; Kayal, 1999). The areas of convergence can be articulated to be characterised by most rapid growth and innovation and based on that so we could anticipate shorter TCT values in converging environments. In addition Bierly and Chakrabarti (1996) have found that TCT is significantly faster for firms that predominantly generate new knowledge internally, and slower for firms that rely on external sources of new knowledge. 3.2 Sample and databases The primary source for acquiring the list of companies and their activities in the RFID field was an independent consultancy company (Das and Harrop, 2008). Companies were categorised by utilising SIC codes (Thomson ONE Banker, 2008) using the three-digit level in the analysis. Based on Das and Harrop (2008) the companies were in further analyses categorised into four groups: value chain s upstream focused players, downstream focused players, vertically integrated firms, and paper and printing companies. Upstream players are involved in developing, manufacturing and selling identifiers such as chips, antennas, tags and labels, or devices such as readers and printers. Downstream focused firms are involved in software and integration, and they operate closely in the end customer interface. Vertically integrated firms operate broadly in the whole value chain with activities both upstream and downstream of the value chain. In addition we separately analysed printing and paper companies operating in the field. The paper and printing companies are operating in the middle of the value chain and have recently entered to the emerging industry. All in all, the two main industrial classes of each group (vertically integrated, upstream focused, downstream focused, paper/printing) were taken to analyse the possible trajectory changes and knowledge spillovers between industries. The concordance (Smoch et al., 2003) between paper and publishing and electronics industry classification and the companies evaluated are presented in Appendix 2.

12 280 M. Karvonen and T. Kässi Information on patenting was drawn from the EPO Worldwide Patent Statistical Database (PATSTAT) (2008) including data on patents, such as citations and IPC codes, and also on patent families based on priority links. The coverage of the database regards documents from more than 80 patent offices worldwide since the 1970s. This worldwide statistical patent database, also known as PATSTAT, was developed by the EPO in 2005 and first distributed in April The elements of PATSTAT are the title and abstract of application, filing, priority and publications dates of the application, applicants and inventors and detailed addresses, the IPC classification symbol, and priority applications. In this study we used these data mainly to identify: 1 in which technology fields the players have patented 2 backward citations of the patents in period analyse the spill-over effects between industries. The limitations of patent data are well known. Not all innovations are patented by firms and the market value of patents is extremely skewed (e.g., Hall et al., 2005). Different technologies are differently patentable and firms may have different propensities to patent their innovations since in some industries patents have a crucial strategic importance. However, patents represent a very homogenous measure of technological novelty and are available for long time series. They also provide very detailed data at the firm and technological class level. For our present purposes, they represent therefore a very valuable and unique source of data on innovative activity and knowledge spillovers. 4 Analysis and results 4.1 Descriptive statistics In the analysis there were altogether 87 firms which were characterised into four different clusters under the following headings: upstream focused players (N = 26), vertically integrated players (N = 23), downstream players (N = 17) and paper and printing companies (N = 18). In the empirical part we analysed each cluster patents in years and their citations Each patent grant is assigned to sub-classes to determine the nature of the patent. One patent can be assigned to more than one sub-class if the patent finds application in various industrial domains. Each company has a few sub-classes to which most of their patents are assigned. These sub-classes are like their core technological competencies. If a company has been granted patents only in a few sub-classes, it can be said technologies employed by the company are highly focused in a narrow range of areas. On the other hand, if all the patents are not concentrated on a few sub-classes, research can be said to be diversified. The way we have measured this type of diversification is to find out what percentage of IPC codes of granted patents are assigned to the top 50 sub-classes for the four categories. We analysed each cluster top 50 IPC classes covering almost 90% of all the patents, so our data represent well each of the four category technological evolution. The players (Table 2) patents distribution shows clearly that downstream players have most focused technological competencies as 93.6% patents are related to computer and communication technologies, especially to computer technology (60.9%) and digital technology (12.1%). Integrated players have also focused patents mostly to computer and communication technologies (64.4%), but have got also

13 Analysis of convergence in paper and printing industry 281 significant share of patents (24.3%) in electronics and electrics. Regardless of the huge growth of the downstream player patents in computer and communications since 2003 the integrated players have stayed marginally as the dominant player. Upstream players have more diversified patents distribution as their patents are more evenly distributed to computer technology (29.4%), semiconductors (17.9%), telecommunications (13.5%) and audio visual technology (12.6%). Most of the patents (46.1%) are related to electronics and electrics (more specifically to semiconductors 32.0%, basic communications processes 8.1%, and electrical machinery 6.0%) and this share has increased over the time period. Upstream players have quite significant patent share also in computer technology (20.9%) and instruments (16.2%). Paper and printing firms have most diversified patent portfolio. Table 2 The players TOP50 IPCs patents distribution Industry/IPC group Paper and Printing (N = 18) Upstream electronics (N = 26) Vertically integrated (N = 23) Downstream players (N = 17) Patent count 77, , ,560 43,518 TOP50 IPC (%/all) 84.7% 87.6% 91.7% 98.0% 1 Electrical engineering 22.3% 80.8% 88.7% 95.8% Electronics and electrics (1,5,8) (7.2 %) (46.1%) (24.3%) (2.2%) Computer and communication (15.1 %) (34.7%) (64.4%) (93.6%) (2,3,4,6,7) 2 Instruments 16.5% 16.2% 6.9% 3.3% Optics (13.4%) (4.7%) (3.3%) (0.4%) Measurement (1.8%) (5.4%) (2.5%) (1.1%) Control (1.2%) (5.4%) (1.1%) (1.8%) Other (0.7%) 3 Chemistry 18.5% 0.5% 1.6% 0.0% Macromolecular chemistry, polymers (3.9%) - (0.5%) - Basic material chemistry (2.8%) Surface technology, coating (11.4%) (0.5%) (0.8%) - Other chemistry (0.4%) (0.3%) 4 Mechanical engineering 41.4% 2.0% 2.9% 0.6% Handling (15.0%) (0.8%) (0.4%) (0.2%) Paper machines (20.7%) - (2.1%) (0.4%) Other special machines (5.7%) (1.2%) (0.4%) - 5 Other fields 1.3% 0.5% 0% 0.3% Paper and printing firms have traditionally been strong in mechanical engineering and chemistry patents. In the sample 41.4% of the paper and printing firms patents were related to mechanical engineering (handling 15.0%, paper machines 20.7%, and other special machines 5.7%) and 18.5% chemistry patents (surface technology, coating 11.4%, macromolecular chemistry, polymers 3.9%, and basic material chemistry 2.8%). The

14 282 M. Karvonen and T. Kässi paper and printing firms patents 22.3% were related to electrical engineering (audio-visual technology 7.0, computer technology 6.7%, and semiconductors 4.5%) and 16.5% were related to instruments (optics 13.4, measurement 1.8%, and control 1.2%). Figure 5 suggests that there has been a change in the focus for the paper and printing industry in the recent years. With the emergence of printed intelligence and RFID there has been a significant increase in the patenting activity into the new fields. The technology integration type of convergence seems to have opened new market opportunities for the paper and printing industry players. Regardless of the dramatic shift in the patenting activities it has to be remembered that relatively paper and printing firm s patent share is still quite marginal compared to the other players. However, their share has increased especially in patents related to audio visual technology, computer technology and semiconductor devices. The results of the analysis indicate that patterns of technological integration seem to have significant influence on the technological cores of the firms operating in the field. Figure 5 Paper and printing firms patents distribution The paper and printing firms patents (N=18) distributution to the technological fields % Electrical engineering 80 % Instruments Electrical engineering Relative share (%) 60 % 40 % Chemistry Instruments Chemistry Mechanical Engineering 20 % Mechanical Engineering 0 % Year 4.2 Citation analysis The analysed firms had made altogether 1,991,602 citations of previous patents, of which 211,110 (10.6%) were self-citations. Table 3 shows that the vertically integrated and downstream players have on average made most of the backward citations. Paper and printing firms have made an average of only 2.14 citation references to their previous patents. We analysed citations made of each cluster distribution top 50 IPC classes in the technological fields.

15 Analysis of convergence in paper and printing industry 283 Table 3 Citations made (backward) Industry/IPC group Paper and printing (N = 18) Upstream electronics (N = 27) Vertically integrated (N = 24) Downstream players (N = 18) Citations made 166, ,321 1,100, ,931 Citations (average) TOP50 IPC (%/all) 75.1% 72.7% 87.2% 95.8% Self-citations 16,496 36, ,791 20,376 (9.9%) (7.5%) (12.7%) (8.6%) Within industry 70.4% 85.4% 89.8% 95.1% Beyond industry 29.6% 14.6% 11.2% 4.9% External citations 150, , , ,555 (90.1%) (92.5%) (87.3%) (91.4%) Within industry 63.1% 82.7% 84.1% 90.7% Beyond industry 36.9% 17.3% 15.9% 9.3% The paper and printing firms backward citations distribution to technological fields show a clear declining trend into mechanical engineering, whereas electrical engineering patents have been rising enormously in recent years (Figure 5). The biggest growth in the electrical engineering patents has been especially in semiconductor and electrical machinery patents indicating the technological convergence view and vertical spillovers between different industries. The downstream players backward citations are very focused and have become over time even more related to computer and communication patents. Upstream players backward citations are mostly concentrated on electronics and electrics, especially on semiconductor patents, whereas integrated players citations have been increasing in the computer and communication technology. As a whole there seems to be quite a lot of horizontal spillovers between competitors in the electronics industry, but the vertical spillover flow between firms in different industries is very low. The mean percentage of self-citations made is 10.6%. The fact that the percentages are much higher for vertically integrated players supports the notion that innovation is concentrated in very large firms (IBM, Nokia, Intel), and hence the likelihood that they will cite internally is higher. In the other groups, innovation is more widely spread among highly heterogeneous assignees. Interestingly, paper and printing firms self-citations are predominantly related to the mechanical engineering and chemistry indicating that the capability and core competencies are still premidominantly in the traditional fields. Almost all the downstream players self-citations are in computer and communication technologies indicating a very strong competitive position in these technologies. Compared to the downstream firms, upstream players have more diversified technological competencies, where a major share of the self-citations is however accumulated in electronics and electrics patents. Integrated players have numerous selfcitations both in electronics and electrics and computer and communication technologies. Although upstream players have increased their share in electronics and electrics patents and downstream firms their share in computer and communication technologies, the

16 284 M. Karvonen and T. Kässi competitive position of integrated players have remained quite strong in these technologies throughout the period. 4.3 Technology cycle time TCT was applied to measure technological innovation progress for the industry players. Overall all the electronics industry players have on average quite similar TCT values (Table 4). Downstream players seem to have on average fastest technological progress, but there is also a lot of variation between different years. Paper and printing companies have a longer median age of cited patents and also there is a lot of variation. On average it is evident that the paper and printing companies prior art is relatively older and only in years 1985 and 1987 it was lower than the other players. Table 4 Average values of TCT Industry player TCT average ( ) Standard deviation Upstream 5, Paper/printing Integrated Downstream All 5, Figure 6 The paper and printing firms backward citations distribution The paper and printing firms (N=18) backward citations % Electrical engineering 80 % Instruments Electrical engineering Relative share (%) 60 % 40 % Chemistry Instruments Chemistry 20 % Mechanical engineering Mechanical engineering 0 % Year

17 Analysis of convergence in paper and printing industry 285 Figure 7 The players TCT trend Technology cycle time (TCT) TCT (years) 10,00 9,50 9,00 8,50 8,00 7,50 7,00 6,50 6,00 5,50 5,00 4,50 4,00 3,50 3,00 2,50 2, Year Paper/printing Upstream Integrated Downstream Figure 7 presents the trend data collected from 1978 to The TCT trend for the paper and printing companies was decreasing in years indicating a fast technological progress and new knowledge was added to the knowledge base. The second faster period for new knowledge was from years These trends are quite in line with the growth of beyond industry citations (see Figure 6) in the same time periods. 5 Discussion Regardless of the growing importance of the concept of convergence there is no consensus among researchers concerning triggers, processes and outcomes of convergence (see e.g., Nyström, 2008). Industry boundaries definition is of high importance for determining competitive arena. Competition in converging environments is coming from several directions well beyond traditionally defined boundaries and thus challenges managerial decision making. Due to its high strategic importance, an early identification of trends of convergence and anticipation of changing industry structures matters to all stakeholders, including managers, academics and regulators. The recognised trends of the paper and printing industry would suggest that technological profiles of firms have become more diversified, but whether the phenomenon of convergence have been pervasive enough to blur industry boundaries is still uncertain. The citation analysis indicates that there seem to be real convergence in technologies, as invention more heavily borrows information or processes from electronics field. We distinguish self-citations from external citations, and further divide external citations into two groups: within the industry and beyond the industry citations. Self-citations typically imply a more competitive position in that field (capability), external citations may suggests that the citing firm is entering a technological competition within (competition) or beyond (diversification) industry. The paper industry players make on average fewer citations than the other players. However, both self-citations and

18 286 M. Karvonen and T. Kässi external citations indicate a quite strong capability diversification to the new fields. TCT seem to be also shorter in most important convergence periods where new beyond industry knowledge were absorbed to the knowledge base. TCT seem to be also faster for firms that generate new knowledge mostly internally. For the paper and printing companies technological convergence seem to be in the application phase by now as companies focuses on developing products and services that leverage the underlying advantages provided by the technological basis. The potential industrial convergence might lead to a collision of business models as the players suddenly become competitors to the existing electronic firms. It should be however noted, that the overlapping technologies resulting in a converged technological base only partly substitutes the conventional electronics in silicon chip, displays and lightning industries. Moreover, the self-citation analysis indicate analysis indicate that the paper and printing firms have made self-citations mainly to their traditional fields of core competencies. Considering the whole electronics industry the paper and printing companies are still quite marginal players, but quite huge new markets should be available when technology finally progress to low cost flexible substrates. It seems that the new rules of competitive environment results in alliances and M&A activity reaching beyond previously established industry boundaries. In this potential industrial convergence phase competing and collaborating at the same time will be a key element through the convergence process, which poses challenges on partnering and alliance capabilities of firms. It seems evident that the importance of external innovation will increase and the winners will be those who succeed to capture of external innovations and learn to use collaborative R&D. The low investments in R&D activities might however decrease the paper and printing firms ability to gain innovation from outside the company. In the transformation also the balance between old and new, ( strategic duality ) equilibrium between incremental and radical innovation will be important. 6 Conclusions There are no doubts about the importance of knowledge spillovers for economic growth and its importance should be especially high in converging industry environments. The process of technological convergence is initiated through evolutionary spill-over between knowledge and technology bases of distinct industries, which eventually can expand to more applied level of convergence. In this paper we have examined the overlapping technologies and spillover effect between the paper and electronics industry players. The results of the study indicate that paper and printing companies still patents predominantly in mechanical technologies and chemicals, but increasingly patents in electronics technologies, suggesting that the industries are becoming more technologically convergent. This asymmetric convergence and the erosion of industry-specific competencies have meant that the paper industry is becoming more technologically diversified. Technologically dominated stage of convergence and domain of similar technologies does not, however, automatically lead to industries (or sub-industries) converging and thus complete industrial convergence. All in all, there is a clear demand for means to anticipate convergence and understand its meaning in the reorganisation and transformation of industries. The backward citation and TCT analysis helps companies to recognise trends early in the industry and can