Scuola di Dottorato in Scienze Economiche e Statistiche Dottorato di Ricerca in Direzione Aziendale XX ciclo
Alma Mater Studiorum - Università di Bologna
0B
Technology Acquisition Through Licensing Implications for firm’s strategy
Maria Isabella Leone
Dipartimento di Scienze Aziendali 2008
Alma Mater Studiorum – Università di Bologna
DOTTORATO DI RICERCA
Direzione Aziendale Ciclo XX Settore scientifico disciplinare di afferenza:
SECS-P/08
TECHNOLOGY ACQUISITION THROUGH LICENSING Implications for firm’s strategy
Presentata da: Maria Isabella Leone
Coordinatore Dottorato Prof. Federico Munari
Relatori Prof. Salvatore Torrisi Prof. Paolo Boccardelli Prof. Raffaele Oriani
Discussione, 4 Giugno 2008
2
TABLE OF CONTENTS PREFACE.........................................................................................................................................................5 INTRODUCTORY SECTION........................................................................................................................8 INTRODUCTION...................................................................................................................................................... 9 Research Background .............................................................................................................................................. 9 Overall aim of the thesis....................................................................................................................................... 13 Structure of the thesis........................................................................................................................................... 16 References................................................................................................................................................................... 18 LITERATURE REVIEW AND RESEARCH QUESTIONS............................................................................21 The rising interest in licensing within the strategic management research............................... 21 Extensive investigation of the Licensor’s Perspective ............................................................................ 23 The Under-investigation of the Licensee’s perspective.......................................................................... 28 Research questions................................................................................................................................................. 34 References................................................................................................................................................................... 35 METHODOLOGY....................................................................................................................................................40 Secondary analysis ................................................................................................................................................. 40 Dataset search.......................................................................................................................................................... 42 The Financial Valuation Group Intellectual Property Dataset.......................................................... 44 Data cleaning and integration.......................................................................................................................... 46 References................................................................................................................................................................... 52 RESEARCH PAPERS.................................................................................................................................. 54 TECHNOLOGICAL EXPLORATION THROUGH LICENSING: NEW INSIGHTS FROM THE LICENSEE’S POINT OF VIEW ...........................................................................................................................55 Introduction............................................................................................................................................................... 56 Theory and Hypotheses ........................................................................................................................................ 57 Method ......................................................................................................................................................................... 63 Results .......................................................................................................................................................................... 71 Descriptive statistics and correlations.......................................................................................................... 71 Discussion and Conclusion.................................................................................................................................. 74 References................................................................................................................................................................... 76 FUEL ON THE INVENTION FUNNEL: TECHNOLOGY LICENSING-IN, ANTECEDENTS AND INVENTION PERFORMANCE...........................................................................................................................84 Introduction............................................................................................................................................................... 85 Hypotheses ................................................................................................................................................................. 87 Method and Data..................................................................................................................................................... 93 Results ........................................................................................................................................................................100 Discussion and Conclusion................................................................................................................................103 References.................................................................................................................................................................109 UNCERTAINTY, FLEXIBILITY AND UPFRONT FEE.............................................................................. 118 IN PATENT LICENSE ........................................................................................................................................ 118 Introduction.............................................................................................................................................................119 Hypotheses Development...................................................................................................................................121 Data and Method...................................................................................................................................................125 Results ........................................................................................................................................................................134 Discussion and Conclusion................................................................................................................................137 References.................................................................................................................................................................140 CONCLUSION AND FUTURE RESEARCH ..........................................................................................147 OVERALL CONCLUSION AND IMPLICATIONS....................................................................................... 148 DIRECTIONS FOR FUTURE RESEARCH.................................................................................................... 151
References.................................................................................................................................................................155 UPCOMING RESEARCH PAPER...........................................................................................................157 DOES LICENSING FOSTER RAPID INVENTION? ................................................................................... 158 Introduction.............................................................................................................................................................159 Technology in-Licensing and Invention Performance .........................................................................161 Data and Method...................................................................................................................................................170 Results ........................................................................................................................................................................182 Conclusions...............................................................................................................................................................186 References.................................................................................................................................................................188 APPENDIX A .............................................................................................................................................199 APPENDIX B .............................................................................................................................................210
1B
4
PREFACE My Ph.D. adventure started in late 2003 while I was developing my Master thesis on the strategic evaluation of intellectual property rights advised by Paolo Boccardelli and Raffaele Oriani for the chair held by Professor Franco Fontana. After I graduated in Management in 2004 these three introduced me to an exciting and challenging academic carrier encouraging me to join the research group at LUISS Business School and to enter the Ph.D. program in management at the University of Bologna. Their belief in my enthusiasm for doing research and carrying out rigorous research together with their confidence in my teaching skills became the beginning of my endeavour into my new academic carrier. They supported me beyond reasonable expectations in my efforts to overcome the challenging tasks of writing a PhD thesis on a topic that is only now being developed internationally. For all these reasons, I would like to thank Prof. Franco Fontana for early encouragements as well as for recognizing my potential as a future researcher and university educationist (a very good mentor!); I would like to express my thanks to Paolo for his positive attitude and friendly nature and his ability to make me do things successfully even without prior experience. I’ve learned a lot from this process (a very good manager!); I would like to thank Raffaele for his passion for research and for encouraging me to do my very best. Without his high standards and expectations, as well as his continuous endorsements the tasks seemed unconquerable (a very good co-author!). Paolo and Raffaele, thanks for your friendship as well. It definitely makes our partnering very effective and pleasant. Thus, just graduated from Management at LUISS Guido Carli University in Rome with lots of enthusiasm and hopes I approached the examination to enter the Ph.D. program in Management at the University of Bologna. My initial doubts about pursuing this Ph.D. program were quickly substituted with commitment and eagerness once I was notified that my application had been accepted and once my Ph.D. adventure commenced in January 2005. From then to mid-2006 my life as PhD Student developed and grew becoming extremely busy and stressful filling my life with seminars, lectures, research project, teaching assistance, students tutoring, exams, pre-proposals, and – last but not least – the development of my research proposal. I would like to express my gratefulness to all the individuals and institutions of the Department of management of the University of Bologna – my supervisor Professor Salvatore Torrisi, first - that hosted me and provided me with the required formation to develop and improve my research ideas and capabilities. During those months, my Ph.D. colleagues played a very important role in providing me with a constant support and thereby helping me through the most difficult times. Sara, Serena, Francesco, Matteo, Michele and Riccardo, for all the time spent together, having fun, sharing efforts, going out and for our reciprocal support, I would like to thank you very much. Another very important part of my Ph.D. experience took place in Copenhagen at the Department of Innovation and Organizational Economics (INO) at Copenhagen Business School (CBS) where I spent the final year of my PhD to finalize my dissertation. It proved to be a very exciting and fruitful research environment at which I enjoyed many advantages. I would like to thanks Professor Keld Laursen who hosted me for an extended period at CBS and provided me with all the necessary facilities to work effectively. He gave me the opportunity to develop and further improve my thesis in general. Our constant exchange of ideas and our growing interaction lead to a very nice friendship paving the way for a long-term partnership, currently in the form of co-authorship of one of my dissertation paper. Likewise,
I would like to thank Professor Toke Reichstein for his amazing kindness and helpfulness introducing me to econometric programming that only some months ago was unknown to me. My countless questions, inquiries and requests on problem-solving in handling data possibly made him realized that we might as well become friends! A very nice friendship that not only helped me to overcome academic difficulties and challenging times, but during my stay in Copenhagen it was feed by cherished moments that we shared with others in Wonderful Copenhagen. I am sure we both hope that our current co-authorship will lead to a long-term partnership that will enable us to reach promising targets. For the above-mentioned sharing moments, I would also like to thank the remaining people of CBS (visiting and non) that made me feel part of their department and with whom I had a lot of funny and relaxing days: Maria Theresa, Christian, Stefanie, Francesco, Monica, Larissa, Christoph, Vikram, Serden, Margret, Mark, Henrik, Lars, Henrich, Kristina, Yen, Christine, Rasmus and Adele. I will never forget my first julefrokost with all of you! Thanks also to all the senior professors at INO for their kind hospitality. I express my sincere thanks to my supervisor Salvatore. In particular I think your very nice and fruitful stay at CBS where we worked together with Keld on our paper was beneficial and rewarding. Not only did we come a long way in our academic undertakings, but I also think our formal relationship turned into an informal friendship that I believe will be essential for our future collaboration. Thanks again to all my co-authors: Keld and Salvatore, Paolo and Mats, Raffaele, and Toke. Your support and commitment have been instrumental for achieving these promising results of my research and thus to collect all these papers for my dissertation. Matteo, with whom I have shared the same path since 2000 where we both started our students experience at LUISS Guido Carli University, has my affection in being my best colleague ever. He is one of the most brilliant guys I ever met and a very good friend whom I trust completely. I would also like to pay thanks to my “young” colleagues, Federica and Francesca and lately Luca and Roberta who took care of all the tasks associated to the teaching assistance that I left while I was visiting the CBS. Hope our relationship will keep growing and will turn into a long-term friendship going beyond the boundaries of academia. I would also like to thank my friends/advisors Francesco Di Ciommo and Massimiliano Granieri for valuable comments and suggestions along this long journey. I would like to thank all the people from the many conferences at which I’ve presented my research, for their useful suggestions and for encouraging me in pursuing this promising field of research focusing on the demand side of licensing. I mainly refer to Alfonso Gambardella, Ashish Arora, Bronwyn H. Hall, Ove Granstrand, Scott Stern, Brian S. Silverman, Markus Reitzig, Marco Giarratana, Fredrik Tell, and Nicolas van Zeebroeck. I would like to express all my respect to Mike Mard from the Financial Valuation Group. He very enthusiastically supported my work from the very beginning by giving me the opportunity to employ the database developed within his company. This dataset acts as the input to all of the empirical tests of my research ideas. I would like to thanks my non-academic friends Francesca, Giuni, Sonia, Luca, Aldo, Manlio, Cristina, Lola, Manuela, Marco, Riccardo, Cesare, Gianluca and Alberto who have always struggled to understand what exactly I have been doing and even so to express interest and
6
support me by providing very useful and objective ways to judge and inform my work from the outside. Last but not least, I would like to thank my be-loved parents who have been very patient and supportive and who have eased my path by providing me with all the understanding and the help I needed. I definitely thank my be-loved brother Carlo for teaching me how to live life easily and for his caring and his kindness. He is always ready to help me out or just to drive me to the station/airport in order for me to take the train/airplane somewhere near or far. I finally own thanks to myself for not having given up never despite many encountered difficulties and to have still kept the passion and the enthusiasm to prove myself and to face new challenges everyday. I hope the interested reader will enjoy the outcome of this long, exciting, difficult but promising research journey! ;-)
Tivoli, April 28th, 2008 Isabella
7
INTRODUCTORY SECTION
2B
8
INTRODUCTION
8B
“The OECD economies are increasingly based on knowledge and information. Knowledge is now the recognized as the driver of productivity and economic growth, leading to a new focus on the role of information, technology, learning in economic performance. The term “knowledge-based economy” stems from this fuller recognition of the place of knowledge and technology in modern OECD economies.” (OECD, 1996)
Research Background
18B
In the 1980s, the concept of Knowledge-based economy started being circulating within the academic community as a way to address the third industrial revolution, founded on the important role of new information and communication technologies that would characterise the coming century (Harris, 2001). The intellectual origin of the Knowledge-based Economy stems from the recognition that economic wealth rests upon knowledge and its useful application (Teece, 1981, 2007). Knowledge creation thus becomes the most relevant mean to generate profits and growth in living standards. Granstrand (1999) provided a broader interpretation of the concept, by introducing the term Intellectual Capitalism referring to “an economic system based on capitalist institutions in which productive assets and process as well as commercial transactions and products, are predominantly intellectual or immaterial rather than physical in nature.” (Granstrand, 1999:360) At the level of the single firm, this revolution implied a “new ecology of competition” (Rivette and Kline, 1999:4) in which the firm’s competitive advantage fundamentally rests on the deployment, development, combination and utilization of intangible assets, including knowledge and intellectual property rights. Accordingly, the most valued companies in the U.S. market, such as IBM, Lucent, Dow Chemical, General Electrics, Microsoft, Intel, Merck and Novartis have built over time their market position by leveraging, empowering and valuemaximizing their immaterial resources, including know-how, customer loyalty, relational capital, reputation, and their IPRs’ portfolios. As a consequence, they have displayed a growing gap between their market and book values suggesting the relevance of the stream of revenues associated with the exploitation of intangible assets within the overall strategy of the firm (Granstrand, 1999). Among other intellectual assets, patents have displayed the greatest potential as a source of revenues and as a mean to improve the competitive position of leader firms (Rivette
9
and Kline, 1999). They, in fact, represent a very powerful weapon upon which firms leverage in order to be on the forefront of market competition. This is true for two reasons. First, they are by definition the most tangible forms of intellectual property ensuring patentees a strong legal protection. Indeed, the legal monopoly granted by patents, while establishing a proprietary market advantage, allows firm to reward R&D efforts and to focus on complementary activities (e.g. branding) that are meant to sustain the products once they have been launched in the market. Second, patent datasets can be treated as a virtual “Alexandrian library of information” (Rivette and Kline, 1999:4) that can be used according to different purposes. Indeed, consistent with the patent literature 1 , patent statistics can be employed as a F
F
measure of differing technology indicator. Specifically, patents may be used either as a measure of technological change with relation to economic development (Schmookler, 1996), or as a measure of technological flow across countries (e.g. Jaffe and Henderson, 1993), or even as a measure of output of research activity (e.g. Griliches, 1984). In this last case, the focus of analysis is the innovation/patenting activity of the firm and the information contained in patents is conceived as a rich source of competition intelligence. Indeed, on the one hand, it may stir firms’ R&D by avoiding the risks to be sued in courts by competitors for infringements; on the other hand, it may help to anticipate technology shifts in the industry, pointing out promising trajectories of search and research to embark upon. Initially the knowledge-based economy literature argued it to be imperative for firms to fiercely protect and exclusively rely upon their own patents and through that establish a monopoly in the market. From the mid 1990s, the trend has changed in favour of a more open attitude towards the firm’s strategic management of its Intellectual Property Rights. Indeed, the recent diffusion of markets for technology - virtual spaces where innovations are exchanged in the form of intellectual property rights, products and services (Arora, Fosfuri and Gambardella, 2001) – has confirmed this trend. According to Arora et al. (2001:5) the existence of these virtual markets makes more likely the profitable exploitation of any intangible resource (e.g. technologies) that have been developed in-house. As such, they “can improve efficiency by reducing duplicative R&D and by matching technology producers and users” (Arora et 1
For an exaustive review see Basberg (1987). More recently, instead, works by (among others) Trajtenberg (1990), Albert, Avery, Narin and McAllister (1991), Lanjouw and Shankerman (2001), Hall, Jaffe, Trajtenberg, (2002), Harhoff and Reitzig, (2002) and Harhoff, Scherer and Vopel (2003) represent the new stream of industrial economics on technology and innovation research that has extensively employed patent data and patent counts.
10
al., 2001:9). In this scenario, patents provide an enhanced liquidity to knowledge that has not been previously experienced in the markets (The Economist, 2005). This is primarily due to the nature of patents that by definition make technology boundaries be legally well-defined and, as such, more easily and effectively transferred among organizations. Indeed, “in facilitating technology exchange, patents may be self-correcting: a stronger legal right to exclude others from using an invention generally provides a stronger economic incentive to include them through licensing” Gallini (2002: 141). As such, they represent one of the most relevant factors in triggering the further diffusion of markets for technology by easing their functioning with the reduction of transaction costs associated to the negotiation and enforcement of contractual agreements. Available information suggests that market for technologies are recently growing at increasing pace (Rivette and Kline, 1999; Gans and Stern, 2003; Kline, 2003; Mendi, 2007; Athreye and Cantwell, 2007). According to several authors (e.g. Rivette and Kline, 1999; Kim and Vonortas, 2003; Sheehan, Martinez and Guellec, 2004, Athreye and Cantwell, 2007) indeed licensing is the most visible vehicle for the transfer of technological knowledge among firms operating in high-tech industries, such as chemicals and pharmaceuticals 2 , electrical, F
F
software and ICT 3 , as shown, for instances, in Grindley and Teece (1997), Rivette and Kline F
F
(1999), Annand and Khanna (2000), Gu and Lev (2001), Arora et al. (2001a), Arora and Fosfuri (2003), Vonortas (2003) and Kim and Vonortas (2006a, 2006b). Specifically, patent licensing revenues have skyrocketed from only 15 billion of dollars annually some decades ago to more than 100 billion of dollars today (Kline, 2003; Litchtenthaler, 2007). Big players like IBM, Texas Instruments, DuPont and Merck have recorded hundred of billions in licensing revenues and stunning rate of profit growths thanks to their increasing licensing activity (Gu and Lev, 2004). Although it is widely recognized that a large portion of technological assets are unexploited, thus creating large potential for trade (Rivette and Kline, 1999; JIII, 2003, 2004;
In the chemical and pharmaceutical industries, preponderance of licensing (generally patent licensing) is related to the ability of the inventors in these industries to appropriate the benefits of patent and defend ownership against infringement (Gu & Lev, 2001). 3 In the electrical, software and ICT industries the ascendant of intellectual property rights, and thus of licensing as a mean to transfer them, is relatively recent. This is primary because of the slow diffusion of the DOJ/FTC 1995 Antitrust Guidelines for the Licensing of Intellectual Property (IP) that reflect more modern and updated thinking about IP, Among other things it includes recognition of and considerations on the potential efficiency benefits of licensing and crosslicensing. 2
11
Razgaitis, 2003, 2004; Kamiyama et al., 2006; Cockburn, 2007), the extent of this explosion in licensing activity, however, is remarkable. What is even more stunningly is the nature of this trend. Indeed, very recently firms have started to shift from a tactical to a strategic attitude towards licensing and this is becoming the usual practice in every industry (Litchtenthaler, 2007). There are indeed an increasing number of firms that perceive technology licensing as a fundamental part of firms’ business strategy. Licensing is not any longer perceived as a shortterm maneuver triggered by immediate need. It is becoming an integrated part of a longerterm plan of proactive management of intellectual property rights. It thus no longer the responsibility of the firm’s legal department rather it is managed as an independent business unit that should get involved early and continuously with other business units in technology commercialization decisions and in shaping and directing internal research and development (McCurdy and Phelps, 2002). As a consequence, until recently companies limited their licensing to technologies that were peripheral to their core business or licensed core technologies to firms that were not direct competitors. Now firms are putting efforts into extensive licensing programs of their “crown-jewel technologies” (Kline, 2003: 90). This trend spans both traditionally-known “protective” firms, like Procter & Gamble and “open-oriented” firms, like IBM that have recognized both the potential of technology licensing as a mean of maximizing the return on their R&D investments. The former one, for instance, known for being fiercely protective of its proprietary innovation, announced in 2000 a very extensive indiscriminate licensing strategy of their main patents and an unprecedented joint venture with one of its strongest competitors (Kline, 2003). The latter one, instead, by putting licensing at the very heart of its strategy for the diffusion of core products has earned millions of revenues from its licensed patents. The discussion above has indirectly emphasized the supply-side of markets for technologies. In fact, in general, there is a tendency among firms to pay more attention to selling their intellectual property rights to others (the financial benefits are quite evident) than buying intellectual property rights from outsiders (Chesbrough, 2003). However, it is not without relevance to also pay attention to the demand-side of markets for technologies. For each licensor I also observe a licensee that chose to in-license external technologies for reasons that could be financial (typically to reduce the R&D exposure) or strategic (basically to get access to technologies otherwise not available in-house). Indeed, understanding innovation and innovation activity as an open and distributed process provides momentum and relevance
12
to studying the perspective of the licensee. According to Chesbrough (2003:155) “the Open Innovation paradigm assumes that there is a bountiful supply of potentially useful ideas outside the firm and that the firm should be an active buyer and seller of intellectual property”. Specifically, the open attitude is deemed to challenge the so-called “not-invented-here syndrome” (Katz and Allen, 1982) that affects the more traditional firms and emphasizes the myth of internal research and development supremacy. This new model of innovation, instead, requires firms to leverage on external technologies to unlock the potential of firms’ internal innovative efforts. In other words, “instead of restricting the research function exclusively to inventing new knowledge, good research practice also includes accessing and integrating external knowledge” (Chesbrough, 2003:51). In this context, firms’ competitive advantage depends both on their ability to recognize available opportunities inside and outside their boundaries and on their readiness to exploit them in order to fuel their innovation process dynamically. According to Parr and Sullivan (1996:6) “[i]nterdependence is at the root of the paradigm shift taking place. Technology management in the future will centre on leveraging technology that a company owns to gain access to technology that it needs”. This new completely reversed logic underlying the innovation behaviour of the firm has been triggered more recently by the increasing dynamics and changing nature of the competitive environment and the increasing complexity of technology and the amount of resources required to come out with innovations. In this scenario, firms are required to “multiply the building blocks of innovation” (Rigby and Zook, 2002:82) and thus technology inlicensing stands as a way to pursue this strategy and improving the cost, quality and speed of innovation by importing new ideas from the outside. By in-licensing external technologies, the licensee takes advantage of the new division of innovative work (Arora et al., 2001) promoted by the diffusion of markets for technologies. This means that as long as firms can rely on external sources of knowledge to feed their innovative capacity, internal technological constraints become much less critical while the ability to exploit the increasing amount of external sources of knowledge is much more relevant for them.
Overall aim of the thesis
19B
The original idea of my thesis draws on the research background just disclosed. Specifically I am interested in investigating the licensee behavior as an integral part of the
13
overall technology strategy of the firm. I focus only on pure technology licensing agreements involving patents. 4 As already underlined, patents represent the best way to transfer and F
F
exploit technologies for two reasons. First, they ensure the highest level of enforceability of the granted right. Second, they allow the desired level of disclosure that permits follow-on innovations. However, while it is widely recognized that licensing may entail some strategic dimension from the licensor’s perspective, I argue that licensing-in also involves strategic considerations reflecting corporate objectives. From the licensee’s viewpoint, the licensing decision is concerned with the search, acquisition, integration, assimilation, exploitation of external technologies, including learning from them. As such, in my opinion, the main questions that the license firm is required to address can be summarized as follows:
Which technologies am I searching for?
Which technology outcome will I get from?
How much am I willing to pay?
The aim of my research project is to deal with these questions. Aiming at this, I investigate licensing-in in three different fashions: as a way to diversify firms’ technological portfolio by importing new technologies from outside while saving time and cost of development; as a new mean to learn from licensed technologies and thus to foster the innovation process; and as analogous of an option-to-exploit the licensed technologies in the future as soon as the associated market and technological potential becomes more secure and the exploitation of the patented technologies more profitable. In uncovering the potential of licensing-in, I am aware of the relevance of the contractual scheme of licensing (including, for instance, the exclusivity and the grant-back clauses as well as all the components of the remuneration structure) that (un)directly affects the exploitation of the licensed patents and that in turn is strongly related to the business model of the potential licensee. Consistently, the overall intention of my thesis is to provide a sharper picture of the demand-side of markets for technologies that can be framed within the new Open Innovation Paradigm. According to that, technology licensing together with other external technologies vehicles and coherently to internal research and development activity is a mean to design the innovation roadmap
4
We then exclude any other technology transaction that involves a different relationship between the licensor and the licensee besides the one-way transfer of technology. We refer to collaboration or settlement agreements, cross-licensing, technology purchases and plans of merger that for their specific features may display different patterns than pure licensing agreements.
14
(Chesbrough, 2003) of the firm with the final aim to filling gaps and overcoming blind spots as shown in the following figure.
Figure 1. Technology licensing within the firm open innovation strategies
Source: Chesbrough (2003:183)
With this purpose, my thesis contributes to the debate about the strategic relevance of licensing, by enriching our understanding of the licensee behavior that has been generally under-emphasized by the literature 5 and apparently by firms themselves that seem to be more F
F
concerned with licensing-out than licensing-in decisions, as underlined by Chesbrough (2003). For this reason, I believe that my research topic merits great relevance and interest for different audiences. It indeed may inform both strategic management scholars and technology management on the features of firms’ in-licensing decisions as well as on the new opportunities that patent licensing may offer to firms (e.g. technological learning), enabling them to achieve new forms of competitive advantage in the marketplace. But it can also provide insightful suggestion to policy makers on the drivers of the innovation strategies of firms and on new ways to collaborate strategically.
5
In the next chapter I will explain this point, based on an extend literature review.
15
Structure of the thesis
20B
The thesis is organized as a collection of three papers 6 . Each of them provides F
F
different insights of the same phenomenon - technology in-licensing - within different theoretical frameworks of reference. With the exception of the second paper – that only provides a conceptual model - they share the same methodology of analysis - regression analysis on secondary data - however they employed different econometric model specification according to their research questions. In the first work, I investigate the links between licensing and the patterns of firms’ technological search according to the framework of references of the Search and Absorptive Capacity literature. In the second paper - that continues where the first one left off – I analyze the new concept of learning-by-licensing, in terms of development of new knowledge inside the licensee firms (e.g. new patents) some years after the acquisition of the license, according to the Dynamic Capabilities and Learning perspective. Finally, in the third study, I deal with the determinants of the remuneration structure of patent licenses (form and amount), and in particular on the role of the upfront fee from the licensee’s perspective according to the predictions of the Real Options Theory. All together the three works aim at providing a refined picture of the licensee’s behavior by investigating very interesting and relevant angles through which technology inlicensing can be interpreted. As such, they each contribute substantially to attaining the overall aim of the thesis. In the following section the abstract of each paper will be presented. 1. Technological exploration through licensing: New insights from the licensee’s
point of view. The market for technology plays a crucial role in firms’ technology strategy as a way to undertake search in the available technological space. Drawing on innovation search theory and the literatures on licensing and absorptive capacity we address the issue of the factors that affect how technologically distant from the existing technological portfolio in-licensing firms are able to move when they in-license externally developed technologies. We posit that a long technological distance reflects the outcome of more 6
Plus one, entitled “Does Licensing Foster Rapid Invention?” which is inlcluded in the conclusive section, as upcoming reaserch paper.
16
exploratory search, while a short distance reflects the outcome of exploitative search. We conjecture two distinct dimensions of absorptive capacity in terms of the firms’ stock of knowledge (“assimilation capacity”) and the degree to which firms have searched broadly in the past (“monitoring ability”) to affect the distance of exploration from the existing technological portfolio. Furthermore, we compare firms that explore through licensing and firms which do not explore through licensing, but do so through search reflected in own patenting activities. We propose that the effects of assimilation capacity and monitoring ability should be more pronounced for licensees. Combining data on 176 license agreements and related patent information and while using a control sample of non-licensing firms we find—with exceptions—support for these ideas.
2. Fuel on the invention funnel: technology licensing-in, antecedents and invention performance. In this paper, we examine the impact of technology licensing-in on firm invention performance. Studying a sample of 266 licensees and matched non-licensees using a two-part model specification, we find that licensees are more likely to introduce inventions than their non-licensee counterparts. This holds both if we consider invention in general, and invention in the licensed technological class only. We also show that familiarity with the licensed technology and technological specialization drives licensees to pursue a narrow invention strategy primarily focusing on the technological class specified in the license agreement.
3. Uncertainty, flexibility and upfront fees in patent licensees. One of the main challenges of licensees and the licensors when attempting to agree on the contractual specifications of a technology license agreement is the payment structure. Utilizing a dataset covering 175 technology license agreements, we take a real option approach to investigate the role of contractual flexibility, market uncertainty and technical uncertainty in shaping the upfront fee. Controlling for the relative bargaining power of the involved parties and that some contracts don’t specify any upfront payments at all, our analysis indicates that contractual flexibility and market uncertainty indeed is associated with higher upfront fees in the contractual specification. Technical uncertainty does not have any direct effect. However, contractual flexibility seems to be valued higher under technical uncertainty. These results thereby insinuate that licensees not necessarily
17
consider uncertainty to be a negative attribute of a patent license but instead judge uncertainty to represent a high potential upside value and therefore willing to pay a higher upfront payment. Each paper stands as a single independent chapter of my thesis. However, in order to get introduced in their stream of though, two sections are provided before. The theoretical section comes first. A general overview of licensing literature is provided with the aim of identifying the three research questions and to highlight the relevance of using the selected theoretical framework to investigate the same phenomenon from different angles. Then the methodology section follows. It provides a very detailed description of the process of search, cleaning and integration of my dataset that I employed to run the econometric models of each of my research papers. The three research papers are then developed in the third, forth and fifth chapter. The last chapter includes the conclusion of my work and a snapshot of my future research in licensing issues, including one of my upcoming research papers – “Does Licensing Foster Rapid Invention?”. Appendix A, describing a real complex licensing agreement, and Appendix B, describing a real patent document (involved in the same licensing agreement) , are then attached as an integral part of the overall understanding of my research work.
References
21B
Anand, B.N. and T. Khanna (2000) “The Structure of Licensing Contracts”, The Journal of Industrial Economics, 48, 1, pp.103-135. Arora, A. and A. Fosfuri (2003) “Licensing the Market for Technology”, Journal of Economic Behaviour & Organization, 52, pp.277-295. Arora, A., A. Fosfuri and A. Gambardella (2001) Markets for Technology. Cambridge, MA: The MIT Press. Athreye, S. and J. Cantwell (2007) “Creating competition? Globalization and the emergence of new technology producers”, Research Policy, 36, pp. 209-226. Chesbrough, H. (2003) Open innovation. Cambridge, Massachusetts: Harvard University Press. Cockburn, I.M. (2007) “Is the Market for Technology Working? Obstacles to Licensing Inventions, and Ways to Licensing Inventions, and Ways to Reduce Them”, Paper prepared for the Conference on Economics of Technology Policy, Monte Verità, June
18
Gallini, N.T. (2002) “The Economics of Patents: Lessons from Recent U.S. Patent Reform”, Journal of Economic Perspective, 16, 2, pp.131-154. Gans J. S. and S. Stern (2003) “The Product Market and the Market for “Ideas”: Commercialization Strategies for Technology Entrepreneurs”, Research Policy, Vol. 32, Issue 2 , pp. 333-350. H
H
Granstrand, O. (1999) The Economics and Management of Intellectual Property. Cheltenham, UK: Edward Elgar Publishing. Griliches, Z. (ed.) (1984) R&D, Patents and Productivity. Chicago: University of Chicago Press. Grindley, P.C. and D.J. Teece (1997) “Managing Intellectual Capital: Licensing and CrossLicensing in Semiconductors and Electronics”, California Management Review, 39, 2, pp. 8-41. Gu, F. and B. Lev (2004) “The Information Content of Royalty Income”, Accounting Horizons, 18(1), pp. 1-12. Harris, R. (2001) “The knowledge-based economy: intellectual origins and new economic perspectives”, International Journal of Management Reviews, Vol. 3, Issue 1, pp. 21-40. Jaffe, A. and R. Henderson (1993) “Geographic Localization of Knowledge Spillovers as Evidenced by Patent Citations”, Quarterly Journal of Economics, Vol. 108, pp. 577-598. Japan Institute of Invention and Innovation (JIII) (2003) Survey on Patent Valuation System In Patent Licensing Market, Tokyo Japan Institute of Invention and Innovation (JIII) (2004) Survey on Verification and Evaluation of Patent Valuation System Based on Patent Licensing Contracts, Tokyo Kamiyama, S., J. Sheehan and C. Martinez (2006) “Valuation and Exploitation of Intellectual Property”, OECD STI Working Paper 2006/5, June. Katz, R. and T. Allen (1982) “Investigating the Not Invented Here (NIH) syndrome: A look at the performance, tenure, and communication patterns of 50 R&D projects”, R&D Management, 12(1), pp. 7-19. Kim, Y. J. and N. S. Vonortas (2006a) “Determinants of Inter-firm Technology Licensing: The Case of Licensors”, Managerial and Decision Economics, 27(4), pp. 235-249. Kim, Y. J., and N. S. Vonortas (2006b) “Technology Licensing Partners”, Journal of Economic and Business, 58, pp. 273-289. Kim, YJ. and N.S. Vonortas (2003) “Strategy and Cost in Technology Licensing”, Working paper, The George Washington University, August. Kline, D. (2003) “Sharing the Corporate Crown Jewels”, MIT Sloan management review , Vol. 44, Nº 3, pp. 89-93. H
Lichtenthaler, U. (2007) “The Drivers of Technology Comparison”, California Management Review, 49, pp. 67–89.
H
Licensing:
An
Industry
19
Mendi, P. (2007) “Trade in disembodied technology and total factor productivity in OECD countries”, Research Policy, Vol. 36, Issue 1 , pp. 121-133. H
H
Parr, R. L. and Sullivan, P. H. (1996) Technology Licensing: Corporate Strategies for Maximizing Value. New York: John Wiley & Sons. Razgaitis, R. (2003) Valuation and Pricing of Technology-based Intellectual Property. Hoboken, NJ: Wiley. Razgaitis, R. (2004) “U.S./Canadian Licensing in 2003: Survey Results”, Journal of the Licensing Executives Society, 39(4), pp. 139-151. Rivette, K. and D. Kline (1999) “Discovering New Value in Intellectual Property”, Harvard Business Review, pp. 55-66. Schmookler, J. (1996) Invention and Economic Growth. Boston, MA: Harvard University Press. Sheehan, J., C. Martinez and D. Guellec (2004) Understanding Business Patenting and Licensing: Results of a Survey, Patent Innovation and Economic Performance, 4, pp. 1-89. Teece, D. (1981) “The Market for Know-how and the Efficient International Transfer of Technology”, The Annals of the Academy of Political and Social Science, 458, pp. 81–96. Teece, D. J. (2007) “Explicating Dynamic Capabilities: the Nature and Microfoundations of (sustainable) Enterprise Performance”, Strategic Management Journal, 28, pp. 1319-1350. The Economist (2005) “A Market for Ideas”, (K. Cukier), 22th of October. Vonortas, N.S. (2003) Technology Licensing, Final Report, The George Washington University, October.
20
LITERATURE REVIEW AND RESEARCH QUESTIONS
9B
The rising interest in licensing within the strategic management research
22B
As widely described in the introduction, nowadays there is an increasing trend in licensing practices, suggesting of a more strategic attitude of firms in this area, especially in those industries characterized by high technology intensity. This recent explosion in inter-firm agreement practices, alongside the explosion in patenting, has revitalized past debates about the economic as well as strategic determinants underlying such trends. Accordingly, Gambardella (2008)7 , suggesting that theoretical inquiry is generally driven by empirics, F
F
confirmed that the stunning rate of growth of markets for technology and thus of licensing activities has stimulated the interest of scholars and boosted research production in this field over the last few decades. Licensing activity and practices have been the subject to theoretical investigation since the 1980s, when Caves Crookell and Killing (1983), in their widely-cited work titled “The Imperfect market for technology Licenses”, provided a relevant empirical evidence about licensingout and licensing-in strategies in the US, UK and Canada. As suggested by the title of their article, their main aim was to investigate the determinants of the imperfections of such dealings, according to the theoretical lenses provided by the contemporary literature of institutional economics – Transaction Cost Theory (Williamson, 1975). Their findings were however primary referred to international technology transfers. By then, indeed, licensing contracts were primary considered as one of the most relevant vehicles allowing for international expansion (Teece, 1981). Nevertheless, it is un doubtful that this paper, together with the work by Taylor and Silbertson (1973), by providing a very interesting and relevant evidence on the patterns of real licensing agreements, has motivated the theoretical literature that has been developed until the 1990s (e.g. Kamien and Tauman, 1984; Gallini, 1984, Katz and Shapiro, 1986). In the late 90s, Grindley and Teece (1997) and Rivette and Kline (1999) conducted clinical and historical studies of (cross) licensing practices, namely in the chemical industry, in
7
Keynote speaker at the DRUID-DIME Academy Winter 2008 PhD Conference, Aalborg, Denmark, January 17-19, 2008.
21
the electronics industry and in the computer industry. However, these works only provide small evidence of the phenomenon that was taking the lead in those years. In recognition of this lacuna with the turn of the century several different works have been developed with the intention of “amassing a detailed dataset on licensing contracts” (Annand and Khanna, 2000: 104). The Licensing Executives Society (LES) has been playing a relevant role in this respect. It has commissioned or stimulated several extensive surveys conducted by LES member all over the world, with the aim to collect data on this increasing phenomenon of interest. Among others 8 , F
F
the two big project entitled “An international survey on Technology Licensing Practice” and “The Diversity of Technology Licensing Agreements and their Clauses” coordinated by Brousseau at the University of Paris in 2002, supply a wide evidence not only of the extent of markets for licensing in Europe but mainly on the heterogeneity of these contractual agreements that in reality entail very complex governance structures. Different aim, instead, underlines the yearly survey-based studies performed by LES members that identify and explore difficulties encountered in markets for technology, as perceived by licensors and licensees (Razgaitis, 2004; Cockburn, 2007). Recently, the European Commission has fund the PatVal-EU research project 9 , involving many research fellows from different European universities with the intent F
F
of compiling a dataset on the features of patenting/licensing activities of European inventors. The project was accomplished in 2003. It provided new insights on the difference between the willingness to license European patents and whether a licensing agreement actually takes place or not. This empirical evidence has contributed to ignite the interest of many scholars and thus to directly or indirectly foster the increasing literature on licensing over the last decades (among others 10 , Arora Fosfuri and Gambardella, 2001; Arora and Ceccagnoli, 2004; Fosfuri, F
F
2006; Kim and Vonortas, 2006; Litchtenthaler and Ernst, 2007; Gambardella, Giuri and Luzzi, 2007).
See also, for instance, 2003 Compensation Survey Report, Licensing Executive Society (U.S. and Canada), 2003 9 This European project., founded by the DG Science & Technology of the European Commission, was conducted between 2001-2003 by 6 research units in 6 countries (France, Germany, Italy, Netherlands, Spain, UK) with the aim to create database from a survey of a sample of EPO patents with priority date 1993-1997. 10 The PatVal-EU dataset has been already employed in many academic works. See Research Policy (2007), Volume 36. 8
22
Extensive investigation of the Licensor’s Perspective
23B
As licensing issues have become centre-stage in firm’s strategic agenda, the identification and the analysis of strategic as well as economic rationales and determinants lying behind licensing strategy have gained momentum within the economics and strategic management literature. The starting point of this literature is that licensing entails a large set of incentives that make this strategic decision be profitable in comparison with the traditional decision to internalize the commercialization of in-house developed innovations. The economic (revenues-based) and strategic (competition-based) incentives come into being because of firms interplay in the markets for technology and for product.
57BU
The case of Monopolist technology-holder Until the 90s, the majority of work were developed by economists within stream of the
Industrial Organization literature and based on the game theory modelling. The focal and starting point of this literature is the recognition of the importance of the potential of an incumbent firm to retain its dominant position in a market that could be eroded by new entrants engaging in R&D investments for a new technology (Gallini, 1984). The main assumption is that the incumbent firm is the only technology-holder. In order to face and overcome the threat of new potential competitor, two are the alternative strategic choices available to the incumbent: to pre-empt potential entrants by inventing a new technology slightly earlier than would its rivals (Gallini, 1984; Kats and Shapiro, 1987); or to license its old technology to its rivals before they embark on new R&D trajectories leading to a new (possibly better) technology. According to Gallini (1984:931) “in contrast to previous model in which R&D activity deters entry into the product market, firms are encouraged into the product market – via licensing – as a way of deterring them from R&D activity”. In other words, other factors being equal, licensing-out is generally driven by the desire of the licensor to reduce/chose the competition in the markets for technology and the product market where the product technology is produced and sold. This represents an ex-ante or strategic incentive for licensing (Gallini and Winter, 1985) that reduces the would-be licensee’s incentives to do research and consequently prevents the possible erosion of the low-cost firm’s market position by its rival’s discovery of a superior technology. By contrast, an ex-post incentive for licensing reflects rents from the replacement of inefficient production. (Salant, 1984) This incentive is strong when firm’s cost
23
production under existing technologies are close, and very firm have interest in embark on additional R&D investment in order to achieve the lowest cost-position in the market and to gain additional rents in the form of royalties from licensing. Thus, while strategic incentive could significantly discourage R&D investments, by hindering the technological progress in the long-run, ex-post incentives, instead, could enhance it, by empowering the innovation race among competitors. However, in both cases, licensing strategy may help to increase industry profits, by reducing industry costs, since the higher-cost technologies are crowd out by the lower-cost technology (Shepard, 1987). This is consistent with the emphasis on process innovation that is characteristic of this literature. Further, Eswaran (1994) investigate the incentive an incumbent might have to license its technology not to its rivals but to outsiders with no technology in their own. In this case, incumbent firm not only choose the competition, as it do if it licenses a weak rival to crowd the market and deter entry by a stronger competitor (Rockett, 1990), but also it can elude opportunistic behaviour by licensee that have no technological capability to invent around the incumbent’s technology and that is willing to pay royalties in order to exploit licensed technology. In some other cases, incumbent firm may face capacity constraints that don’t allow for the widest diffusion of the new technology. Thus, one of the available decisions at hand would be to license his innovation to rivals in order to expand the scale of use of the new technology (Kats and Shapiro, 1985), primary in those markets where he has limited expertise that would rather be necessary. In other words, licensing could be considered as a means to allow for licensor’s market differentiation and licensor’s technology demand enhancement. According to Shepard (1987:360), “it is common practice in [the semiconductor] industry for an innovating firm – a firm that has developed and produces a new, proprietary product – to license one ore more competing firms, thereby creating multiple sources of supply. […] [This is due] to the innovating firm’s desire to expand product demand.” Consistent with the second sourcing, creating competition allows seller to commit to a level of product quality higher and to a level of product price lower than it is consistent with a monopolistic position hold without licensing (Shepard, 1987; Farrell and Gallini, 1988). Using the more recently established concepts within the recent strategic literature, this means that licensing practice may allow the incumbent to build reputation and stimulate customer’s commitment in the long-run. According to Arora et al. (2001: 175) “this is particularly important when buyers must make a substantial specific investment to use the new technology”. Moreover, by licensing their rivals, incumbent firm not only achieves the so-called Perfect Commitment with customers, but it also could achieve another favorable outcome, that is the establishment of its
24
technology as a de-facto standard (Arora et al., 2001).
58BU
Markets for Technology Over the last years, there has been a revitalized interest in licensing strategy that has
59B
stimulated the recent focus of strategic literature on the analysis of the so-called markets for technology (Arora, 1995, 1996, 1997; Arora and Fosfuri, 2003; Arora et al., 2001; Fosfuri, 2006; Arora and Ceccagnoli, 2006). Markets for technology are virtual spaces in which technology, in the forms of intellectual property rights (patents), intangibles (software, knowhow), product or services are traded. Overall they refer to the “transactions for the use, diffusion and creation of technology” Arora, Fosfuri and Gambardella (2000: 5).With their increasing diffusion, implying the presence of several firms owning substitutable technologies (technology holders), strategic scholars have reject (as it might be sometimes implausible) the strong assumption made in the economic literature of a monopolist technology holder (Fosfuri, 2006). Starting out from Arora et al. (2001), literature on licensing and markets for
60B
technology (MfT) has been focused on the so-called licensing dilemma (Fosfuri, 2006) of the licensor – whether to license out technologies or commercialize them in-house. According to Gambardella and Giarratana (2007:4) “[t]he debate in the MfT literature [has centred] on the role of transaction costs and IPRs”. Accordingly, most of works have dealt with the analysis of the effect of strong Intellectual Property Regime on the diffusion on markets for technology, relying on the theoretical reasoning provided by the Transaction Cost Theory (Williamson, 1975). The fundamental argumentation that has been tested is that the strength of patent system may be pivotal to an innovator’s decision to license-out new technologies rather than to use them exclusively (Gallini, 2002). The reason for this is twofold. First, because strong patents make patented technologies boundaries be well-defined and thus their transfers be feasible and effective (by avoiding the undesired imitation of licensed technology from the licensee); second, because strong patents enhance the inventor’s bargaining power that could make licensing agreements be faster concluded. Gans, Hsu and Stern (2007) found, for instance, that the licensing of a patent occurs primary within a small time period around the date of its grant, since the legal uncertainty of the right fades away and the transaction costs associated to the license are in part resolved. Thus, by reducing transaction costs of negotiating contractual
25
agreements, they encourage the diffusion of technological knowledge and consequently the development of markets for technology. This is consistent with the findings of Fosfuri, Giarratana and Luzzi (2008) suggesting that patents open up the market for trade in technological information. Indeed, by enjoying the desired protection provided by software patents, firms are more likely to exchange information in open source environment. In other words, “in facilitating technology exchange, patents may be self-correcting: a stronger legal right to exclude others from using an invention generally provides a stronger economic incentive to include them through licensing” (Gallini, 2002: 141). Although patents may solve part of the transactions costs involved in the exchange of technology, however, they cannot be the only response unless they are coupled with other assets or conditions easing the transfer or motivating the exchange of technologies. First of all, patents only refer to that part of knowledge that is articulable and then codified. However, there is a large amount of knowledge that is not codified either because it is costly to do, or because it is not strategically convenient to do. In these cases, bargaining firms could face the so-called double-side moral hazard problem (Arora, 1996; Arora et al., 2001). It consists on the fact the both licensor and licensee - double-side - have some incentive to behave opportunistically in agreeing on of know-how clauses. On the one hand, licensor may license their technology without providing the required know-how to exploit it, on the other hand, licensee, given the possibility of moral hazard on the part of the licensor, would like to make the bulk of the payments only after being satisfied that the full technology, including the tacit part, has been transferred. In addition, as highlight by Arora et al. (2001: 114), “as the licensee has learned the know-how, she cannot be forced to unlearn it. Hence, a licensee may refuse to pay the agreed upon amount in full after the know-how is transferred”. Thus, in order to solve this kind of problems, players can trust in their reciprocal reputations (e.g. in gentlemen agreements), or they can use any other technology input that is transferred alongside know-how (e.g. patents), as hostage. Accordingly, Arora (1996) found that there would be a two-way relationship between knowhow and patents and their effect on licensing effectiveness. That is, the more is the know-how transferred with patented technologies the higher is the probability that licensee would understand and integrate that technology and that licensing would lead to successful outcomes. By contrasts, the higher is the part of technology protected by patents, the higher is the probability that licenses involving also technological know-how may be successfully completed. Thus, the complementarity between know-how and patents, increase the value of
26
each only in conjunction with the other. Thus, the patent/know-how works as hostage that can be withdrawn by the licensor, every time the licensee behaves opportunistically. According to Arora and Ceccagnoli (2006) the propensity of licensing-out technologies depends on the interplay between the appropriability regime and the presence of complementary assets. “When patent protection is weak, an innovator may forgo patenting and rely upon its complementary assets. On the other hand, patenting does not preclude the use of complementary assets: patents can be used to exclude competitors and prevent imitation, allowing the innovator to leverage its complementary assets” (Arora and Ceccagnoli, 2006: 4). This line of reasoning is deeply rooted in Teece’s seminal work (1986) in which it is argued that licensing propensity increases if the innovator, lacking complementary assets - marketing and manufacturing capabilities - enjoys strong patent protection. According to Arora and Fosfuri (2003) and Fosfuri (2006) an important but yet little understand determinant of licensing is competition in the market for technology that comes up as a direct consequence of relaxing of the ‘unique’ technology holder assumption. The author suggests that licensing decision results from the interplay of two effects: the revenue effect that is equal to the present value of the flows of rents in the form of licensing payments (royalties, fees and lump-sum); and the dissipation effect that is given by the erosion of profits due to another firm competing in the downstream market. The licensing-out decision is based on the attempt to strive the right balance between the two effects. The underlying assumption is that the licensor does not compete with his licensees in the downstream markets but only with other technology-holders in the markets for technology and the friction-less nature of markets for technology (Cesaroni, 2006). The latter one implies that there are no limitations in technology transfer (sell and exchange) among firms. Licensor’s market share and the product differentiation in the downstream markets negatively affect the propensity to license-out technologies respectively. Recently several authors have provided renew insights on the reasons underlying the licensing-out decisions of markets for technologies. Several authors have in fact identified and analyzed a certain number of determinants of the licensor decision related to the licensor’s reputation of being a valuable technology provider (Litchtenthaler and Ernst, 2007), the features of the licensor-licensee relationship (Kim and Vonortas, 2006), and the generality nature of technology and thus the fragmentation of downstream markets (Gambardella and Giarratana 2007). The last two works offer an attempt to include the other side of the licensing agreement – the licensee - in their argumentation. Nevertheless the emphasis is still
27
directed towards the understanding the factors affecting the likelihood of licensing-out technologies. Starting from the same model developed by Fosfuri (2006), Kim and Vonortas (2006) suggested that the technological proximity and the experience in partnering between licensee and licensor raise the likelihood of concluding a contract among this precise pair of firms. Gambardella and Giarratana (2007), instead, aimed at assessing the overall probability of licensing when product market is fragmented and if the licensed technology is general purpose, accounting for the probability that a potential licensee demands the technology.
The Under-investigation of the Licensee’s perspective
24B
The precedent review of the relevant works on licensing and MfT has revealed the bias affecting this literature. While the licensing dilemma of the licensor (Fosfuri, 2006) has been and it is still extensively investigated, the licensing decision of the licensee has been generally under-investigated so far. In other words, licensor’s active role has been generally overemphasized in comparison with the alleged licensee’s passive role (Atuahene-Gima and Patterson, 1993). The omitted assumption underlying these works is the technology-push nature of the innovation process. Put differently, they overall assume the frictionless nature of markets for technologies (Cesaroni, 2005). Nevertheless, some authors (e.g. Arora et al., 2001) have already pointed out that behind the non-perfectly functioning of markets for technology there are besides the so-called transactional constraints due to the nature of technology - that has been centre-stage in the interest of economic and strategic scholar so far (Gambardella and Giarratana, 2007) – there are also the so-called cognitive constraints that are related to the nature of the process of technological innovation. They refer to the difficulty to assimilate the technology from the recipient firm (the licensee) depending on the firm-specific nature of the innovation process. Being aware of the existence of this kind of limitations, it implies a renew way to investigate the same problem of missing markets for technologies from the recipient firm’s viewpoint (Arora et al., 2001): the Not-Invented-Here syndrome (Katz and Allen, 1982). It also calls for a better understanding of the demand-side of markets for technologies. Indeed, for each licensor we also observe a licensee that chose to in-license external technologies for reasons that could be financial (typically to reduce the R&D exposure) or strategic (basically to get access to technologies otherwise not available in-house).
28
In recognition of this scant attention paid to the licensee’s perspective, some academics (e.g. Caves et al., 1983; Atuahene-Gima, 1992, 1993; Atuahene-Gima and Patterson, 1993; Lowe and Taylor, 1998; Arora et al., 2001; Cesaroni, 2004) began to highlight the necessity to take into account also the demand side of technology transfers. However, with the only exception of Caves et al. (1983) and Cesaroni (2004), these attempts have been not framed within the MfT literature. Mostly, they refer to the stream of research focused on the new product development process (e.g., Atuahene-Gima, 1992, 1993; Atuahene-Gima and Patterson, 1993) or the international technology transfer from developed countries to developing countries (e.g. Kim, 1999). The first empirical – albeit only descriptive – investigation of licensing practices accounting for the licensee’s perspective is provided by Caves et al. (1983) that showed an extensive evidence about licensing-out and licensing-in strategies in the US, UK and Canada. In particular, the survey of licensee covered 21 companies operating in Canada and 13 in the UK. Through the theoretical lenses provided by the Transaction Cost approach, the main aim of this paper was to investigate the determinants of the imperfections of such dealings. According to the authors, by analyzing these imperfections, it is possible to derive different classes of behavioural predictions (Caves et al., 1983) about the circumstances in which technology-holders of technology will license-out, the circumstances in which potential licensees will enter such contracts and the configuration of contractual terms these contract will display as result of the negotiation between parties. They also stated that their empirical evidence was collected in order to address the latter two classes since “[s]ubstantial research has already addressed the first class, while much less it is known about the second and the third” (Caves et al., 1983: 252). Indeed, this work provided seminal insights about the behaviour of licensee firms with respect to their decision to license current technologies - for which some degree of inhouse technological capabilities is required since the licensee is not allowed to access future improvements made by licensor- in order to pursue a closely related diversification or even only to strengthen their current activities 11 . They also shaded light on the effect of different F
F
legal clauses, such as exclusivity, technology flow-back, included in licensing agreements that are settled in order to prevent any opportunistic behaviour by the licensee to the detriment of 11
Based on the same empirical evidence, Killing (1978) provided further insights on the link between licensing and diversification, in the attempt to investigate the appropriateness of licensing-in to carry out a program of product diversification. He found that they found that the majority of licenses were concluded to acquire technology related to products the companies were already producing.
29
rent extraction in the licensor’s behalf. In few words, their findings suggested that the obvious advantage that licensee has to enter licensing agreement is to “secure technology at a cost lower than by developing it afresh” (Caves et al, 1983; 265). However they also found that licensing-in may be conceived as a means to get access to technologies according to firms’ diversification strategy in which they utilize other assets that are already available to them. The most visible lack of this paper depends on its descriptive nature that by definition does not include any empirical analysis. Virtually agreeing with the three authors, after almost twenty years, Cesaroni (2004) assessed that “[since] recent studies [focused on the influences of markets for technology on firm’s corporate strategy 12 ] have mostly explored the supply side of markets for technology [.] [i]n this F
F
study we take the opposite perspective” (Ibidem, 2004: 1547). He analyzed the impact of the diffusion of such markets on firms’ technology outsourcing decisions and their propensity to enlarge their product portfolio. According to a transaction-costs argumentation, he suggested that “in the presence of markets for technology, transaction costs associated to technology trade become less severe, and firms have more incentives to substitute internal technology development with external technology acquisition” (make-or-buy decision) (Cesaroni, 2004:1550). In order to test this hypothesis, he defined the source of technology (internal or external) as dependent variable that he regressed against a measure of markets for technology (number and size of available licensors), a measure of technological capabilities available to and innovative effort provided by the firms at the time of license, the size of the licensee and several technological sector dummies. Consistently he also advanced a second hypothesis that link the efficiency of markets for technology to the product diversification strategy of the firm. His argumentation can be stated as follow. The more easily technologies can be traded on the markets, the higher the probability that internal technological constraints that prevents firm from entering into new product markets can be overcome. In this respect the decision to diversify will be driven by the availability of other non-technological assets that can be shared between current and new product markets (Prahalad and Hamel, 1990; Teece, Pisano and Shuen, 1994). In order to test this hypothesis, he chose as dependent variable the novelty of firms’ production (new chemical compound) and employed the same independent variables of the first model specification. In both regression analyses he found confirmation of his hypotheses. However, although insightful,
12
He referred, for instance, to the works by Arora, Fosfuri and Gambardella (2001) and by Gans and Stern (2003).
30
this paper leaves some open issues that in my opinion are worth investigating in more details. First of all, in terms of work setting, the focus on the chemical industry that displays very industry-specific factors that may in fact hinder the generalization of the results; second, in terms of theoretical background, the author relies on the Transaction Cost Theory that in fact may not account for several determinants affecting firms’ decision besides the widely recognized economic rationales; third, and consequently, it leaves undisclosed some issues related to the relationship between the licensor and the licensee and the nature of technology that may affect the decision to in-license technologies according to a diversification strategy. Some other works have provided useful insights on the licensing decision of the licensee firm. They span different features associated to the licensee’s behavior, however, without positioning their hypotheses within a specific theoretical framework of references. The work by Atuehene-Gima (1992, 1993), for instance, aimed at investigating some factors affecting the firm’s intention or propensity to adopt technological inward licensing (ITL). He classified these are classified into four different groups, namely firm’s characteristics, management characteristics, benefits and costs of ITL and external factors. He tested his hypotheses by multiple regression analysis based on data drawn by a survey directed to CEOs or manager directors of Australian engineering firms. Among other, the most interesting result he found was about the effect of the lack of “internal new product development capabilities”, the “satisfactory ITL experience” and “the perceived benefits of the ITL” on firms’ propensity to adopt ITL 13 . Specifically, licensee firms were found to be more prone to adopt ITL the lowest the F
F
level of their NPD capabilities, the higher the level of perceived benefits from the license and the higher the satisfaction tied with the previous ITL experience. Following the same reasoning and based on the same survey-based evidence, Athuene-Gima and Patterson (1993) focused their attention on the examination of the perceptions of managers in their decisions to license technology from unaffiliated organizations (defined as organizations independent from the licensee firm) according to their new product development strategy. Starting from the review of previous work (e.g. Killing, 1978) that simply provided a list of the benefits that accrue to the firm from licensing, such as speed of market entry, access to licensor support to reduce technology development risks, diversification or ready adoption of an industry standard 13
According to the author, the propensity to adopt ITL referred to the firm´s attitudinal orientation towards ITL. This construct was measured by foru items: the strength of the need to license-in; the likelihood the firm will in-license in the next two years; the likelihood the licensee will enter new product markets through licensing; the likelihood that licensee will expand their current product markets by licensing rather than by internal development (Atuehene.Gima, 1993: 233).
31
(Atuahene-Gima and Patterson, 1993: 328), the main objective of their work was to understand which benefits, costs and risks are involved and which ones have the greatest impact on the firm’s decision to license. The study revealed that firms use licensing to acquire new products mainly to meet the more immediate need to gain competitive advantage in the short run rather than having access to future technology. This result is consistent with the traditional explanation of technology licensing that is to get rapid access to proven/mature technology by reducing their financial exposure (Roberts and Barry, 1985; Chatterji, 1996). This rationale reflects the typical short-run objective driving the prospective licensee’s decision. However, as underlined by Lowe and Taylor (1998: 265) “[b]uying-in technologies may deplete long term capability if it reduces internal research skills and knowledge but it can also help develop capability by speeding the process of knowledge acquisition and building complementary assets”.
The
traditional puzzle about the complementary or substitute relation between R&D and technology acquisition strategy was indeed the focus of their research. Based on a sample of 128 UK licensors and licensees, they defined eleven propositions about some key determinants of technological acquisition through licensing that refer to transfer costs and the pace of technical change, technology gaps, internal and external orientation of the firm and complementary assets. Their findings allowed them to argue that the two strategies, composing the dilemma of make or buy decision, are in fact complementary rather than substitute and that the use of licensing require substantial complementary assets to be in place. This result led to two relevant considerations. First of all, licensing may be conceived as a strategy to unlock the potential of internal capabilities in the long-run as long as licensee firms learn instead of being passive in the relationship. Second, the alleged complementarity-relation implies that licensing may not represent a significant diversification strategy if firms deviate from internal existing core capabilities (Lowe and Taylor, 1998: 275). To sum up the three reviewed works share indeed the same objective (analysis of the determinants of technologylicensing), the same type of data-collection procedure (survey-based) and the same type of sample composition (two different groups of firms operating in different industries with and without licensing agreements). The main contribution of the paper by Lowe and Taylor (1998) was to highlight the desirable active role of the licensee against the formerly established passive role that had been generally attributed to them (Ibidem: 265) and to point out further research on the learning organization concept and its relationship to extensive sources of knowledge (Ibidem: 275).
32
Finally, another way to look at the licensee’s perspective has been proposed by Ziedonis (2007). In his work he analyzed the use of option contracts and the adoption of real option reasoning by those firm acquiring rights to commercialize university technologies. Consistent with the basic premise of Real-Option theory, he tested the probability that a potential licensee purchases option contracts, instead of engaging in licensing agreements in the first place, depending on the extent of technological uncertainty related to the commercialization of the licensed technologies. Although his first objective, as reflected in the title of his article, “Real options in technology licensing”, was to test the suitability of a real-option logic for analyzing licensing-in decision in a particular institutional setting (that per se involves option agreements), he provided interesting insights on the determinants of the licensee’s choice conceived as a fundamental of her technology strategy as a whole. Indeed, after successfully testing the first hypothesis, arguing a positive relationship between the level of technological uncertainty and the likelihood of the firm to purchase an option agreement before licensing, he shifted the attention of his argumentation to the analysis of the effect of firm characteristics on the level of technological uncertainty which she makes her decision upon according to the real-option logic. In this respect, recalling the (relative) absorptive capacity argument (Cohen and Levinthal, 1990; Lane and Lubatkin, 1998; Zahra and George, 2002), he suggested that both firm’s ability to evaluate external technologies, based on its extant technological knowledge relevance, and the relatedness between firm’s technological knowledge base and licensed technologies, decrease the likelihood to purchase an option before licensing because the level of uncertainty characterizing that license from the buyer’s point of view is lower. This also implies other more subtle reasons for exercising option in technology licensing, as highlight by the authors: “on the one hand, firms appear to benefit from their ability to learn about technologies during the option period [,] [o]n the other hand firms that are more able to absorb the technology during the contract period may have reduced incentives to subsequently license the invention” (Ziedonis, 2007:1) The provided insights are very relevant and, in my opinion, useful to partially fill the gaps I found in the work by Cesaroni (2004), however there are still some other questions put aside that rather display a great potential for future analysis. First, the author focuses on the licensing decision process while the real option framework deploys its greatest potential in the valuation of innovation activities and outcomes (Pennings and Lint, 1997; Perlitz, Peske and Shrank, 1999; Schwartz and Moon, 2000). Second, he only considers technological uncertainty, while previous literature suggests that both technological and market uncertainty can have distinct effect that should be taken into account in the analysis (McGrath and Macmillan, 2002; Oriani and
33
Sobrero, 2007; Anand, Oriani and Vassolo, 2007). Third, he analyzes only academic licensing agreements that by definition show specific characteristics (Jensen and Thursby, 2001) that may not allow for generalization of the results.
Research questions
25B
As shown in the previous section, literature about licensing also accounting for the licensee’s perspective is very heterogeneous and currently poorly developed. The aim of this research work is to contribute to the understanding of the conditions and determinants affecting the licensing behaviour of the licensee firm. The background perspective informing my analysis is the so-called Open Innovation Paradigm that has been developed by Chesbrough (2003, 2006). This paradigm suggests a completely reversed logic underlying the innovative behaviour of the firm that challenges the so-called “not-invented-here syndrome” (Katz and Allen, 1982) affecting the more traditional firms emphasizing the myth of internal research and development supremacy. This new model of innovation requires firms to leverage external technologies to unlock the potential of firms’ internal innovative efforts. Indeed, “instead of restricting the research function exclusively to inventing new knowledge, good research practice also includes accessing and integrating external knowledge” (Chesbrough, 2003:51). As a consequence, technology licensing together with other external technologies vehicles and coherently to internal research and development activity is a mean to design the innovation roadmap (Chesbrough, 2003) of the firm with the final aim to filling gaps and overcoming blind spots. Building on this frame of references, the licensing-in decision is concerned with the search, acquisition, integration, assimilation, exploitation of external technologies, including learning from them. As such, in my opinion, the main questions that the license firm is required to address can be summarized as follows:
Which technologies am I searching for?
Which technology outcome will I get from?
How much am I willing to pay?
34
The aim of my research project is to deal with these questions. Aiming at this, I investigate licensing-in in three different fashions: as a way to diversify firms’ technological portfolio by importing new technologies from outside while saving time and cost of development; as a new mean to learn from licensed technologies and thus to foster the innovation process; and as analogous of an option-to-exploit the licensed technologies in the future as soon as the associated market and technological potential becomes more secure and the exploitation of the patented technologies more profitable. For each of them, I decided to rely upon differing theoretical perspective allowing for dealing with these issues from different, complementary angles of analysis.
References
26B
Anand, B.N. and T. Khanna (2000) “The Structure of Licensing Contracts”, The Journal of Industrial Economics, 48(1), pp. 103-135. Anand, J., R. Oriani and R. Vassolo (2007) Managing a Portfolio of Real Options, Advances in Strategic Management, 24. Arora, A. (1995) “Licensing Tacit Knowledge: Intellectual Property Rights and the Market for Know-How”, Economics of Innovation and New Technology, 4, pp.41-59. Arora, A. (1996) “Contracting for Tacit Knowledge: the Provision of Technical Services in Technology Licensing Contracts”, Journal of Development Economics, 50, pp.233-256. Arora, A. (1997) “Patents, Licensing, and Market Structure in the Chemical Industry”, Research Policy, 26, pp.391-403. Arora, A. and A. Fosfuri (2003) “Licensing the Market for Technology”, Journal of Economic Behaviour & Organization, 52, pp.277-295. Arora, A., A. Fosfuri and A. Gambardella (2001) Markets for Technology. Cambridge, MA: The MIT Press. Arora, A., and M. Ceccagnoli (2006) “Profiting from licensing: The role of patent protection and commercialization capabilities”, Management Science, 52(2), pp. 293-308. Atuahene-Gima, K. (1992) “Inward Technology Licensing as an Alternative to Internal R&D in New Product Development: A Conceptual Framework”, Journal of Product Innovation Management, 10, pp. 156-167. Atuahene-Gima, K. (1993) “Determinants of Inward Technology Licensing Intentions: An Empirical Analysis of Australian Engineering Firms”, Journal of Product Innovation Management, 10, pp. 230-240.
35
Atuahene-Gima, K. and P. Patterson (1993) “Managerial Perceptions of Technology Licensing as an Alternative to Internal R&D in New Product Development: an empirical investigation”, R&D Management, 23 (4), pp. 327-336. Caves, R. E., H. Crookell and J. P. Killing (1983) “The Imperfect Market for Technology Licenses”, Oxford Bulletin of Economics and Statistics, 45, pp. 249-267. Cesaroni, F. (2004) “Technological outsourcing and product diversification: do markets for technology affect firms' strategies?”, Research Policy, 33(10), pp. 1547-1564. Cesaroni, F. (2006) Strategie Tecnologiche e Competitività dell’Impresa. Milano: Franco Angeli Editore. Chatterji, D. (1996) “Accessing External Sources of Technology”, Research Technology Management, Vol. 39, No. 2, pp 48-56. Chesbrough, H. (2003) Open innovation. Cambridge, Massachusetts: Harvard University Press. Chesbrough, H. (2006) Open Business Models: How to Thrive in the New Innovation Landscape. Boston: Harvard Business School Press. Cockburn, I.M. (2007) “Is the Market for Technology Working? Obstacles to Licensing Inventions, and Ways to Licensing Inventions, and Ways to Reduce Them”, Paper prepared for the Conference on Economics of Technology Policy, Monte Verità, June Cohen, W.M. and D.A. Levinthal (1990) “Absorptive Capacity: A New Perspective on Learning and Innovation”, Administrative Science Quarterly, Vol. 35, pp 128-152. Eswaran, M. (1994) “Licensees as Entry Barriers”, The Canadian Journal of Economics, 27, 3, pp. 673-688. Farrell, J. and N. T. Gallini (1988) “Second-Sourcing as a Commitment: Monopoly Incentives to Attract Competition”, The Quarterly Journal of Economics, Vol. 103, No. 4, pp. 673-694. Fosfuri A. (2006) “The Licensing Dilemma: Understanding the Determinants and the Rate of Technology Licensing”, Strategic Management Journal, 27, pp. 1141-1158. Fosfuri, A., M. S. Giarratana and A. Luzzi (2008) “The Penguin has entered the building. The Commercialization of Open Source Software Products”, Organization Science. Gallini, N.T. (1984) Deterrence by Market Sharing: A Strategic Incentive for Licensing. The American Economic Review, 74, 5, pp. 931-941. Gallini, N.T. and R.A. Winter (1985) “Licensing in the Theory of Innovation”, Rand Journal of Economics, 16, 2, pp. 237-252. Gambardella, A. and M. Giarratana (2007) “ General Technologies, Product-Market Fragmentation, and the Market for Technology: Evidence from the Software Security Industry ”, Working Paper, November. H
H
36
Gambardella, A. and M. Giarratana (2007) “ General Technologies, Product-Market Fragmentation, and the Market for Technology: Evidence from the Software Security Industry ”, Working Paper, November. H
H
Gambardella, A. P. Giuri and A. Luzzi (2007) “The market for patents in Europe”, Research Policy, Vol. 36, Issue 8 , pp. 1163-1183. H
H
Gans JS, D.H. Hsu, S. Stern (2007) “The impact of uncertain intellectual property rights on the market for ideas: evidence from patent grant delays”, Management Science, forthcoming. Grindley, P.C. and D.J. Teece (1997) “Managing Intellectual Capital: Licensing and CrossLicensing in Semiconductors and Electronics”, California Management Review, 39, 2, pp. 8-41. Jensen, R. and M. Thursby (2001) Proofs and Prototype for Sale: The Tale of University Licensing, American Economic Review, 91, pp. 240-259. Kamien, M. I. and Y. Tauman (1984) “The Private Value of a Patent: A Game Theoretic Analysis”, Journal of Economics. Katz, M.L. and C. Shapiro (1986) “How to license intangible Property”, The Quarterly Journal of Economics, pp. 567-589. Katz, M.L. and C. Shapiro (1987) “R&D Rivalry with Licensing or Imitation” The American Review, 77, 3, pp. 402-420. Katz, R., and T. Allen (1982) “Investigating the Not Invented Here (NIH) syndrome: A look at the performance, tenure, and communication patterns of 50 R&D projects”, R&D Management, 12(1), pp. 7-19. Killing, J. P. (1978) “Diversification through licensing”, R&D Management, 8(3), pp. 159-163. Kim, YJ. and N.S. Vonortas (2006) “Determinants of Inter-firm Technology Licensing: The Case of Licensors”, Managerial and Decision Economics, 27, pp. 235-249. Kim, L. (1999) “Building technological capability for industrialization: analytical frameworks and Korea's experience”, Industrial and Corporate Change, Vol. 8, N. 1, pp. 111-136. Lane, P.J. and M. Lubatkin (1998) “Relative Absorptive Capacity and Interorganizational Learning”, Strategic Management Journal, Vol. 19 No. 8, pp. 461-477. Lichtenthaler, U. and H. Ernst (2007) “Developing reputation to overcome the imperfections in the markets for knowledge”, Research Policy, 36, pp. 37-55. Lowe, J. and P. Taylor (1998) “R&D and Technology Purchase through Licence Agreements: Complementary Strategies and Complementary Assets”, R&D Management, Vol. 28 No. 4, pp 263-278. MacMillan, I. C. and R. G. McGrath (2002) “Crafting R&D project portfolios”, Research Technology Management, 45, pp. 48-60.
37
Mowery, D.C., J.E. Oxley and B.S. Silverman (1996) “Strategic Alliances and Interfirm Knowledge Transfer”, Strategic Management Journal, Vol. 17, Special Issue, pp 77-91. Mowery, D.C., J.E. Oxley and B.S. Silverman (1998) “Technological Overlap and Interfirm Cooperation: Implications for the Resource-based View of the Firm”, Research Policy, Vol. 27, pp 507-523. Oriani, R. and M. Sobrero (2008) Uncertainty and the Market Valuation of R&D within a Real Options Logic, Strategic Management Journal. Pennings, E. and O. Lint (1997) “The option value of advanced R & D”, European Journal of Operational Research, 103(1), pp. 83-94(12). Perlitz, M., T. Peske and R. Schrank (1999) “Real Options Valuations: the New Frontier in R&D Project Evaluation?”, R&D Management, 29 (3), pp. 255-269. Pitkethly, R. (2001) “Intellectual Property Strategy in Japanese and UK companies: Patent licensing decisions and learning opportunities”, Research Policy, Vol. 30, pp 425-442. Prahalad, C.K. and G. Hamel (1990) “The Core Competence of the Corporation”, Harvard Business Review, May-June. Razgaitis, R. (2004) “U.S./Canadian Licensing in 2003: Survey Results”, Journal of the Licensing Executives Society, 39(4), pp. 139-151. Rivette, K. and D. Kline (1999) “Discovering New Value in Intellectual Property”, Harvard Business Review, pp. 55-66. Roberts, E.B. and C.A. Berry (1985) “Entering New Businesses: Selecting Strategies for Success”, MIT Sloan Management Review, Vol. 26 No. 3, pp 9-34. Rockett, K.E. (1990) “The quality of licensed technology”, International Journal of Industrial Organization, 8, pp. 559-574. Salant, S.W. (1984) “Pre-emptive Patenting and the Persistence of Monopoly: Comment”, American Economic Review, 74, pp. 247-250. Schwartz, E. and M. Moon (2000) “Evaluating Research and Development Investments”. In M. Brennan & L. Trigeorgis, (Eds.), Project Flexibility, Agency, and Product Market Competition: New Developments in the Theory and Application of Real Options. New York, NY: Oxford University Press. Shepard, A. (1987) “Licensing to Enhance Demand for New Technologies”, RAND Journal of Economics, 18, 3, pp. 360-368. Teece, D. (1981) “The Market for Know-how and the Efficient International Transfer of Technology”, The Annals of the Academy of Political and Social Science, 458, pp. 81–96. Teece, D. J. (2007) “Explicating Dynamic Capabilities: the Nature and Microfoundations of (sustainable) Enterprise Performance”, Strategic Management Journal, 28, pp. 1319-1350.
38
Teece, D., G. Pisano and A. Shuen (1997) “Dynamic Capabilities and Strategic Management”, Strategic Management Journal, Vol. 18 No. 7, pp. 509-533. Williamson, O.E. (1975) Markets and Hierarchies: Analysis and Antitrust Consideration. New York: The Free Press. Zahra, S.A. and G. George (2002) “Absorptive Capacity: a Review, Re-conceptualization, and Extension”, Academy of Management Review, Vol. 27, pp. 185-203. Ziedonis, A. A. (2007) “Real Options in Technology Licensing”, Management Science, 53, pp. 1618-1633.
39
METHODOLOGY
10B
Secondary analysis
27B
In order to address the three research questions I rely on a secondary analysis of patent and license data. Secondary Analysis is a quantitative research method that consists of analyzing existing data. This type of analysis is increasingly being employed in social research for two reasons. First, because of the continuous theoretical developments. Second, due to an increasing reliability of statistical elaboration tools allowing researchers to employ the same data to answer new questions or to answer old questions with new tools of analysis. For this same reason, there is also a rising number of public agencies 14 and private companies 15 whose F
F
F
F
business consists in collecting data and then provide them in exchange of a monetary value (Corbetta, 1999). Availability of licensing datasets remains very limited and the few that do exist are poorly developed. The reason for this is widely recognized and it depends largely on firmpolicy. Firms in fact have been always considering information about their licensing activity as very competition-sensitive and thus they have been reluctant to disclose it (Cockburn, 2007). Nevertheless, thanks to the increasing amount of transaction concluded every day in the virtual markets for technology and thus thanks to the rising interest of academic scholars in licensing and licensing-related topics, as shown over the last decades, datasets on licensing or on technology transactions in general is blooming. However, the information available in these datasets is still limited and often not suitable for carrying out extensive analysis across industries and years. Among other pieces of information generally undisclosed, exchanged patents and the features of the remuneration scheme are the most relevant. Consistently, the provider 16 of one of the dataset I have checked up for my research wrote to me the following F
F
considerations: “We do not proactively search for patent numbers of products to be licensed by companies unless they are explicitly mentioned in publicly-disclosed news articles. Most of the time, companies participating in an alliance merely informs the public
For instance, the Economic & Social Research Council Data Archiv (University of Essex) in United Kingdom (HUhttp://www.esrc.ac.uk/ESRCInfoCentre/Support/access/UH) 15 Among others, Thomson Financial Group (HUhttp://www.thomson.comUH) 16 Corresponding distributor of the TFSD Joint Ventures and Alliances database (2006) 14
40
about the kind of products they would license without giving out the patent numbers or even financial details of the agreement” (2006)
In addition, the level of details achieved by existing datasets is very low compared to the stunning amount of information possible to obtain from the text document of any single licensing agreement. We refer to the legal clauses and provisions that, in most cases, are only known to people that write these contracts (e.g. licensing officers or Intellectual Property lawyer). Indeed, besides the information about the licensee and the licensor names and business, the statement of their willingness to be committed in the agreement, the duration of the agreement and the clause of exclusivity, license entail a series of interesting and relevant pieces of neglected information that represent great potential for analysis. They include the field of use of the license, the possibility to sublicense, the territory in which the licensed product can be sold, the confidentiality regime underpinning the agreement, the duties of the parties in terms of technology/know-how furnishing by licensors or due payments by the licensees – on the bases of a very complex remuneration scheme, the rights on future improvements of the licensed technology (grant-back clause), the ownership of the inventions and future patent applications - with the identification of the part that is responsible of bearing the cost of filing, prosecuting and maintenance – just to mention a few. Indeed, only very recently, some scholars (e.g. Anand and Khanna, 2000; Bessy, Brousseau and Saussier, 2002) have realized how much variety and diversity characterize empirically observed license agreements. In fact, they are more complex and sophisticated than the simple market-based exchange depicted by the theoretical literature. In Appendix A you find a very insightful example of a complex licensing agreement settled by IBM Corporation and Intuitive Surgical Inc. in 1997 17 . F
F
Based on these considerations and since I decided to use patent-based measures 18 for F
F
the construction of the most relevant variables of our analysis, it was clear that our secondary analysis was far more challenging than expected and in fact required us to commit a lot of resources to data integration and cleaning. We had to rely on multiple sources of information 19 F
F
As available at HUhttp://www.secinfo.com/dr6nd.51J7.2.htmUH, retrieved in Febraury 2008. We follow the insights provided by the literature in industrial economics on technology and innovation that has extensively employed patent data and patent counts (e.g. Scherer, 1965; Comanor and Scherer, 1969; Basberg, 1982, 1987; Trajtenberg, 1990; Albert, Avery, Narin and McAllister, 1991; Henderson and Cockburn, 1994, 1996; Lanjouw & Shankerman, 2001; Hall, Jaffe, Trajtemberg, 2002; Harhoff and Reitzig, 2002; Harhoff, Scherer and Vopel, 2003; Reitzig, 2004; Ziedonis; 2007) 19 The most important ones being the NBER dataset (Hall, Jaffe and Trajtemberg, 2002), the OCED STAN dataset that are both available online respectively at HUhttp://www.nber.org/patents/UH 17 18
41
that allowed us to build up a very complex dataset including records at the level of any single licensee, licensor, licensed patent, license contract (e.g. clauses) and the industry involved. The aim of the next paragraphs is to describe stepwise the procedure that I followed leading to a final dataset of 227 license agreements. This dataset represents the common ground for each of the three research papers. However, the number of observations finally selected varies between each of the research papers depending on research question and model specification – that defines the variables, the econometric approach and tools employed.
Dataset search
28B
Initially I searched for available dataset that provide info about licensing agreements. Among others, I basically found four alternatives that I considered worth being checked up. They are listed as follows: 1. the SDC (Given Security Company) database of the Thomson Financial Group; 2. the TFSD (Thomson Financial Given Securities) Joint Ventures and Alliances database 20 of the Thomson Financial Group; F
F
3. the RoyaltySource® database 21 Agreements of the Licensing Economics F
F
Review; 4. The FVGIP (Financial Valuation Group Intellectual Property) Dataset 22 of F
F
the Financial Valuation Group.
The first database has been extensively employed in the literature on licensing until now (e.g. Arora, Fosfuri and Gambardella, 2001; Kim and Vonortas, 2003; Fosfuri, 2006; Kim and Vonortas, 2006a, 2006b). It records all publicly announced alliance deals worldwide coming from the Security Exchange Commission filings, press releases, trade magazines, & HUhttp://elsa.berkeley.edu/~bhhall/bhdata.htmlUH; HUhttp://www.oecd.org/document/15/0,2340,en_2649_201185_1895503_1_1_1_1,00.htmlUH. 20 HUhttp://library.dialog.com/bluesheets/html/bl0554.htmlU 21 HUhttp://www.royaltysource.com/royaltyrates.htmlU 22 The Financial Valuation Group (FVG) is one of the leading business valuation consulting and litigation service firms in North America. (HUhttp://www.fvginternational.com/index.htmlUH, accessed June 2007).
42
professional journal and the like. However, this dataset only provide data that dates back to 1998. Also, it has recently been substituted with the Thomson One Banker® dataset that collects only corporate financial information. The TFSD (Thomson Financial Given Securities) Joint Ventures and Alliances database, instead, has the main advantage to record data from the 1990 till today. However, after using the trial version of the dataset, as provided by the European distributor of the Thomson Financial Group, I realized that the majority of the recorded transactions refer to alliances agreements instead of pure licensing agreement. The third dataset compiles only data about license contracts. The main feature that makes this dataset appreciable is that it reports detailed information on the remuneration structure of the contract (specifically the royalty-rate) and that it displays this information in a very userfriendly format. However, it only provides a brief summary of other information concerning the contract. Also, in the negotiation phase I was only allowed to use a very limited initial sample (only 2 hundred transactions) for our analysis, without having checked them before. On top of these reasons, all three of the above sources scarcely reported information on patents exchanged through licensing (if any). Finally, these three data sources represented financial constraints as none of them are freely available. After discarding the previous alternatives, in one of our last attempts of search I found the Intellectual Property database that has been developed by the Financial Valuation Group with the aim to conduct empirical research on intellectual property. After evaluating its characteristics, I decide to choose this database as the starting point for our analysis. The reasons for this choice are as follows. This database is a compilation of intellectual property transactions gleaned from publicly available documents. The selection of transactions is based on three criteria: 1) each license involves the exchange of an intellectual property; 2) the transaction took place; 3) a certain payment structure is agreed upon by the parties and it was (will be) paid, even if that amount is not disclosed (Financial Valuation Group, 2007). All three criteria match the requirements of my investigation including the three research questions and consistent with the econometric model specification I decided to rely on. Finally, I was able to get this dataset for free on the bases of a well-defined confidentiality agreement and thanks to the open attitude of the Financial Valuation Group that as “a firm encourages empirical research” (Mike Mard 23 , 2006) in the field of licensing and intellectual property valuation. F
23
F
Managing Director of the Financial Valuation Group, Tampa Office (Florida)
43
The Financial Valuation Group Intellectual Property Dataset
2
The FVGIP database records data on approximately 3,000 transactions. According to the editor of the dataset, while analyzing the data during the data collection procedure, they discovered that it would require at least 40 individual data fields to capture the basic information required to economically interpret the various transactions including 24 : F
F
1. Licensor Name and Demographics 2. Licensee Name and Demographics 3. Type of Agreement (i.e., Patent, Trademark, Copyright, etc.) 4. Geographic Region of License 5. Detailed Description of Licensed Intellectual Property 6. Royalty Payment Structure 7. Percentage Royalty Amount (Fixed, Low, High) 8. Dollar Royalty Amount (Fixed, Low, High) 9. Basis for Payment of Royalty (i.e., Net Sales, Annual Fee, Per Unit, etc.) 10. Guaranteed Annual Royalties 11. Maximum Lifetime Royalty 12. Term of License Agreement 13. Original or Amended 14. Exclusivity 15. Other Considerations of the Agreement The list above is suggestive of the level of details contained in this dataset. It in fact summarizes the most important characteristics of license agreements with a particular focus on the components – e.g. milestones, minimum royalties, royalty rate – of their complex remuneration structure. The following figures and tables contain summary statistics revealing the nature of the dataset. Figure 2. Distribution of transactions by industries
24
This information is drawn by the FVG internal report.
44
Source: FVGIP (2007)
Figure 2 provide evidence that the dataset is very consistent with the empirics of previous work on licensing agreements suggesting that the majority of such contracts are concluded in high-tech industries, such as Chemicals and Pharmaceuticals, Electronics, Software and ICT (e.g., Anand and Khanna, 2000; Arora and Fosfuri, 2003; Arora, et al., 2001; Fosfuri, 2006; Grindley and Teece, 1997; Gu and Lev, 2004; Kim and Vonortas, 2006a, 2006b; Rivette and Kline, 2000; Vonortas, 2003).
Figure 3. Type of IP agreements
Source: FVGIP (2007)
45
According to Figure 3 among several distinctly different intangible assets that can be transferred among firms, patents tend to be the most frequently exchanged intellectual property, with trademarks, technology and products following respectively.
Data cleaning and integration
30B
For the purpose of our analysis – I am interested in pure licensing agreements that involve the exchange of patents - I initially selected 1052 technology agreements out of almost 3000 as recorded in the original dataset, classified as “patent”
25 F
F
and “technology”
26 F
F
transactions. As shown in the previous chart they represent the majority of Intellectual Property Transactions recorded in the original dataset. Among these I was able to find the original document (License Agreement) or some references in other filing data (e.g. S1, 8K, 10K) from the Security Exchanged Commission (SEC) website of about 600 licenses ranging from 1970 to 2001. The reason is twofold. First of all, not all the available documents are online anymore. Second, with respect to licensing agreements per se, the undesirable drop is due to the strategic relevance and sensitive nature of information included in these contracts which parties are very reluctant to disclose (Cockburn, 2007). In this case, the information collected by the FVG is the only one available at the time of the license. Among these in only 101 cases the FVGIP database records the USPTO identification number of the patents involved (patno) 27 . Among those, 94 were patent transactions while 7 were classified as F
F
technology transactions. Consistent with our interest in using patent-based measures, I tried to maximize the number of transactions reporting the patno of patents. To achieve this I decided to read through all the documents available of the (almost) 500 licenses left. This activity allowed us to verify whether a real exchange of patents had been executed in the remaining licenses and to exactly identify them through their patno. We indeed was able to directly find We consider patent licenses those whose primary type of agreement is classified as “patent” by FVG. We consider patent licenses those whose primary type of agreement is classified as “technology” by FVG. 27 Very few licenses reported only the identification number of non-US patents. For these patents we always took the corresponding USPTO patents making the analysis omogeneous in this respect. Also, in cases where the license involved the exchange of USPTO patents and its corresponding foreign patent applications or issued patents we did not track their identification number. However, we accounted for them by including the variable familysize (=number of jurisdctions in which the patent is in force) in our final dataset. 25 26
46
the desired information or, whenever impossible, to get that by browsing the USPTO dataset according to the information available in the text of the contracts (e.g., the application number or the title of the issued patents included in the description of the transaction) or in specific cases, by searching for the name of the assignee in the same database in the focal year together with the key-word provided in the description of the licensed technology. In addition, by reading each contract I was
able to integrate the information provided in the original
database or at least to make it sure that all kinds of possible mistakes (e.g. misspellings, missing part and so on) would be minimized. After this long process, I kept only 301 patent agreements. Finally, given the specific purpose of our analysis — according to which licensing is indeed considered as a mechanism to access external technology — I only included those transactions that have been filed as (pure) licensing or assignment agreements in our final sample. We then exclude all other transactions that refer to collaboration or settlement agreements, cross-licensing, technology purchases and plans of merger that for their specific features may display different patterns. At the end of this cleaning activity I came up with a final sample of 227 patent licenses involving almost 900 USPTO patents exchanged among licensor and licensee firms. Apart from using the data already available in the original dataset, I integrated it with information from multiple other data sources. We then developed a very exhaustive relational database including records at the level of any single license contract (e.g. clauses), licensee, licensor, licensed patent(s), and the industry(ies) involved. The common key-field for any type of stored information is the ID-number of each license as reported in the original FVG IP database. Figure 2 provides a simplified representation of the final dataset.
47
Figure 4. Relational Patent-License Database
LICENSORS PATENTS LICENSORS ID #
LICENSED PATENTS
LICENSES
LICENSEES
INDUSTRIES
61BU
LICENSEES PATENTS
License level From the original available documents I extrapolated the following information for
each license: 1.
Year of the license
2.
Scope (=number of patents involved in the license)
3.
Country (=name of the territories in which the licensee is allowed to exploit
the licensed technology) 4.
Term (=number of years in which the licensee is allowed to exploit the
licensed technology) 5.
Exclusivity of the license
6.
Sublicense of the licensed rights
7.
Techfurnishing (if the licensor is required to furnish technology/know-how to
the licensee besides the licensed technology 8.
Grantback (if the licensee is required to grant back to the licensor the
improvements of the licensed technology that she will be developing on her own) 9.
Not-to-compete (if the licensor is required not to compete with the licensee in
the exploitation of the licensed technology) 10.
Trademark (if the patent license involve the exchange of the trademark
associated to the licensed product)
48
11.
Remuneration structure
By reading the original documents, I was able to make a distinction between licenses that did not involve upfront payments and those that included this form of payment in their remuneration structures. Among them, I also identified those that omitted this value for reasons of confidentiality 28 . We found the same for royalty-rate, milestones payments, F
F
minimum royalty and per-unit royalty. We also made the distinction between upfront payments based either on the exchange of a certain amount of money or shares.
62BU
Licensed patents level In order to collect all the relevant statistics on each licensed patent I matched our
database with the National Bureau of Economic Research (NBER) dataset 29 (Hall, Jaffe, and F
Tratjenberg, 2001) and its 2002 update.
F
Among others, I could rely on the following
information:
28
29
1.
Patent Number
2.
Grant Year
3.
Grant Date
4.
Application Year
5.
Claims (numbers)
6.
Main Patent Class (3 digit)
7.
Technological Category
8.
Technological Sub-Category
9.
Number of Citations Made
10.
Number of Citations Received
11.
Measure of Generality
12.
Measure of Originality
These licenses report the following standardized statement “Confidential Information Omitted and Filed Separately with the Security Exchange Commission. Asterisks denote such Omissions” These data comprise detail information on almost 3 million U.S. patents granted between January 1963 and December 1999, all citations made to these patents between 1975 and 1999 (over 16 million), and a reasonably broad match of patents to Compustat (the data set of all firms traded in the U.S. stock market). For further information see HUhttp://www.nber.org/patents/UH & HUhttp://elsa.berkeley.edu/~bhhall/bhdata.htmlUH.
49
We then aggregated these measures at the level of the license by computing various moments of these data (e.g. averages, maxima, minima, standard deviations) depending on the nature of the data and to the meaning that these indices may need to express at the aggregated level.
63BU
Licensee and licensor level As already advanced, the FVGIP dataset reports licensors’ and licensees’ name and
demographic, respectively. Specifically, the following information is included in the original dataset: 1.
Licensor Name
2.
Licensor Country
3.
Licensor Industry (brief description)
4.
Licensor SIC code of the industry
5.
Licensor NAICS code of the industry
6.
Licensee Name
7.
Licensee Country
8.
Licensee Industry (brief description)
9.
Licensee SIC code of the industry
10.
Licensee NAICS code of the industry
With respect to licensors, after having read the original document of the license, I was able to distinguish individual, non-profit or for-profit (company) licensors. Among these, we also controlled whether they were the licensee’ s parent company or not. With respect of licensees that represent our main focus of analysis, we dedicated a large amount of time searching for information about their research activity, their size and stage of development at the time of the license. Accordingly, we gathered data on their R&D expenditures, net sales, number of employees and year of foundation from proprietary or publicly-available data source, namely Thomson Research, Osiris, Comp tech, Google Finance and, whenever necessary, company websites or other online available data sources. This
50
activity was not without difficulties. First of all, the considerations about the disclosure of licensing information, including exchanged patents, also apply for firms’ R&D expenditures, since the decision whether or not to disclose this information rests upon the discretion of the firm (Hall and Oriani, 2006; Hall, Thoma and Torrisi, 2007). Second, it was difficult to find data at the precise time of the license, since some of these licenses date back to the early 90s. Third, for some firms it was impossible to find any information. This was suggestive not only that they are non-public firms but also that they are very small or currently inactive firms. In order to get information on both the licensors and the licensees patenting activity we relied on the NBER dataset. Specifically, before the match we made sure to find all the possible corresponding names of our firms that may identify them in the NBER dataset as assignees of patents. Our concerns were based on the following consideration. Due to the increasing frequency of Merger and Acquisition operations, it’s very likely that firms change their name over time. Patents are accordingly preferably assigned to the parents company according to the policy of the group which the licensor/licensee belongs to. For this reason, by searching through the same data-sources from which we drew information on firmscharacteristics, we found the following corresponding names of our 227 firms: 1.
licensor’s name
2.
licensor’s former name
3.
licensor’s current name
4.
licensor’s parent company name
5.
licensed patent assignee name (if different from the licensor name)
6.
licensee’s name at the time of license
7.
licensee’s former name
8.
licensee’s current name
9.
licensee’s parent company name
On this basis, we matched the name of our firms with the assignee names recorded in the NBER dataset and, through the corresponding assignee code obtained, we made the link with our dataset and the file recording all patent characteristics. By doing so, we found 182 licensors-assignees that overall have been granted about 188,000 USPTO patents and 173 licensees-assignees with almost 108,000 USPTO patents. The same match also allowed for getting all the relevant statistics for their patent. However, as we did for the licensed patents,
51
we aggregated the single patent-based statistics at the level of the firms by computing various statistical moments reflecting the composition of their patent portfolios.
64BU
Industries The FVGIP database provides the description of the industry which the licensor and
the licensee belong to with the identification of both the SIC and the NAIC code, but also the description of the industry in which the license agreement can be classified. However, in this last case, the original dataset doesn’t report the corresponding SIC or NAIC. We believed that it would have been useful to get this information in order to allow further analysis and enable possible match based on these codes instead of a qualitative description of the industry. For this reason, by finding the match of this description among those available for the licensees and licensors we identified the corresponding SIC for each license. In order to collect data at the industry level, we refer to the OECD STAN database 30 . F
F
For the purpose of our analysis 31 , we focused on measures of output associated to the F
F
industries involved in any license agreements. Specifically we selected the “value-added” variable as a measure of economic activity for all the OECD countries plus Japan and Korea. Since we observed 15 different combinations of geographical areas (countries/continents) involved in our licenses, we decided to calculate for each combination the average and volatility of the growth rate within three and five years from the license at the industry level. However, since our sample is based on SIC codes while the OECD STAN dataset relies on ISIC codes, we made the match by the means of the ISIC-SIC correspondence table as available online 32 F
F
References Anand, B.N. and T. Khanna (2000) “The Structure of Licensing Contracts”, The Journal of Industrial Economics, 48(1), pp. 103-135.
30The
OCED STAN database provides analysts and researchers with a comprehensive tool for analyzing industrial performance at a relatively detailed level of activity. It is accessible at: HUhttp://www.oecd.org/document/15/0,2340,en_2649_201185_1895503_1_1_1_1,00.htmlUH. 31 We specifically need this information for the third paper, coauthored with Raffaele Oriani, in which we investigate the effect of market’s uncertainty on the up-front value paid by the licensee. 32 For instance, see HUhttp://www.macalester.edu/research/economics/PAGE/HAVEMAN/Trade.Resources/Trad eConcordances.html#FromusSICUH.
52
Arora, A., A. Fosfuri and A. Gambardella (2001) Markets for Technology. Cambridge, MA: The MIT Press. Bessy, C., E. Brousseau and S. Saussier (2002) “The Diversity of Technology Licensing Agreements”, Working Paper, Université de Paris 1, June. Cockburn, I.M. (2007) “Is the Market for Technology Working? Obstacles to Licensing Inventions, and Ways to Licensing Inventions, and Ways to Reduce Them”, Paper prepared for the Conference on Economics of Technology Policy, Monte Verità, June. Corbetta, P. (1999) Metodologia e Tecniche della Ricerca Sociale, Il Mulino: Bologna. Fosfuri A. (2006) “The Licensing Dilemma: Understanding the Determinants and the Rate of Technology Licensing”, Strategic Management Journal, 27, pp. 1141-1158. Grindley, P.C. and D.J. Teece (1997) “Managing Intellectual Capital: Licensing and CrossLicensing in Semiconductors and Electronics”, California Management Review, 39, 2, pp. 8-41. Gu, F. and B. Lev (2004) “The Information Content of Royalty Income”, Accounting Horizons, 18(1), pp. 1-12. Hall, B. H., G. Thoma and S. Torrisi (2007) “The market value of patents and R&D: Evidence from European firms”, NBER Working Paper Series. Hall, B.H. and R. Oriani (2004) “Does the market value R&D investment by European Firms? Evidence from a panel of manufacturing firms in France, Germany and Italy, International Journal Of Industrial Organization, 24, pp. 971-993. Hall, B.H., A. B. Jaffe and M. Trajtenberg (2002) “The NBER Patent Citations Data File: Lessons, Insights and Methodological Tools”, NBER Working Paper Series. Kim, Y. J. and N. S. Vonortas (2006a) “Determinants of Inter-firm Technology Licensing: The Case of Licensors”, Managerial and Decision Economics, 27(4), pp. 235-249. Kim, Y. J., and N. S. Vonortas (2006b) “Technology Licensing Partners”, Journal of Economic and Business, 58, pp. 273-289. Kim, YJ. and N.S. Vonortas (2003) “Strategy and Cost in Technology Licensing”, Working paper, The George Washington University, August. Rivette, K. and D. Kline (2000) “Discovering New Value in Intellectual Property”, Harvard Business Review, pp. 55-66. Vonortas, N.S. (2003) Technology Licensing, Final Report, The George Washington University, October.
53
3
RESEARCH PAPERS33
33
This thesis encapsulates papers that are still progressing with the aim of submitting to - and thus publishing in - prominent journals in the next months. Therefore, fine-tuning of analysis and rewording may have been done since the day of the doctoral thesis defence, 4th of June 2008.
54
TECHNOLOGICAL EXPLORATION THROUGH LICENSING: NEW INSIGHTS FROM THE LICENSEE’S POINT OF VIEW 34 F
Forthcoming in Industrial and Corporate Change (ICC), Special Issue on “Markets for Technology and Industry Evolution” 35 F
Presented at:
CBS Conference on Organizing for Internal and External Knowledge Creation and Innovation: Looking within or Searching Beyond?, Copenhagen, Denmark, 30-31 October, 2008
Markets for Technology: Strategy and Industry Evolution Conference, Madrid, Spain, September 19-20, 2008
R&D Management Advanced Workshop, Linköping, Sweden, September 15-16, 2008
2008 Annual Meeting of the Academy of Management, Anaheim, California, August 8-13, 2008
The 25th DRUID Celebration Conference, Copenhagen, Danmark, June 17-20 2008
CCC Doctoral Consortium , Pittsburgh, Pennsylvania, April 11-13, 2008
DRUID-DIME Academy Winter 2008 PhD Conference, Aalborg, Denmark, January 17-19, 2008
Abstract. The market for technology plays a crucial role in firms’ technology strategy as a way to undertake search in the available technological space. Drawing on innovation search theory and the literatures on licensing and absorptive capacity we address the issue of the factors that affect how technologically distant from the existing technological portfolio in-licensing firms are able to move when they in-license externally developed technologies. We posit that a long technological distance reflects the outcome of more exploratory search, while a short distance reflects the outcome of exploitative search. We conjecture two distinct dimensions of absorptive capacity in terms of the firms’ stock of knowledge (“assimilation capacity”) and the degree to which firms have searched broadly in the past (“monitoring ability”) to affect the distance of exploration from the existing technological portfolio. Furthermore, we compare firms that explore through licensing and firms which do not explore through licensing, but do so through search reflected in own patenting activities. We propose that the effects of assimilation capacity and monitoring ability should be more pronounced for licensees. Combining data on 176 license agreements and related patent information and while using a control sample of non-licensing firms we find—with exceptions—support for these ideas. Keywords: Licensing, Market for technology, Search Activity, Technological Exploration
The paper in its current version is coauthored with Professor Keld Laursen - Department of Innovation and Organizational Economics, Copenhagen Business School, Copenhagen,
[email protected] – and Professor Salvatore Torrisi - Department of Management, University of Bologna, Bologna,
[email protected]. 35 This is an earlier version of that one accepted for pubblication in Industrial and Corporate Change. 34
55
Introduction
32B
The increasing importance of markets for technology is challenging the traditional model of organizing innovation, where R&D and the complementary assets required for innovation are largely integrated within the firm (Teece, 1988; Arora et al., 2001a). The emergence of these markets has offered a new window of opportunity to firms that are more open to the outside and that are engaged in permanent search activity (Arora et al., 2001b; Laursen and Salter, 2006). Indeed, given the fact that firms can rely on external sources of knowledge to feed their innovative capacity, the ability to explore the increasing amount of external sources of knowledge becomes more and more relevant for them. Recent empirical studies have found that increasing efficiency of markets for technology, and the associated declining transaction costs, make technology outsourcing an important alternative to in-house R&D in various industries (e.g., Silverman, 1999; Arora et al., 2001b; Fosfuri, 2006; Lichtenthaler and Ernst, 2007). The literature on licensing behavior has, however, mostly focused on the supply side of the technology market. Although this literature has greatly enhanced our understanding of the licensing phenomenon, the demand side of the market has generally been overlooked, with licensees assumed to play a passive role. As pointed out by Henry Chesbrough: “[b]oth the buying and the selling perspectives are necessary to improve the management of IP.” (Chesbrough, 2003: 158). A small number of previous studies have examined the licensee perspective (e.g., Atuahene-Gima, 1993; Atuahene-Gima and Patterson, 1993; Lowe and Taylor, 1998)
and only few of them have shed light on technology in-licensing as a
diversification option (Killing, 1978; Caves et al., 1983). These studies on technology inlicensing have found that the acquired licenses were most often closely related to the focal firm’s technological competencies. However, from these contributions it is not clear to what extent markets for technology provide innovating organizations with greater strategic flexibility and a larger number of feasible options as compared to in-house search. We attempt to remedy this research gap by comparing the behavior of in-licensing firms to the behavior of comparable non-in-licensing firms. Our analysis draws on the idea that firms can undertake two types of technological search and diversification: local search (or exploitation) and distant search (or exploration) and that firms’ managers need to balance these two types of search (prominent examples include, March, 1991; Levinthal, 1997; Rosenkopf and Nerkar, 2001; Katila and Ahuja, 2002; Benner and Tushman, 2003; Gupta et al., 2006). In this paper, we assume that exploitative and
56
exploratory search represent a continuum with “exploitation” and “explorations” as two extremes. From this perspective, we posit that a long technological distance reflects the outcome of more exploratory search, while a short distance reflects the outcome of more exploitative search and we identify characteristics of in-licensing firms that may lead them to explore or exploit through in-licensing. 36 We show that two specific dimensions of absorptive F
F
capacity (AC) affect firms’ distance from the existing technological portfolio to a given technology acquisition through in-licensing. Cohen & Levinthal (1990) have defined absorptive capacity as the “ability to recognize the value of new information, assimilate it, and apply it to commercial ends” (p. 128). We focus on two dimensions of AC, the monitoring/valuation/identification ability and the assimilation capacity. These two dimensions would both be included in what Zahra and Georg (2002) later dubbed “potential AC” (defined as the ability to acquire and assimilate externally generated knowledge)—as opposed to “realized AC” that has to do with how firms transform and exploit externally acquired knowledge to commercial ends. Several later studies have refined and extended the AC construct (see, Jansen et al., 2005; Lane et al., 2006). To our knowledge, however, no previous work has tried to distinguish and operationalize these different dimensions of the AC construct in the context of licensing strategy. Our empirical analysis draws on a sample of 176 firms with license agreements over the period 1974-2001. We combine licensing information with related patent information and a number of other types of data. By using a control sample of non-licensing firms, we conduct an analysis with difference-in-difference characteristics that allows to assess the existence of significant differences in firms’ ability to undertake technological exploration of varying degrees of distance from the firms’ existing technological portfolio—through licensing-in or own patenting activity—across the two samples of licensing and non-licensing firms.
Theory and Hypotheses
33B
The organizational theory of innovation and the theory of search have found that exploration and exploitation are pursued in different organizational settings, ranging from inhouse search activity to alliances, acquisitions and licensing-in. Here we draw on Koza and 36
Note that in-licensing is a way to acquire already existing external technology in the market. In that sense, we examine firms’ ability to explore technologies that are new to the firm, not new to the world.
57
Lewin (1998: 260) who argue that: “In licensing and franchising (from the point of view of the licensee), absorptive capacity determines the rate and effectiveness through which technology, brands, and the like may be internalized.” We propose that pure internal search imposes stronger constraint on distant exploration as compared to “external market mechanisms” like inter-firm alliances and licensing and that strong AC is particularly important in allowing for distant exploration using these external market mechanisms. We center on in-licensing because it represents the form of contractual arrangement which is closest to pure transactions in the market for technology. Many studies have explored the antecedents and consequences of AC. However, few scholars have tried to understand the implications of different dimensions of the AC construct for exploratory and exploitative search. For instance, in the context of strategic alliances, Mowery et al (1996) measured AC in terms of the pre-alliance level of technological overlap with partner firms and found the extent of a firm’s absorption of technological capabilities from its alliance partners to be positively related to its pre-alliance level of technological overlap with partner firms. Following Zahra and George (2002), subsequent research has tested the construct empirically by digging deeper into the organizational mechanisms, such as cross-functional interfaces, participation in decision-making and job rotation that affect potential and realized absorptive capacity (Jansen et al., 2005). This view of AC then emphasizes the firm’s ability to exploit external knowledge (Lane et al., 2006: 855). Later attempts at refining and extending the concept of AC have distinguished between explorative learning (monitoring), transformative learning (assimilation) and exploitative learning (apply it to commercial ends) (Lane et al., 2006). In this paper, we measure firms’ monitoring ability as the degree to which they have searched broadly in the past. Lane et al (2006) claim that transformative learning (or assimilation) results from the combination of new knowledge with existing knowledge. Given that firms’ exact combination of new knowledge with existing knowledge is very difficult to gauge, we use the concept of assimilation capacity from a different perspective and measure it in terms of the breadth of the firms’ stock of knowledge. It can be noted that Jansen et al. (2005) measured assimilation in terms of firms’ aptitudes to react to market signals. However, as we are dealing with the specific case of technological exploration— rather than with a multi unit service firm as in the case of Jansen et al. (2005)—we need to use an aspect of assimilation capacity in the technology rather than in the market environment (in line with Mowery et al., 1996).
58
Technological exploration and assimilation capacity Technological exploration can be viewed as a search process by which firms learn and expand their knowledge base. Scholars of organizational learning and strategic management of innovation distinguish between local search or exploration and distant search or exploitation. As discussed above, while local search is an activity aiming at “refinement and extension of existing competencies, technologies, and paradigms” (March, 1991:85), distant search refers more specifically to learning or acquiring new knowledge and therefore innovation. Scholars recognize that both exploration and exploitation are innovative activities that bring about different levels of organizational levels “…exploitative innovations involve improvements in existing components and build on existing technological trajectory, whereas exploratory innovation involves a shift to a different technological trajectory (Benner and Tushman, 2003: 679). From this perspective, the difference between exploration and exploitation lies mainly in the distance from the firms’ core capabilities. As noted by Rosenkopf and Nerkar (2001), exploitation is the most localized form of exploration. However, other scholars have made the point that the differences between exploration and exploitation are more substantial and concern learning, resources and routines (see Gupta et al., 2006 for a survey). Exploration increases variety, helping firms to minimize the risk of obsolescence which is particularly high under conditions of rapid environmental change (Sorensen and Stuart, 2000; Jansen et al., 2005). Through exploration organizations can regenerate their existing knowledge and develop new capabilities (March, 1991; Gavetti and Levinthal, 2000). Exploration may also lead to path breaking innovations (Nelson and Winter, 1982; Galunic and Rjordan, 1998; Fleming and Sorenson, 2001; Nerkar and Roberts, 2004; Miller et al., 2007). Since both types of search are interactively self-reinforcing, it is possible that firms are driven to excessive exploitation or exploration although, in practice, firms often thrive to balance exploration and exploitation (Katila and Ahuja, 2002; Gupta et al., 2006; Lavie and Rosenkopf, 2006). The costs and risk of exploration induce firms to start searching for solutions close to their current knowledge and competencies and to explore alternative solutions thereafter (March and Simon, 1958; Helfat, 1994; Rosenkopf and Nerkar, 2001). The literature has pointed to the importance of cognitive obstacles to exploratory search, such as existing shared knowledge and organizational routines, communication channels and information filters, that makes it difficult for an organization to recognize and assimilate knowledge outside the scope
59
of its core competencies (Nelson and Winter, 1982; Cohen and Levinthal, 1990; Henderson and Clark, 1990; Miller et al., 2007). From this perspective, entry into new technologies or businesses appears to be different from a random walk because the access to new knowledge is costly and the cost of entry into a new technology increases with the distance from the firms’ core knowledge and competencies (e.g., Granstrand et al., 1997; Gambardella and Torrisi, 1998; Piscitello, 2000). Moreover, exploration of new technologies requires AC, an important antidote against the myopia of learning (Cohen and Levinthal, 1990). As mentioned above, here, we use the concept of assimilation capacity and measure it as the breadth of the firms’ stock of knowledge. As Cohen and Levinthal (1990) have pointed out “the ability to assimilate information is a function of the richness of the pre-existing knowledge structure: learning is cumulative, and learning performance is greater when the object of learning is related to what is already known … diversity of knowledge plays an important role … a diverse background provides a more robust basis for learning because it increases the prospect that incoming information will relate to what is already known.” (p. 131). In this context, Lavie & Rosenkopf (2006) have noted that absorptive capacity facilitates exploration through alliances. They also posit that a “broad absorptive capacity” accumulated by interacting with a heterogeneous group of partners is important to explore new alliances (p. 803). We follow this line of argumentation and posit that a large assimilation capacity is associated with a broad absorptive capacity in terms of a diversified technological background and that this capacity is important for future exploration. Accordingly: Hypothesis 1. The broader the knowledge firms have accumulated (assimilation capacity) in the past, the more distant technological exploration from their technological portfolio they will (be able to) undertake in the future. Technological exploration and monitoring ability The literature on licensing has shown that the information that a firm receives by screening external technologies is important to drive the choice of licensing sources. For example, drawing on patent citations, Link & Scott (2002) found that the probability of a licensing agreement between two firms increases with the number of licensee’s citations to licensor’s patents before the agreement. The theoretical explanation they provide is twofold. First, citations imply that the prospective licensee has already gained access somehow to the prospective licensor’s pool of technological assets (e.g., by inspecting patent documents published by the patent office). Second, citations also mean that the potential licensor’s
60
technology is valuable to the potential licensees (e.g., Trajtenberg, 1990; Albert et al., 1991; Harhoff et al., 2003 ; Hall et al., 2005; Hall et al., 2007). According to Cohen and Levinthal (1990) and Lane et al. (2006), another important dimension of AC—in addition to the assimilation capacity—is the ability to recognize, identify and evaluate the potentiality of external knowledge. This capability is accumulated by screening the technological landscape. Katila & Ahuja (2002) developed the concept of search scope and operationalized it by using the share of new citations to total patent citations reported in a focal firm’s patent stock. In this context, Katila & Ahuja (2002) argue that that search scope signals a firms attempt at exploring the technological landscape. We posit that backward citations in general indicate the firm’s exploration of the technological space. Moreover, past exploratory activity, in our view, should enhance the firm’s ability to screen and evaluate future external knowledge. Our proposition is in line with the idea discussed earlier; that past exploration induces more future exploration: “explorative tendencies, guided by absorptive capacity, intensify with firms’ prior exploration experience” (Lavie and Rosenkopf, 2006:803). These arguments lead us to conjecture: Hypothesis 2. The broader firms have searched in the past (monitoring ability), the more distant technological exploration from their technological portfolio they will (be able to) undertake in the future. Assimilation capacity and monitoring ability between licensing and non-licensing firms Based on earlier insights (Cyert and March, 1963), evolutionary economists such as Dosi (1982), Nelson and Winter (1982) and Helfat (1994) have argued that search through firm-internal processes are almost always highly localized in that firms search along established trajectories created by past experiences, routines and heuristics. As expressed by Pavitt (1988), “…the search process of industrial firms to improve their technology is not likely to be one where they survey the whole stock of technological knowledge before making their technical choices. Given its highly differentiated nature, firms will instead seek to improve and to diversify their technology by searching in zones that enable them to use and to build upon their existing technological base.” The constraints to exploration and diversification that are typical of in-house search are likely to be less stringent when exploration is pursued through licensing because of smaller upfront costs and lower technological and market risk (the licensed knowledge may have been used by the licensor before licensing). More generally, organizations operating in complex environments characterized by a diversified knowledge base or firms that enter a new
61
knowledge domain have to draw on alliances and other inter-organizational learning mechanisms to increase their absorptive capacity (Powell et al., 1996; Lane et al., 2006). However, earlier studies on licensing have suggested that technological trajectories that firms pursue when license-in new technologies are closely related to their pre-existing technological background (Killing, 1978; Caves et al., 1983; Chatterji, 1996; Lowe and Taylor, 1998; Kim and Vonortas, 2006). In the words of Caves et al. (1983: 254): “…licensees are not randomly drawn firms that acquire license and then procure the complementary assets; only if those assets lie at hand and command low internal shadow-prices, do they seek the missing technology in the licensing market.” In addition Roberts and Berry (1985) suggest that licensing is more suitable when firms have to acquire technologies that are new, but familiar. Within their framework of reference, familiarity with a technology is defined as “the degree to which knowledge of technology exists within the company, but it is not necessarily embodied in [its] products” (Roberts and Berry, 1985: 3). Combined, all this shows that both licensing-in and internal technological exploration (or diversification) are constrained by firms’ previous experiences and overall suggests that licensing-in follows similar trajectories as internal exploration. To be sure, when a firm wishes to explore into a new patentable technology that is more or less distant to what it already does, the firm has—in many, if not most cases—the possibility of exploring through in-licensing. It can also explore internally—either through searching in a different direction or by inventing around an existing patent. Concerning the latter, the empirical literature has shown that because innovations are typically complex and difficult to define completely and precisely in a patent, it is possible to invent around existing patent protections and, for firms that want to, do so successfully and fairly quickly (Mansfield et al., 1981). However, the important point is that licensing-in (like other inter-organizational mechanisms) entails a higher potential for more distant exploration than exploration through internal search due to the smaller upfront costs and lower technological and market risk in the former case. Accordingly, a large assimilation capacity and monitoring ability are likely to be able unleash this potential. So the idea is that while these aspects of AC are important in the case of pure internal exploration they become important a fortiori in the case of exploration through in-licensing. In other words, external linkages like licensing-in moderate the effect of AC on exploratory search. This leads us to conjecture: Hypothesis 3.a. Licensing-in reinforces the positive effect of a larger assimilation capacity on the distance of future technological exploration from firms’ existing technological portfolio.
62
Hypothesis 3.b. Licensing-in reinforces the positive effect of past exploration on the distance of future technological exploration from firms’ existing technological portfolio.
Method
34B
Data and Sample In order to test our hypotheses we developed a research design based on multiple sources of information on license and patent data. We started from the Intellectual Property database maintained by the Financial Valuation Group.
This database records IPRs
transaction agreements concluded from the 1970s to the present, including the exchange of software, know-how, technology, copyright, patent and products. For the sake of our analysis we extracted only “patent” and “technology” transactions, identified in the database as such. This led us to an initial set of 1052 observations. For each transaction we could originally rely on basic information on the document source, the date of the event and the source which reported the event, the names of the licensor and the licensee, their respective Standard Industrial Classification (SIC) and North American Industry Classification System (NAICS) industry codes with a qualitative description of industries, a brief synopsis of the transaction, a description of the remuneration structure and,—whenever available—the identification number of patents involved in the transaction. This information allowed us to cross-link the original dataset with many other sources of information that were deemed useful and necessary for our analysis, at the level of the license agreement, the licensing parties and the licensed patents. The first additional data source we used was the Security Exchanged Commission (SEC) website. We first searched for the original contractual document (License Agreement), or at least for some more detailed references concerning the transactions in other filing data (e.g., S1, 8K, 10K) in order for us to check both the reliability of data drawn from the FVG IP dataset, and to include these data in the dataset. However, the unavailability of key information caused the sample to drop substantially. Because of their strategic relevance and sensitiveness, several important details concerning licensing contracts are usually protected by confidentiality agreements between the parties. In fact, in some cases we could not find the contractual document, while in others, information about the technology that had been exchanged was missing. In the latter case—in order to check whether a real transfer of patents had taken place and to exactly identify the type of technology involved in the agreement, we
63
browsed the US Patent Office (USPTO) dataset according to the information available in the text of the contract (e.g., the application number or the title of the issued patents included in the description of the transaction). In specific cases we searched through the names of the assignees in the focal year together with the keyword provided in the FVG IP dataset (e.g., ruby laser hair). After this search and integration activity, we ended up with 301 patent agreements. However, given the specific purpose of our analysis, we included only those transactions that have been filed originally by the parties as (pure) licensing or assignment agreements, implying a one-way technology/IPRs transfer whereby the licensor maintains the ownership of the licensed/assigned technology. Subsequently, we then exclude all other transactions that refer to collaboration or settlement agreements, cross-licensing, technology purchases and plans to merge. At the end of this process we relied on a sample of 224 licenses involving almost 900 USPTO patents exchanged among licensor and licensee firms. In order to collect all relevant statistics on each licensed patent, we then matched our database with the National Bureau of Economic Research (NBER) dataset (Hall et al., 2001) and its 2002 update. The NBER patent database is widely used in research relying on patents information and statistics. It is a collection of all patents granted by the US patent office until 2002 and contains information on year of application, year of grant, number of claims, citations received and made and primary International Patent Classification (IPC) class. Given that we use patent citation to create the measure of monitoring ability, using USPTO patents is important because, unlike— for instance—European Patent Office (EPO) patents, a large number of citations are assigned by the patent applicant. This observation makes the use of citations a reasonably good indicator of search conducted by the applicant during the inventive process. At the firm level, we matched the available information on licensees’ name and industry, with data on firm size measured by the number of employees drawn from proprietary or publicly available data sources, namely Thomson Research, Comp tech, Google Finance and, whenever necessary, company websites or other online available data sources. We then matched the name of our licensees with the patent assignee’s names recorded in the NBER dataset (on some occasions, the patent was assigned to the parent company) to obtain their patent portfolios and related statistics for each patent. We found that our licensees overall have been granted about 108,000 USPTO patents. At the very end of the data collection, integration and cleaning procedure our dataset was enriched substantially by adding information on the characteristics of all licensed patents, the licensee’s main business (SIC
64
code) and size, and detailed information about the licensee’s patents already issued at the time of the license, about the licensor, and on the license contract itself. In order to explore whether there are significant differences in searching behavior between licensees and non-licensees and to understand the full implications of licensing activity and patterns, we also constructed a control sample consisting of non-licensee firms whose profile is similar to our treatment sample of licensees. Since our research design builds on patent data, we decided to compare licensees and non-licensees that display a history of patenting activity over time. We picked up potentially matching firms from the whole set of USPTO patent assignees. By doing so we ensured that both licensees and matched firms have dealt with patentable inventions. In order to get all the relevant information on patents filed by these matching companies automatically, we relied again on the NBER patent dataset (2002 version) as we did for our licensee sample. For the same reason, we first imposed a discriminatory condition according to which the matched company should have applied for a patent in the same four-year time-span as the treatment company licensed-in a technology. Once we ensured the existence of this condition, we combined three extra criteria for matching. These are: same SIC code at the 2-digit level; same region of the world (Asia, Europe or USA); and same patent portfolio size. Indeed, firm-size (in our case measured as the size of the patent stock), geographical localization and industry affiliation are all accepted as matching criteria in the literature (see for instance, Fosfuri et al., 2005). The matching procedure allowed us to identify a sample of potential matched nonlicensees. We then manually checked (through the Thomson Research Database) whether the each of the potentially matched firms had, in fact (i.e., missed by the FCGIP database), licensed in any technology around the year of the licensing firm’s license. When no technology licensing activities were found in the Thomson Research Database, we proceeded by “googeling” the name of the company combined with the term “license agreement” to get all the possible publicly available information for that firm. Only in the cases where we were sure that the potential matching firms did not acquire a license in the relevant period, we included the firm in the matching sample. We considered ourselves to be sure when we detected material about the firms that included information about technological activity (typically, the firms’ annual reports)—when we found no information on the given firm or the firm had licensing activity (this happened frequently), we deemed the potential match an unfit match and we went on to examine another potential match. We repeated this procedure eight times
65
to increase the number of matched firms. In this way we ended up with 183 licensees that were matched up with 183 non-licensees, matched on a one-to-one basis. The large difference in the average patent stock between the two samples led us to drop five outliers from the treatment sample and their matched firms. The average patent stock for the treatment sample before dropping these outliers was about 319 patents against 34 patents of the matched sample. The patent stock of outliers ranged between 3,664 (Rosetta Inpharmatics) to over 25,000 patents (IBM). Our final sample then consists of 176 licensees and 176 non-licensees. Since firms can look for external knowledge using different channels than licensing-in, we also examined the patent documents of the two samples to find cases of co-assignment. Measures Dependent variable. Our dependent variable is the distance of technological exploration from the firms’ existing technological portfolio pursued by firms through licensing (for licensing firms) or own patenting activity (for the matched non-licensing firms). Here distance should be understood as how overlapping the existing patent portfolio and the new patent (licensed-in or own patent) are. When the existing patent portfolio contains only a small fraction of patents in the same IPC code as the new patent, the distance is considered high and when the existing patent portfolio contains a large fraction of patents in the same IPC code as the new patent, the distance is considered low. When the entire existing patent portfolio is within the same IPC code as the new patent, the distance is considered ultimately low and when the entire existing patent portfolio is outside the same IPC code as the new patent, the distance of technological exploration is considered ultimately high. Inspired by the focus index introduced by Ziedonis (2007)—the citation-weighted share of patents granted within the last six years that are in the same 4-digit IPC codes as the licensed patents—we computed a measure of distance between the licensed patent and the licensee existing patent portfolio composition. More precisely, we took the citation-weighted sum of licensee I’s granted patents that were applied for within 6 years of the time t of the license and are in the same primary IPC class as the class of the licensed patent (or at least in one of them if patents are more than one) divided by the citation-weighted sum of all patents issued to the licensee that were applied for by the time t of the license. Our measure of distance of technological exploration is 1 minus the focus index:
66
t ~ Distance of technological exploration = 1- ∑∑ C i pì c t −6 J
t ∑∑ C~i p ì t −6 J
t ~ where ∑∑ Ci pì is the citation-weighted (Ci) sum of firm i’s patents (pi) that were c t −6 J
applied for within six years at the time of the licensing decision t and are in the same primary t ~ patent class as the class c of the [licensed] patent under consideration”; and ∑∑ Ci pì is t −6 J
the citation-weighted sum of all patents issued to the firm (the licensee) that were applied for by date t (year of the license)”. The greater the value of our index, the higher is the distance of technological exploration from the firms’ existing technological portfolio. Based on our matching procedure criteria, described above, we identified comparable firms that did not license any technology from external sources of knowledge in the relevant years of investigation. For these firms we computed the same index of exploratory search above calculated with internal patenting activity rather than licensed-in patents. That is, we computed the complement to one of the citation-weighted sum of non-licensee I’s granted patents that were applied for within 6 years of the time t of the license and are in the same primary 4-digit IPC class as the class of the newly in-house developed patent, divided by the citation-weighted sum of all patents issued to the firms that were applied for by the time of license, t. Independent variables. Consistent with the aim of our paper—which is to further the understanding of the main factors affecting the distance of technological exploration from the firms’ existing technological portfolio pursued through licensing (or own patenting activity)— we focus our attention on two key regressors that, we believe, make the difference when comparing licensee and non-licensee firms. They refer to licensees’ characteristics, reflecting either their assimilation capacity or monitoring ability. We operationalized assimilation capacity by measuring firms’ patent portfolio dispersion across technological classes. This choice is rooted in organizational research and innovation management. For example, Lane, Koka, & Pathak (2006) have noted that “the breadth of knowledge that a firm understands determine how far its exploratory learning can venture from its existing knowledge base” (p. 855). The greater the dispersion of a firm’s technological background, the higher its ability to assimilate external technologies in distant, unfamiliar knowledge domains. We measure this
67
ability as the complement to one of the Herfindahl index applied to licensees’ patent portfolio composition as recorded at the time of license. This index reflects the degree of dispersion of the licensee’s patents across different 4-digit IPC classes and varies between 0 and 1. The higher the index, the broader the scope of the licensee’s technological expertise and therefore, the more likely it will be able to enter new technologies. Firms with a dispersed patent portfolio have learned to manage different technologies and therefore should display a greater ability to enter into a new technological field compared with firms endowed with a narrow technological portfolio. We created the same variable for non-licensee firms. In doing so, we took care of excluding the patent(s) that have been applied in the same year of the corresponding license, t, in order not to include it/them as it/they represent(s) the benchmark for the construction of the dependent variable. The second variable reflects firms’ past exploratory search activity and thereby firms’ monitoring ability. The expected impact of past exploration (and monitoring ability) on future exploratory search is in line with the literature on organizational learning which posits that “exploration often leads to more exploration, and exploitation to more exploitation” (Gupta et al., 2006: 695). Here, we follow earlier studies which have relied on backward citations as a measure of technological search (Katila and Ahuja, 2002). Our proxy for past monitoring activity is the average number of backward citations reported in the focal firms’ patent stock before the license announcement. As mentioned before, backward citations signal a firms attempt at exploring the technological landscape over time. The higher the average number of citations, the broader is the firm’s technological search activity, and the higher is the variety of knowledge sources the firm is aware of. Accordingly, we assume that an intensive citation activity enhances the monitoring ability of the firm. We computed the same variable for nonlicensee firms, but we did not include in the count of backward citations those referring to the newly in-house developed patent as it stands as our baseline for the construction of the dependent variable. An alternative measure of past exploratory search activity would have been a variable reflecting past in-licensing experience. However, we have to rely on citations, given that our sample contains only very few firms with repeated license-in experiences. Control variables. To control for the effect of exploratory mechanisms other than licensing, we generated a dummy called co-patenting that takes value 1 if the firm has at least one co-assigned patent in its patent portfolio. We obtained this data from a recent addition to the NBER patent data set provided by James Bessen at http://www.nber.org/~jbessen/. In addition, since we believe that firms’ experience in patenting activity may have an impact on
68
its distance of technological exploration from the existing technology portfolio at a given point of time, we account for this by introducing three different variables, including licensee’s patent stock, licensee’s patent experience and patent activity. The licensee’ patent stock was obtained by counting the number of patents applied for by the licensee before the license announcement. We use it to control for the scale of innovative activities.
Like the
diversification of the patent portfolio described above, this could be viewed as another proxy for absorptive capacity. However, it is a quite crude proxy of absorptive capacity because it does not account for the composition of the firms’ patent stock. Patent experience takes into account the lag between the license year and the year of issue of the licensee’s first patent. This measure is supplemented with a dummy (patent activity), that takes value one if the licensee has been granted at least one patent all its life, and zero otherwise. We got this information from
the
USPTO
database
and
the
patent
genius
database
available
online
(www.patentgenius.com). Moreover, we control for the degree of generality of the technology that the recipient firm in-licenses from outside, as generality is an important characteristic of technology (Bresnahan and Trajtenberg, 1995; Hall and Trajtenberg, 2004). The market for general purpose technologies (e.g., IT and software) is generally more efficient than that of other technologies. This is demonstrated by the large share of these technologies in technology transactions (Arora et al., 2001b). For this purpose we rely on the Generality index reported in the NBER dataset for
1 − ∑ j sij2 ni
USPTO patents. The generality index is calculated as follows:
where
sij
is is the
percentage of citations that a patent in technological class i receives from patents in class j. The index is a complement to one of the Herfindahl index of patents’ forward citations across (IPC) technological classes. The index varies between 0 (minimum generality, all citations received are concentrated in one technological field) to 1 (max generality, citations are highly dispersed across different fields) (Hall and Trajtenberg, 2004). For our purposes here, and given the fact that firms may have licensed more than one patent, we use the highest value of the generality index among all patents exchanged through each transaction. We generated an equivalent measure of generality by applying the generality index above to the newly in-house developed patent of non-licensees in order to allow comparison across samples. According to the overall research design, we created the same variables for the nonlicensee sample. We did not include the patent activity dummy since by definition it is equal to
69
1, given that non-licensees were sampled among the overall set of USPTO patent assignees. For the construction of the equivalent measure of the patent stock and patent experience we made sure not to include the focal patent that has been filed in the year of the corresponding license agreement. Also, as underlined in the description of the matching procedure, we include the same variable—patent stock—in the list of the matching criteria we employed to identify the non-licensee firms. We also control for firm size by creating a categorical variable based on the number of employees—less than 100 (small firm), between 100 and 1000 (medium firm) and above 1000 (large firm). We could have relied on the patent stock measure as proxy for size, but we decided not to do this, because very small firms may display an extensive patenting activity as compared to large firms. Finally, we controlled for industry dummies based on the SIC-code of at the 2-digit level as attached to the licensee. SIC classes have been aggregated into 10 broad industries. We have generated a dummy variable for each of these industries. For instance, Communication, Business, Engineering, Accounting, Research, Management, and Related Services were grouped in a sector called knowledge and information-based services (KIBS). In our setting, industry dummies may account for unobservable environmental conditions, like technological and market uncertainty or appropriability that may affect the degree of explorative search of the firm. Econometric method As our dependent variable is continuous we use ordinary least squares as the means of estimation. In the first part of our analysis we focus on search through licensing among licensees alone and non-licensees alone. Thereby, we can cast light on Hypotheses 1 & 2. To utilize our control group we conduct an analysis with difference-in-difference characteristics that also allows for assessing the existence of significant differences in technological search patterns across the two samples that is needed to test Hypotheses 3.a and 3.b (and to cast further light on Hypotheses 1 & 2). However, as we focus on the search behavior through licensing and how far an acquired license is from the existing technology-base of the focal firm, our research design is inherently cross-sectional. For that reason, we do not strictly rely on the difference-in-difference estimator that analyses effects of a treatment over time (Wooldridge, 2002: 128-132). Here, we consider licensees to be our treatment group and nonlicensees to be the control groups. Our key independent variables are then considered to be the treatment. Concretely, and in order to allow us to compare the two samples we created a
70
dummy called licensee that takes the value 1 if the 176 observations refer to the licensee firms, and 0 otherwise. By interacting this variable with the main regressors, we can assess the differences in technological search patterns across the two samples. We also interacted the copatenting variable with the main regressors to see whether the effect of licensing on the distance of technological exploration remains significant beyond that of other external sources of knowledge.
Results
35B
Descriptive statistics and correlations
36B
Table 1 summarizes descriptive statistics for each variable included in our regression analysis. Descriptive statistics are available for the pooled sample, the sample firms (licensees) and the control sample (non licensees). Apparently, there are no significant differences between the two groups in terms of the distance of technological exploration (our dependent variable) and assimilation capacity (measured by the breadth of the firm’s patent stock). However, it can be noted that the average number of citations in patents (our measure of monitoring ability) held by the treatment group is significantly smaller than the average citations in patents held by the control group (t statistics=4.133; p-value=.000). Firms in the treatment group show on average a larger stock of patents before licensing, but the difference with the stock of the control sample is not statistically significant. ------------------------------------------Insert Table 1 about there ------------------------------------------Table 2 shows the Pearson correlation coefficients of the variables included in the analysis. The correlations do not warrant further examination with respect to multicollinearity. ------------------------------------------Insert Table 2 about there ------------------------------------------Table 3 reports the results of the OLS regressions analysis. Our dependent variable across the six models reported in this table reflects the distance of technological exploration from the firm’s existing technological portfolio through licensing-in or own patenting. The
71
first two columns report estimation results on the treatments sample and the control sample respectively. The third column reports the baseline model in which the two samples are pooled together and only control variables are entered, including the dummy for co-patenting. Model IV includes our main explanatory variables along with controls. It should be noted that this latter model includes interactions between the license dummy and the key regressors, namely assimilation capacity and monitoring ability. Model V includes the interactions between the co-patenting dummy (instead of the license dummy) and the key regressors while Model VI includes all variables and interactions. ------------------------------------------Insert Table 3 about there ------------------------------------------For the sake of simplicity, we focus our discussion on the full model in column VI. Hypotheses 1 and 3.b find support in our data, whereas Hypotheses 2 and 3.a are not supported. In Hypothesis 1, we conjectured that a greater assimilation capacity, measured by the breadth of the patent stock, increases the firm propensity to engage in more distant technological exploration, thus departing from its current knowledge base. In Table 3 the coefficient for assimilation capacity is positive and remains significant below the 5 per cent level when all controls are included in the equation. We should note, however, that this effect is stronger within the licensee sample (see Models I and II). One may wonder whether this difference between the two groups is driven by the group of licensees. We will turn to this issue in the discussion of Hypothesis 3.a below. Contrary to our expectations (as stated in Hypothesis 2), the coefficient for past exploratory search (monitoring ability) is negative and statistically significant at the 1 per cent level. If we look at Models I and II, however, we notice that this effect is driven by the control group. We recall that this group has a significantly larger monitoring ability than the licensee group. Accordingly, it is plausible that a more intense past exploration leads non-licensing firms to concentrate technological activities on the core competencies rather than exploring further the technology space. Pertaining to Hypothesis 3.a we test the idea that licensing-in reinforces the positive effect of a large assimilation capacity (a broad patent portfolio) on the distance of future exploration from the firm existing patent portfolio. This amounts to say that we expected a combined positive effect of assimilation capacity and licensing on the distance of exploration. The coefficient on this interaction effect in Models IV and VI is positive as expected, but it is
72
not significant at the conventional levels. Hence, our results do not lend a strong support to Hypothesis 3.a, suggesting that firms with a strong assimilation capacity have greater propensity to explore the technology market regardless of their involvement in licensing-in. In other words, we do not find evidence to suggest that the result obtained concerning assimilation capacity in affecting technological exploration (Hypothesis 1) is driven by the group of licensees. In Hypothesis 3.b, we predict that the interaction between licensing-in and a large monitoring ability (accumulated through past exploratory search) will lead firms to explore more distantly from their existing patent portfolio. While a large monitoring ability as such does not lead firms to explore far from the existing technological background, the combination between monitoring ability and the use of the technology market (licensing) yields a positive and significant effect on the distance of exploration. This latter result is robust to controlling for co-patenting (Models IV and VI). This suggests that licensing-in offers firms endowed with monitoring ability an important opportunity to explore the technological space beyond and above other mechanisms such as co-patenting. The latter has an insignificant effect on the distance of exploration. Interestingly, the interaction between copatenting and our two key regressors yields effects that are similar to that of licensing-in. In particular, co-patenting reinforces the positive effect of a large monitoring ability on exploration. Concerning other control variables, all models show that the size of patent stock has virtually no effect on the distance of technological exploration. This means that firms that have more familiarity with patents do not necessarily explore far from their existing technological background to capture the opportunities available in the market for technology. Firms with longer patent experience before the focal year (license-in or new patent application) and firms owning at least one patent (measured by our patent activity variable) all their life show a lower propensity to engage in exploration as suggested by the negative coefficients for these variables. These effects, however, become insignificant when all controls are included in the regression analysis (Model VI). Finally, larger firms show a higher propensity to explore far from their technological background as compared to smaller firms. The weak association between firm size and assimilation capacity (Table 2) suggests that the effect of size is not mediated by technological breadth. By the same token, the weak association between size and monitoring activities (Table 2) suggests that the positive effect of size on exploration is not mediated by a greater
73
monitoring ability (number of citations per patents). Larger firms, then, are probably in a better position to explore the technological landscape for organizational reasons such as a better division of labor in innovation and related activities.
Discussion and Conclusion
37B
This paper began by observing that the existing literature on technology licensing mostly focuses on firms’ choices as to whether to produce an innovation in-house or to license it to another firm, while only few studies examine how firms use in-licensing as a part of their overall technology strategy. In this paper, we have focused on the distance between the technology acquired by licensing-in and the firm’s existing technological portfolio. We posited that a long distance indicates the outcome of exploratory search whereas a short distance reveals the outcome of more exploitative technological search. The underlying idea is that the degree of exploration through licensing-in is shaped by two distinct dimensions of absorptive capacity (AC)—assimilation capacity and monitoring ability. Our empirical analysis showed that assimilation capacity is an important determinant of the ability to explore distantly from the firms’ existing technological portfolio. The negative sign of monitoring ability was, however, unexpected. Although our cross sectional design does not allow for a dynamic explanation of this result, we can speculate that firms alternate phases of exploration, whereby they monitor the external technological space, with phases of exploitation during which they assimilate and further develop what they have learned from past exploration. This reasoning is in line with the proposition that exploration and exploitation are complements in the long run but are likely to be substitutes at a given point in time (they are synchronically substitutes). Moreover, our findings showed that licensees with large monitoring ability (acquired through past exploratory search) tend to explore more distantly from their existing technological portfolio as compared to similar firms that do not rely on licensing-in. The positive effect of the interaction between monitoring ability and licensing as a mode of technology acquisition points to the importance of markets for technology in the exploration of the technological landscape in the search for new knowledge. It also suggests that gaining access to distant, unfamiliar, technologies through the market for technology requires prior investments in monitoring ability. This result is in line with the idea put forth by Cohen and Levinthal (1990) stating that knowledge is not a public good and
74
requires specific investment to be absorbed. Markets for technology can reduce, but not eliminate, the costs of access to external knowledge. Our analysis suggests that the more distant is the knowledge that a firm seeks to acquire in the market for technology, the greater the total cost of acquisition which consists of an explicit component (the license fees) and an implicit component (the cost of AC). The latter is difficult to measure, but it is important to recognize for R&D and IP managers. The reason why firms endowed with strong assimilation capacity (measured by the breadth of their patent portfolios) do not rely on licensing in particular to explore new technological fields is less clear. We can speculate that firms with strong assimilation capacities do not need the market for technology to gain access to new technological fields—these fields can be reached through in-house exploration. As mentioned before, the interaction between co-patenting and the two dimensions of AC considered in this study (assimilation capacity and monitoring ability) are similar to those of licensing-in. This suggests that ex-ante R&D collaboration (resulting in co-patenting) also requires a prior effort in the accumulation of AC. Our study extends previous research in the following directions. First, we contribute to the literature on the markets for technology (Arora et al., 2001b; Fosfuri, 2006) by having focused on the demand-side of this market; an issue that has been generally under researched so far. Second, our paper makes a contribution to the literature on technological search and open innovation (e.g., Katila and Ahuja, 2002; Laursen and Salter, 2006) by having shedded new light on the role of licensing-in as a strategy to capture new technological opportunities outside the boundaries of the firm. Several earlier studies have analyzed the role of various types of alliances as a learning mechanism, particularly when firms explore new businesses (e.g., Kogut and Zander, 1992; Khanna et al., 1998). However, there is still limited research concerning the antecedents of licensing-in as a mechanism of exploration. Third, our analysis focuses on different dimensions of AC as antecedents to technological exploration. Several previous studies have further developed the notion of absorptive capacity following the seminal paper by Cohen & Levinthal (1990); (see, Mowery et al., 1996; Zahra and George, 2002; Jansen et al., 2005; Lane et al., 2006). However, the papers with an empirical component have either not dealt with the dimensions of AC relevant to technological exploration (Jansen et al., 2005) or have claimed to measure AC in general (Mowery et al., 1996). By having made the distinction between assimilation capacity and monitoring ability, we see this paper as a first step towards breaking down the multi-level AC concept into components relevant to the technology dimension. More in particular, however,
75
our contribution regarding AC lies in the fact that no previous work—to our knowledge—has attempted to examine the implications of AC in the context of licensing-in. This is a significant contribution given the rising importance of the market for licensing (Arora et al., 2001b; Robbins, 2006). We acknowledge that interpreting the complex interactions among assimilation capacity, monitoring ability, and technology licensing vs. internal development is a very difficult task which deserves further scrutiny, possibly based on further analysis and in-depth case-studies. One possibility is that assimilation capacity and monitoring ability are not independent drivers of the distance of exploration, but one of them is a mediator through which the other translates into more distant exploration. For example, one could argue that monitoring ability will translate into increased exploration through developing assimilation capacity. Another limitation to this study is that we have focused on a cross-section of licensing agreements. This prevents us to account for firm-specific unobserved heterogeneity. Collecting information on licensing agreements to obtain a panel dataset with a significant longitudinal dimension remains the object of future research. We have used only rough control variables (in particular industry-level dummies) that can somehow be said to reflect environmental pressures including market or technological uncertainty. Although our dependent variable is reflecting something that goes on in the technological domain—and not successful commercialization of the technology—it would also have been desirable to control for downstream complementary assets. Despite our matched sample approach can alleviate some of these problems, future research would benefit from the use of finer grained controls. Nevertheless, we believe that this paper is a significant first attempt at better understanding the trajectories of firms’ in-licensing behavior and their association with the firms’ learning capabilities.
References
38B
Alcácer, J. and M. Gittelman. (2006), 'Patent Citations as a Measure of Knowledge Flows: The Influence of Examiner Citations,' The Review of Economics and Statistics, 88(4), 774–779. Arora, A., A. Fosfuri and A. Gambardella. (2001a), 'Markets for Technology and their Implications for Corporate Strategy,' Industrial and Corporate Change, 10(2), 419-451.
76
Arora, A., A. Fosfuri and A. Gambardella. (2001b), Markets for Technology: The Economics of Innovation and Corporate Strategy. The MIT Press: Cambridge, Massachusetts. Atuahene-Gima, K. (1993), 'Determinants of Inward Technology Licensing Intentions: An Empirical Analysis of Australian Engineering Firms,' Journal of Product Innovation Management, 10, 230-240. Atuahene-Gima, K. and P. Patterson. (1993), 'Managerial Perceptions of Technology Licensing as an Alternative to Internal R&D in New Product Development: an empirical investigation,' R&D Management, 23(4), 327-336. Benner, M. J. and M. L. Tushman. (2003), 'Exploitation, exploration, and process management: The productivity dilemma revisited,' Academy of Management Review, 28(2), 238256. Bresnahan, T. F. and M. Trajtenberg. (1995), 'General purpose technologies: Engines of growth,' Journal of Econometrics, 65, 83-108. Caves, R. E., H. Crookell and J. P. Killing. (1983), 'The Imperfect Market for Technology Licenses,' Oxford Bulletin of Economics and Statistics, 45, 249-267. Chatterji, D. (1996), 'Accessing External Sources of Technology.,' Research Technology Management, 39(2), 48-56. Chesbrough, H. (2003), Open innovation. Harvard University Press: Cambridge, Massachusetts. Cohen, W. M. and D. A. Levinthal. (1990), 'Absorptive Capacity: A New Perspective of Learning and Innovation,' Administrative Science Quarterly, 35, 128-152. Criscuolo, P. and B. Verspagen. (2008), 'Does it matter where patent citations come from? Inventor vs. examiner citations in European patents,' Research Policy, 37(10), 1892–1908. Cyert, R. M. and J. G. March. (1963), A Behavioral Theory of The Firm. Prentice-Hall: Englewood Cliffs, N.J. Dosi, G. (1982), 'Technological Paradigms and Technological Trajectories: A Suggested Interpretation of the Determinants and Directions of Technical Change,' Research Policy, 11(3), 147-162. Fosfuri, A. (2006), 'The Licensing Dilemma: Understanding the Determinants of the Rate of Technology Licensing,' Strategic Management Journal, 27(12), 1141-1158. Fosfuri, A., M. S. Giarratana and A. Luzzi. (2008), 'The Penguin Has Entered the Building: The Commercialization of Open Source Software Products,' Organization Science, 9(2), 292-305. Fosfuri, A. and J. A. Tribó. (2008), 'Exploring the Antecedents of Potential Absorptive Capacity and its Impact on Innovation Performance,' Omega, 36(2), 173-187. Gambardella, A., D. Harhof and B. Verspagen. (2008), 'The value of European patents,' European Management Review, 5, 69–84.
77
Gambardella, A. and S. Torrisi. (1998), 'Does Technological Convergence Imply Convergence in Markets? Evidence from the Electronics Industry,' Research Policy, 27, 445-463. Granstrand, O., P. Patel and K. L. R. Pavitt. (1997), 'Multi-Technology Corporations: Why They Have 'Distributed' Rather than 'Distinctive Core' Competencies,' California Management Review, 39(4), 8-25. Gupta, A. K., K. G. Smith and C. E. Shalley. (2006), 'The interplay between exploration and exploitation,' Academy of Management Journal, 49(4), 693-706. Hall, B. H., A. B. Jaffe and M. Tratjenberg. 2001. The NBER Patent Citation Data File: Lessons, Insights and Methodological Tools. 8498, NBER. Hall, B. H. and M. Trajtenberg. 2004. Uncovering GPTS with Patent Data (Working paper). Helfat, C. (1994), 'Evolutionary Trajectories in Petroleum Firm R&D,' Management Science, 40, 1720-1747. Henderson, R. and K. B. Clark. (1990), 'Architectural innovation: The reconfiguration of existing product technologies and the failure of established firms,' Administrative Science Quarterly, 35(1), 9-30. Jansen, J. J. P., F. A. J. van den Bosch and H. W. Volberda. (2005), 'Managing Potential and Realized Absorptive Capacity: How do Organizational Antecedents Matter?,' Academy of Management Journal, 48, 999-1015. Katila, R. and G. Ahuja. (2002), 'Something old, something new: A longitudinal study of search behavior and new product introduction,' Academy of Management Journal, 45(6), 11831194. Khanna, T., R. Gulati and N. Nohria. (1998), 'The Dynamics of Learning Alliances: Competition, Cooperation and Relative Scope,' Strategic Management Journal, 19, 193-210. Killing, J. P. (1978), 'Diversification through licensing,' R&D Management, 8(3), 159-163. Kim, Y. J. and N. S. Vonortas. (2006), 'Determinants of Inter-firm Technology Licensing: The Case of Licensors,' Managerial and Decision Economics, 27(4), 235-249. Kogut, B. and U. Zander. (1992), 'Knowledge of the Firm, Combinative Capabilities, and the Replication of Technology,' Organization Science 3(3), 383-397. Koza, M. P. and A. Y. Lewin. (1998), 'The Co-Evolution of Strategic Alliances.,' Organization Science, 9(3), 255-264. Lane, P. J., B. Koka and S. Pathak. (2006), 'The reification of absorptive capacity: A critical review and rejuvenation of the construct,' Academy of Management Review, 31(4), 833 - 863. Laursen, K. and A. J. Salter. (2006), 'Open for Innovation: The role of openness in explaining innovative performance among UK manufacturing firms,' Strategic Management Journal, 27(2), 131-150.
78
Lavie, D. and L. Rosenkopf. (2006), 'Balancing exploration and exploitation in alliance formation,' Academy of Management Journal, 49(4), 797-818. Levinthal, D. A. (1997), 'Adaptation on rugged landscapes,' Management Science, 43(7), 934-950. Lichtenthaler, U. and H. Ernst. (2007), 'Developing reputation to overcome the imperfections in the markets for knowledge,' Research Policy, 36(1), 37–55. Lowe, J. and P. Taylor. (1998), 'R&D and Technology Purchase through License Agreements: Complementary Strategies and Complementary Assets.,' R&D Management, 28(4), 263-278. March, J. G. (1991), 'Exploration and Exploitation in Organizational Learning,' Organization Science, 2, 71-87. Miller, D. J., M. J. Fern and L. B. Cardinal. (2007), 'The use of knowledge for technological innovation within diversified firms,' Academy of Management Journal, 50(2), 308-326. Mowery, D. C., J. E. Oxley and B. S. Silverman. (1996), 'Strategic alliances and interfirm knowledge transfer,' Strategic Management Journal, 17, 77-91. Nelson, R. R. and S. Winter. (1982), An Evolutionary Theory of Economic Change. Harvard University Press: Cambridge, Massachusetts. Pavitt, K. L. R. (1988), 'International patterns of technological accumulation,' in N. Hood and J.E. Vahlne (eds.), Strategies in Global Competetion. Croom Helm: London. Piscitello, L. (2000), 'Relatedness and coherence in technological and product diversification of the world's largest firms,' Structural Change and Economic Dynamics, 11, 295-315. Powell, W. W., K. W. Koput and L. Smith-Doerr. (1996), 'Interorganizational collaboration and the locus of innovation: Networks of learning in biotechnology,' Administrative Science Quarterly, 41(1), 116-145. Robbins, C. A. (2006), Measuring Payments for the Supply and Use of Intellectual Property. Bureau of Economic Analysis U.S. Department of Commerce: Washington D.C. Rosenkopf, L. and A. Nerkar. (2001), 'Beyond local research: boundary-spanning, exploration, and impact in the optical disk industry,' Strategic Management Journal, 22(4), 287-306. Scotchmer, S. (2004), Innovation and incentives. MIT Press: Cambridge, Mass., USA. Silverman, B. S. (1999), 'Technological Resources and the Direction of Corporate Diversification: Toward an Integration of the Resource-Based View and Transaction Cost Economics,' Management Science, 45(8), 1109-1124. Sorensen, J. B. and T. E. Stuart. (2000), 'Aging, obsolescence and organizational innovation,' Administrative Science Quarterly, 45(1), 81-112. Teece, D. (1988), 'Technological change and the nature of the firm,' in G. Dosi, C. Freeman, R. Nelson, G. Silverberg and L. Soete (eds.), Technical Change and Economic Theory. Pinter Publishers: London, 256-281.
79
Wooldridge, J. M. (2002), Econometric Analysis of Cross Section and Panel Data. The MIT Press: Cambridge Massachusetts. Zahra, S. A. and G. George. (2002), 'Absorptive capacity: A review, reconceptualization, and extension,' Academy of Management Review, 27(2), 185-203. Ziedonis, A. A. (2007), 'Real Options in Technology Licensing,' Management Science, 53(10), 1618-1633.
80
TABLE 1 Descriptive Statistics Pooled sample Obs.
Mean
S.D.
Min
Max
Distance of T echnological Exploration
352
0.84
0.31
0
Assimilation Capacity
352
0.72
0.35
0
1 1
Monitoring Ability
352
5.77
10.54
0
79
Co-patenting
352
0.10
0.30
0
1
Patent Stock
352
24.01
158.28
0
2231
Generality
352
0.49
0.40
0
1
Patent Experience
352
5.54
8.26 -11
37
Patent Activity
352
0.86
0.35
0
1
Size Licensee
352 352
0.84 0.50
0.81 0.50
0 0
2 1
Licensees (split sample #1) Obs.
Mean
S.D.
Min
Max
Distance of T echnological Exploration
176
0.85
0.31
0
Assimilation Capacity
176
0.73
0.34
0
1 1
Monitoring Ability
176
3.50
7.19
0
37
Co-patenting
176
0.07
0.26
0
1
Patent Stock
176
35.03
221.15
0
2231
Generality
176
0.58
0.35
0
1
Patent Experience
176
5.10
7.74 -11
37
Patent Activity
176
0.72
0.45
0
1
Size
176
0.60
0.79
0
2
Licensee
176
1.00
0.00
1
1
Non-licensees (split sample #2) Obs.
Mean
S.D.
Min
Max
Distance of T echnological Exploration
176
0.83
0.31
0
1
Assimilation Capacity
176
Monitoring Ability
176
0.72
0.36
0
1
8.05
12.68
0
79
Co-patenting
176
0.12
0.33
0
1
Patent Stock
176
12.97
33.17
0
211
Generality
176
0.40
0.42
0
1
Patent Experience
176
5.98
8.75
0
34
Patent Activity
176
1.00
0.00
1
1
Size
176
1.08
0.76
0
2
Licensee
176
0.00
0.00
0
0
TABLE 2 Correlation Matrix 1.
2.
3.
4.
5.
6.
1.
Distance of Technological Exploration
2.
Assimilation Capacity
3.
Monitoring Ability
4.
Patent Stock
-0.05
-0.00
5.
Generality
0.06
0.02 -0.01 0.02
6.
Patent Experience
-0.13* -0.11* 0.18* 0.21* 0.06
7.
Patent Activity
-0.21* -0.32*
8.
Size
9.
Co-patenting Licensee
10.
7.
8.
9.
1 0.33*
1
-0.28* -0.22*
1 0.01
1 1 1
0.22* 0.06 -0.10 0.27*
1
0.11*
0.09
0.01 0.16* -0.02 0.27* 0.15*
1
0.03
-0.02
0.02 0.28* 0.07 0.11* 0.13*
0.17*
0.03
1
0.01 -0.22* 0.07 0.22* -0.05 -0.41* -0.29* -0.08
82
TABLE 3 OLS regressions: Determinants of the Distance of Technological Exploration
Variable Assimilation Capacity Monitoring Ability Co-patenting Patent Stock Generality Patent Experience Patent Activity Medium Firm Large Firm Co-patenting
Model I Treatment Model 0.283 [0.082] 0.002 [0.003] -0.050 [0.123] 0.000 [0.000] 0.120 [0.056] -0.001 [0.003] -0.122 [0.059] -0.015 [0.058] 0.108 [0.041] -0.050 [0.123]
Model II Control Model
Yes
0.742 *** [0.112] 176 0.33
0.806 *** [0.107] 176 0.27
1.036 *** [0.071] 352 0.14
0.894 *** 0.834 *** 0.905 *** [0.097] [0.094] [0.097] 352 352 352 0.25 0.24 0.26
***
Licensee × Assimilation Capacity Licensee × Monitoring Ability Co-patenting × Assimilation Capacity Co-patenting × Monitoring Ability
Constant N r2 *:p0) with (1-f1(0)) to ensure that the probabilities of the outcomes sum to unity. The hurdle model is reduced to the standard count model in cases where f1(.)=f2(.). We follow McDowell (2003) and use a complementary log-logistic specification for the first part and a truncated Poisson specification to model the positive outcomes in the second part. In the two regression specifications we employ the same variables as regressors. The Huber-White sandwich estimation technique is used to correct standard errors for possible heteroskedasticity. To test hypotheses 3 and 4, we develop a model specification that explores the type of licensee, which, subsequent to the license agreement, tends to invent narrowly in the licensed technologies. This is done by examining only the inventive licensees and regressing the explanatory variables used in the previous regressions against the number of inventions in the licensed technologies relative to the total number of inventions introduced by these licensees. This measure is truncated at zero for those observations that did not invent in the licensed technologies, and at 1 for those subjects that only invent in the licensed technologies. We use the two-limit Tobit regression specification to model this, as prescribed by Tobin (1958).
Results
42B
Descriptive Statistics Descriptive statistics and correlations between variables are presented in Table 1. The coefficients reveal that the analysis is unlikely to suffer from multicollinearity, with only few correlations approaching high levels. This is confirmed by a variance inflation factor analysis. Table 1 also shows that approximately 59% of firms engaged in some patenting activity within the five years following the license agreement, and about 24% of firms patented in the IPC codes of the patents included in the reference license agreement. ***INSERT TABLE 1 ABOUT HERE*** Table 2 presents the distribution of observations across the three categorical variables included in our model. The sample is split between licensees and non-licensees. In terms of
100
firm size, Table 2 shows that the sample contains a majority of small firms. However, taking the skewness of the population size distribution into consideration, it is evident that licensingin is more likely to be a large firm strategy. Although the size variable is not used in the matching procedure, there is an overwhelming similarity between the licensee and nonlicensee samples in terms of firm size distribution. Table 2 also shows that there are geographical differences between licensees and non-licensees. In our sample the relative number of firms categorized as Non-North American is higher for non-licensees. This may be a by-product of licensing being more widely used as an integrated activity in the invention strategies of North American firms, making it more difficult to find a North-American match that is also a non-licensee. This is also consistent with evidence on the markets for technology which suggests that American firms lead in terms of technology exchange, compared to the rest of the world (e.g. OECD, 2007). 44 Also, and not surprisingly, we observe an equal number F
F
of licensees and non-licensees in the six different technological classes attributable to the technological class used in the exact matching procedure. However, we can also see that a substantial number of observations patent primarily in the “Drugs and Medical” and the “Chemicals” technological classes. ***INSERT TABLE 2 ABOUT HERE*** The propensity score matching procedure matches not on regressors, but on estimated likelihoods given regressors. We cannot be certain, therefore, that the matching variables are distributed equally across the licensee and non-licensee samples. We ran a probit regression using the licensee dummy as a dependent variable, and propensity score matching variables and the technological class dummies used in the exact matching procedure as explanatory variables. We found no significance, and a pseudo R-square only marginally above 1. Also, our Chi-square statistic suggests that it is likely that all the estimated parameters are equal to zero, indicating that the matching procedure was successful in finding matching firms based on these input variables. These results also suggest that any significance found in the hurdle models with respect to the matching variables can be attributed to within-group correlations rather than between-group differences. 44
According to the 2007 OECD Science, Technology and Industry Outlook (OECD, 2007: 12) “Royalty receipts from outward licensing have been estimated at 6.0%, 5.7% and 3.1% of total R&D spending for US, Japanese and European firms, respectively, suggesting that technology licensing markets are better developed in the United States than elsewhere”.
101
Licensing and the Invention Performance of the Licensee The results from the two hurdle models are reported in Table 3. The columns to the left refer to the regression for inventions in general, where the dependent variable is a dummy for whether the firm introduced any inventions at all (complementary log-logistic regression) and the total number of patents after the firm invented at least one (truncated Poisson). The columns on the right show the regression results for inventions in the licensed technology with a similar set-up, based on inventions in the same technological class as that of the licensed technology. ***INSERT TABLE 3 ABOUT HERE*** Hypotheses 1a and 1b are strongly endorsed by the results. The licensee variable is highly significant in explaining both the likelihood of inventing and the number of inventions introduced after the signing of the reference license agreement. Our analysis also provides support for hypothesis 2a, by suggesting that licensing-in is affiliated to a higher chance of introducing new inventions in the licensed technology. We find no support for hypothesis 2b. Once non-licensees have overcome the hurdle of inventing in the technology acquired by the licensees, they seem to be equally well equipped to invent extensively in this particular technology. These regressions provide evidence that licensing agreements allow licensees to enjoy spill-over effects from the licensed technology and the licensor in the form of knowledge flow and learning, facilitating patenting in the licensed technology as we as in other technology classes than those specified in the license agreement. Among the other results presented in Table 3, we highlight the following as being integral to our analysis. We find that search depth and search scope increase the firm’s likelihood of introducing new inventions in general, but that search scope, at best, has only a weak effect on raising the probability of inventing in the licensed technology. However, the results do suggest that firms employing a search depth strategy are hampered in terms of the number of inventions filed in the licensed technological class, while firms employing a search scope strategy engage in greater invention activity in general, and are more active in invention in the technology specified in the license agreement. Technological specialization seems to increase the likelihood of inventing generally, as well as becoming successful in producing at
102
least one invention in the IPC codes covered in the reference license agreement. The firm’s patent stock prior to the license agreement increases the extent of invention regardless of whether it is invention generally or invention in the licensed technology. Furthermore, the regression results suggest that technological experience hampers the extent of invention in general. Specialization, Familiarity and Narrow Invention Strategy Table 4 presents the results of the Two-Limit Tobit regression investigating whether particular types of licensees tend to invent relatively more often in the IPC codes of the licensed technology. Table 4 reveals that 98 of the 133 licensees did invent in the five years after the license agreement. It also shows that 38 of the licensees did not invent in the licensed technology at all, and that 13 licensees invented only in that technological class. ***INSERT TABLE 4 ABOUT HERE*** Hypotheses 3 and 4 suggest that technologically specialized licensees, and licensees familiar with the licensed technologies, tend to be more likely to exploit the narrowly specified technologies covered in the license agreement and to search the invention landscape in the area of the licensed technologies. We provide support for both these hypotheses in finding significant positive estimates for both specialization and familiarity. The coefficients of our two dummies for technological classes – namely the “Drugs and Medical” and the “Electric and Electronics” – are also significant. This suggests that inventions in these technology classes are relatively less used for general technological advancement than technologies licensed in the chemical technology class (benchmark category).
Discussion and Conclusion
43B
This paper was motivated in part by the recent trends in firms’ approaches to increase their invention rate. Firms are adopting more open models of innovation, thereby taking advantage of the opportunities that the markets for technology may present, to foster and unlock the potential of firms’ internal R&D efforts. In particular, firms are embracing inward technology licensing as a means to realize their invention objectives. Indeed, licensing has
103
become one of the most visible mechanisms of knowledge transfer among firms and one of the more accessible means of tapping into the knowledge bases of other firms. Licensing-in strategies are traditionally considered to be driven by the licensee’s desire to get rapid access to proven/mature technology. However, such practices are increasingly being recognized as strategic and for the pursuit of other goals, such as technological learning, which, in turn, leads to the development of new technological capabilities. To the best of our knowledge, very few studies have attempted to address whether and to what extent licensingin generates new technological advances. This is despite the increasing empirical evidence that licensees gain competitive advantage, and even achieve technological leadership, by leveraging and exploiting the learning opportunities that licensing practices enable. This paper fills a gap in the existing knowledge by providing strong quantitative evidence of licensing functioning as a catalyst for developing and introducing new inventions. The results suggest not only that licensing increases the likelihood of introducing new inventions, but also that it increases the number of inventions that the licensee is able to introduce. Thus, licensing promotes invention activities in general, enabling the firm to enter technology fields beyond those included in the license agreement. However, we have also shown that licensing increases the likelihood of introducing a new patent in the technological class embedded in the license agreement. In addition, our results suggest that there are particular antecedents that drive the licensee to focus on the technological class specified in the license agreement. Technological specialization and familiarity with the licensed technology promote focused invention activities. Licensees with high levels of specialization and/or technology familiarity tend to invent primarily in the regions of the technological landscape of the licensed technology. Accordingly, diversified firms and licensees relatively unfamiliar with the licensed technologies tend to explore the invention landscape more widely than the license agreement technology focus. This indicates that there is a need to focus on learning patterns and their related invention and innovation effects when deciding about licensing-in. This observation is highly relevant to a better understanding of the path dimensions of dynamic capabilities (see e.g. Teece et al., 1997), an aspect that has received far less attention than, for example, resource constellations, and the managerial and organizational processes involved in changes to the firm’s resource base over time. Our empirical observations indicate that licensing-in not only
104
is a means for deepening already existing knowledge sets, but also induce new search patterns, which ultimately may lead to a broadening of the firm’s patenting activities and new combinations of existing and new knowledge. Hence, technology licensing could reduce the path dependency of established firms and trigger the pursuit of new invention endeavors. This is in line with the suggestions in Granstrand (1998), which stress that technological opportunities can be created through the combination and cross-fertilization of old and new technologies, and that combinatorial possibilities grow exponentially when new technologies are added to the firm’s existing technological base. Thus, the introduction of new technologies via licensing-in constitutes a potentially important vehicle for generating new opportunities for invention by inducing new search patterns and technology combinations. This may lead to a widening of the technological trajectory that influences the firm’s strategic maneuvering space. The different search patterns triggered by different licensing approaches can also be important for understanding the flexibility of a firm’s resource base, an aspect seen as important for firms to remain competitive over time (see e.g. Wernerfelt (1984) and Sanchez (1995)). Thus the inherent flexibility of a resource base, which, in technology-based businesses, may play a significant role in challenging relentlessly dynamic and entrepreneurial markets, may be affected by licensing-in activities. These activities may enable firms to open up the box of learning opportunities for diverse or more consistent uses of their knowledge bases, depending on the technological characteristics of recipient firms and the search behavior adopted by licensees. Limitations This paper suffers from a number of limitations. Technology licensing-in does not necessarily need to be an integrated part of a invention strategy, but may be pursued for other reasons. Among other motives, firms may license a given technology as a part of a broader R&D partnership or cross-licensing agreement, or simply because they are forced to do so as they have previously infringed a licensor’s property rights (settlement agreement). We are aware of these motives, but since we are concerned in this paper only with technology exchange agreements - which imply a one-way technology/intellectual property rights transfer, from licensor to licensee - we included only those transactions that were originally filed by the parties as (pure) licensing or assignment agreements. Thus, we excluded all other transactions that refer to R&D collaboration, cross-licensing, settlement agreements, and also technology
105
purchases
and
merger
plans 45 F
F
that
were
incorrectly
listed
under the
heading
“Transaction/Patent Licenses” in the original dataset. Nevertheless, we are not able to separate agreements that were signed as a way to pre-empt a violation of a licensor’s intellectual property rights, from those that are an integral part of a licensee’s invention activity. Given this, the positive correlation between licensing-in and invention performance may need to be re-evaluated as it points to an unobserved omitted variable that may influence invention performance and promote the decision to license a given technology. However, we consider this case to be unlikely in context of our dataset. We believe that firms engaging in license agreements with the sole purpose of avoiding legal litigation would previously have developed the technology they considered to be in danger of violating the intellectual property rights of the licensor. Hence, we would expect the licensee, in this case, to apply for a patent immediately after signing the license agreement. We studied the time it took for a licensee to apply for a new patent after having licensed the technology and found it to be on average 109 months after signing the license agreement. This long time period suggests that at least the majority of our licensees do not license as a way to avoid legal litigation. The invention performance of the firm is a time-dependent issue. The cross-sectional nature of our sample limits the scope of the analysis. It may exclude some relevant insights on the dynamics of the licensing-in decision and internal R&D efforts over time. We look only at the effects of a specific decision to license-in on the future inventive outcome of the firm. However, decisions should be framed within the overall strategy of the firm, which is developed and modified across the years. Thus, our analysis is limited in the sense that we assume that the invention strategy of the firm is fixed across time. However, our licensees’ and non-licensees’ approaches to technological change may diverge at the point of licensing beyond signing a license agreement. We have no immediate way of controlling for this potential source of bias in the analysis. The sample of license agreements under investigation all provide information on the technology embedded in the publicly available contract. However, in some circumstances, secrecy may be of major importance to the licensee. Revealing the contents of license agreement may signal the technological strategy of the licensee, thereby providing competitors with information that may be disadvantageous in the invention race. Disregarding license 45
Original documents downloaded from the SEC website generally indicate type of transaction – license, settlement agreement, or the like – in headings. We were careful to check for any information suggesting that the contract referred to another type of agreement than a technology/patent license.
106
agreements whose contents are not disclosed may lead to a bias in our estimates. However, we contend that this potential bias would go against our primary hypotheses. Consequently, our results should be considered conservative estimates of the relationship between invention performance and taking the decision to license-in a technology. In addition, it is possible that our matching procedure suffers from unobserved heterogeneity and hence may produce bias due to omitted variables. Our results rest on the specification of the matching procedure. The propensity score-matching procedure has been criticized for bias based on the number of variables used in the matching procedure. We did find some indications that our licensees and non-licensees matched on other dimensions when we considered the control and explanatory variables suggesting the matching process to be robust.
Future Research The results of this study indicate clearly that licensing puts the licensee in a favorable position compared to a matched non-licensee, in terms of ex-post licensing inventive performance. The license agreement provides potential learning which extends beyond the specifications in the agreement, for instance, a patent application, and which drives the inventive performance of the licensee. We hypothesize that signing a license agreement also opens up other channels for information flow between licensee and licensor, creating a mutually beneficial collaborative scenario. According to the literature, this applies mostly to patent licensing for a very obvious reason. As patents encompass knowledge that is formally codified and legally enforced, they make licensors more likely to provide the more tacit part of technological knowledge (know-how) which is relevant to understanding and fruitful exploitation of the licensed technology. Follow-up research might investigate whether these channels also help firms to introduce inventions more quickly by accelerating the speed at which they can progress in the innovation landscape, identify opportunities, and overcome innovation barriers. Future work might also look more deeply at the dynamics of crossorganizational collaboration induced by licensing and whether it contributes extensively to our understanding of the role played by formal agreements compared to informal channels of information and knowledge flow. By disentangling the relative importance of these two, we
107
would achieve a scholarly grasp of the true relationship between signing a license agreement and invention performance. We would then understand better whether license agreements are direct drivers of technological change or have an indirect effect driven by the formation of informal network ties that promote knowledge sharing and thereby increase the number of potential new combinations of existing knowledge bodies. Our study has also investigated the antecedents to a firm’s engaging in a licensing agreement in terms of benefiting from the technology it licenses-in. However, given that formal and informal linkages in the formation of channels of information and knowledge between licensee and licensor are extremely important, we propose that future research should investigate the nature of the relationship between licensee and licensor. This is important in two dimensions. First, the overlap between competences and capabilities of licensees and licensors defines the scope of potential knowledge combinations and resulting invention opportunities. Second, following the arguments of Li, Eden, Hitt and Ireland (2008), partner selection is of great importance in the formation of R&D alliances. Trust between partners, and protection of intellectual property rights play a major role in defining the boundaries of the information flow between the parties involved. Similarly, the knowledge and information flows between licensee and licensor may suffer in the presence of mistrust, and restrict inventiveness based on the license agreement. A related stream of research could focus on improving our understanding of the role of certain contractual clauses –such as grant-back – in the partner selection. Finally, we propose that future work should investigate whether firms experienced in cross-organizational collaboration are also better equipped to draw advantages from licensing activities. Experience may help firms to select the best partners and also develop their abilities to manage partnerships, maximizing the benefits in terms of knowledge and information flows. Indeed, the facility to define the boundaries of a collaborative partnership is an acquired capability that involves defining the intellectual property rights of traded and produced technological assets as well as setting the scope of the license agreement in terms of knowledge sharing. The right settings may facilitate a more relaxed and more fruitful contractual partnership by promoting trust and mutual understanding. Repeated technology licensing partnerships, for instance, may provide suggestions about the aspects that facilitate and promote a mutually beneficial contractual relationship, leading to a deeper understanding of the dynamics of technology licensing.
108
References
44B
Ahuja, G., & Lampert, C. M. 2001. Entrepreneurship in the Large Corporation: A longitudinal Study of How Established Firms Create Breakthrough Inventions. Strategic Management Journal, 22(6-7): 521-543. Anand, B. N., & Khanna, T. 2000. The Structure of Licensing Constracts. The Journal of Industrial Economics, 48(1): 103-135. Arora, A. 1996. Contracting for tacit knowledge: the provision of technical services in technology licensing contracts. Journal of Development Economics, 50(2): 233-256. Arora, A., Fosfuri, A., & Gambadella, A. 2001. Markets for Technology: The Economics of Innovation and Corporate Strategy. Cambridge, Massachusetts: The MIT Press. Arora, A., & Gambadella, A. 1990. Complementarity and External Linkages: the Strategies of the Large Firms in Biotechenology. Journal of Industrial Economics, 38(4): 361-379. Arrow, K. 1962. Economic Welfare and the Allocation of Resources for Inventions, The Rate and Direction of Inventive Activity: Economic and Social Factors: 609-625. Princeton, N. J.: National Bureau of Economic Research. Atuahene-Gima, K. 1993. Determinants of Inward Technology Licensing Intentions: An Empirical Analysis of Australian Engineering Firms. Journal of Product Innovation Management, 10: 230-240. Bessy, C., Brousseau, E., & Saussier, S. 2002. The Diversity of Technology Licensing Agreements. Working Paper: Université de Paris 1. Cassiman, B., & Veugelers, R. 2006. In Search of Complementarity in Innovation Strategy: Internal R&D and External Knowledge Acquisition. Managament Science, 51(1): 68-82. Chatterji, D. 1996. Accessing External Sources of Technology. Research Technology Management, 39(2): 569-596. Chesbrough, H. 2003. Open innovation. Cambridge, Massachusetts: Harvard University Press. Chesbrough, H. 2006. Open Business Models: How to thrive in the New Innovation Landscape. Boston: Harvard Business School Press. Cohen, W. M., & Levinthal, D. A. 1989. Innovation and Learning: The Two Faces of R&D. The Economic Journal, 99(397): 569-596. Cohen, W. M., & Levinthal, D. A. 1990. Absorptive Capacity: A New Perspective of Learning and Innovation. Administrative Science Quarterly, 35: 128-152.
109
Cyert, R. M., & March, J. G. 1963. A Behavioral Theory of the Firm. Englewood Cliffs, NJ: Prentice-Hall. Daneels. 2002. The Dynamics of Product Innovation and Firm competences. Strategic Management Journal, 23(12): 1095-1121. David, P. A., & Olsen, T. E. 1992. Technology Adoption, learning Spillovers, and the optimal Duration of Patent-based Monopolies. International Journal of Industrial Organization, 10(4): 517-543. Fleming, L. 2001. Recombinant uncertainty in technological search. Management Science, 47(1): 117-132. Fosfuri, A. 2006. The Licensing Dilemma: Understanding the Determinants of the Rate of Technology Licensing. Strategic Management Journal, 27(12): 1141-1158. Gambadella, A. 1992. Competitive Advantages from In-house Basic Research: The US pharmaceutical industry in the 1980s. Research Policy, 21(5): 391-407. Granstrand, O. 1998. Towards a theory of the technology-based firm. Research Policy, 27(5): 465-489. Granstrand, O., Bohlin, E., Oscarsson, C., & Sjöberg, N. 1992. External Technology Acquisition in Large Multi-technology Corporations. R&D Management, 22(2): 111-133. Hall, B., Jaffe, A., & Trajtenberg, M. 2001. The NBER Patent Citations Data File: Lessons, Insights, and Methodological Tools. NBER Working Paper, 8498. Hall, B. H., & Trajtenberg, M. 2004. Uncovering GPTS with Patent Data: National Bureau of Economic Research, Inc. Heckman, J. J. 1979. Sample Selection Bias as a Specification Error. Econometrica, 47(1): 153161. Heckman, J. J., Ichimura, H., & Todd, P. 1997. Matching as an Econometric Evaluation Estimator: Evidence from Evaluating a Job Training Program. Review of Economic Studies, 64(4): 605-654. Henderson, R., & Cockburn, I. 1994. Measuring Competence? Exploring Firm Effects in Pharmaceutical Research. Strategic Management Journal, 15(s1): 32-59. Johnson, D. 2002. Learing-by-Licensing: R&D and Technology Licensing in Brazilian Invention. Economics of Innovation and New Technology, 11(3): 163-177. Katila, R., & Ahuja, G. 2002. Something Old, Something New: A Longitudinal Study of Search Behaviour and New Product Introduction. Academy of Management Journal, 45(8): 1183-1194. Katz, R., & Allen, T. J. 1982. Investigating the Not Invented Here (NIH) syndrome: A look at the performance, tenure, and communication patterns of 50 R&D Project Groups. R&D Management, 12(1): 7-19.
110
Kim, Y. J., & Vonortas, N. S. 2006. Technology Licensing partners. Journal of Economics and Business, 58(4): 273-289. Lane, P. J., & Lubatkin, M. 1998. Relative Absorptive Capacity and Interorganizational Learning. Strategic Management Journal, 19(8): 461-477. Lanjouw, J. O., & Schankerman, M. 1999. The Quality of Ideas: Measuring Innovation with Multiple Indicators, Vol. 7345. Working Paper Series: NBER. Laursen, K., Leone, M. I., & Torrisi, S. 2008. Technological Search and Diversification Through Licensing: New Insights from the Licensee’s point of view, Markets for Technology: Strategy and Industry Evolution Conference. Madrid, Spain. Levinthal, D. A., & March, J. G. 1993. The Myopia of Learning. Strategic Management Journal, 14(s2): 95-112. Li, D., Eden, L., Hitt, M. A., & Ireland, E. D. 2008. Friends, Acquaintances Or Strangers? Partner Selection in R&D Alliances. Academy of Management Journal, 51(2): 315-334. Lin, B.-W. 2003. Technology Transfer as Technological Learning: a Source of Competitive Advantage for firms with limited R&D sources. R&D Management, 33(3): 327-341. Lowe, J., & Taylor, P. 1998. R&D and Technology Purchase through License Agreements: Complementary Strategies and Complementary Assets. R&D Management, 28(4): 263-278. Lubatkin, M. 1983. Mergers and the performance of the acquiring firm. Academy of Management Review, 8(2): 218-225. March, J. G. 1991. Exploration and Exploitation in Organizational Learning. Organization Science, 2(1): 71-87. March, J. G., & Simon, H. A. 1958. Organizations. New York: Wiley-Blackwell. Mathews, J. A., & Cho, D.-S. 1999. Combinative capabilities and organizational learning in latecomer firms: The case of the Korean semiconductor industry. Journal of World Business, 34(2): 139-156. McDowell, A. 2003. From the Help Desk: Hurdle Models. The Stata Journal, 3(2): 178-184. Miller, D. 1993. The architecture of simplicity. Academy of Management Review, 18(1): 116138. Miller, D., & Chen, M.-J. 1996. The Simplicity of Competitive Repertoires: An Empirical Analysis. Strategic Management Journal, 17(6): 419-439. Motohashi, K. 2006. Licensing or Not Licensing?: Empirical Analysis on Strategic Use of Patent in Japanese Firms: Research Institute of Economy, Trade and Industry (RIETI). Mowery, D. C., Oxley, J. E., & Silverman, B. S. 1998. Technological overlap and interfirm cooperation: implications for the resource-based view of the firm. Research Policy, 27(5): 507523.
111
Narula, R. 2001. Choosing Between Internal and non-internal R&D activigties: Some Technological and Economic Factors. Technology Analysis and Strategic Management, 13(3): 365-387. OECD. 2007. Innovation and Performance in the Global Economy, OECD Science, Technology and Industry Scoreboard, Vol. October 2007. Paris: OECD. Pitkethly, R. 2001. Intellectual Property Strategy in Japanese and UK companies: Patent licensing decisions and learning opportunities. Research Policy, 30(3): 425-442. Pohlmeier, W., & Ulrich, V. 1995. An Econometric Model of the Two-Part Decision-making Process in the Demand for Health Care. Journal of Human Resources, 30(2): 339-361. Prahalad, C. K., & Hamel, G. 1990. The Core Competence of the Corporation. Havard Business Review, 68(3): 79-91. Rivera-Batiz, L. A., & Romer, P. M. 1991. Economic integration and endogenous growth. The Quarterly Journal of Economics, 106(2): 531-555. Roberts, E. B., & Berry, C. A. 1985. Entering New Businesses: Selecting Strategies for Success. MIT Sloan Management Review, 26(Spring). Rosenbaum, P. R., & Rubin, D. B. 1983. The Central Role of Propensity Score in Observational Studies for Causal Effects. Biometrika, 70(1): 41-55. Rothaermel, F. T., & Deeds, D. L. 2004. Exploration and exploitation alliances in biotechnology: A system of new product development. Strategic Management Journal, 25(3): 201-221. Sampat, B. N., Mowery, D. C., & Ziedonis, A. A. 2003. Changes in University Patent Quality After the Bayh-Dole Act: A Re-examination. Internation Journal of Industrial Organization, 21(9): 1371-1390. Sanchez, R. 1995. Strategic Fexibility in Product Competition. Strategic Management Journal, 16(Summer): 135-159. Santaló, J., & Becerra, M. 2006. The dominance of diversified versus specialized firms across industries Journal of Business Research, 59(3): 335-340. Teece, D. 2008. Explaining Dynamic Capabilities: The Nature and Microfoundations of (sustainable) Enterprise Performance. Strategic Management Journal, 28(13): 1319-1350. Teece, D., Pisano, G., & Shuen, A. 1997. Dynamic Capabilities and Strategic Management. Strategic Management Journal, 18(7): 509-533. Tobin, J. 1958. Estimation of Relationships for Limited Dependent Variables. Econometrica, 26(1): 24-36. Tsai, K.-H., & Wang, J.-C. 2007. Inward Technology Licensing and Firm Performance: a longitudinal study. R&D Management, 37(2): 151-160.
112
Veugelers, R., & Cassiman, B. 1999. Make and buy in innovation strategies: evidence from Belgian manufacturing firms. Research Policy, 28(1): 63-80. Villalonga, B. 2004. Diversification Discount or Premium? New Evidence from the Business Information Tracking Series. Journal of Finance, 59(2): 479-506. Wernerfelt, B. 1984. A resource-based view of the firm. Strategic Management Journal, 14: 412. Zahra, S. A., & George, G. 2002. Absorptive capacity: A review, reconceptualization, and extension. Academy of Management Review, 27(2): 185-203. Ziedonis, A. A. 2007. Real Options in Technology Licensing. Managament Science, 53(10): 1618-1633. Zollo, M., & Winter, S. G. 2002. Deliberate learning and the evolution of dynamic capabilities. Organization Science, 13(3): 339-351.
113
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.
Mean
s.d.
1.
2.
3.
4.
Patent Activity 0.59 0.49 Nunmber of Patents 10.65 52.77 0.17 Patent Activity in Licensed technology 0.24 0.43 0.47 0.27 Number of Patents in Licensed Technology 2.92 16.70 0.15 0.85 0.31 b Share of Patents in licensed Technology 0.22 0.35 . 0.04 0.78 0.21 Licensee 0.50 0.50 0.30 0.17 0.49 0.17 Familiarity 0.16 0.31 0.05 -0.01 0.30 0.07 Technological Specialization 0.54 0.39 0.07 -0.11 0.08 -0.05 Patent Stock 1.20 1.12 0.17 0.54 0.19 0.39 Average Number of Cites 9.69 14.96 0.09 0.05 0.01 0.02 Average Time Between Patents 21.58 34.04 -0.15 -0.09 -0.23 -0.09 Technological Diversity 3.05 15.00 0.09 0.74 0.14 0.40 Technological Collaborator 0.06 0.25 0.06 0.37 0.03 0.31 Search D epth 0.42 1.46 0.19 0.14 0.29 0.09 Search Scope 0.32 0.42 0.24 0.14 0.20 0.13 Technological Experience 59.06 75.71 0.09 0.21 0.04 0.12 Patent Stock Generality 0.55 0.39 -0.04 0.12 -0.04 0.07 Average Number of Claims 0.80 2.43 -0.07 -0.04 -0.02 -0.04 Medium Sized Firm 0.28 0.45 0.06 -0.01 0.10 -0.01 Large Sized Firm 0.16 0.37 0.15 0.24 -0.05 0.18 North American Firm 0.86 0.35 0.07 0.06 0.15 0.07 a The data have 266 observation. Coefficients greater in magnitude than 0.12 are significant at the 0.05 level b N umbers associated with this variable are only based on 149 observations
Variable
Means, Standard Deviations, and Correlationsa
TABLE 1
0.42 0.49 0.31 -0.10 -0.06 -0.17 -0.04 -0.09 0.15 0.02 -0.11 -0.16 0.03 0.03 -0.20 0.20
5.
0.12 -0.29 0.07 0.01 -0.01 0.08 0.11 0.18 0.02 0.17 -0.11 -0.11 0.07 -0.04 0.24
6.
0.16 -0.10 0.07 -0.08 -0.06 0.04 0.12 0.05 -0.16 -0.06 0.18 -0.04 -0.11 0.03
7.
-0.34 0.11 -0.03 -0.14 -0.09 0.08 0.03 -0.34 -0.20 0.21 -0.06 -0.09 -0.10
8.
0.11 -0.10 0.65 0.33 0.20 0.15 0.70 0.57 -0.09 0.10 0.26 0.01
9.
0.08 -0.02 -0.01 0.04 -0.03 0.06 0.31 0.34 0.08 -0.04 0.12
10.
-0.07 0.03 -0.06 -0.21 0.42 0.02 -0.12 0.01 0.09 0.02
11.
0.33 0.13 0.06 0.39 0.18 -0.04 -0.02 0.23 -0.01
12.
0.11 0.05 0.25 0.12 -0.07 0.04 0.10 -0.24
13.
0.24 0.04 0.13 -0.04 0.10 -0.00 0.07
14.
0.02 0.02 0.17 0.12 0.15 -0.04
15.
0.39 -0.19 0.07 0.27 -0.00
16.
0.07 0.06 0.05 0.03
17.
19.
-0.27 -0.01
114
0.04 -0.00 0.05
18.
-0.09
20.
TABLE 2 Firm Size, Country of Residence and Primary Technological Class of Nonlicensees and Licensees
Variable Firm Size Small Medium Large Geography North American Not north American Primary Technological Class Chemicals Computers and Communications Drugs and Medical Electrical and Eletronics Mechanical Others Tota l
Licensees
Non-Licensee
Total
73 41 19
77 33 23
150 74 42
103 30
125 8
228 38
34 13 48 12 6 20
34 13 48 12 6 20
68 26 96 24 12 40
133
133
266
115
TABLE 3 Determinants of general and targeted invention performance, results of hurdle modelsa General Invention Regression Complementary Log-logistic T runcated Poisson Explanatory Varibales Licensee Familiarity T echnological Specialization Matching Variables Patent Stock Average Number of Cites Average T ime Between Patent s T echnological Diversity T echnological Collaborator Cont rol Variables Search Depth Search Scope T echnological Experience Patent Stock Generality Average Number of Claims Firm Size Large Medium Small North American Firm Primary T echnology Dummies Computers and Communications Drugs and Medical Electrical and Eletronics Mechanical Others Chemicals Constant Number of Observations Log-Likelihood Wald Chi-Square Pseudo R-Square
1.408 *** [0.252] 0.194 [0.323] 1.632 *** [0.329]
Invention in Licensed T echnology Regression Complement ary Log-logistic T runcated Poisson
1.136 *** [0.282] -0.124 [0.446] -0.441 [0.379]
3.625 *** [0.865] 2.174 *** [0.459] 2.000 *** [0.564]
-0.329 [0.641] -0.462 [0.672] 1.670 ** [0.792]
[0.254] [0.008] ** [0.004] *** [0.008] [0.526]
0.610 *** [0.145] 0.011 * [0.006] -0.004 [0.008] 0.002 [0.004] 0.293 [0.334]
0.818 ** [0.324] 0.001 [0.011] -0.021 * [0.011] -0.005 [0.009] -1.629 ** [0.719]
0.710 ** [0.298] -0.010 [0.010] -0.023 [0.019] -0.010 [0.007] 2.235 *** [0.651]
0.694 *** [0.256] 0.518 ** [0.245] 0.002 [0.003] -0.515 [0.355] -0.151 *** [0.053]
-0.094 [0.064] 0.942 *** [0.223] -0.006 *** [0.002] -0.513 [0.318] -0.040 [0.061]
0.111 0.661 * -0.003 -0.962 -0.046
[0.128] [0.391] [0.004] [0.634] [0.069]
-0.246 *** [0.088] 1.747 *** [0.514] 0.000 [0.003] -1.857 ** [0.775] 0.115 [0.094]
0.855 *** [0.295] 0.205 [0.240] Benchmark 0.194 [0.327]
0.909 *** [0.305] 0.680 ** [0.288] Benchmark 0.693 [0.630]
-0.133 [0.525] 0.332 [0.350] Benchmark 0.418 [0.929]
-0.030 [0.779] 0.494 [0.604] Benchmark 12.649 .b
-0.097 [0.349] 0.114 [0.268] 0.052 [0.436] 0.628 [0.514] 0.072 [0.331] Benchmark
0.107 [0.509] 0.207 [0.281] 0.344 [0.279] -0.064 [0.525] -0.013 [0.398] Benchmark
-0.691 [0.664] 0.218 [0.414] 0.628 [0.623] 0.749 [0.816] 1.208 ** [0.574] Benchmark
-0.115 [1.342] 0.452 [0.843] 0.189 [0.643] -0.637 [0.847] -0.209 [0.809] Benchmark
0.517 0.022 -0.011 -0.029 0.106
**
***
-0.197 [0.632] -2.584 *** [0.527] 266 157 -129.074 -950.969 76.452 *** 7995.502 *** 0.808
-6.470 *** [1.145] -12.459 *** [0.963] 266 64 -68.060 -325.627 84.804 *** 1560.044 *** 0.705
*: p
