STUDY ON EVALUATING THE KNOWLEDGE ECONOMY WHAT ARE PATENTS ACTUALLY WORTH?

THE VALUE OF PATENTS FOR TODAY’S ECONOMY AND SOCIETY
Tender n° MARKT/2004/09/E

Final Report for Lot 1

May 9, 2005

1

Authors and acknowledgements
This Report is the output of a research project (ETD/2004/IM/E3/77) conducted for the European Commission, Directorate-General for Internal Market. The project was administered by the CERM Foundation (Siena, Italy) with the following partner institutions to develop the project: LEM, Sant’Anna School of Advanced Studies, ZEW and Inno-tec, Ludwig-Maximilians-Universität, Universitat Pompeu Fabra, Eindhoven University of Technology, SPRU, University of Sussex, Université Lyon2, INSEAD, University of Pécs. The research leading to this Report was conducted by the following people: Marco Ceccagnoli, Alfonso Gambardella, Aldo Geuna, Walter Garcia Fontes, Paola Giuri, Dietmar Harhoff, George Licht, Myriam Mariani and Bart Verspagen. Alfonso Gambardella was the Project Coordinator. This Report was prepared by Marco Ceccagnoli, Alfonso Gambardella, Paola Giuri, Georg Licht and Myriam Mariani with the precious assistance of Serena Giovannoni. The research teams gratefully acknowledge support by Claudia Colla at the European Commission.

2

TABLE OF CONTENTS

1. 2. 2.1 2.2

ORGANIZATION OF THE REPORT ...................................................................... 10 LITERATURE SURVEY.......................................................................................... 11 Literature search and classification ......................................................................... 11 Analysis of the Literature by Theme........................................................................ 12

Theme A1. Economic Value of Patents ..................................................................................... 12 Theme A2: Utilization of Patents and Potential for Enhancing their Value ................................. 18 Theme B1: New Firm Creation and Employment ...................................................................... 20 Theme B2: The impact of patent protection on knowledge spillovers and productivity............... 21

3. 3.1 3.2

DESCRIPTIVE STATISTICS II: PATVAL DATASET ............................................. 24 The Patval Dataset.................................................................................................. 24 The Value and Social Costs of Patents................................................................... 25

The value of patents.................................................................................................................. 25 The cost of patents.................................................................................................................... 34

3.3

The Economic Use of Patents................................................................................. 38

The use of patents .................................................................................................................... 38 Patent value and Patent use ..................................................................................................... 43 The importance of different motives for patenting...................................................................... 44

3.4 3.5

Creation of new businesses from the patented invention........................................ 46 Collaboration, Spillovers and the Sources of Knowledge in the Invention Process. 49

Collaboration amongst inventors in the innovation process ....................................................... 49 Formal and informal collaborations in research ......................................................................... 54 The role of geographical proximity for collaboration .................................................................. 56 The Sources of Knowledge ....................................................................................................... 59

4. 4.1 4.2 4.3

THE EPO PATENT EXPLOSION – SOME DESCRIPTIVE STATISTICS .............. 63 The Data Source ..................................................................................................... 63 Overall trends in EPO applications.......................................................................... 64 Contributions by to Patent Explosion by Technology Areas and by Sector............. 70 3

...........2.......................................................................................................................e............... The Monetary Value of Patents and the Use of Multiple Indicators ........... An empirical analysis of the effectiveness of different means of knowledge spillovers from patents .......................2......................1... 79 Study A................... fields in the Access table) on each publication of the reference list..............................4............................................................................................................................................................................................ other patent uses.. On the use of patents: a quantitative analysis based on the PatVal-EU data ...........................100 Theme & Type of contribution .................. Spillovers through labour mobility from academia to business .......2... 85 Datasets............................................................ Patent uses and patent spillovers............ 82 (B1) New Firm Creation and Employment ..2... 75 (A1) The Economic Value of Patents .............................................. 81 Study A............................1........2........... Value of Patents and Relationships with R&D................................................. The Value of Patents from Survey Data .............. 86 REFERENCES ..... 83 Study B.......... CONCLUSIONS AND RESEARCH PLAN FOR LOT2 .............................................................................1................................. 99 Themes & Technologies ......................................................... 80 Study A.........1........ 85 Study B....................100 Themes & Inventors’ organizations ...... 78 Study A..............................5.... The Determinants of Licensing vs...............................................................................................................1....... 84 Study B................................................................................................... 97 ANNEX II.............................101 Countries & Technologies .....................................102 Inventors’ organizations & Technologies ...........................................1.......................................3.............1................... 82 Study B.........................................1.................3...........1..................................................................................................................... Spillovers at the level of the research unit........ 85 C) Policy Implications....................2............................................. 80 (A2) The economic Use of Patents and the Potential for Enhancing It ......................... 81 Study A.................................... 96 Information (i.................. Firm creation and the value of patents ... 78 Study A............................. An empirical investigation from European Survey Data............................................................. The Value of Patents from Indirect Indicators.......................................... 88 ANNEX I.............................. 96 Criteria used for classification of the publications .............................................2.............................................. 99 Themes & Countries ......................2. 83 (B2) Knowledge Spillovers from Patents............ COVERAGE OF THE LITERATURE SURVEY .....................................................101 Countries & Inventors’ organizations ................... DATASET OF THE LITERATURE..102 4 .......

....... ADDITIONAL DESCRIPTIVE STATISTICS FROM THE PATVAL-EU DATASET.........123 The Sources of Knowledge ..........................................119 The role of geographical proximity for collaboration .......................................................................................................................................................................................................................................................... 103 A III.............................................. 106 The value of patents............................................ 109 The use of patents ................................ Spillovers and the Sources of Knowledge in the Invention Process116 Collaboration among inventors in the innovation process...................................................................................ANNEX III......................... 112 A III..........................................4 Collaboration...........................................................110 The importance of different motives for patenting..........................116 Formal and informal collaborations in research .............................2 The economic use of patents ................................126 5 ...........................................................................110 A III..........................................................1 The value and social costs of patents....................................................................................................109 AIII...................3 The creation of new businesses from the patented inventions .........................107 The cost of patents...........

. 66 Figure 4.. Distribution by “macro” technological class55 Figure 4......6 Development of EPO Patent Application by Broad Technology Areas ......... Number of countries in which the applicants are located............. The man-months required by the patents’ invention process across countries............. 1 Share of new firms created by type of applicant: individual inventors vs.............................5 Share of new firms from patented inventions by type of inventors’ employer..... Informal collaborations amongst institutions............12 Contribution to Overall Patent Growth by Country Technology Areas – Other Technologies ..........11 Contribution to Overall Patent Growth by Country Technology Areas – Mechanical Technologies72 Figure 4.... organisations.....Chemicals.................Electronics........7 Development of EPO Patent Application by Broad Technology Areas -Other Technologies ............................................Mechanical Technologies68 Figure 4. 65 Figure 4....... 35 Figure 3... Communication.............. 72 Figure 4........................................... 115 6 ................. 49 Figure 3... 71 Figure 4........... Distribution by country ...................................... 70 Figure 4...............................3 Share of new firms from patented inventions by country ...................................................1 The value of European patents across the EU countries......... Electrical Engineering ....4 Share of new firms from patented inventions by macro technological class ......1 Development of EPO Patent Application By Inventor Country ...... 27 Figure 3.... 48 Figure 3........6 Formal vs........ 114 Figure A.2................. 3 Share of new firms by country of the applicants .......... 67 Figure 4........ 65 Figure 4..LIST OF FIGURES Figure 3.......8 Contribution to Overall Patent Growth by Country Technology Areas...... 73 Figure A..............10 Contribution to Overall Patent Growth by Country Technology Areas – Electronics.9 Contribution to Overall Patent Growth by Technology Areas – Drugs and Health ............................. 55 Figure 3.. 46 Figure 3.. 2 Share of new firm formation........... Communication and Electrical Engineering ................................5 Development of EPO Patent Application by Broad Technology Areas ..........................................................4 Development of EPO Patent Application by Broad Technology Areas .3 Development of EPO Patent Application by Broad Technology Areas ............. Informal collaborations amongst institutions.............................................................. 68 Figure 4...................................2 Development of EPO Patent Application By Applicant Country............. 114 Figure A.......7 Formal vs...........Health and Drugs ................. 67 Figure 4.......

............................................ 45 Table 3........................21 Affiliation of the inventors involved in the development of the patents......................................................... 57 7 ..24 External co-inventors............................ 57 Table 3...............12 The distribution of patent uses by micro technological class ................10 The distribution of patent uses by country.....6 The distribution of the value of European patents by type of inventors’ employer... 36 Table 3. share of opposed patents and average number of forward citations by “micro” technological class ...............7 The man-months required by the patents’ invention process by “macro” technological class.............................................. 43 Table 3. 37 Table 3................................................ ...18 Share of new firms from patented inventions by macro technological class..........11 The distribution of patent uses by macro technological class ................. Distribution by country..........................................25 External co-inventors....... 47 Table 3...............................................17 Share of new firms from patented inventions by European regions (NUTS2) ................................... Distribution by “macro” technological class and by affiliation of the “external” inventors ..16 Importance of different motives for patenting............. 33 Table 3.................................. 50 Table 3.......................................................... 54 Table 3.....................14 The distribution of patent value by patent use.............. .............. in %...................................... Distribution by type of employer of the interviewed inventors and by affiliation of the “external” inventors ................................................................. 40 Table 3.................................. 45 Table 3................ Distribution by European regions (NUTS2)...... 53 Table 3......................................................8 The man-months required by the patents’ invention process by “micro” technological class .... Distribution by type of inventors’ employer.......20 Number of applicant institutions in a patent.....................................3 The value of European patents by “macro” technological class.....2 The value of European patents across European regions (NUTS2)...... Distribution by type of primary inventors’ employer..1 The value of European patents across the EU countries.. 29 Table 3............. 38 Table 3................... 51 Table 3.............. .......................... 42 Table 3....................................... Distribution by macro technological class...........23 External co-inventors........19 Affiliation of the inventors involved in the development of the patents.......... 28 Table 3............LIST OF TABLES Table 3................... 30 Table 3........ in %.. ..................................................... 48 Table 3.................. 34 Table 3........... 40 Table 3.. 44 Table 3.............. 53 Table 3...15 Importance of different motives for patenting.............................................................27 Importance of the interaction with other people while developing the innovation....... Distribution by country and by affiliation of the “external” inventors.............................................................................. Distribution by “macro” technological class....... Distribution by country..........................26 Importance of the interaction with other people while developing the innovation...... 32 Table 3........................................22 Affiliation of the inventors involved in the development of the patents......................................................13 The distribution of patent uses by type of inventors’ employer ...9 The man-months required by the patents’ invention process by type of inventors’ employer ........................................................ 51 Table 3....................4 The value of European patents by “micro” technological class..................................... 52 Table 3.......................... Distribution by country.................5 The value of European patents: average value....

7 The man-months required by the patents’ invention process... 112 Table A......................................................... 112 Table A............................................... 17 Affiliation of the inventors involved in the development of the patents...... Domestic vs......... 13 The importance of different motives for patenting.......................................................................................... Domestic vs........ Distribution by applicants’ country109 Table A............... 108 Table A................... 118 Table A..... Number of countries in which the applicants are located................. 107 Table A.... 18 Collaboration among inventors............................. 15 Share of new firms from patented inventions by European regions (NUTS2) ................................ 21 Number of applicant institutions in a patent.......... Number of countries in which the applicants are located............................................ Distribution by “micro” technological class ............................ Number of countries in which the applicants are located ....................... 118 Table A....31 Importance of different sources of knowledge.. Distribution by type of inventors’ employer ........1 List of Macro and Micro technological classes...........110 Table A.28 Importance of the interaction with other people while developing the innovation.. foreign applicants ..............2 Contribution to Total Patent Growth by Sector (in %) – sorted by size of sector contribution to patent growth ......Table 3.....................30 Importance of different sources of knowledge....... 2 Composition of the sample by macro technological class and type of inventors’ employer .................... 20 Number of applicant institutions in a patent...... Distribution by type of inventors’ employer ....................... Distribution by applicants’ country ...... Number of applicant countries .......... Distribution by “macro” technological class .... 60 Table 3................................................. 107 Table A.. 12 The importance of different motives for patenting by micro technological class ..................................... 119 8 ................... 113 Table A..................................... 117 Table A........................... 116 Table A.. 61 Table 3............................................................................. 14 The importance of different motives for patenting by nationality of the applicants ........................29 Importance of the interaction with other people while developing the innovation...... 104 Table A...................... 5 The value of European patents. 110 Table A................ Number of applicant countries............... 74 Table A.................................... 16 Affiliation of the inventors involved in the development of the patents..................... 6 The value of European patents across European regions (NUTS2) .......... 8 The man-months required by the patents’ invention process. ................................................. 10 The economic use of European patents..........................33 Importance of different sources of knowledge.. foreign applicants............. 69 Table 4.. Number of countries in which the applicants are located ................................... 22 Number of applicant institutions in a patent.. 110 Table A.. Distribution by “micro” technological class...... 59 Table 3... 105 Table A.......................................................... 19 Number of applicant institutions in a patent............................... 119 Table A.............. 111 Table A.. 59 Table 3....... 4 The value of European patents....32 Importance of different sources of knowledge..................... 3 Composition of the sample by micro technological class and type of inventors’ employer ............... Distribution by country............................... Distribution by “macro” technological class ..... Distribution by “macro” technological class......... 117 Table A........ Distribution by “micro” technological class .................................... 58 Table 3................ 104 Table A.............................................................................. 109 Table A................ 62 Table 4.............. Distribution by applicants’ country ...........1 Country Shares in Total EPO Application 1996-2002 by Technology Area (%) . 11 The importance of different motives for patenting by country......... 9 The economic use of European patents...........................

................ 35 Importance of different sources of knowledge............ 126 Table A.. Domestic vs............... . foreign applicants... 125 Table A.... 121 Table A........ 122 Table A..................................................... 121 Table A.. Distribution by type of inventors’ employer..................... 28 Formal collaborations.................. Domestic vs....... Informal collaboration amongst institutions................... 119 Table A......... 29 Formal collaborations.................. 125 Table A........ 33 Importance of the interaction with other people while developing the innovation.. Number of applicant countries.................. foreign applicants..... Distribution by type of employer ...................... Distribution by “micro” technological class............ Number of applicant countries......................... 124 Table A............................................ ........................................... .............. Distribution by European regions (NUTS2) .. Domestic vs...................... 30 Formal collaborations.. 23 Number of applicant institutions in a patent................ Informal collaboration amongst institutions...... Number of applicant countries................................... 127 9 ........... ......... 32 Importance of the interaction with other people while developing the innovation............ 25 Formal vs... Affiliation of the inventors involved in the development of the patents................. 34 Importance of the interaction with other people while developing the innovation.......... 36 Importance of different sources of knowledge............................................... . foreign applicants............ Informal collaboration amongst institutions.............Table A..... 120 Table A............ Domestic vs.............. Affiliation of the co-inventors involved in the development of the patents................................................ 126 Table A......................... 121 Table A......... 26 Formal vs.................. 31 Importance of the interaction with other people while developing the innovation.. Informal collaboration amongst institutions.......... 121 Table A............................ 24 Formal vs......... foreign applicants.......... 27 Formal vs........................................ Distribution by “micro” technological class120 Table A.............. Number of applicant countries............ Distribution by “micro” technological class in %... .................................. Affiliation of the co-inventors involved in the development of the patents....................

and the database on the surveyed scientific publications that we built by country. technology and type of applicant organization.1. such as the effects on employment and the knowledge spillovers that they produce. EU15. (B1) The creation of new businesses from the patented innovation. 10 . The indirect benefits of patents. B. country. New Member States. Annex III reports additional Tables of descriptive statistics based on the PatVal-EU dataset. technology. It provides a complete survey of the existing scientific literature in economic and business studies. We provide Tables of descriptive statistics by country. (B2) The impact of patent protection on knowledge spillovers and productivity. Section 4 finally reports descriptive statistics on the patent explosion at the EPO by showing the patent counts and the contribution to the growth of EPO patents from 1986 to 2001 by broad regions (EU25. technologies and type of applicants’ organization. both actual and potential effects. theoretical or empirical). the third and fourth Themes deal with “The indirect effects of patent protection”. The Annexes at the end of the Report provide additional empirical evidence on the issues discussed in this work. spillovers and the sources of knowledge in the innovation process. Section 2 is organised in 4 Themes: (A1) The economic value of patents. Section 5 summarises the results and concludes with a detailed plan of research for Lot2 of this project. US and Japan) and by “macro” technological classes. (B2) Collaboration. (A2) The economic use of patents and the importance of different motives for patenting. The direct benefits of patents. and it uses original data to present descriptive statistics and elaborations on the value of European patents.e. type of inventors’ employer and type of contribution to the literature (i. Annex I describes the structure of the dataset that has been built in order to develop a comprehensive literature survey on the four themes of this tender. Annex II includes a set of Tables showing the systematic analysis of the coverage of the literature by theme. (B1) New firm creation and employment. The first and second Themes go under the heading “The direct economic effects of patent protection”. Section 3 presents and discusses selected Tables of descriptive statistics based on the PatVal-EU dataset on 4 research Themes that are closely related to those explored in the survey of the literature: (A1) The value and social costs of patents. (A2) The utilization of patents and the potential for enhancing their value. Section 2 describes the methodology adopted for the survey of the literature. The survey of the literature and the empirical analysis are organised around two main headings: A. ORGANIZATION OF THE REPORT This Report deals with the economic value of European patents and offers a comprehensive analysis of the “state of the art” on this issue.

the European Commission. We collected articles. recent articles receive a smaller number of citations than older papers. collection and classification of the literature relied both on the expertise of the team members. Analysis of the literature by theme 2. The methodology employed for the survey follows three steps: a. technologies. Annex 1 describes the structure of the database. As far as the use of citations is concerned. we constructed a dataset that classifies the surveyed publications by using the following criteria: research Theme. Social Sciences Index. 11 .2 presents the Review of the literature that we collected. In addition. Literature search b. books and book chapters.g.2. Helecon. We also searched for publications and reports of the OECD. Business Source Premier. and case studies). technological area. Second. we had to solve two problems. We select the relevant publications and we assess their quality by using the team expertise and the codified rules of evaluation based on the number of citations received by the articles (source: Social Science Citation Index). ABI Inform. This ensured a comprehensive coverage of the “state of the art” on the topics under investigation. We therefore checked for the quality of the publisher. International Bibliography of the Social Sciences. empirical vs. they are not available for reports. In order to limit this problem we used the impact factor of the Journal. and other international organizations. and on the systematic search for references. main contribution of the publication. country. Section 2. sectors and technologies. Jstor. We also found contributions that analyse specific themes by means of disaggregated levels of analysis (e. Classification of the literature c. LITERATURE SURVEY The survey of the literature provides a comprehensive overview of the existing studies on the direct and indirect impact of patents in different countries.1 Literature search and classification The methodology for the search. the impact factor of the journal and the quality of the publisher. working papers and reports that deal with broad themes and aggregate analysis. theoretical paper. abstract. By using these papers we identified a number of key-words and conceptual maps for a systematic search of the literature. We performed the latter by browsing computerized and printed searching tools available at Sant’Anna School of Advanced Studies and at INSEAD: EconLit. First. specific countries. The Georges and Edna Doriot Library. NBER Working Paper Series and the Research on Innovation Web site. due to a “truncation” effect. books. First of all. the team experts provided us with a list of relevant papers on each research Theme.

and thus the inefficiencies associated with monopoly protection. The survey findings of Levin. such as first mover advantages or secrecy. 2002 for an excellent and succinct review on this topic) suggest that policies that broaden the scope of patents do not however unambiguously stimulate innovation in industries characterized by cumulative technical change. Nordhaus economic framework has been extended. Taylor and Silberston (1973). (1959). Gallini (1992) operationalises this concept as the cost that rivals must incur to imitate an invention without infringement. (2000) suggest that in most industries patents are less featured than other means of protecting innovations. These extensions relate to the design of optimal patent policy. by providing restrictions to the use of patented inventions. stimulate the introduction of improvements and protect early innovators.1 Overview on the Economics of Patenting and some Recent Theoretical Contributions The classical theoretical contribution on the basic economics of IP is that of Nordhaus (1969). No matter how scope is defined. A1. the empirical literature on the impact of patent scope on R&D incentives is surprisingly limited. Important extensions of these theories during the 90’s (cf. a broader patent increases the market price for the invention. especially during the 90s. Klevorick. The main social cost is the restriction in use. and have focussed mainly on two dimensions of patent protection: its length and scope (or breath). The theory highlights that the main benefits of patent protection are to stimulate innovative investments and promote the diffusion of technological knowledge. as evidenced in the following sections. The main conclusion from this literature is that in world where both licensing and commercialisation are feasible alternatives to profit from the invention. patent law provides the ability to recover the investment needed to introduce technological innovations. Economic Value of Patents A1. Indeed. (1981). 1995) and Japan (Goto and Odagiri.. The evidence has been interpreted as suggesting that the inducement provided by patents for innovation is small in most industries. 12 . Cohen et al. Nelson and Winter (1987) and. 1997). however. a case can be made for narrow and long patents. Despite its theoretical importance. The studies by Scherer et al.2 The Value of Patent Protection from Survey Data Prior empirical work has tried to evaluate the effectiveness of patent protection in stimulating innovation largely through survey-based studies. Similar results were obtained for Europe (Arundel et al.2 Analysis of the Literature by Theme Theme A1. Indeed. broad and short patents are indicated as superior policy in the case of cumulative innovations because they prevent duplicative R&D. Mansfield (1986) suggest that patent protection may not be an essential stimulus for the generation of innovation in most industries. more recently. in exchange of the disclosure of the technical details of the patented inventions to the public. Gallini.2. Mansfield et al. A broader scope refers to a broader area of technology space from which others are excluded.

the patent renewal models cannot accomplish such a task. viz. Pakes and Putnam. a concept closely related to the value of patent rights computed by the patent renewal models. Pakes. it is necessary to pay a renewal fee each year in order to keep a patent in force.e. where the focus is on explaining the value of patent rights.” Without a model that links such values to R&D investments. Pakes (1986) found that in France one percent of the patents has a value higher than 70. Schankerman (1998) uses patent renewal data to estimate the value of the cash subsidy to R&D conferred by patent protection in France. and then a way to compare the benefits from that response. an 11 to 16 percent increase in returns does not seem small”. Arora. Estimates suggest returns around 15%. Schankerman. The ESR is calculated by dividing the total value of patent rights by total R&D. Interestingly enough. To question whether this is enough an incentive for innovation. 1986. In many countries. (2003) used the number of patent applications. are not effective. Hence. when the renewal fee is not paid. Lanjouw. in the sense that they provide a positive expected premium – i.000 US dollars. The rationale for using the patent renewal fees as an indicator of the value of patent protection is that they are paid only if the expected future returns of the patents are higher than the costs of keeping the patent rights.000 US dollars. Putnam (1996) used patent applications data to uncover the value of international patent protection. However. with the costs of patent protection. This literature finds differences across European countries. Archibugi and Pianta. An important exception is given by the studies that use patent renewal data. In the next section we will survey the literature that more recently has tried to accomplish such a task. the renewal fee itself. this research has focused on them. and R&D data to compute the patent premium. on average.3 The Value of Patent Protection from Patent Renewal and Application Data The literature has traditionally used patent data as indicators of inventive success of the underlying innovative effort (Griliches. The objective was to uncover the causes and consequences of inventive activity. given that he used only patents for which an international extension was applied for. the patent has expected returns lower than the threshold (Pakes and Schankerman. 1998). 1998. Schankerman and Pakes. Pakes and Simpson (1989) argue that: “Compared with other institutionally created incentives (such as tax breaks). 1986. since data on patent renewal fees were available only for the European patents. 2003. on average. one would need an estimate of the R&D response to the increase in returns. greater net expected returns from patenting an innovation relative to not patenting it – for only a small fraction of innovations. and not to directly evaluate the impact of patent protection on innovation. while in Germany one percent of the patents have value higher than 120. They also add that “Of course. the relative magnitude of benefits and costs suggests that firms expect to loose about 50% of the 13 . It also reflects the average return to R&D conditional on patent protection.A1. which he calls the equivalent subsidy rate (ESR) to company-funded R&D due to patent protection. to judge the effectiveness of this incentive. 1996). 1990. For example. 1984. the propensity to patent. and he showed that this is higher than the one found by using patent renewal data (see also Deng. In fact. results are not directly comparable to those obtained by the patent renewal models. Ceccagnoli and Cohen. plus whatever benefits there are from publicizing the content of the patent. They find that patents. It corresponds to the subsidy that firms would need in order to maintain R&D at current levels in the absence of patents. for a recent extension of this approach).

there is only one study (Sakakibara and Branstetter. Their model. among others.on average . also its competitors will benefit from it. this study finds that there is only a small positive effect of increasing patent scope on R&D investments. Despite substantial theoretical work on the impact of patent scope on innovation and social welfare. This literature focussed on the estimation of patent production functions. For a recent survey and extensions of this work see. firms expect to earn . this could have important implications for innovation. Their model further recognizes that if one firm benefits from stronger patent protection in a specific area. A1. (2003). did not allow them to address the impact of patent propensity on R&D. After controlling for firm R&D. In semiconductors and communications equipment the premium and the incentive effect are much lower. By using firm-level data from the Carnegie Mellon Survey (CMS). however. with R&D and knowledge spillovers as the main inputs. although confirming that most patented innovations have small conditional premia due to the skewness of the distribution of the premia conditional on patenting. also presents evidence that the average expected premium for patented innovations is much larger than the average expected premium for all innovations. and Griliches. including the appropriability incentive of patenting and the offsetting role of patents in producing R&D spillovers.1 This latter point was addressed by Arora et al.value of an innovation by patenting it. these authors tried to disentangle the impact of supply-side factors (such as technological opportunities) and demand factors on the rate of technical change. By using panel data on a sample of manufacturing firms in Japan.4 The Impact of Patent Protection on R&D and Innovation Among the first studies that focussed on the relationship between patents and R&D are the contributions by Pakes (1985). In considering the impact of patenting on the R&D conducted by the industry incumbents. 14 . for the innovations that firms choose to patent. and computers. See also the work by Licht and Zos (1996) for the case of Germany. medical instruments. However.a 50% premium over the no-patenting case. changes over time in the unobserved factors that affect the number of patents granted to a firm were interpreted as capturing supply-side changes in R&D activity. this is the first study on the impact of changing the value of patent rights on innovation. Cincera (1997). measured as the number of patent applications or grants. or firms’ propensities to patent. to the extent that entry is associated with innovation. 2001) that directly analyses the impact of patent scope on innovation. They could either represent technological opportunities. However. The main objective of these earlier studies was to uncover the determinants of inventive activity. The results of this work show that patents have the greatest positive incentive effect on research and development (in the sense that an increase in the premium generates a positive a substantially positive response in R&D) in pharmaceuticals. although still positive and not negligible. Their model takes into account the fact that patenting and R&D decisions are driven by common factors. which seems to be a key determinant of the expected returns to R&D as conditioned by the existence of the patent system. In particular. biotechnology. The lack of direct measures for patent propensity. 1 A related literature has examined the determinants of patenting. this analysis does not consider the impact of the uses of patents on industry entry. Hall and Pakes (1991).

g. Jaffe and Trajtenberg (2005) for the US. several studies in the literature used them as indicators of the value of patents. Bloom and Van Reenen (2002) found that doubling the citation weighted patent stock would increase the value of UK public firms per unit of capital by about 35%.6 Patent Citations and the Value of Patents Quite a few studies have found that the economic value of patents is correlated with their citations. backward and forward citations. it is not clear whether this study captures the impact of innovation on performance as opposed to the benefits of patent protection over and above the profits derived from alternative appropriation strategies. However. 1999. who computed a measure of social returns to innovation in the computer-tomography scanner industry. The traditional contribution is Trajtenberg (1990).g. They found that patent citation counts are positively associated with patent renewal fees and with the private 15 . Kanwar and Evenson. has made an explicit link between patent citations and the social and private value of patents. e. Jaffe and Trajtenberg. though. 2001). None of the subsequent studies. number of claims. A1. 2002. Hall. A positive. for the US. for example. This is also the case of Lerner (1994). or at least of their importance. Jaffe and Trajtenberg. Scherer and Vopel (1999) used survey data obtained on 964 inventions made in the US and Germany.7 Multiple Indicators for Patent Value Recent contributions found that several indicators can be used as proxies for the value of patents. see Hall. Since citations to US and European patents became widely available (e. an extra citation per patent boosts market value by 3%.All the contributions that used cross-national aggregate data have instead found a positive and significant effect of the strength of patent protection on R&D (Eaton and Kortum. 1985. 2003). They also report that. and on which German patent renewal fees were paid to full-term expiration in 1995. there seems to be a need for additional empirical work to evaluate the impact of patent policies on R&D incentives within specific countries and over time along the lines of Sakakibara and Branstetter (2001). This literature has consistently estimated a positive and significant marginal value of the patent stock. Given the results of the existing studies on this issue. but somewhat lower response was found by Hall. who looks at the impact of patent scope on the market value of a sample of US biotech companies. A1. He also found a positive and significant correlation between this measure and patent citations. 2003. He finds that a one standard deviation increase in the average patent scope is associated with a 21% increase in the firm's value. oppositions. Harhoff. Bloom and Van Reenen. 2005). Most of this literature suffers from the inability to disentangle the impact of patent protection on firm performance from the impact of innovation itself. family size.5 The Impact of Patent Protection on Firm Performance One approach to assess the value of patents estimates the impact of the patent stock of firms on their stock market value after controlling for their stock of R&D and physical capital (Pakes. Maloney and Lederman. This captures the change in the market expectation of the discounted rents from the patented innovations. A1.

an important question is whether the quality of such patents. One rigorous survey is the one conducted by Harhoff. and thus their value. Scherer and Harhoff (2000). Harhoff. and Harhoff. as a consequence. 2003b). Scherer and others (Scherer and Harhoff. Along this line. Survey data can provide direct measures of the value of patents as one can ask the inventor or any relevant individual in the applicant organization direct information on patent value. Narin. The empirical literature focuses mainly on the US. A1. Their results suggest that the number of references to the patent literature as well as the citations a patent receives are positively related to its value. and Vopel (2003b). If patents do not meet the novelty and utility thresholds. A1. Hall. and in particular that part of the change due to changes in the legal environment. 2001). Hall and Ziedonis (2001) conjectured that much of the increase in patenting in semiconductors reflected “harvesting” behaviour—that is the patenting of inventions that would have been invented in any event. Finally. Harhoff and Reitzig (2004) analyse the determinants of oppositions (the most important mechanism by which the validity of a European patent can be challenged) to biotechnology and pharmaceutical patents granted by the EPO between 1978 and 1996. which is interpreted as signalling the value of the invention. There are only few studies that employ survey information on the economic value of patents. measured as the average 16 . Research on the skewness of the patent value distributions includes Harhoff. Harhoff. Longland and Bosworth. They find that the likelihood of opposition increases with the value of the patent.8 Patent Value from Survey Data The literature has usually used indirect measures of the economic value of patents. 2000. but not for the EU in general (Bloom and Van Reenen. Methodological and empirical investigations are available for the US and the UK. These studies have also shown that the distribution of patent values is highly skewed. Scherer and Vopel (2003) combined estimates of the value of patent rights from a survey of patent-holders with a set of indicator variables in order to model the value of patents. These studies cover only German and US patents. with a spike at zero. the social costs due to monopoly protection are more likely to outweigh the benefits conferred by patents. Sherry and Teece (2004) use the success rate of patent lawsuits in the US to measure the change in the patent value over time. 2001. either because they do not know with precision the actual economic returns of their patents. Scherer. and that. Duysters and Verspagen (2002) that also include some European countries. has declined. Scherer and Vopel (1999). Harhoff and Hall (2002) and Bekkers. Scherer and Vopel.economic value of patents. one would have expected a decline in quality. Jaffe and Trajtenberg. Guellec and van Pottelsberghe (2002) present systematic international comparisons for the OECD area on the determinants of the probability for a patent application to be granted. or because they do not want to disclose it. Hall and Ziedonis did not.9 Value of Patents and their quality: Trends over time Given the rapid growth in patenting. find a clear decline in quality. Greenhalgh. 2002. Exceptions are Harhoff and Reitzig (2004). however. One limitation of survey data is that the inventors or the assignees may provide a subjective answer to the question on the value of their patents.

see section A2. Approximately 95% of all patent suits settle either before or during trial. Olivastro and Hamilton (2001) found an increase in patent quality over time. computers. hence generating a social loss of resources.number of citations per patent in semiconductors. very few patent suits actually go to trial. whether these costs have grown due to any increase in the strategic use of patenting ( i. the risk of being suited for patent infringement and hence incurring the direct legal costs associated to a trial creates a barrier to entry. about half of the estimated legal costs of litigation are incurred before the end of the discovery phase (AIPLA. Consistent with this finding. The statistics on litigations show that the crude number of patent suits filed in the US has constantly grown during the last two decades – although with some differences among different technological areas – following the constant increase of patenting over the same period of time. apart from direct costs.10 Other Social Costs of the Patent System There is some evidence suggesting that litigation costs have increased in the US. these tests of the quality of patents that rely on forward citations do not make the case that the patents that have been issued over the recent past have declined in terms of the standards of novelty. In information technology more broadly. non-obviousness. In industries where patent protection is stronger. On this issue. In particular. defensive or blocking patents). 1985-90 and 1991-95. and 1% settle during trial. settlements before trials are likely to lead to collusive outcomes. and utility (Cohen. the study shows that no increase in litigation (in relative terms) occurred in any technological field. 2004. On this topic. Cesaroni and Giuri (2005) have recently reviewed the empirical literature on the extent and the costs of litigation. and the costs per case have grown (National Research Council. Moreover.” A1. In the literature survey of Cesaroni and Giuri (2005) it is also stressed that. patent litigations generate indirect costs whose relevance is even more important for the society on the whole. However. electronics. 2004). mechanical.2. biotechnology. and once the growth in patenting is taken into account. It is not clear. Often. Hicks. There is also evidence that the costs of prosecuting patents have grown rapidly. They show that the propensity to litigation varies among fields. thus making litigation extremely costly. rather than the actual amount of legal costs associated to litigations. it is the risk of incurring in such costs to create negative effects and to induce firms to modify their behaviours accordingly. For instance. Breitzman. Even though most patent suits end up with a settlement before or during trial. p. 78% settle before the pre-trial hearing. however. an additional 16% settle before trial. However. from a social perspective.e. and prevents firms from investing in innovations. chemicals. Lanjouw and Schankerman (2003) analyse the filing rates by different technology fields (drugs. other health. In turn. in a recent study. and by using a normalized measure of the number of times the previous five years patents are cited in the current year. a system of strong property rights might reduce the overall level of investments in R&D and innovative 17 . especially per dollar of R&D spending. Lanjouw and Schankerman (2003) actually find a positive relationship between portfolio size and the number of forward citations per patent. 2001). a recent study by the US National Research Council (2004) states that “…the claim that quality has deteriorated in a broad and systematic way has not been empirically tested. Furthermore. contrarily to policy expectations. 31). when time trends are considered. and miscellaneous) during three time periods: 1978-84.

2004). stronger patents can reduce transaction costs in technology licensing contracts (Arora. However. Hsu and Stern (2002) find that the presence of patents increases the likelihood that an inventor will license to an incumbent rather than enter the product market by commercialising the invention. Anand and Khanna (2000) come the closest to a systematic cross-industry comparison. systematic and direct empirical support is limited. though plausible. it is plausible that by making possible the market exchange of new knowledge. technical services are more likely to be bundled into licensing contracts if patents are also present. These studies tend to have small samples. 1996. However. Hsu and Stern. Nakamura and Odagiri (2003) find that stronger IP protection tends to decrease transaction costs and to stimulate technology transactions between Japanese manufacturing firms. the patent system contributes to a more efficient use of new knowledge. There are a handful of papers. this encourages firms to offer technologies for licensing or technological capability for hire (Gans. as reflected by the studies described in the previous section. especially for small firms that are most exposed to the risks of patent litigations. for a sample of technology import agreements signed by Indian firms. 2002. By using a sample of MIT inventions. They find that in chemicals there are many more technology deals than in other sectors. Insofar as stronger patents also enhance bargaining power of the technology holder. Lamoreaux and Sokoloff (2001). they under-invest in those areas where patent protection of large firms is higher. point to the information disclosure aspect of patents. both of which are critical for the successful commercialisation of new knowledge. Cassiman and Veugelers (2002) do not find that the patent strength encourages Belgian firms to enter collaborative R&D agreements. and many of them have limitations as well.activities. Since small firms often lack the required financial resources to sustain long and costly litigation causes. The empirical evidence can be found in Lerner (1995). limited or no variation in patent strength.1 Impact of patent protection on markets for technology IPRs and especially patents have been thought of primarily in terms of providing incentives for innovation. Lanjouw and Lerner (2001) and Lanjouw and Schankerman (2003). Arguably. a 18 . 2003). In many instances. semiconductors. Thus. Theme A2: Utilization of Patents and Potential for Enhancing their Value A2. and they are industry specific. Mixed empirical evidence on the impact of patenting on international licensing using cross-national data has been recently presented by Fink and Maskus (2005). following Coase. In other words. a large fraction of them involve arms length licensing deals. Gans. instruments and chemicals have used their intellectual property as a means to obtain financing and corporate partners. Arora and Fosfuri. start-up firms in industries such biotechnology. Arora and Merges. unused technologies find more willing buyers and innovators incapable of exploiting their innovations (or unwilling to do so) can appropriate the rents from their innovation by licensing or selling their innovation to others. Arora (1996) shows that. a large fraction of them involve exclusive licenses. economists have also argued that the definition of property rights in innovation will also make them easier to exchange.

there is no empirical study that has evaluated the R&D incentive effect of patent licensing.g. for the basic theoretical piece). 2000). Goto and Odagiri (1997) for Japan. 2000. 2000. There are also many patents applied for by large firms that are never used. are used more often when markets are concentrated and populated by large and research intensive firms. rather than rights on existing technologies). Cohen. Shapiro. strategic entry deterrents. Indeed. or it can be licensed out. or it can be used for strategic purposes. They find that an increase in the effectiveness of patent protection increases licensing propensity when complementary assets required to bring new technologies to market are absent or unimportant. Early empirical evidence on such use is reported by Bunch and Smiley (1992) for the US. This is often the outcome of strategic behaviour. when firms are better positioned to bring new technologies to the market. Anand and Khanna (2000) speculate that this is due to two inter-related factors. small firms. 2001. individual inventors. Earlier studies on this issue focused on the use of patents to build strategic entry barriers and to preserve monopoly power (Gilbert and Newberry. Nelson and Walsh. Many patents are not used commercially simply because the inventors or the assignees do not have adequate assets to exploit them (e. systematic evidence in support of a positive effect of patent protection on technology licensing is found in Arora and Ceccagnoli (2005). biotechnology and pharmaceuticals have stronger links to science so that the underlying knowledge base allows for technological knowledge to be articulated easily (Arora and Gambardella. Cohen. and especially in biotech and pharmaceuticals. By contrast. scientific institutions). as it is also confirmed by the skewness of their value distribution. Note that this effect of the patent system is closely linked to that of encouraging the generation of innovations. Cohen. however. but that its impact is critically mediated by the ownership of specialized complementary assets. Nelson and Walsh (2000) for the US. and there is a large fraction of ex-ante deals (where the contract is about future rights. the impact of patenting on licensing is a secondary key channel through which patent protection stimulates further investments in new technology. as a patent can be exploited commercially by using own complementary assets. To our knowledge. Second. A2. Other studies on the extent and the effectiveness of such strategic uses across manufacturing industries and countries are: Bessen and Hunt (2004) for software patents.2 Impact of patent protection on other dimensions of firms’ technology strategies Only very few patents yield economic returns. They show that for newly developed products. 1994). increases in patent effectiveness increases patenting propensity but reduces the propensity to license. van de Paal and Soete (1995). The patenting option opens up such strategic opportunities in addition to licensing (Hall and Ziedonis. patents are more effective in chemicals. 1992. Rivette and Kline. Nelson and Walsh (2000) show that it is quite 19 . This is not surprising. Finally. First. Blind and Thumm (2004) and Reitzig (2004) for Europe. such as patents. In particular. to the extent that firms can more easily profit from patented innovations by licensing them to other firms that have the complementary commercialisation assets. They present systematic cross-industry empirical support for the US manufacturing sector in favour of the proposition that patent protection is a key determinant of the market for technology. Arundel.small fraction of deals are among related firms.

the strengthening of IPRs stimulates innovation and growth of the North and hurts the imitator when the rate of imitation is high. 2004. Systematic empirical evidence on whether patent thickets. Theme B1: New Firm Creation and Employment As discussed by several studies. From a theoretical point of view a classic study is that of Helpman (1993). Nelson and Walsh (2000) and Shapiro (2000) suggest that the use of patents in portfolio exchanges among large incumbents may actually deter entry into the semiconductor and in other complex product industries. is still weak. 1996) supports the idea that strong patent protection is associated to higher economic growth. Walsh. smaller firms specialised in the production of new technologies can be a major factor in enhancing the employment and economic performance of specific regions. especially on the part of universities. and there is more patenting of upstream discoveries since the Bayh-Dole Act. impeding subsequent innovation and commercialisation. Also. The major reason is that firms and other institutions have developed “working solutions” that limit the effects of the intellectual property complexities that exist. who shows how in a North-South endogenous growth model. they found no evidence of breakdowns in negotiations over rights. and university and government personnel. Chen and Dahlman (2004) find that a 20 percent increase in the annual number of patents held is associated with an increase of 3. To assess the degree to which either restrictions on access to upstream discoveries or anticommons are indeed hampering biomedical innovation. where the North innovates and the South imitates. or firms avoiding projects due to the prospect of an anticommon. Arora and Cohen (2003) conducted 70 interviews with scientists and executives employed by firms. at least not yet. Cohen. for commercially worthwhile projects. Gould and Gruben. especially around the impact of the patenting and licensing of “research tools. However. The available empirical evidence (Chen and Dahlman. The possibility that access to a patent on a key upstream technology may be blocked. electronics and semiconductors -. 1998). “blocking patents”).” which include any tangible or informational input into the process of discovering a drug or any other medical therapy or method of diagnosing disease (cf. 20 . that these developments are impeding the development of drugs or other therapies in a significant way.common in industries such as computer.8 percentage points in annual economic growth. there are a few studies at the macro level on the relationships between patent protection and economic growth. such as licensing and occasional litigation. They found that patents are indeed now associated with new therapeutic products.create or reinforce barriers to entry. This is why a patent system that allows for the formation of such firms can raise employment. They did not find. however. Strategies of patent portfolio expansion as responses to hold-up problems in the US semiconductor industry have been recently analysed by Ziedonis (2003).e. intellectual property practitioners. Heller and Eisenberg. extensive cross-licensing and the associated accumulation of large patent portfolios among incumbents -. has attracted particular attention recently in biomedicine. Scholars such as Hall and Ziedonis (2001).common to patent technologies around a certain invention to avoid that others use it even if they do not plan to exploit that invention (i.

Ziedonis (2003) reports that. from their first appearance in 1983. and Sharfstein (2004). since the 1980’s there was rapid entry in the semiconductor industry by design firms that relied heavily on patents to protect their intellectual property. Many start-ups in industries like biotechnology. There is also consensus that in industries where patents are especially effective. as in biotechnology (e. as demonstrated by the limited amount of studies which have attempted to do so. Orsenigo and Pisano. and the market for technology discussed earlier. But if they do not have the resources and the capabilities to invest in such downstream assets. they reduce the need for duplicate investments in R&D.g. 1999). By contrast. as well as to identify their beneficiaries and their sources. 21 . both of which are critical for the successful commercialisation of new knowledge. Sometimes patents are crucial for such firms to arise. Henderson. we observe entry based at least partly on the strength of patent protection (Cohen. which suggests that in the very industry where concerns have been raised over the effect of patent portfolio races and cross-licensing on entry. intellectual property enables them to sell the rights on the invention to other firms that own development and commercialisation assets. and this provides opportunities for real cost reductions and increases in total factor productivity. and the study on venture capital backed spin-offs of Gompers.. 2004). like developing the innovations or producing and selling the final goods. most often they do not produce the invention in the first place. instruments and chemicals have used their intellectual property as a means of obtaining financing and corporate partners. Lerner. systematic evidence on the formation of new firms from patented inventions is still missing. they have provided the basis for raising capital and stimulating entry. empirically it is hard to measure spillovers. Studies on this topic include the empirical analysis on the spin-offs from patented inventions by university scientists of Shane (2004) and Shane and Khurana (2003).Another way to assess the impact of patents on employment is to evaluate the creation of new businesses based on patented inventions. the number of semiconductor design firms grew to over 40 by 1994. Yet. Theme B2: The impact of patent protection on knowledge spillovers and productivity To what extent the reduced social value of the inventions produced by the protection offered by the patent is compensated by the fact that patents enhance social welfare by encouraging knowledge spillovers? Theoretical models have shown that knowledge spillovers are an important determinant of economic growth. Even more challenging is to evaluate the impact of patents in facilitating these flows. semiconductors. the study on large firms of Klepper (2001). This encourages the formation of these firms. For example. This is because without them these firms cannot appropriate the returns from their innovations unless they carry out further downstream steps in the innovation process. However. According to Ziedonis (2003).

This issue is at the core of the indirect effects of patents. A major source of information on the empirical importance of patents in facilitating knowledge diffusion has been survey evidence presented by Cohen et al. and thus greater competition. but it requires active participation in a network of knowledge exchanges. By analysing the US semiconductor industry. For a recent methodological survey on how to identify and measure research spillovers see Garcia Fontes (2005). A problem which permeates the studies of spillovers that use patent citations is related to the fact that a larger number of citations may also capture greater technological activity in a particular field. which confirm that patent citations reflect spillovers as perceived by the participants.There is a well established literature using patent citations to assess whether different patent holders rely on each other knowledge bases. it is not obvious what their impact on industry R&D is. Fogarty and Banks (1998) find that two-thirds of citations to patents of NASA-Lewis' Electro-Physics Branch were evaluated as involving spillovers. Trajtenberg and Henderson (1993) have used patent citations to assess the importance of spillovers across geographically close inventors. and it has not been adequately explored in the existing empirical literature. suggesting that knowledge spillovers are primarily intra-national in scope. but not sufficient for transferring knowledge across firms. they show that in the US the examiners add 40 per cent of all citations and two-thirds of citations on the average patent are added by examiners. Once one recognizes this dual role of spillovers in general. Almeida and Kogut (1999) find that knowledge flows are embedded in regional labour networks. and appear to be a key reason for greater intra-industry R&D spillovers there. Breschi and Lissoni (2005) provide evidence that geography is not a sufficient condition for accessing a local pool of knowledge. Similarly. (2002). One of the main results of this literature is that patent citations are higher the shorter the geographical distance between inventors of the cited and citing patents. Similar studies have been carried out for Europe (Verspagen. 2004). 1999. and therefore are not ‘true’ externalities. This is confirmed by Maskus and 22 . suggesting that patent policy can importantly affect information flows. Branstetter (2001) provides estimates of the relative impact of intra-national and international knowledge spillovers on innovation and productivity at the firm level. who suggest that patents play a more central role in diffusing information across rivals in Japan relative to the US manufacturing sector. implying that patents are important. supporting the idea that knowledge flows are geographically concentrated. A second major problem for the use of citations as indicator of spillovers has been recently discussed by Alcacer and Gittelman (2004). Trajtenberg and Fogarty (2000). Convincing survey evidence on the validity of patent citations as a measure of spillovers is provided by Jaffe. Furthermore. 40 per cent of all patents have all citations added by examiners. Verspagen and Schoenmakers. albeit with substantial noise. In particular. 1997 and Verspagen and De Loo. using patent citations for a large sample of US multinational firms. Jaffe. Similarly. for a sample of Italian firms belonging to multiple industries. Other recent studies on this topic reinforce the limit of patents as channels of information flows. This has been recently confirmed by Singh (2005). Jaffe. using a panel data from the US and Japan. who exploit recently available data for the US suggesting that an important fraction of citations contained in a patent are included by the examiners. Breschi and Lissoni (2001) suggest however that a large fraction of such patent-based spillovers are market mediated.

This is true for both the US and Europe. They show that the probability of using patent databases increases with firm size and R&D investments. They also find that across industries. a greater propensity to patent is positively associated with a greater use of patents as a channel of information on other firms’ innovative activities. however. and the causal link between such flows and firms’ R&D productivity. Cross-section survey evidence on the use of patents as information channels in Europe is provided by Arundel and Steinmueller (1998). very few studies have been able to assess the extent to which knowledge spillovers are conditioned by patent protection. who provide econometric evidence that the technology diffused through the Japanese patent system had a significant and positive impact on post-war growth in Japanese total factor productivity. In general. 23 .McDaniel (1999). and it is higher among firms that also find patents as a valuable appropriation mechanism.

Spain and the United Kingdom (hereafter. The distribution of the surveyed patents across countries is the following: 3. The number of patents surveyed in each country mirrors the relative size of the country population. Electronics (including Semiconductors). Computing. and it was directed to the inventors of 27.124 in the Netherlands.1 DESCRIPTIVE STATISTICS II: PATVAL DATASET The Patval Dataset This Section uses the Patval-EU database to present Tables of descriptive statistics on the value of European patents. The Patval-EU dataset was constructed by collecting information from the inventors of European patents applied at the EPO in 1993-1997. The targeted number of patents for which we expected the inventors to respond was 10. Mechanical engineering. “EU6 countries”). the motives for patenting.486 in France. like their age.531 EPO patents with priority date in 1993-1997 located in France. 1. Pharmaceutical. This is because the extent to which the variables used in this Report suffer from missing observations differs. The PatVal-EU survey covers representative samples of patents in the following technological areas: Biotechnology. 1. The survey provides also information about the type of the patent applicant: it indicates if the applicant is a small firm (less than 100 employees). and they may differ across Tables and Figures. and 269 in Spain. their economic use. The PatVal-EU survey. and the setting up of formal or informal collaborations 2 The dataset was built within the PatVal-EU project of the European Commission (contract number HPV2-CT-200100013). 1.017 patents3.2 The full scale PatVal-EU survey was conducted from May 2003 to January 2004.624 questionnaires covering 9.542 in the UK.346 patents are invented in Germany. Chemicals. 1. 24 . a university. a public or private research institution.000. however. 3 The number of observations in the Tables and Figures presented in this Report may be lower than 9017. In the end the European inventors responded to 9. technology and type of applicants’ organization. the process that led to the innovation such as the sources of knowledge used in the research process. Communications. produced other interesting and unique data on: the characteristics of the inventors. and the extent of knowledge spillovers that arise from different sources during the innovation process.250 in Italy. a medium-seized firm (100-250). Electrical engineering. or an individual inventor. the Netherlands. a large firm (<250). Germany. the institutions to which they are affiliated.3. Italy. 3. We provide these statistics by country. The composition of the sample by technological classes and types of inventors’ employers is reported in Tables A2-A3 in Annex III. Instruments. the importance of patents for setting up new firms. The primary goal of the PatVal-EU survey was to gather information on the economic value of the European patents. the educational and work background. Production engineering.

such as the licensing behaviour of firms. We also use information about the strategic motives for patenting. or if it was licensed. this Section provides a number of Tables and descriptive statistics to analyse the 4 themes of our study: A1. We use the PatVal-EU data on whether the patent produced economic returns. We gathered information on the existence of any form of collaboration among the inventors who developed a patent. This literature describes the positive effects of patents that might offset some of the social costs of 4 For detailed information on the methodology adopted for the PatVal-EU survey and for descriptive statistics of the main variables see the PatVal-EU final report (European Commission. We also employ data on whether the research leading to the patent was based on formal and informal collaborations among different institutions. etc. and it helps derive policy implications for the European innovative and economic performance4.2 The Value and Social Costs of Patents The value of patents There is a long literature on the impact of patents on innovative and economic competitiveness. and whether the patent gave rise to opposition procedures in Courts. We employ data on whether the patent right was used for commercial or industrial purposes. B2. The Economic Use of Patents. Collaboration. the motivations to patent and the use of property rights. The Value and Social Costs of Patents. 2005). such as the reaction to the behaviour of rivals or the willingness to block competitors. and we checked for the importance of the geographical distance among the parties involved in the exchange of knowledge.among individual inventors and organisations. the number of forward citations they receive after the publication date. and we explored whether the inventors are employed in the same organisation or in different organizations. - - - 3. We use data on whether the patent gave rise to a new firm. B1. the strategic reasons to patent. A2. Specifically. The combination of this information provides a good understanding of the relationship between the input and the output variables in the innovation process. We employ the inventors’ estimate of the monetary value of the patents. We also use the inventors’ estimate of the man months required by the research activity leading to the patents. Spillovers and the Sources of Knowledge in the Innovation Process. 25 . The creation of new firms from the patented innovation. In the present context the PatVal-EU survey gives the opportunity to investigate a number of issues related to the value of patents and their economic exploitation that the literature has partially neglected because of the lack of suitable information that the patent documents alone could not provide.

26 .intellectual property protection. 2003. which is typically estimated in the literature by using indirect measures. only few patents produce high economic returns. To obtain a measure of the present value of the patent we asked inventors to give their best estimate of the value of the innovations that they contribute to develop. Figure 3.8 % have a value higher than 3 million Euros. Although these indirect measures are useful when the actual monetary value of the innovations cannot be observed directly. Pakes. Harhoff et al. 2004).5 This Section reports the patents’ monetary value obtained through the PatVal-EU survey. However. which is 6-7 years after the application year of the latest patents in the survey (1997). other organisations. 2005).4% has a value between 1 and 3 million Euros. 1988). In particular. 1984. The Tables below show the distribution of the value of European patents by country. the largest share of patents falls in the left-end of the distribution. 1999). they have a number of limitations (Griliches. Another concern we had is that the inventors may not be the most informed respondents about the value of the patents. 1999a. and 16. and its determinants. 1990. Very few studies use survey-based information on the economic value of patents in specific countries (Harhoff et al. More precisely. 2003. whether the firm. see European Commission. 1999a. and Scherer and Harhoff 2000) only 7. A manager.2 % of the patents in our sample are worth more than 10 million Euros.e. 2004). Schankerman and Pakes. up to patents that are estimated to produce more than 300 million Euros. Since we were aware of this problem. or the inventor himself. Multiple measures are also employed to construct composite indicators of the quality of patents (Lanjouw and Schankerman. for a survey see Hall et al. This is a sufficient time span for a good deal of information about the use and value of the patents to become available. and the number of opposition and annulment procedures incurred by the patents (Harhoff and Reitzig. the renewal fees paid by the patent holders to extend the patent protection (Pakes and Schankerman. As the review of the literature pointed out. A share of 15. Scherer and Harhoff. technological class. 2000). for example. 2001). 1986. and type of organisation in which the inventors were employed at the time of the invention.. inventors were asked to estimate the minimum price at which the owner of the patent.. 1999b. the number of backward citations to other patents and to the non-patent literature (Harhoff et al. especially in the case of large firms would be a more suitable person to ask. we monitored whether the inventors actually knew about the value of their invention (for details and for the tests we performed on this issue. ranging from those that are worth less than 30 thousands Euros. About 5 There could be differences in the amount of information available to the inventors about the patent value. the questionnaire was answered in 20032004. the number of citations that patents receive after their publication (Trajtenberg. these indicators include. We constructed 10 classes for the value of the patents. To improve the precision of this “best estimate” we asked the inventor to assume that he/she had all the information available at the moment in which responded to the questionnaire. the number of countries in which the patent is asked for protection. However. This Section focuses on the monetary value of European patents. Our PatVal-EU survey provides new data on the monetary value of patents. Consistently with the well-known skewness of the distribution of the patent value (i. 1986). the inventors might have less information for more recent patents.1 shows the value of the PatVal-EU patents in each of the six European countries involved in the survey. would have sold the patent rights on the very day in which the patent was granted.

00% 25.1 The value of European patents across the EU countries 35.68%. Spain. Germany and France are in the bottom of the list with 5. 8. The share of patents with a value higher than 10 million Euros is 12. The value of each patent in the sample is approximated by the mean value of the monetary class in which it falls.00% 15.79% in Spain. Italy follows with 7.00% 5. This is due to the additional data cleaning and coverage of missing observations that we performed after we delivered the final PatVal-EU report to the European Commission.00% Share of patents 20.58 % of patents with a value higher than 10 million Euros.1. Symmetrically. the UK (53.1 reports the average monetary value of the patents invented in each country along with the standard deviations.00% 0. 27 10 0 30 m -3 >3 0 k1 0m m m m .12% in the UK.00% 10. Figure 3.86% in the Netherlands and 11. Italy comes fourth with 67.19% and 5. the average monetary value of the patents is higher in Spain. Consistently with Figure 3.83%. the Netherlands. and about 8 % have a value lower than 30 thousand Euros.04 %) compared to the EU6 share (67.8%). the Netherlands and the UK have a share of high value patents larger than the EU6 share. the share of patents whose value is lower than 1 million Euros is lower in Spain (54.73%) and the Netherlands (61. while Germany (74. and the UK compared to the average EU6 value.9%) have the largest share of comparatively lower value patents.00% 30.00% <3 0k 00 k 00 k 3m -1 0m k1 m -1 00 1m -3 00 m 0m 10 0k -3 10 m -3 30 30 0 Monetary value: classes DE ES FR IT NL UK 6 Some statistics presented in this report might slightly differ from those presented in the PatVal-EU report.95%) and France (74. Table 3.68% of all our patents produce less than 1 million Euros. Italy.6 At the country level.13%).

20) 11.Table 3.80) 8.7 Table 3.248 NL 6.63 (1.318) 0.542 Total 6.008 (21.610) 0.63 citations on average compared to 0. Italy and France follow with an average of 0.397) 1.97) 6. but when we perform specific tests on the ranking correlation between the three indicators.e.675 (47.e.6 in Annex AIII.91 of the EU6 average. This is probably due to the presence of large companies in Germany that because of their large patent portfolios tend to cite their own patents more than smaller companies. Leicestershire. The average value of patents is highest for a group of regions in UK (i.46) 9. Table A. Piemonte e Lombardia).33 (1.02) 4.124 UK 9.010) 0.49% 267 FR 3. This is so both at the country level and at the NUTS2 level. The number of observations shown in this Table refers to the number of forward citations and opposed patents. citations and oppositions for the regions with an average patent value below the EU6 average.63% 1. Rutland and Northants.10% 1.49% 1. Friuli-Venezia Giulia.407) 0. and the average number of citations received by the patents after their publication (i.60% 9. Table 3.486 IT 10.63 (2.98% 3.000) 0. and we used it to assign each innovation to the European region in which it was invented. the results are inconclusive. For the average patent value the number of observation is 7.20 (0.89 (1.754.358 (30. 7 28 . France and Italy compared to the other countries and to the EU6 share.1 complements Table 3.640 (16. Also the share of opposed patents is higher in Germany.767 (28. Outer London.210 (36. The geographical classification used for this exercise is the Nomenclature des Unités Territoriales Statistiques at the second level of disaggregation (NUTS2) as provided by Eurostat.2 lists the European regions whose patents have an average monetary value higher than the overall EU6 average.e. regions (NUTS2) and provinces (NUTS3). Cheshire) and for a group of Italian regions (Toscana. Surrey and Sussex.64) 3. The research in the second part of this project (i. Cataluña in Spain is the third region in this rank.1 also shows the share of patents that have incurred in opposition procedures. The average number of citations received by the patents after their publication date is particularly high for patents invented in Germany where they receive 1. Lot2) will specifically look at the correlation between different indicators of the patent value by using multiple correlation analysis that controls for many other We calculated the statistics at the regional level only for the NUTS2 regions in which 30 or more patents of our sample were invented.22) 10. For some of these regions the average number of forward citations and the share of opposed patents is also above the average. We also collected information on the address of the inventors of our PatVal-EU patents.2 by showing the average monetary value.049 (58.63 citations per patent.89 and 0.013 Note: standard deviations in parenthesis. The Nomenclature des Unités Territoriales Statistiques (NUTS) is a Eurostat classification that subdivides the European Union in groups of regions (NUTS1). Kent.27 (0.629) 0. while France appears with only one region with an average patent value above the overall EU6 average.1 The value of European patents across the EU countries DE Average patent value Average number of forward citations Share of opposed patents Number of observations 4. forward citations).346 ES 16. It also reports the average number of forward citations and the percentage of patents that incurred in litigation procedures.71% 1. Germany and the Netherlands are listed with a few regions in the lower part of the ranking.91 (1.

Table 3.49 (0.47 (52816.33 (0.30) 0. This Table shows the European regions in which the average value of patents is above the EU6 average.21) 9070.03) 12445.45 (1.70% 1. This classification is based on the International Patent Classification (IPC) and distinguishes among 30 “micro” technological fields and 5 “macro” aggregated technological areas.80) % Opposed patents 4.08% 15.12 (1.00) Average Forward Cits (st dev) 0.24 (0. just below the EU6 average.79% 3.754.60% in The technological classification uses the framework elaborated by the German Fraunhofer Institute of Systems and Innovation Research (ISI) together with the French patent office (INIPI) and the Observatoire des Science and des Techniques (OST).11% 10.18 (0.05 (3.66 (2.71%) compared to the overall sample (7.42 (1.50) 7077.27% 1. Rutland and Northants Cheshire Friuli-Venezia Giulia Piemonte Lombardia Bedfordshire. The correspondence between the “macro” and “micro” technological classes is reported in Table A1 in Annex III.45% 2.44) 8556.81) 0.49) 1.71) 13263.03% 5.51% 8.47) 0.91 (1.82% 7.59 (22431.00% 3.26) 8016.24 (0.35 (30407.36) 0.22% in Electrical Engineering. 6.02 (44760. We classified our sample patents in five “macro” technological classes: Electrical Engineering.89 (49158. 49 57 116 44 93 37 53 76 205 424 44 33 127 37 194 337 147 46 62 90 88 114 58 51 9013 Note: the number of observations reported in the Table refers to the patents we used to calculate the average number of forward citations and oppositions.53 (35023. 8 29 .factors that potentially affect the value of the European patents.45 (40003. Process Engineering.48) 0.23% 3.11) 12772.77 (34892.23 (48821.03 (1.64% 12.34 (52226.69) 10720.73 (56739.04) 14616.56 (2.69) 2.83 (1.28 (0.59 (51165. Process Engineering comes second with a share of 6.49 (21434. east and west Sussex Leicestershire.23%).47 (50512.45) 14573.90) 0.82) 0.00% 3.70) 1. The share of topvalue patents is 6.89) 0.40) 1. and the number of observations in each region is ≥30. Instruments.62% 0. Innovations that are worth more than 10 million Euros are more frequent in Chemicals and Pharmaceuticals (11.79 (37445.89 (43238.10% in Mechanical Engineering and 5. Bucks and Oxfordshire Gießen Rheinhessen-Pfalz Noord-brabant Karlsruhe Derbyshire and Nottinghamshire West Yorkshire Freiburg Hannover Limburg Provence-alpes-Côte d'azur West Wales and the valleys Average value (st dev) 25829.97) 6751.95) 0.80% 9.85) 8313.08% 2.89 (76980.53) 1.65 (37321.22 (52056. such as the characteristics and the education of the inventors.59) 10391.67% 0.51) 0.35 (1.99) 10336.24 (72344.3 shows the distribution of the patent value in these five macro-technologies.46) 7002.08) 0. the company specificities and the technological classes in which the patents were invented.33% 6. Chemicals-Pharmaceuticals.25 (0. and Mechanical Engineering.38 (1.29) 0.86 (51231.89% 8.8 Table 3.89 (49308.93) 0. For the average patent value the number of observations is 7.05) 0.49) 11779.06) 6358.60% Number of obs.64) 12037. Hertfordshire Eastern Scotland Berkshire.09) 1.42 (38528.09 (0.17) 8886.31% 5.35 (0.89% 22.75% patents that are worth more than 10 million Euros.2 The value of European patents across European regions (NUTS2) Country NUTS 2 UK IT ES UK UK UK UK IT IT IT UK UK UK DE DE NL DE UK UK DE DE NL FR UK Total Kent Toscana Cataluña Outer London Surrey.16 (0.84) 0.65 (1.56 (23305.59 (18498.74) 14197.81) 10641.24% 0.31) 11660.00% 8.73) 19727.21 (0.08 (0.23) 1.98) 6369.06 (0.37% 7.02) 0.73) 0.

Petrol & Basic Material Chemistry (12.74%.82) 6. Macromolecular Chemistry & Polymers) are in the 30 .474) 0.3 The value of European patents by “macro” technological class Value intervals <30k 30k-100k 100k-300k 300k-1m 1m-3m 3m-10m 10m-30m 30m-100m 100m-300m >300m Total Average patent value Average number of forward citations Share of opposed patents Number of observations Electrical Engineering Instruments Chemicals & Pharmaceuticals Process Engineering Mechanical Engineering Total 9.90%).425 8.75) 11. and reports the distribution of the patent value in each class.10% 2.70% 22.40% 3.65% 21.86 (1. The average value of a Chemical-Pharmaceutical patent is much higher than the average value of a patent in any other technological class.93% 3.00% 9.00% 6.784 (28.04% 3.75% 0.57% 2.3 other two chemical technologies (Agriculture & Food Chemistry.075) 1.88% 17.00% 5. the share of opposed patents is the highest in Process Engineering (11. Consistently. Electrical Engineering. For the average patent value the number of observation is 7.72% 100. Semiconductors (12.20% 14.62% of patents that incurred in opposition procedures. Again.58% 22.36% respectively).73% of opposed patents.00% 0.91 (1.04 (2.Instruments.14% 22.359 (30.581 (37. Differently.578 (24.93% 9. Instruments.08% 7.39% 20.48) 8. Process Engineering and Mechanical Engineering have a share of about 70% of patents with a value lower than 1 million Euros (72.863) 0. These data are mirrored in the average monetary value of the patents. Table 3.57%).31% 0.78% 18. Symmetrically.27% 15.61% 18.87 (1.64% 5.21% 1.36% 12.21% 21.79% 0. the technological sectors with the highest share of patents in this class are: Pharmaceuticals & Cosmetics (17.48%). Table 3.65%.54%).07%).14% 14. 68.84 (1. 58.14% 100.690 7.80) 8. and Material Processing.410) 0.60% 9.248 8.85% 10. and Mechanical Engineering is third with a share of 10.92% and 70. This is so also for the average number of forward citations (1.41% 100.61% 0.99% of opposed patents. The number of observations shown in this Table refers to the number of forward citations and opposed patents.33% 1.45% 9. Electrical Engineering and Instruments are at the bottom of the list with 5.119) 0. Chemicals-Pharmaceuticals follows with 8.38% 18.68% 1.00% 4.48% 100.91 of the overall EU6 average).74% 0.47% 2. Textile & Paper (9.69% 1. If we consider the innovations that are worth more than 10 million Euros.69) 5.57% 3.00% 21.68% 1. Chemical. Organic Fine Chemistry (13.00% of the Chemical and Pharmaceutical patents are in the left-end tail of the distribution where patents generate less than 1 million Euros.809 (24.354 (31.48% 0.00% 6.07% 9.013 Note: standard deviations in parenthesis.94 (1.04 compared to 0.70% 2.58% 3.76% 0.66% 0.43% 1. there are technologies in which the probability of inventing valuable patents is higher than in others.4 shows the technological classification of our patents in 30 “micro” technological classes. with Table 3.72% 0.55% 21.91% 18.71% 18.80% 15.73% 981 6.77% 100.886) 0.00% 4.88% 21.41% 20.81%).90% 100.40% 1. 67.752.669 6.93% 17.64% 11.34) 8.68% and 6.49% 14.69% 20.99% 1.

Agricultural Machinery and Processing (74. large chemical and semiconductor companies have larger patent portfolios that can cite their own patents. Textile & Paper.60%).17%). Table 3.top positions. Macromolecular Chemistry and Polymers follow with 1. and Petrol & Basic Material Chemistry are ranked in the top positions. The lowest shares of low value patents are in Pharmaceuticals and Cosmetics. and Agriculture & Food Chemistry. Compared to the European overall share of patents that incurs in opposition procedures (8. By means of cross-licensing these firms share their patent portfolios and avoid costly litigations. Semiconductors are also the top sector in terms of the average number of forward citations (1.e. patents that generate less than 1 million Euros) the share of patents in this class is high in Consumer Goods and Equipment (76. below the overall European average (7.87%) and Surface Technology and Coating (11. In Biotechnology. As far as the left-end tail of the distribution is concerned (i.16% patents being opposed (i. This might depend on the fact that citations include also selfcitations. Materials and Metallurgy. biotechnology stands in the lower ranking: the share of patents that are worth more than 10 million Euros is 3. Compared to smaller firms in biotechnology. Pharmaceuticals & Cosmetics. The low share of opposed patents in Semiconductors might be explained by the frequent use of cross-licensing agreements among the large semiconductor firms. Table 3. i. citations made by patents developed by the very same company that applied for the cited patent.89 of the overall European average).56%). Material Processing. Unexpectedly. Semiconductors are at the bottom of the list with only 1.06%).5 shows that the share of opposed patents is larger in Materials and Metallurgy (16.83%). 31 .33 forward citations on average. one patent).e. Agricultural & Food Processing (12.5 shows the average monetary value of the PatVal-EU patents in the 30 “micro” technological classes. there is a large share of patents with a value between 1 and 10 millions Euros (about 29. Semiconductors.67%).23%).69%).35%).e. Organic Fine Chemistry.5%.21%).38 compared to 0. Macromolecular Chemistry & Polymers (13. however. Biotechnology patents receive the lowest average number of citations (0. Electrical Devices (75.23). Patents in Organic Fine Chemistry. Material Processing. Textile & Paper (13.

62 12.50 13.51 0.45 16. Machinery & Apparatus Transport Nuclear Engineering Space Technology Weapons Consumer Goods & Equipment Civil Eng.60 7.00 0.59 17.41 5.76 1.25 0.22 11.75 10.36 0.03 0.92 16.92 18.28 4.70 5.19 0.55 20.33 3.56 0.00 2.42 6.00 100.00 24.63 1.14 24.30 12.51 0.63 0.00 100.00 100.49 19.35 3.28 4.88 5.76 2.39 14.85 0.67 22.92 1.84 9.76 4. Building & Mining Total <30k 9.28 2. & Food Chemistry Chemical & Petrol Industry.42 1.23 25.03 12.90 16.80 8.92 3.90 13.00 100.77 18.71 20.40 14.81 5.94 6.94 16.69 4.97 3.00 12.45 3m-10m 6.75 28.00 100.49 6.86 19.65 14.88 5.93 9.89 7.79 17.90 5.00 100.00 100.49 5.31 9.34 19.68 7.60 0.75 3.24 12.00 100.36 24.77 26.71 21.00 100.06 10.84 15.81 20.85 22.40 24.73 9.71 12. Textiles & Paper Thermal Processes & Apparatus Environmental Technology Machine Tools Engines.00 100.02 21.08 20.00 3.00 100.95 6.25 1.99 22.21 19.94 13.69 17.20 9.68 17.67 0.87 0.00 100.24 2.65 0.00 100.92 17.00 19.29 21. Basic Materials Chemistry Chemical Engineering Surface Technology & Coating Materials Processing.36 20.32 13.17 7.15 21.69 4.00 2.00 100.76 8.61 1.71 14.00 100.95 16.44 3.00 100.82 2.48 6.07 8.43 6.45 2.42 23.87 1.91 0.91 0.00 100.54 19.36 18.30 7.48 1.82 11.00 100m300m 0.46 4.26 10.80 4.53 21.Table 3.48 21.26 13.80 4.15 0.86 1.00 0.93 10.79 1.33 5.16 0.58 10m-30m 2.00 100.56 2.70 1.76 0..43 0.00 0.59 27.20 28.53 16.87 11.82 15.71 14.42 11.26 11.47 15.50 11.07 8.39 9.17 2.77 14.75 22.51 5.46 17.35 0.07 24.73 0.22 21.75 0.30 7.43 15.00 0.00 1. Electrical Eng.01 0.56 11.82 18.74 0.29 3.13 12.01 25.75 2.4 The value of European patents by “micro” technological class.44 15.32 2.47 21.00 100.00 0.60 14.64 2.48 2.70 30m100m 0.00 100.08 5.07 17.13 19.66 0.43 9.08 15.20 4.05 1.00 12.46 0.00 0.39 0.67 21.51 8. Pumps & Turbines Mechanical Elements Handling & Printing Agricultural & Food Processing.09 20.08 1.10 1.00 0.71 10.21 17.89 15.00 100.23 14.57 16.99 18.80 0.73 0.60 1.67 0.28 0.00 32 .00 100.04 21.59 0.61 1.67 3.61 16.81 12.64 9.12 19.81 1.03 18.30 2.20 20.59 8.05 17.76 0.77 Total 100.42 7.66 28.88 30k-100k 19.78 22.48 15. Measurement.00 100.88 0.78 17.87 12.00 2. & Electrical Energy Audio-visual Technology Telecommunications Information Technology Semiconductors Optics Analysis.00 100.93 11.09 0.00 0.36 21.82 1.45 5.82 4.00 100.83 2.21 7.93 10.00 100.55 2.00 100. & Control Technology Medical Technology Organic Fine Chemistry Macromolecular Chemistry & Polymers Pharmaceuticals & Cosmetics Biotechnology Materials & Metallurgy Agriculture.33 14.16 0.74 9.00 12.57 21.00 0.85 0.64 1.78 10. in % ISI Technological Classes Electrical Devices.55 0.59 0.39 100k300k 24.41 21.25 13.16 24.03 18.83 26.96 20.05 7.00 100.41 1.26 14.67 3.00 100.48 10.99 14.65 300k-1m 22.39 1.05 10.60 4.48 22.77 20.33 9.23 10.09 26.62 21.00 1.53 18.95 21.88 2.80 1m-3m 13.76 >300m 0.25 20.53 19.00 0.39 18.35 6.56 1.31 16.30 2.52 0.38 22.72 0.00 0.08 19.00 100.25 6.

41% 4.72) 1.82% 12.520 (26.43) 0.51%) and in the “other” residual category (10.91 (1.22) 0. private non-profit research institutions (9. 678 176 286 199 86 169 537 233 549 463 172 70 308 115 300 290 136 487 191 145 317 256 390 680 202 627 42 57 468 384 9013 Note: standard deviations in parenthesis. Machinery & Apparatus Transport Nuclear Engineering Space Technology Weapons Consumer Goods & Equipment Civil Eng.360 (30.89% 8.280 (23.66 (1. & Food Chemistry Chemical & Petrol Industry.45 (1.16% 5.87 (1.808) 5.962) 3. Table 3.352 (33.345) 4.990 (30.473) 4.815) 7.812) 4. Building & Mining Total Average value (st dev) 3.95) 0.26% 3.00%) compared to the overall European share (7.36% 7.69% 8.00%).99 (1.67) 0. small firms.370 (23.63% 4.204 (28.903) 6.74 (1. public research organisations.302) 8.56% 10. government laboratories.77) 0.17) 0. Pumps & Turbines Mechanical Elements Handling & Printing Agricultural & Food Processing.79 (1.752.81) 0.46% 8.586 (22.103 (39.98 (1. private non-profit research institutions (like Foundations and Hospitals).217) 3.91 (1.17% 11.01 (2.88 (1.62% 11.62) 0. The Table reports the distribution of the patent value for each type of inventors’ employer. Textiles & Paper Thermal Processes & Apparatus Environmental Technology Machine Tools Engines.95 (1.603) 5.082 (33.945) 1.448) 5.53) 1.929 (33.619 (27.75) 0. and other employers like the inventors themselves or other institutions.5 The value of European patents: average value.411) Average Forward Cits (St dev) 0.80 (1.81 (1.597 (26.15% 4.52) 0.67% 8.54% 2.72) 1.019 (34.07 (2.42% 8.409) 12.771) 7.118) 4.50) 0. universities.82 (1.64) 0.47% 11.98) 1.25%). Basic Materials Chemistry Chemical Engineering Surface Technology & Coating Materials Processing.77 (1.754) 6.875) 5.806 (24.Table 3.73) 0.22%).. & Electrical Energy Audio-visual Technology Telecommunications Information Technology Semiconductors Optics Analysis. Public research organisations and universities develop the smallest shares of very high value patents 33 .80) % Opposed patents 8.29% 16.6 lists different types of organisations in which the PatVal-EU inventors were employed at the time of the invention: large firms.33% 6.52) 1.376 (9.27 (0.498 (37.98 (1.73) 0.090 (18.802) 8. The number of observations shown in this Table refers to the number of forward citations and opposed patents.51 (1.59) 0.830) 7.345) 7.38% 3.07 (1. For the average patent value the number of observation is 7.07% 8.559) 8.112 (40.610) 11.40) 0.808) 4.60% Number of obs.04) 0.84 (1.520) 6.51% 7.908 (27.76% 13.02% 1.25) 0.471 (18. Measurement.230 (48.489 (17.058 (29.04) 1. medium-sized firms. & Control Technology Medical Technology Organic Fine Chemistry Macromolecular Chemistry & Polymers Pharmaceuticals & Cosmetics Biotechnology Materials & Metallurgy Agriculture.74 (1.35% 10.42) 1.988 (45.791 (3.562) 12.01% 13.478 (44.09 (2.707) 8.722 (34. The share of innovations that are worth more than 10 million Euros is higher for government laboratories (10.52) 0.68 (0.589 (24.93) 0.87% 6.67 (1.72% 11.683) 9.20% 3. Electrical Eng.477) 4.36 (1.163 (19. large companies (7.11 (1.87 (1.43% 10.251 (38.76) 0.88) 0.95 (1. share of opposed patents and average number of forward citations by “micro” technological class ISI Technological Classes Electrical Devices.35 (2.583) 5.371) 5.98) 0.

00 6.90% 775 4.00 5.00 10.53 20. As expected. Table 3.00 16.77 17.17% 59 7.00 100.84 15.85 (1.96 3.00 1. The cost of patents Consistently with the literature on the value and social costs of patents. in % Large companies <30k 30k-100k 100k-300k 300k-1m 1m-3m 3m-10m 10m-30m 30m-100m 100m-300m >300m Total Average patent value Average number of forward citations Share of opposed patents Number of observations Medium sized companies Small companies Private Research Institutions Public Research Institutions Universities Other Governm.46 0.00 100.49 24.81 23.00 0.61) 0.486) 0.73 20.67 8.00 8.22 11.10% 1211 11.00 6.27 0.37 22.641 (27.37 16.47 1.00 5.36 20.48% 67 10.99) 9.99 0.68% and 4.00 20.98 22.23%).30) 8.79 100.41) 5.015 (33. This ranking is mirrored in the average monetary value of the patents in the first raw of the bottom part of Table 3.17 100. are opposed more than the European average.67 20.87% 284 20.69 1.38 0.00 0.6.759) 0. government laboratories and private non-profit research institutions develop only 83 patents in our sample.67 0.70 1.65 2.70% 8808 Note: standard deviations in parenthesis.00 6. 34 .36 (0.38 9. However. patents invented in private companies. public research institutions and private non-profit research institutions have a lower probability of being opposed compared to the overall EU6 share of opposed patents.08 (1.14 15.6 The distribution of the value of European patents by type of inventors’ employer.15 16.73) 10. followed by large companies and the “other” residual category. Institutions Others Total 8.11 12.37 9.04 25.07 19.00 14. The number of observations shown in this Table refers to the number of forward citations and opposed patents.53 8.77 100.51 (1. and that the social loss of 9 However.00 1.52% 181 8.92 17. a system of strong intellectual property protection that leads to patent oppositions.851) 1.00 10.54 3.25 14.29 (0.33 12.00 10.43 17.00 22.77 15.27 100. also implies that the enforcement of property rights is costly. and especially in large companies. from a policy perspective.561 (25. Inventors employed by private firms also invent highly cited patents.051) 0.67 0.894 (34.00 28. as they have large citing patent portfolios.87 0.00 6.12) 7. 2004).00 100. By contrast.73 11.64 0.68 1.00 2.37 (0.681) 0.00% 14 6. universities.03 4.65 1.655) 0.50 1.81) 8.(4.00 14.00 4. For the average patent value the number of observation is 7. the share of opposed patents is considered as an indicator of the value of patents (Harhoff and Reitzig.26 3.291) 0.04 9.00 0.85 3.00 30.52 (1.59 20.64 (1.00 8.00 2.64 100.9 Patents invented in government research institutions.587.70 0.970) 0.00 0.00 16. suggesting a lower social cost of patent protection.92) 4.92 (1.96 20.76 0. The most valuable patents are developed by government laboratories and private non-profit research institutions.05 19.73 21.21 3.35% 6217 6.00 12.67 100.523 (46.577) 0.00 0.775 (27.402 (30. this is particularly matching for patents developed by large companies.00 21.22 0.615 (30.511 (29.94) 3.67 0.11 12.67 2.

00% 5. while only 5. Figure 3. “macro” and “micro” technological classes. these costs are unevenly sustained by different actors (individuals vs.63% of patents involve more than 48 months (the overall EU6 shares are respectively 52.2 shows that for the overall EU6 countries. firms. The man-months required by the patents’ invention process across countries 30. 52. We analysed the length of the research process by country. and that large and medium companies incur in patent suits more than small companies.00% 0. Table 3. This section uses the amount of time devoted to the innovation process that leads to the patent as an indicator of its monetary cost. At the country level. Precisely.25% of the patents require up to 6 man-months for the invention process to be completed. firms with different characteristics).40 % of the patents require more than 48 man-months. By contrast. Figure 3.00% <1 1-3 4-6 7-12 13-24 25-48 49-72 >72 Man Months DE ES FR IT NL UK Total At the level of “macro” technological classes.00% 20. In Chemical and Pharmaceutical technologies only 36% of the patents involve less up to 6 man-months for the invention process and 9. The Tables in the previous section showed that the share of opposed patents differs across countries. 2005).56% of patents in Germany entails less than 6 man-months.00% 25. The large share of short research projects that lead to a patented invention suggests that there is a lot of patenting concerning “small” innovations that may actually be part of a group of intertwined patents developed within the same organisation. ranging from 1 man-month (or less) up to more than 72 man-months.resources due to litigations in legal disputes is relevant (Cesaroni and Giuri. private and public research institutions or individuals.00% 15.7 highlights interesting differences in the amount of resources requested by the invention process leading to the patent. in Electrical Engineering and Mechanical Engineering the share of patents requiring less than 6 months is about 60% while the share of 35 . regions and technological classes. 66. In all the other countries the share of patents that takes more than 6 man-months to be invented is larger than in Germany.40%). large firms vs. we constructed 8 categories with the number of manmonths required by the research project leading to the patent.25% and 5. and types of inventors’ employer.00% 10. Furthermore.2. small firms.

57% 24.8 shows that the share of patents whose research process requires less than 6 man-months is larger than the EU6 average in Electrical Devices.45%).12% 3.16%) and Civil Engineering. By contrast.09% 17.43% 2.00% 13.67% 13. the patents invented in small firms are skewed towards a large number of man-months required to develop them as compared to large firms’ patents.17% 2.52% 5.7 The man-months required by the patents’ invention process by “macro” technological class Man-months <1 1-3 4-6 7-12 13-24 25-48 49-72 >72 Total Electrical Engineering 15.56%).00% Total 12.25% of the EU6 average.28% 100. By contrast.37% 5.28% 2.43%.57% 17.13% 21. This also suggests that “small” patents are less frequent in small firms.81% 18.26% 13.27%).96%) and Environmental Technology (10. It ranges between 11.51% 19.88% 19.62% 20. in private and public research institutions.87% 100. Handling & Printing (65.22% 15.66% 1. Table 3.36% 3.04 and 21.00% Mechanical Engineering 16. the patents produced by research institutions and universities are the output of larger research projects compared to those invented in the private sector. large.19%).89% 100. Transport (62. Materials & Metallurgy (11. on average.47% 14. The distribution of patents across categories in Instruments and Engineering are instead closer to the EU6 average. on average.42% 100. and research institutions and universities on the other hand.37% 1. Table 3. the technological classes where a high share of patents requires more than 48 months for the invention process to be performed are Organic Fine Chemistry (12. small and medium companies) is similar to the overall distribution. Moreover.40% 9. Pharmaceuticals & Cosmetics (10.31% 17.27% 12.47% 18. The number of man months required in the private sector to develop a patent (i.18%). Agriculture & Food Chemistry (12.10% 3.37% 19.69% 5.9) shows interesting differences among companies on the one hand.26%.48% 24.56% 8.94% 100.patents requiring more than 48 months are respectively 3. This suggests that. the share of innovations produced in less than 6 man-months is sensibly lower than the EU6 average.12% 18. Mechanical Elements (66.70% and 3.31% 1. universities and other government organisations. Biotechnology (12.71 %.53%).21%). The distribution of patents across the 8 man-months categories and type of inventors’ employer (Table 3.58% 2.52%). Electrical Engineering & Electrical Energy (63.27% 16. 36 .94% 3.00% Chemical & Pharmaceutical tech.00% Process Engineering 13.78% 19.17% 100.71 to 35.31%).02% 20. Building & Mining (63.83% 17. Telecommunications (60.43% 18.00% At the level of the 30 “micro” technological classes.72%). while the share of patents that require more than 48 man-months is far above 5.25% 20.19% 12. 3. It ranges from 25.e.00% Instruments 12. in which the legal costs of preparing and applying for a patent is often greater than in large firms.

28% Total 100. Textiles & Paper Thermal Processes & Apparatus Environmental Technology Machine Tools Engines.95% 1.56% 3.05% 21.54% 18.18% 12.63% 17.59% 20.37% 18.68% 2.56% 20.42% 16. & Food Chemistry Chemical & Petrol Industry.99% 3.45% 2.43% 13.04% 16.04% 12.24% 12.00% 100.00% 100.50% 1.70% 19.17% 18.17% 18.00% 100.23% 16.00% 100.00% 100.79% 20.48% 20.33% 8.13% 24.00% 100.51% 0.00% 100.09% 2.79% 1.08% 1.30% 0.50% 21.20% 16.00% 100.33% 18.50% 0.78% 6.85% 5.90% 4.00% 100.25% 14.06% 14.86% 12.83% 16.84% 2.62% 2.10% 1.04% 14.20% 26.78% 18.84% 14.14% 12.21% 5.29% 15.51% 16.85% 25.00% 100.86% 15.94% 20.50% 25.00% 100.92% 1.00% 20.97% 2.80% 2.41% 7. Machinery & Apparatus Transport Nuclear Engineering Space Technology Weapons Consumer Goods & Equipment Civil Eng.22% 13-24 13.42% 2.50% 17.10% 1.04% 21.70% 12.25% 18.19% 3.55% 18.00% 100.50% 5.48% 1.23% 18.14% 5.50% 7.68% 20.69% 17.00% 100.26% 0. Building & Mining Total 17.11% 10.82% 17.68% 16.03% 9.00% 100.38% 17.13% 14.85% 20.54% 2.8 The man-months required by the patents’ invention process by “micro” technological class <1 Electrical Devices.55% 4.69% 7.89% 6.58% 4.97% 15.69% 19.31% 26.61% 1.70% 7.49% 12.82% 22.00% 100.95% 15.12% >72 1.58% 19.71% 17.35% 9.01% 10.01% 16.58% 12.46% 14.00% 18.71% 18.01% 10.74% 20.95% 20.97% 3.00% 100.03% 18.03% 18.88% 14.81% 16.86% 11.81% 24.99% 19.75% 17.00% 1.78% 0.50% 7.19% 9.01% 25.18% 14.29% 20.27% 12.66% 22.94% 2.67% 22.58% 15.27% 3.95% 17.77% 13.49% 18.27% 19.00% 100.47% 2.77% 16.25% 2.99% 20.00% 16.02% 2. Electrical Eng.33% 3.21% 14.21% 1.66% 6.00% 100.07% 9.34% 18.60% 18.09% 12. & Electrical Energy Audio-visual Technology Telecommunications Information Technology Semiconductors Optics Analysis..07% 21.34% 12.00% 100.07% 15.08% 5.53% 24. Basic Materials Chemistry Chemical Engineering Surface Technology & Coating Materials Processing. Measurement.59% 7.58% 15.39% 12.98% 10.19% 17.84% 14.23% 7.91% 15.00% 100.71% 10.00% 100.49% 4.84% 7.99% 3.41% 2.18% 13.86% 6.08% 4.57% 12.58% 49-72 1.62% 15.18% 19.79% 14.35% 15.81% 27.48% 4.50% 18.00% 100.57% 1.51% 4-6 22.00% 16.12% 7-12 16.17% 5.40% 6.00% 100.62% 1-3 23.19% 11.66% 19.46% 4.05% 2.37% 2. Pumps & Turbines Mechanical Elements Handling & Printing Agricultural & Food Processing.38% 21.00% 2.96% 3.52% 21.55% 1.20% 13.56% 5.00% 37 .84% 18.15% 2.35% 25.46% 2.14% 4.56% 25-48 4.54% 18.57% 18.74% 19.58% 12.00% 100.14% 12.37% 0.14% 15.57% 20.00% 26.01% 12.88% 13.71% 25.49% 12.74% 21.05% 13.29% 12.00% 100.67% 9.75% 22.00% 100.27% 0.00% 10.46% 14.55% 20.38% 16.10% 1.94% 1.21% 10. & Control Technology Medical Technology Organic Fine Chemistry Macromolecular Chemistry & Polymers Pharmaceuticals & Cosmetics Biotechnology Materials & Metallurgy Agriculture.57% 4.17% 9.70% 14.00% 100.00% 100.15% 10.75% 24.28% 23.83% 20.36% 19.14% 13.42% 12.00% 100.55% 4.14% 4.Table 3.00% 100.27% 27.89% 13.

2001).00% Other Governm.00% Medium sized companies 11.48% 1.50% 21.75% 2. and the motivations for patenting.49% 22. and the evidence on whether patents are used internally by the applicant company for economic purposes or for strategic reasons.90% 22.40% 19.07% 5. 2003).00% Private Research Institutions 4.Table 3.82% 1.13% 4.14% 100.61% 15.91% 100.93% 19. however.29% 14. Nakamura and Odagiri.84% 22.68% 100.29% 14. and type of applicant organisations.00% 21. 2000. 2002. 2005.54% 6. Only a few studies analyse these issues more broadly by using US and Japanese data.3 The Economic Use of Patents The literature on the economic use of patents suggests that large firms tend to under-use their patents for strategic reasons.00% Annex A III.00% Public Research Institutions 6.84% 14. and the number man-months required for developing the innovations according to the number and type of countries in which the applicant organisations are located. 38 .16% 21. Hall and Ziedonis. The empirical work on these topics. In general.00% Small companies 12.54% 19.91% 100.41% 100.64% 16.29% 21.64% 23. It can be used in the production processes or it is incorporated in marketed products.61% 20. and on the licensing activity of universities (Arora and Ceccagnoli.65% 8. This is the aim of the present Section that presents data by EU country. By providing the necessary information on these issues. technological classes.39% 2. firms.39% 100. or if the patents are simply “sleeping” within the companies. Institutions 14.08% 17.64% 14. al (2000) use US firm data to show the motivations for patenting by US companies.40% 17.9 The man-months required by the patents’ invention process by type of inventors’ employer Large companies <1 1-3 4-6 7-12 13-24 25-48 49-72 >72 Total 13. The patent is exploited internally for commercial or industrial purposes.36% 1.94% 100. The use of patents The PatVal-EU questionnaire defines six possible uses of patents: Internal use.00% 14.79% 7.86% 16.67% 9.00% Total 12.25% 100.66% 10.21% 17.75% 19.84% 8.57% 19.13% 11.00% Universities 5. Gans et al. Rivette and Kline. the PatVal-EU survey provides a unique opportunity to explore the different uses of European patents. Fosfuri and Gambardella.10% 3.43% 0.83% 10.14% 15.92% 12.95% 8.74% 14.45% 5. Licensing and the development of markets for technology are means to enhance the exploitation of patents (Arora.29% 0. Likely.13% 7. 2001. is limited.43% 100. This is particularly true for the work done on European patents and firms.49% 17. Cohen et.92% 9.50% 2.18% 10.95% 1.11% 2.22% 14.79% 24.10% 18. is very poor.1 adds five tables of descriptive statistics on the value and costs of patents.50% 20.75% 1.67% 20.00% Others 8. 3. information on the actual use of patents is not available. the average number of forward citations and oppositions. They describe the distribution of patent value.43% 18.56% 17. There are contributions on the extent and motivations for licensing in specific industries.

Rather. Also the share of licensed and licensed & used patents in France is lower than the EU6 average. patents are licensed and licensed & used more than in the overall EU6 countries.29% are “sleeping” patents.61%) and “sleeping” patents (8. Fifteen percent of the patents are exchanged in the market for technologies: 6. 18. The UK shows a distribution similar to the Netherlands.46%). and 3.48% and 2.62%) is larger than in the Netherlands. In Italy inventors patent mainly for internal use (55.83%). In Germany the share of sleeping patents is the largest amongst the EU6 countries (25. The patent is not used (neither internally. Only 37. and 17. About 35% are not used: specifically.52%) or for blocking competitors (23. In Spain. and it is also used internally by the applicant organisation. and the share of licensed patents (10. More than half of Process Engineering and Mechanical Engineering patents are used for internal purposes (54.62% respectively. By contrast.35%). Table 3. it is held unused within the applicant organisation in order to block competitors. Blocking competitors. The patent is licensed to another party in exchange for another innovation.10 shows the use of patents. nor for licensing).50% respectively).25%). and 22. The patent is not used internally by the applicant.inventors and applicant organisations that use the patent internally have the downstream complementary assets to perform production and commercialisation activities. Also the exploitation of patents in the market for technologies is less diffused in Italy than in the other countries. while about 50% of the patents in this class are held unused: 28. licensed & used patents (4.11 shows the distribution of the six patent uses by “macro” technological classes. Table 3. France has the largest share of internally used (64.38% are licensed. The patent is “sleeping” in the sense that it is not employed in any of the uses described above. Licensing & use.44%).62% and 56.97% are both licensed and internally used.24% are used for blocking competitors.60%) and cross-licensed patents (7.03% are used in cross-licensing agreements. 3.67%) and cross-licensed patents (3. Cross-licensing.10 also highlights the differences in the use of patents across the countries involved in the survey. In line with the existing evidence on licensing activities (Anand and Khanna.93% of Chemical and Pharmaceutical patents are used internally. half of the patents are used internally (50. The Netherlands patent less than the other countries for internal use. but it is licensed out to another party.49%). together with a large share of blocking patents (23.53%). Table 3. Only the share of cross-licensed patents (4.57%). Licensing. The share of cross-licensing is above the average in Electrical 39 .90%) are much smaller than the EU6 shares.69% are applied for strategic reasons.44% are “sleeping” patents. while the share of sleeping patents is the smallest amongst the EU6 (9. The patent is both licensed out to another party. At the overall EU6 level. Licensing & use and cross-licensing are instead used less than the average: 2. Sleeping patents. 2000) Chemical and Pharmaceutical patents are licensed frequently (6.57%). while the shares of “blocking” (11.21%) is the largest among the six EU countries. the Netherlands shows large shares of licensed patents (7.

74% 8. Hall and Ziedonis.29% 3.09% 4. Table 3.00% 100.57% 10.00% 4.40% 19.74% 14.00% 100. in Medical Technologies.97% Blocking Competitors 14.714 Table 3.35% 14.) in Pharmaceuticals & Cosmetics.711 Table 3.21% 47.09% and 4. seems to have underestimated the importance of patenting in Process Engineering and Mechanical Engineering: Table 3.24% 15.03% Licensing & Use 3.90% 9.00% 100.00% Number of observations = 7.50%): Agricultural & Food Processing.97% 18.44% Total 100.00% 100.00% 100.52% 47.11% 11.62% 56.61% 23.29% 15.29%).21% 6.44% 7.83% 4.52%).62% 3.35% 1.39% 18.52% 37.55%).79% 3. there are technologies in which the share of internally used patents is above the overall average (50.54% 5. and Civil Engineering more than 5% of patents are both licensed & used.48% 4.42% 5.97% 17.60% 55.28% 2.67% 3.Engineering and Instruments (6. Textiles.46% 23.45% 18.00% 100.39% 28.03% 7.44% 64.38% Crosslicensing 2.66% 50. Process Engineering Mechanical Engineering Total 49.00% 100. 40 .94% 5. the largest shares of licensed patents are in Biotechnology (18.95% 2.59% 52.57% 13. Pharmaceuticals & Cosmetics. 2000. Between 9% to 11% of the total number of patents in our sample are licensed also in Medical Technologies. and Environmental Technology.10% 3. and in Biotechnology.50% Licensing 3.25% 2.50% 50.09% 7.23% 3.30% 17. and Chemical Engineering (12.00% 100. Information Technologies and Semiconductors).81% 22.36% 12.38% 17. 2001). Moreover.13% 5.60% 8.10% 45.69% Sleeping Patents 18.85% 4. Thermal processes & Apparatus.36% 5.12 describes the distribution of the patent uses by “micro” technological classes.92% 9.11 shows that the share of unused patents in these technologies is the lowest among the 5 technological classes.25% 12. in Instruments (like Optics.11 The distribution of patent uses by macro technological class Internal use Electrical Engineering Instruments Chemical & Pharmaceutical tech.00% 100.10 The distribution of patent uses by country Internal use DE ES FR IT NL UK Total 49. As far as the internal use of patents is concerned. & Papers.38% Crosslicensing 6. Material Processing.95%).49% Licensing 4. Cross-licensing is frequent in Electrical Engineering technologies (like Audio-visual Technologies.59% 4.80% 6.00% Number of observations = 7.08% 6. Consistently with Table 3.93% 54. Materials & Metallurgy.11. followed by Agriculture and Food Chemistry (14.53% 23.02% Licensing & Blocking Use Competitors 3.00% 100.04% 1.85% 19. therefore confirming the findings by earlier contributions for the US firms in electronics and semiconductors (Grindley and Teece.69% Sleeping Patents 25. Surface Technologies & Coatings. Medical Technologies and Nuclear Engineering.62% 2.44% Total 100. 1997.00% 100.08% 2. The literature. Anand and Khanna. however.

due to financial constraints. with the largest share of blocking patents in Organic Fine Chemistry (37.62% of their patents for internal production processes and products. In Chemicals and Pharmaceuticals there are large shares of unused patents. Machine Tools (62.91%). (70. and they do not use about 40% of their patent portfolio.32%). As expected. while they exchange less than 10% of them.e.27%) and in Information Technologies. 6.00%).13 shows the use of patents by type of organisation in which the inventors were employed at the time of the innovation. the latter are more likely to apply for patents with a high expected value. and 3. Medium-sized firms use 65. Organic Fine Chemistry (30. More than half of the unused inventions are patented to block competitors. their average quality might decrease leading to larger shares of used patents (i. 14.78% of their patents.58%). Handling and Printing (64. Small firms use 55. Telecommunications. but also for less valuable ones.22%) and in the Chemical and Petrol industry (27. Indeed. sleeping patents) compared to smaller companies. and about 26% are licensed (specifically.89% are cross-licensed). Semiconductors and Optics (from 25% to 28%). Large share of sleeping patents are applied in Environmental Technologies (31. The patent applicants in these sectors are also more likely to invest in complementary assets for producing and distributing the final products compared to other sectors where licensing is more diffused and the “strategic” reasons for patenting are also important. public or private research organizations and university license a large fraction of their technologies and do not use them internally. The large share of unused patents by large firms might also be due to the fact that large companies have the financial strength to apply for patent protection not only for important innovations.90% are licensed & used. Table 3. Only 18% of patents are not used (for blocking and “sleeping” reasons).96% are only licensed.Machinery and Apparatus. 41 . Therefore. as the number of patent applications increases.79%). Large firms use internally half of their patents.14%) and Consumer Goods and Equipment (62.

97% Blocking Competitors 17.90% 8.09% 44.00% 3.23% 17.56% 1.00% 100.78% 6.47% 24.00% 100.70% 6.00% 100.25% 16.27% 21.74% 23.00% 100.58% 8.00% 100.64% 9.07% 17.01% 3.42% 11.00% 100.45% 20.63% 16.66% 5.88% 4.63% 25.76% 7.52% 21.00% 12.23% 3.74% 17. & Electrical Energy Audio-visual Technology Telecommunications Information Technology Semiconductors Optics Analysis.08% 37.79% 58.60% 2.55% 11.52% 6. Pumps & Turbines Mechanical Elements Handling & Printing Agricultural & Food Processing.20% 1.62% 50.93% 13.90% 5.36% 4.00% 100.00% 100.60% 14.12 The distribution of patent uses by micro technological class ISI Technological Classes Electrical Devices.00% 100.97% 16.69% Sleeping Patents 14.13% 5.27% 8.97% 10.57% 4.44% 4.52% 5.00% 100.96% 18.00% 100.04% 1.07% 21. & Food Chemistry Chemical & Petrol Industry.00% 100.86% 31.67% 3.88% 8.97% 2.00% 100.74% 2.66% 14.00% Number of observations = 7.57% 3.55% 13.38% Crosslicensing 3.14% 47.63% 18.81% 1.62% 7.91% 11.88% 3.94% 0.00% 100.21% 47.35% 1.10% 7.27% 3.22% 24.20% 10.58% 11.28% 15.93% 15.80% 49.31% 3. Measurement.77% 54.60% 1.06% 51.00% 100.02% 5.56% 30.39% 8.21% 0..24% 15.03% 1.39% 20. Textiles & Paper Thermal Processes & Apparatus Environmental Technology Machine Tools Engines.00% 100. Machinery & Apparatus Transport Nuclear Engineering Space Technology Weapons Consumer Goods & Equipment Civil Eng.83% 43.87% 2.76% 14.00% 4.85% 18.00% 100.13% 3.48% 17.00% 2.89% 15. & Control Technology Medical Technology Organic Fine Chemistry Macromolecular Chemistry & Polymers Pharmaceuticals & Cosmetics Biotechnology Materials & Metallurgy Agriculture.18% 12.00% 100.05% 38.21% 48.29% 34.38% 25.43% 50.11% 3.22% 37.00% 100.00% 1.71% 46.00% 100.41% 27.91% 52.28% 20.00% 12.59% 14.00% 100.71% 23.27% 9.98% 7.03% 3.56% 6.19% 15.90% 8.96% 16.02% 49.02% 2.58% 3.48% 6.80% 0.96% 51.22% 14.06% 21.25% 64.35% 3.52% 50.01% 3.00% 100.90% 3.711 42 .00% 100.00% 100.90% 20.37% 24.02% Licensing & Use 5.04% 1.19% 26.38% 9.29% 3.07% 3.89% 44.06% 11.85% 38.Table 3.94% 11.00% 100.90% 5.44% Total 100.67% 2.01% 7.09% 9.83% 15.00% 100.22% 62.53% 3.17% 2.59% 62.55% 2.04% 1.81% 9.05% 3.33% 5.00% 100.00% 100.84% 50.27% 11.00% 37.93% 6.70% 16.88% 4. Basic Materials Chemistry Chemical Engineering Surface Technology & Coating Materials Processing.50% Licensing 2.01% 0.29% 25.00% 100. Building & Mining Total Internal use 55.71% 3.01% 7.32% 70.61% 1.62% 7.52% 3.80% 19.54% 14.22% 24.59% 1.79% 1.42% 47.32% 6.55% 27.29% 4.42% 17.71% 56.39% 6.48% 11.00% 100. Electrical Eng.

13% of the EU6.00% 100. Institutions Other Total 49.14 shows the distribution of the value of European patents across the six patent uses.00% 100. Also the share of opposed patents is high for patents that are licensed (9. cross-licensed and produced to block competitors.67% 34.00% 100.38% 14. As expected.77% 18.00% 100.00% 100.19% 22.40% 17.24%.62% 18.75% 10.06% Licensing & Use 3.00% 4.31% 8.96%).78% 16. It is 9.50% Total 100.88%. 43 . Also the average value of a blocking patent is above the overall EU6 average. The overall EU6 share of patents with a value of 10 million Euros or more is 7.59% 6.e. Additional work with multiple correlation analysis is needed to understand the potential value and the social costs of the unused patents.04% 50.15).67% 17.84% 22.06% 27.25% 5.4%.03% 1. and it is 9.35% 5.Table 3.67% 21. It increases to 9. compared to 8.92% Blocking Competitors 21.89% 0. sleeping patents).33% 8.90% 9.50% 25.00% Number of observations = 7.74% 26.00% 100. the average number of forward citations is higher than the EU6 average in the case patents are sleeping (1. Differently.22% 3.87% 13.26% 3.00% 23.41% for patents that are both licensed & used.00% 4. The average value of a licensed and cross-licensed patent is higher than the average value of a patent in any other possible use.90% 6.62% 55.91% for patents that are involved in cross-licensing agreements.00% 0.03% 5. This distribution of the data is reflected in the average monetary value of the patents in the six patent uses.42% 23.72% 13.80% 5. blocking patents) or as a source of knowledge spillovers to other inventors (i.25% 41. The share of blocking patents is 7.97% 35.00% 8.93% 65. which seem to be quite important in terms of their monetary value (i.02% 6.20% 3.53% Licensing 3.11% for patents that are both licensed & used.21% for patents that are only licensed. the share of sleeping patents with a high monetary value (10 millions Euros or more) is only 4.556 Patent value and Patent use Table 3.50% 16.00% 100.75% 8.00% 100.92% 34.06% 10. and it reaches a peak of 11.33% 12.13 The distribution of patent uses by type of inventors’ employer Internal use Large companies Medium sized companies Small companies Private Research Institutions Public Research Institutions Universities Other Governm.17% Crosslicensing 3.83% Sleeping Patents 19.e.51% 3.

Table 3.76% 9.54% 22.207) 1.00% 5.00% 6.80% 100.04% 22.85 (1.81% 4. Another reason for patenting is to block competitors that might patent similar innovations.96% 24. which suggests that patents are important for competitive reasons more than for evaluating or motivating people working in the organization.98% 21.62% 4.57% 1.70% 13.35% 1.87% 100.84% 4. Also licensing and cross-licensing are not considered among the most important motivations for patenting. organizations do not consider reputation as being a very important reason for patenting.54% 6.60 (1.551) 0.17) 6.42% 100. this Section focuses on the motives that led inventors and their organisations to ask for patent protection.83% 0.72) 11. This is consistent with the low share of licensed patents as compared to those that are used internally and to the share of unused patents discussed in the previous Section.47% 20.87% 3.580 (31.996 (22.42% 100.72% 1. prevention from imitation.87% 17.60% 0.845 (38.79% 12.71% 22.33% 3.14 The distribution of patent value by patent use Patvalue <30k 30k-100k 100k-300k 300k-1m 1m-3m 3m-10m 10m-30m 30m-100m 100m-300m >300m Total Average patent value Average number of forward citations Share of opposed patents Number of observations Internal use 5.80% 2.472 (39.00% 9.2 shows the distribution by country).29% 0.11% 306 Crosslicensing 5.40% 100. and reputation.44% 16.72% 21.72% 20.12) 5.450) 0.00% 9.42% 0.03% 21.13% 7710 Note: standard deviations in parenthesis.00% 3.04% 3894 Blocking Competitors 10.35% 100.59% 5. licensing. By patenting the “inventions around” they prevent others to imitate their valuable innovations.39% 10.98% 20.93 (1.94 (2.23% 3.15 (2.15% 7.15 shows the average importance of these six motives for patenting by “macro” technological class (Annex A III. The most important reasons for patenting are the commercial exploitation of the innovations and the prevention from imitation.69% 1.94% 1.57) 8.903.31% 15.532 (30. The number of observations shown in this Table refers to the number of forward citations and opposed patents.66) 9.07% 13.79% 16.06% 0.66% 1441 Sleeping Patents 11.00 (2.79% 0.95 (1.635 (39. The six motives for patenting are the following: commercial exploitation of the innovation.88) 8.80% 19.23% 16.28% 14.04% 2.039) 0.57% 1.34% 10.15% 233 Licensing & Use 1. 44 .73% 1344 Total 7.24% 15.00% 7.70% 1. Table 3.67% 0. inventors and organisations patent because they seek exclusive rights to exploit economically. In other words.70% 0.23% 20. The importance of different motives for patenting While the previous Section described the actual use of patents.72% 2.00% 6.68% 19.43% 17.93% 15.67% 12.126) 0.439) 0.70% 20.35% 0. In the PatVal-EU survey we asked inventors to assign a score from 1 to 5 to the importance of six different motives for patenting within those organizations in which they were employed at the time of the invention.19% 11.96% 492 Licensing 2.896) 1.86% 27.00% 100.452 (17. Indeed.71% 17.48% 0. blocking rivals. cross-licensing. For the average patent value the number of observation is 6.90% 2.41) 9.

29) (0.10) 2.54) 1.86 (1.60) 1.10 (1.75) (1.57) 2.65 3.57) Instruments 3.91 respectively) compared to the overall importance (3.35 1. Table 3. It also shows the distribution of our six motives for patenting across the micro technological classes in which the patents are classified.26) 2.74) 2.56) 1.94 (1.65) Other Governm.15 (1.59 (1.55) Note: standard deviations in parenthesis 45 .46 (1.54) 1. Instruments.85 (1.78 (1.63) 2.01) 1.77 (2.44 2.69) 1.05 (1.27 2.56) 2.82) (1.82 (1.54 (1.87 2.70) 2.78 (1.54 (1.86) (1.68) 2. compared to the others.41) 2.05 (1.42 (1. Distribution by macro technological class.55) (1.78 (1.16 Importance of different motives for patenting.26 3.58 2.05 (1.56) Note: standard deviations in parenthesis Table 3.00 (1.91 4.53) Total 3.57) 1.14 2.12 in Annex A III.18 (1.80 (1.16 shows that the commercial exploitation of a patent is more important for small and medium firms (4.68) 2.60) 2.98 (1.52) 1.80 (1. Large and medium firms also consider prevention from imitation and blocking rivals as important motives to ask for patent protection.56 (1.70 (1.19 (1.74 (1.03) 1.76 (1.67 (1. Large companies Commercial exploitation Licensing Cross-licensing Prevention from imitation Blocking patents Reputation 3.52) (1.76 (1.15 Importance of different motives for patenting.2).83 (1.2 presents additional evidence on the use of patents and the importance of different motives for patenting. Reputation is slightly more important in Electrical Engineering (and in particular in Semiconductors.46) 2.54) 1.85) 3.89 (1.70) 2. Cross-licensing is an important reason for patenting for large firms.08 (1.57) (1.22) (1.87 (1. Annex A III.84) (1.23 (1.65) (1.48 (1.66 3.52) Process Engineering 3.38 (2. 3.12 (1.57) Chemical & Pharmaceutical tech.29) 3.44) 2.57) 2. Electrical Engineering Commercial exploitation Licensing Cross-licensing Prevention from imitation Blocking patents Reputation 3.39 (1.59) 3.60) 3.61) (1.76 2.34) 3.06 (1. Licensing is considered more important in Electrical Engineering.65) 1.54) 3.93 (1.74) Others 3.01 (1.66) 2.55) 2. and Chemical & Pharmaceutical technologies.46 (1.58) Mechanical Engineering 3.27 (1.80 (1.00 (1.26 (1.73) 2. It shows data on the use of European patents and the motives for patenting according to the nationality and the number of countries in which the applicant institutions are located. with a level of importance higher than 3 against the overall average of 2. Finally reputation is an important reason to patent for public research organizations and universities.66 (1.79 (1.02 (1.14 2.54 (2. as shown in Table A.66 (1.77) (1. Different types of employers have different motivations to patent.44) 3.06 (1.83) 1.87 3.67) 2.43) 3.46) 3.55) 3.34) 3.17 (1.75) 2.65) 3.75) (1.15 (2.63) 3.23 (1.91 (1.42) 3.09 2.32 1.69) (1.42 (1.66 (1.53) 1.56) 2.64) 2.50) Medium Private Public Small sized Research Research companies companies Institutions Institutions 3.98 (1.56) 2.72) 1.55) 1.85 (1. Institutions 2.26 (1.94 (1.68) (1. Licensing is more important for private and public research organizations.03 and 3. Distribution by type of inventors’ employer.86) (1.25 (1.41) 1.There are not significant differences across “macro” technological classes in the motives that lead the applicant to ask for patent protection.98 (1.76 (1. including universities.03 3. Table 3.42) Total 3.15 (1.59) 2.25 1.50) 3.62) 3.85) 2.96 (1.58 3.96 (1.57 (1.72) 3.57) (1.69) Universities 3.29) 3.67) (1.73) 1.79).59 (1.54 (1.

This share is the highest in the UK (9.17 shows that the European regions with the largest share of new firms created by using the patented invention are in the UK.00% 8. Figure 3.00% 1. However the empirical evidence on this issue is limited.17 are British.00% 1. It is the lowest in Germany (2. Table A. The evidence for Europe is inexistent. 3 are Italians.56%) and Mechanical Engineering (5.3. The literature focuses on the creation of firms by university inventors.00% 2. Sixteen out of 25 regions in Table 3.69% 4.3 Share of new firms from patented inventions by country Total UK NL IT FR ES DE 0. 2 are Spanish and 2 are Dutch. 2002).97% 9.00% 9.4 Creation of new businesses from the patented invention Recent contributions highlight the role of intellectual property rights to encourage the creation of new firms.00% 2.08% of patents are used for creating a new firm.63% 5.63%). The PatVal-EU survey provides information about the creation of new ventures by using the patented innovations. A patent system that increases the probability of formation of such firms can therefore produce beneficial economic effects.38%).27% 6. specialised and innovative firms can be important ingredients for enhancing the employment and the economic performances of specific regions (see Eurostat.00% 7. At the overall EU6 level. 5.3 shows the regions with a share of start-ups lower than the EU average. In Instruments the share of new firms is the largest (7.17 shows the European regions at the NUTS2 level in which the share of new firms originated from a patent is above the EU6 average (5.75% 5.00% 4.50%).4 shows the share of patents that give rise to new ventures by macro-technological classes.00% 10. Small. and ultimately to increase the rate of employment. Table 3. Consistently with the share of start-ups by country.3 reports the share of patents in the PatVal-EU database used to start-up a new firm. and investigates the phenomenon in the US. 46 .69%) and in Spain (9.27%). only 2 regions are German. followed by Process Engineering (5.00% 5.72% 3.00% Table 3. Figure 3. In Chemicals & Pharmaceuticals only 3. Figure 3.15 in Annex AIII.13%).13% of patents give rise to a new firm.72 %) and in France (1.13% 9.

46% 5. 47 . The least active sectors in this sense are Semiconductors.70% 7.78% 9.53% and 10.26% 5.67% 16. In many of the other technologies the share of patents that led to new firm formation is close to the overall average.73% 10.17 Share of new firms from patented inventions by European regions (NUTS2) Country UK UK UK UK UK UK ES UK UK UK IT UK UK UK NL DE UK ES IT UK NL UK IT DE UK NUTS2 Eastern Scotland Leicestershire.25% 5. Figure 3.34% 10.18 shows detailed information about the creation of new firms by “micro” technological classes. Bucks and Oxfordshire Gloucestershire.09% 8. The share of new firm formation for universities and private research organization is comparatively large. Machinery & Apparatus.75% 16.32% and 14.13% N 32 36 48 62 43 38 39 55 122 87 110 92 41 105 117 33 46 104 163 31 128 85 403 133 78 Total 5.95% 13. Consumer Goods & Equipment follow with shares above 8% of new firm creation from the patented innovations.75% 6. Macromolecular Chemistry & Polymers where the share of new firms is 2% or smaller.00% respectively. In Space Technology Weapons and in Medical Technology the share of patents that gave rise to a new firm is 10. Agricultural & Food Processing.5 shows the creation of new firms by type of inventors’ employers. Civil Engineering.76% 9.Table 3. Audio-visual Technology. as it is 16.66% 10.00% 9.16% 12.67% 14. Organic Fine Chemistry. and it drops to 1.06% for medium-sized firms.69% 6. Rutland and Northants Inner London West Yorkshire Essex Merseyside Comunidad de Madrid Hampshire and Isle of Wight Berkshire.45% 6.91% for large firms.88% 5. This share falls to 5. Wiltshire and North Somerset Veneto Outer London Bedfordshire. Table 3.52% 9.52% 13. About 17% of the patents applied by small firms are used to create a new business.58% respectively.40% 9. Biotechnology. Hertfordshire East Anglia Noord-Holland Gießen Derbyshire and Nottinghamshire Cataluña Emilia-Romagna Tees Valley and Durham Gelderland West Midlands Lombardia Karlsruhe Greater Manchester % New firms 18.82% 12.13% 7394 Note: This Table includes the European regions in which the share of new firms from patented inventions is above the EU6 average and the number of observation in each region is ≥30.

52% 8.38% 5.13% 5.00% 3.13% 4. & Food Chemistry Materials & Metallurgy Total % 10.53% 10. Pumps & Turbines Surface Technology & Coating Macromolecular Chemistry & Polymers Organic fine chemistry Semiconductors 48 .50% 4.27% 4.20% 8. Basic Materials Chemistry Electrical Devices.00% 8.91% 4.08% 5. Measurement.06% 7. & Control Technology Nuclear engineering Materials Processing.67% 8.00% Table 3. Building & Mining Consumer Goods & Equipment Chemical engineering Analysis.86% 2.. & Electrical Energy Mechanical Elements Machine tools Engines.00% 1.00% 7.10% 3.46% 7. Machinery & Apparatus Audio-visual technology Biotechnology Civil Eng.00% 6.13% 5.18 Share of new firms from patented inventions by macro technological class Technological Class Space technology weapons Medical technology Agricultural & Food Processing.00% 5.56% 7.37% 6.97% 3.02% 2.45% 6.62% 8.63% 2.4 Share of new firms from patented inventions by macro technological class Total Mechanical Engineering Process Engineering Chemical & Pharmaceutical tech. Instruments Electrical Engineering 3.00% 1.27% 6.19% 5.69% 3.72% 4.43% Thermal Processes & Apparatus Optics Environmental technology Pharmaceuticals & Cosmetics Transport Telecommunications Handling & Printing Chemical & Petrol Industry.00% 4. Electrical Eng.78% 0.72% 4.86% 4.00% 9.17% 4. Textiles & Paper Information technology Agriculture.84% 4.00% 2.Figure 3.77% 1.38% 5.

The Netherlands (33. This share varies across countries.99% 14.06% 0. The share of individual inventors’ patents is the largest in Spain (57. 1997).27%). 1993.00% 14. whether they are generated within the firm. 49 .19 shows the extent to which individual inventors collaborate in the research activity leading to a patent.14%).36%) is developed by “individual” inventors. Spillovers and the Sources of Knowledge in the Invention Process This Section describes the sources of knowledge used in the innovation process leading to the PatVal-EU patents. The UK and Italy follow with 40.00% 18. They also show the share of patents used to create a new firm as a function of the number of countries in which the applicants are located. or if they are accessed from external sources.91% 4.e. being these individual inventors. The PatVal-EU data provides original and direct information about the sources of such spillovers.24%) are below the EU6 average. public or private research organizations.Figure 3.00% Annex A III.00% 4. Jaffe. Verspagen.58% 17.11% 5. only one third of the PatVal-EU patents (37.34%) and Germany (35. Patent citations have typically been used in the literature to analyse the extent to which knowledge is transferred among individuals.5 Collaboration. other companies. At the overall EU6 level.09% 16.32% 8.3 shows additional Figures on the creation of new businesses from the patented inventions. 3.00% 6.00% 2. Collaboration amongst inventors in the innovation process Table 3.00% 12. universities.68% and 40. scientific meetings or earlier patents.00% 16.15% 14.5 Share of new firms from patented inventions by type of inventors’ employer Total Others Other Governm. and applicant organisations.00% 10.63% 9. They distinguish between individual inventors who applied for the patent. France (36.00% 8. Institutions Universities Public Research Institutions Private Research Institutions Small companies Medium sized companies Large companies 1. organisations and geographical locations (i.15% of patents developed by individual inventors.

34% 40. France and Spain respectively.16 to A.19 also indicates whether the multiple inventors (i. Table 3.06% Patents developed with external co-inventors 23. or to other organizations. Annex A III.86% with little variation across countries.82% 21. Patents developed with internal co-inventors DE ES FR IT NL UK Total 76.86% 35.05%).88% 19. In these countries.36% Table 3. Distribution by country.Table 3. By distinguishing between internal and external co-inventors. It falls to 16.24% 57.20 investigates whether the collaboration among inventors described in Table 3. By comparing these data with those in Table 3.29% and 21.95% 76.29% 16. Table 3.13% 80.14% 64. This also confirms that the collaboration among inventors belonging to different institutions is not visible in the standard information provided by the patent document.14% 36.27% 40.e.52%. Table 3.94% Total 100% 100% 100% 100% 100% 100% 100% Share of Individual inventors’ patents 35. In the UK this share reaches 35. 19.48% 76. This information is difficult to obtain from existing – and “official” – data like the patent document. the networks of inventors tend to be within the same organization. The EU6 share of patents developed by multiple inventors affiliated to different organizations is 23.23) shows additional data on the patents applied by single applicants versus those applied by multiple applicants either independent or from the same corporate group.19 reveals the occurrence of knowledge spillovers between inventors located in different organizations.19 shows up also in the collaboration among the organizations that apply together for a patent. with a limited role of external linkages. and distinguishes between patents applied together by organizations belonging to the same group.94%.4 (Tables A. therefore.68% 37. Related to this. The EU6 share of patents applied by single applicants is 93.19 Affiliation of the inventors involved in the development of the patents.63% in Spain. co-inventors) involved in the development of the patents are affiliated to the primary inventor’s organization. 50 .71% 83.88% in Italy.15% 33.05%. The share of co-applied patents reaches a peak of 96.19 it appears that the extent to which inventors affiliated to different organizations collaborate does not show up in the number of patent applications filled out by multiple applicants. The Netherlands and Germany are close to the EU6 average. and patents co-applied by independent companies.05% 23.18% 78.20 shows the share of co-applied patents. while it is the lowest in France (67. This is a useful piece of information to understand the extent of collaboration among institutions that occurs by means of the collaboration among individual researchers.52% 23.

99% 2.00% 100. It is the largest in Mechanical Engineering where individual inventors develop 49.99% 38.88% 0.21% 77.63% 67. Table 3. Table 3. The sector in which the networks of inventors tend to be more frequently among people from the same organisation is Electrical Engineering where 19. Distribution by country.22 shows the share of individual inventors’ patents.69% 14.95% Total 100% 100% 100% 100% 100% 100% Share of Individual inventors’ patents 38.16% 0.29% 2.21%) compared to the overall EU6 average (37.95% 91.81% 92. By using these data from the PatVal-EU survey we have the possibility to investigate the extent of collaboration between different organisations in the innovation process.67% 4.86% Patents applied with multiple organisations from the same group 1. Differently.00% 100.00% 100.89% 4.05% Patents developed with external coinventors 19.00% 3. most of these collaborations are among researchers employed in the same organisation.Table 3.88% 3. when patents are developed by teams of multiple inventors. Patents applied with single applicant DE ES FR IT NL UK Total 95. internal co-inventors’ patents. and external co-inventors’ patents for different 51 . If one conditions upon the multiple inventors’ patents.58% Total 100.20 Number of applicant institutions in a patent.85% 3.e.36% 3.15% 93.21 Affiliation of the inventors involved in the development of the patents. Patents developed with internal coinventors Electrical Engineering Instruments Chemical & Pharmaceutical tech Process Engineering Mechanical Engineering Total 80.00% 100.11% 3. inventors affiliated to the same institution) and “external” inventors (i.29% 37.05% 79.55% Patents applied by multiple independent organisations 3.95% 76.19% 74. Instruments and Process Engineering is slightly above the European average. Distribution by “macro” technological class.37% 1. In Instruments (28. and distinguishes between inventors all “internal” to the same organisation (i.36% 71.81% 25.21% 38. about 75% of the collaborations are among inventors affiliated to the same organisation.68%) the share of patents developed in collaboration with external co-inventors is higher then the EU6 average. The share of individual inventors’ patents in Electrical Engineering.64% 28. Table 3.29% of the patents. inventors affiliated to different institutions).00% 100.00% The rest of this section focuses on patents produced by multiple inventors.79% 49.00% 100.e. The share of patents developed by individual inventors is low in Chemicals & Pharmaceuticals (14.64% of co-invented patents are invented by “external co-inventors”.79% 22.05% 23.36%).32% 75.21 shows the share of patents invented by multiple inventors in the five “macro” technological classes.01% 96.79%) and Process Engineering (25.68% 24.36% Let’s now turn to the institution-specific effect on the probability of establishing collaborations among inventors affiliated to different organisations.

types of inventors’ employers.14% 40.21% 45.00% Share of Individual inventors’ patents 31.27% of all multiple inventors’ patents developed by universities and public research institutions involve collaborations between inventors from different organisations.73% 41.81%) compared to the overall EU6.00% 100. Table 3. Small companies and universities come second (20. Table 3.29% 36.72%).00% 100.00% 100.00% 23.40% 76.72%) and in Spain (23.60% 23.00% 76. These big research facilities within the large companies lead to the internalisation of knowledge spillovers by means of firm specific coordination and organisation capabilities. In Germany.00% of those in the UK.00% 100.64%).74%) compared to 18.80% of the patents developed through external collaborations in Germany.00% 100.85% We focus now on the “external co-inventors”.91% of the medium-sized enterprises. within the group of private companies.38%) is larger than the EU6 share. the share of collaborations with inventors employed in large companies (64. As expected.09% 60.91% 39. For the overall EU6.25% 50. 52 . Moreover.56% Patents developed with external coinventors 18.26% 65. Small companies also develop the largest share of patents invented by external co-inventors (39. they are important partners in Italy (26.58%).48% 42.27% 58. and 23.64% 38.81%) compared to the EU6 share.79% 54. inventors in large firms develop the smallest share of individual inventors’ patents and the largest share of internal co-inventors’ patents.25% and 16. Medium sized enterprises are particularly important in Spain (23. As far as Universities are concerned.00% 100. Differently.00% 100. small companies have the largest share of individual inventors’ patents (61. the multiple inventors’ patents that they produce are frequently invented by inventors affiliated to external organisations: 58. Institutions Others Total 81.22 Affiliation of the inventors involved in the development of the patents.36%) and third (15. and their distribution across the 6 countries involved in the PatVal-EU survey.44% Total 100.68% 48. large companies are the research partners in about half of the collaborations.41% 61.85%). Inventors employed in small companies are the partners of 25. followed by Italy (51. Universities and public research institutions have the lowest share of individual inventors’ patents (18. Distribution by type of primary inventors’ employer Patents developed with internal coinventors Large companies Medium sized companies Small companies Private Research Institutions Public Research Institutions Universities Other Governm.00% 64.71% 16.60% of the large companies and 23.85% 18.48% and 45.74% 34. This might be due to the presence of large research laboratories in the big corporations where a large number of complementary capabilities are employed.86% 60.00% 100.23 shows their affiliation.52% 57.

66% 10.02%) and with the Universities (14.96% 19.28% 9.33%).20% 10.74% 6.58% Large companies DE ES FR IT NL UK 64. Process Engineering Mechanical Engineering 49.71% 9. Similarly.55% 21.86%).79% Medium sized companies 9.74% 19.66% Number obs.12% 15.10% 6. 438 21 173 116 178 287 1.90% 3.23 External co-inventors.74% 13.96% 12. Table 3.25%) compared to the EU6 share (20.72% 26.05% 16.97% 20.40% 42.34% 26.93% 4.16%) compared to overall EU6 share. Table 3. Also small firms tend to develop linkages with other small firms (41.06% 8.18% 14.71% 51.213 Total Note: The sum of shares by row can be larger than 100% because each patent’s inventor might have 1 or more co-inventors employed in different organisations.32%).58% Large companies Electrical Engineering Instruments Chemical & Pharmaceutical tech.36% 49.18% 8.80% 3.77% 17.67% 23. inventors working in the Academia enter a large fraction of inter-inventors collaborations (25.62% 11.05% 7.01% 8.96% 5.68% 34.89% 8.92% 18.35% 31.03% 6.77% 23.87% 9.08% 12.02% 22. and with large companies (27.00% 20.82% 23.88% 5.51% 49.Table 3.01% 8.25 looks at the type of institutional networks that are developed through the collaborations among “external” inventors.81% 4. By contrast. Table 3.81% 17.76% 6.85% 9. while they have a very limited role in the collaborative networks among inventors in Chemicals and Pharmaceuticals (11. Large companies also develop research collaborations with small companies (14.65% 11. In most of the cases.24 shows that there is little variation across “macro” technological classes in the extent to which inventors employed in large companies are involved in the collaborations with the organisations of the PatVal-EU inventors.36% 4.29% 4.29% 7.18% 5.01% 15.36% 5. 155 158 303 314 283 1.91% 4.08% Small companies Universities 5.41% 8.00% 7.66% Number obs.33% 49. 53 .46% 3. Universities tend to establish formal collaborations with other Universities (31.18% 50.38% 33. The presence of small companies is particularly large in Instruments (31.31%).83% 53.35%). Distribution by “macro” technological class and by affiliation of the “external” inventors Medium sized companies 7.79% Universities 7.21% 16.24 External co-inventors.08% Affiliation of the “external” co-inventors Private Public Small Research Research companies Institutions Institutions 20.72% 12.72% 25. in Chemicals and Pharmaceuticals. large companies collaborate with other large companies (65.76% 15.52% 4.50% 53. Distribution by country and by affiliation of the “external” inventors Affiliation of the “external” co-inventors Private Public Research Research Institutions Institutions 25.80%).50%) and with large companies (37.90%).213 Total Note: The sum of shares by row can be larger than 100% because each patent’s inventor might have 1 or more co-inventors employed in different organisations.

86% 16. however.00% 0. Formal and informal collaborations in research The PatVal-EU inventors were asked to report if they established any formal or informal collaboration with other partners to develop the innovation.85% 19. It is particularly large in France (87.25 External co-inventors.00% 0.46% 31.36% 9.88% 33.23% 0. They cover 76.36% 8. suggesting that there are knowledge spillovers among inventors and institutions that are not mediated by any apparent market mechanism.88% of the collaborations set up by firms and institutions for a common research project leading to a patent. France.Table 3. are informal.31% 12. Institutions Others Total Note: The sum of shares by row can be larger than 100% because each patent’s inventor might have 1 or more co-inventors employed in different organisations.171 Large companies Medium sized companies Small companies Private Research Institutions Public Research Institutions Universities Other Governm.90% 39. By formal collaborations we mean well-defined contracts among the parties to collaborate over an R&D project. This information provides additional data on the extent of collaboration among different organisations while developing the innovation.27% 50.87% 0.62% 7.00% 8.28% Small companies 14.27%).62% 5.45% 0. 692 89 172 12 65 127 3 11 1.33% 11.21% 25.6 in parentheses close to the country names.00% 3. It is higher in the Netherlands (34.44%) and in Spain (86.71% Number obs.32% of German patents are invented in research projects that involve collaboration with external institutions.53%. Figure 3. patents developed in collaboration with other partners) is shown in Figure 3. The UK.08% 9.00% 27. In Italy (65.22% 27.67% 33. The EU6 share of patents produced by using external collaborations is 20. Distribution by type of employer of the interviewed inventors and by affiliation of the “external” inventors Employer Type Affiliation of the “external” co-inventors Large companies 65.67% 36.80% 7.00% 5. Institutions 4.89% Universities 14.59%) the share of formal collaborations is lower than the EU6 share.58% Private Res.05% 5.55% 20.02% 19.67% 4.73% Medium sized companies 3.00% 15.80% 33.33% 27. The share of “collaborative” patents (i.63% 16.34%) and in Germany (71.09% 66. 54 .52%).69% 37.63% Others 7.00% 0.50% 0.33% 18. Only 13. Italy and Spain are close to the EU6 average.10% 41.69% 0. About one fourth of the collaborations.e.00% 23.00% 54.08% 7.6 also shows the share of formal and informal collaborations over the total number of “collaborative” patents.90% 20.87% 12.33% 25.

47% 16.Figure 3.44% 20.03%).88% 81. In terms of formal vs. the share of informal collaborations is larger than the EU6 average in Mechanical Engineering (31.66% 12.19%) Instruments (25. Figure 3.42% 90% 100% Share of formal and informal collaborations over the total number of "collaborative" patents % Formal collaborations % Informal collaborations 55 .53% ) UK (23.65%).56% 79.27% of the total number of collaborations among organizations is done on informal bases.16% 34.12% 18. Distribution by “macro” technological class Total (20.32% ) 0% 10% 20% 76.55% ) DE (13.32% 22.34% 87. Chemicals & Pharmaceuticals (22. At the opposite extreme.53%) Mechanical Engineering (17.27% 71.6 Formal vs.53% 83. (22.7 Formal vs.59% 30% 40% 50% 60% 70% 23.58% 40% 50% 60% 70% 80% 23.88% 68. In the other three technological classes the share of informal collaborations is close to the EU6 average.52) IT (21.53%).90%) and Electrical Engineering (18.03%) 0% 10% 20% 30% 76.02% 77. They are less frequent in Mechanical Engineering (17.19%) and Process Engineering (21. collaborations in Chemicals & Pharmaceuticals tend to be formalised: only 16.44%).73% 28. Informal collaborations amongst institutions.67%) compared to the EU6 share (20.41% 80% 90% 100% Share of f oeal and inf ormal collaborations over the total number of "collaborative" patents % Formal collaborations % Inf ormal collaborations In terms of “macro” technological classes.73% 76. Distribution by country Total (20.33% ) NL (34.12% 31.68% 77. informal collaborations.84% 65.27% 23.90%) Process Engineering (21.56% 13.98% 22. Informal collaborations amongst institutions.68% ) ES (19.67%) Chemical & Pharmaceutical tech.94% ) FR (22.65%) Electrical Engineering (18. collaborations amongst organisations are frequent in Instruments (25.44% 86.

For the overall EU6. that typically takes less than one hour to be reached. (3) interaction with people external to the inventor’s organization. the inventors rank very similarly the importance of the interaction with people from the same organisation and from other organisations (1.4 provides additional data on the extent of collaboration among different institutions to develop a patent.28 to A. less than one hour to reach physically the partner). They show the share of patents invented in collaboration with different external partners (i.27 reports the role of the different types of interactions by European regions at the NUTS2 level. When it takes more than one hour to reach the location of the other researcher. A common pattern to all countries is that the interaction with researchers affiliated to different organizations that are geographically close is the least important form of collaboration. Universities. Only in Spain the affiliation to the same firm seems to be very important to develop linkages among the inventors. This suggests that the geographical proximity among the inventors might not be the key explanation for the importance of external interactions.e.32 respectively). private research institutions.4 focus on formal collaborations. We use a scale from 1 to 5 for the importance of 4 different types of interaction: (1) interaction with people internal to the inventor’s organization. Table 3. and geographically close. by type of inventors’ employer. In a large fraction of these very same regions other forms of interaction are also more important than those highlighted by the EU6 average. The role of geographical proximity for collaboration To explore further the importance of knowledge spillovers among researchers in the innovation process. and geographically distant. and geographically close (i. Again.27 lists the regions in which the importance of interaction with people affiliated to external organizations that are geographically close is larger than the EU6 average. Table 3. medium sized companies. ranks first (3. all the six countries are lined up in the same ranking. being them geographically close or distant. informal collaborations by micro technological classes. (4) interaction with people external to the inventor’s organization. the importance of the interaction with people belonging to the same organization of the inventor (including affiliates). Tables A.e.02).26 indicates that the interaction with other members of the same organization are on average more important than the interaction with people affiliated to other organizations. more than one hour to reach physically the partner). and geographically distant (i. especially if people from the same organization are geographically close. A deeper understanding of the determinants of spillovers in the innovations process due to 56 . this Section highlights the role of geographical proximity in fostering knowledge exchange among the inventors. large companies.e. and to the 30 micro technological class in which the PatVal-EU patents are classified. and “others” as a residual category) according to the nationality and the number of countries in which the patents’ applicants are located. This suggests that in general geographical proximity is not crucial for developing research linkages among individuals that are affiliated to different institutions. small companies.30 in Annex A III. Table 3. by nationality and number of countries in which the applicants are located. The regions in which this type of interaction is evaluated as being most important are listed in the top part of the Table. This is so for all the six countries.31 and 1. The Tables in this Annex look at the distribution of formal vs. (2) interaction with people internal to the inventor’s organization.Annex A III.

92 (1.73 (1.51 (1.83) 1.43) 0.51 (1.30 (1.84) 1.80) IT 2.73) 1.68) 1.88) 3.95) 3.08 (1.97 (1.74) 2.40) 0.64) 1.76) 1.69 (1.76) UK 3.24 (1.73) 1.55) 0.22 (1.75 (1.70 (1.65) 1.09 (2.18 (1.69) 1.44) 0.55) 1.09 (1. 57 .80) 1.63) 3.98 (1.73 (1.43 (1.52) 1.46 (1. This will be performed in Lot 2 of this project.77) N 36 55 61 35 269 37 48 46 47 48 420 67 36 39 43 35 41 40 183 55 80 65 41 86 105 60 120 298 8588 Total Note: This Table lists the European regions for which the importance of Close & External interactions is above the EU6 average.80 (1.79) 3.64) 3.42 (1.33 (1.88 (1.50 (1.75) 3.36 (1.82) 1.66) 1.07 (1.36) 0.13 (1.83) 2.84) 1.20 (1.94) 1.94) 1.44 (1.66) 1.18 (1.14) 1.63) 3.88 (1.93 (2.84) 1.75) 1.34 (1.97 (1.28 (1.09 (1.85) 1.96) 3.64 (1.32 (1.83) Total 3.28 (1.74 (1.46) 0.64 (2.26 Importance of the interaction with other people while developing the innovation.64 (1.92 (1.49) 3.73 (1.43 (1.70) 1.94 (1.83 (1.18) 0.76) 3.48) 1.53 (1.55 (1.87) 1.70) 0.29 (1.75) ES 3.71) 2. Rutland and Northants Bedfordshire.52 (1.23 (1.95 (1.76) 1.00 (1.42 (1.03 (1.90) 1.43 (1.02 (1.44) 1.13) 1.81) 1.02 (1.31 (1.88) 1.70) Close & External 2.28 (1.61) 1.53 (1.61 (1.84) 1.53 (1.69) 1.20 (1.55) 1.61) 1.69) 1.88) Distant & Internal 1.71) 2.69) 1.95) 1.84) 3.58 (1.46 (1.96 (1.76 (1.16 (1.08 (1.53 (1.98 (1.69) 1.97 (1.27 Importance of the interaction with other people while developing the innovation.17 (1.22 (1.78) 1.14 (1.81) 1.76) 1.51 (1.89 (1. Hertfordshire Languedoc-Roussillon Zuid-Holland Unterfranken West Midlands Utrecht Outer London Gloucestershire.53) 1.86 (1.09 (1.33 (1.31 (1.28 (1.77) 2.09 (1.40 (1. Distribution by European regions (NUTS2).40 (1.19 (1.85 (2.91 (1.43 (1.01) 1.28 (1.80) 2.74) 2.20 (1.05 (1.94) 1. Wiltshire and north Somerset East Anglia Alsace Berkshire.07 (1.48 (1.65) 0.83 (1.21 (1. DE Close & Internal Distant & Internal Close & External Distant & External 2.45) Distant & External 1.74 (1.65) 1.42 (1.61) 3.46 (1.77) 1.77) 1.32 (1.05 (1.37 (1.73) 1.87) 3.99) 1. Distribution by country.82) 1. Table 3.67) 3.71) 1.88) 3.69 (1.34 (1.70 (1.90) 1.83) 2.45 (1.44(1.31) 1.32 (1.82) 3.00 (1.69) 1.56) 1.72) 1.80) 1.92) 1.72) 1.65) 0.65) 1.33) 1.91) 1.94) 1.41) 1.84) 0.39 (1.67) NL 3.89 (1.79) 3.59) 1.26 (1.88 (1.44 (1.15 (1.11) 0.57 (1.87) 1.68 (1.64) 1.73) 1.98 (1.45) 1.63) FR 3.92 (1.95) 2.10 (1.25 (1.45) 1.71) 1.31 (1.02 (1.13 (1.pas-de-Calais Île de France Midi-Pyrénées Comunidad de Madrid Picardie Derbyshire and Nottinghamshire Leicestershire.41 (1.geographical proximity is needed through non parametric and multiple regression analysis.44) 1.31 (1.63) 1.88) 1.41 (1.92) 1.74) 3.45) 1.50 (1.96) 1.16) 1.86) 1.83) 2.13 (1.85) 1.24 (1.78) 3.54) 0.95 (1.12 (1.19) 1.82) 3.15 (2.74) 0.26 (1. Bucks and Oxfordshire Düsseldorf Close & Internal 3.70) 1.50) 0.77) Note: standard deviations in parenthesis Table 3.88 (1.85 (1.15 (1.70) 1. Country NUTS2 FR FR FR FR FR UK UK FR FR FR FR FR ES FR UK UK UK FR NL DE UK NL UK UK UK FR UK DE Haute-Normandie Provence-Alpes-Côte d'azur Centre Aquitaine Rhône-Alpes Merseyside Inner London Bourgogne Pays de la Loire Nord .23 (1.85 (1.46 (1.54 (1.91) 3.99) 3.97) 0.32 (1.

93 (1.92) 1. 58 .47) 1. this type of interaction with people affiliated to other organizations that are geographically close ranks second.33 (1.72) Process Engineering 2.75) Total 3. Table 3.67) 0.28. Table 3.84) Chemicals & Pharmaceuticals 3.66) 0. When the researchers are geographically distant.26 and 3. The results are very similar to those in Table 3.88 (1. The importance of the interaction among people who are geographically close and affiliated to the same organization ranks first for the inventors employed in private companies.81 (1.20 (1.28 shows the importance of the four forms of interaction in the five “macro” technological classes in which the patents are classified.76) 1.50) 1.26 (1.93 (1.30 (1. Finally.28 Importance of the interaction with other people while developing the innovation.49) 1.4 provides additional Tables on the importance of geographical proximity for fostering collaborations among inventors in the 30 micro technological classes in which the PatVal-EU patents are classified.93 (1.32 (1.23 (1. The interaction with researchers affiliated to different organizations that are geographically close is the least important form of collaboration. Universities and other government institutions.72) Instruments 2.79 (1. the affiliation to the same organization and to different institutions is ranked similarly. medium and small firms.72) 0.88) 1.02 (1. It also shows the importance of geographical proximity for establishing research interactions in the European regions at the NUTS2 level when the average level of importance of “close & external” interaction is lower than the EU6 average.23 (1.28 confirms that geographical proximity is not a key factor for the collaboration to take place.28 (1. The affiliation to the same organization together with the geographical proximity among the researchers are the most important type of interaction also for the inventors employed in public research institutions. This is so for all the five technologies. and according to the nationality and to the number of countries in which the patents’ applicants are located. Only for the inventors employed by Universities and private research laboratories. being they large.38) 1.76) 0. the interaction with people affiliated to different organizations located in the same geographical area ranks last. Annex A III. Once the researchers are geographically distant. Table 3.93 (1. the importance of the affiliation to the same organization is ranked very similarly to the affiliation of the inventors to other organizations.89) 1.82) Mechanical Engineering 2. Table 3.93) 1.41) 1.77) Note: standard deviations in Italic Finally.83 (1.40 (1.80) 1.46 (1.68) 0.45) 1. Distribution by “macro” technological class Electrical Engineering Close&Internal Distant&Internal Close&External Distant&External 3.90 (1. Like in Tables 3.By using the same scale from 1 to 5.28 (1.26: the interaction with people belonging to the same organization of the inventor that takes less than one hour to be reached ranks first in all the technological classes.31 (1.39 (1.29 shows the importance of the different forms of interaction by type of inventors’ employer.51 (1.70) 0.

58 (1. Distribution by country.42 (1.66) 2. The knowledge coming from university and non-university research laboratories is at the bottom of the ranking with an average importance of 1.67) and the interaction with the firm’s suppliers (1.04 (1.55 respectively).73) 1.35.81 (1.97) 1.07 (1.55 (1.78 (1.25 (1.98 (1.78 (1.79) 2.04 (1.50 (1.70) UK 1.89) 1.26 (2.46 (1.93) 2.16 2.94) 1.60 (1.01) 1.00 (2.77) Note: standard deviations in parenthesis The Sources of Knowledge This Section looks at the sources of knowledge used by the PatVal-EU inventors to develop the innovations.63) 2.00 (1.38) 1.03) 1.71) 0.40 (2.73 (1.41) 1.84) NL 1.35 (1.15).95 (1.58 (2.77 (1.72) FR 1.99 (1.97) 1.72) 2.74) 2.14 (1. By using a scale from 1 to 5 for their importance.25 (1.67 (1.41 (1.88 (1.65) 2.73) 2.57) 1.87) ES 1.90 (1.71 (1.70) 0.31 (1.02 (1.8 (2.87) Private Research Institutions 1.98) 1.71) 2.21 (1.40 (1.50 (1. 7) the firm’s competitors.92) 2.96) IT 0.53) 0.Table 3.83) 2.98 (2.90) 2.84) 1.67 (1.29) 1.55 (1.88) 2.36 (1.04) (2.72) Medium sized companies 2.80) 2.89) (2.55) 0.62) 1.69) 2.62 (1.55 (1.90) 1.7 (1.55 (1.43) 1. The patent literature and the scientific literature are in the second and third positions (2. 4) earlier patents.56) 2. DE 1.64 (2.88 (1.88) 1.59 (1. 6) the firm’s suppliers.91) 1.40 (1.60 and 2.94) 2.80 (1.74) 2.84) Total 1. the participation in technical conferences and workshops (1.76 (1.70) 1. 2) the scientific literature.55).12) Others 1.99) 0.76) Small companies 2.77 (1.23 (1.64) 2.73) 2.02) 1.38 (1.78) 2.65) 1.31 (1.56 (1.30 shows the average importance of each source of innovation.50) 2. we consider the following sources of innovation: 1) the knowledge developed in university and non-university laboratories.00 (1. Table 3.98 (1.23 (1.08 (1.45 (1.29 Importance of the interaction with other people while developing the innovation.47) 1.15 (1.93) 1. At the overall EU6 level. This ranking holds with only small variations in each of the 6 EU countries in the PatVal-EU survey.74 2.87) 0.37 (2.95) 1.87) Laboratories Scientific literature Conferences Patents Users Suppliers Competitors Note: standard deviations in parenthesis 59 .78) 1.74) 1.34 (1. followed by the firm’s competitors (2.83 (1.32 (1.55 (2. Table 3.95) 2.45) 1.71) 2.96 (1.30 Importance of different sources of knowledge.79) 1.25) 1.51 (1. Distribution by type of inventors’ employer Large companies Close&Internal Distant&Internal Close&External Distant&External 3.84 (1.4 (1.38 (1.54 (1.00) 1.88) 1.88).83) 1.03) Public Other Research Universities Governm.03) 1.14 (1. the firm’s users rank first (2.27 (1.02) Total 3.81) 2.35 (1.88 (1.38 (1.25 (1.68) 2.85 (1.72) 0. 3) the participation in conferences and workshops.83) 1.86) 3.20 (1. 5) the firm’s users.26 (1.26 (1.53) 1. Institutions Institutions 3.

and the interaction with the suppliers. Materials and Metallurgy. The firm’s users rank first. Instruments.78) 2. while the university and non-university research laboratories are the least important source of knowledge in the vast majority of the “micro” classes. Table 3.15 (1.31 looks at the importance of the sources of knowledge by macro technological classes.84 (1.35 (1.89) 1.62 (1.27 (1.88 (1.82) Chemicals & Pharmaceutical s 1.56 (1.32.71) 2.75) 2.73) 2.77) 2.20 (1.82) 1.90) 1.84 (1.70) 2.22 (1.77 (1.32 breaks down the “macro” technological classes into 30 “micro” technological classes. the patent literature and the scientific literature are second.41 (1.15 (1. Pharmaceuticals and Cosmetics.36 (1.55 (1.77) 2.31 Importance of different sources of knowledge.87) Note: standard deviations in parenthesis Table 3. There are sectors like Optics.00) 1. and Chemical and Petrol industry where the scientific and patent literature is the most important source of knowledge amongst those listed in Table 3.30 in four out of five classes (i.66 (1. There are.55 (1.91) Mechanical Engineering 1.03 (1. in Biotechnology they rank first together with the scientific literature. however.90) Total 1.92) 3. In most of these classes the ranking of the sources of knowledge reproduces those of the previous tables.91 (1.92) 1.85) Process Engineering 1.36 (1. The knowledge provided by the university and non-university research laboratories is at the bottom of the ranking in most of the “micro” technological classes.28 (1.55 (1.84) 2. Process Engineering.82 (1.60 (1. the competitors. Macromolecular Chemistry and Polymers. and the knowledge coming from the university and non-university laboratories is the last source of knowledge used by the PatVal-EU inventors.20 (1.85) 3. in Chemicals & Pharmaceuticals the patent literature and the scientific literature are the most important sources of knowledge (3.75) 2.92) 3. the participation in technical conferences and workshops. Semiconductors and Information Technology where the scientific literature and the participation in conferences and meetings are of primary importance as sources of knowledge for developing the patent. are still at the bottom of the ranking. 60 . The research laboratories.50) 1. followed by the firm’s users (2.98) 1.30 (1. for examples.64) 2.66) 2.98) 1. and Mechanical Engineering). Organic fine chemistry.98 (1.94) 1.80) Instruments 1. followed by the scientific and patent literature.52) 1.64 (1.Table 3.86) 2.95 (1.72) 2. Electrical Engineering.82 (1. The ranking is similar to that in Table 3.66) 2. Differently.73) 2.73 (1.81) 1.25 (2.89 (1.74) 2.42 (1. a few exceptions.02 (1. Food Chemistry.15 (1.72 (1.33 (1.67 (1. before the knowledge provided by the firm’s suppliers.88) 2.68) 3. technologies like Telecommunications.e.81) 1.46 and 3.16 (1. Distribution by “macro” technological class Electrical Eng Laboratories Scientific literature Conferences Patents Users Suppliers Competitors 1. with the users at the top of the ranking.25). Finally.66 respectively). however. There are.90) 2.56) 1.91) 1.66 (1.46 (1.

53) 1.17 (1.01 (1.88) 2.86 (1.85) 2.13) 2.74) 1.69) 1.95 (2.98) 3.96) 2.88) 2.29 (1.92) 3.44 (1.47 (1.98 (2.90) 1.78) 2.46 (1.72) 3.84) 2.90) 2.63) 1.76 (1.32 Importance of different sources of knowledge.15 (1.90) Users 3.78 (1.87) 3.40) 1.66) 3.88) 2.13) 2.35 (1.21 (1.09 (1.05 (1.66) 1.06 (1.19 (1.20 (2.01 (1.07 (1.84 (1.74) Scientific literature 2.54 (1.94 (1.73 (1.82) 1.21 (1.91) 2.07) 3.35 (1.68 (1.79) 1.74) 1.50) 1.62) 2.70) 2.34) 3.78) 1.78) 2.29 (1.84 (1.19 (1.77) 2.79) 1.80) 2.81 (1.96) 1.64 (1.86) 2. Basic Materials Chemistry Chemical Engineering Surface Technology & Coating Materials Processing.15 (1.30 (1.85 (1.07 (1.74 (1.54 (1.26 (1.79) 2.72) 2.85) 1.77 (1.83 (1.60) 1.61 (2.86) 2.16 (1.87) 2. & Electrical Energy Audio-visual Technology Telecommunications Information Technology Semiconductors Optics Analysis.Table 3.95 (1.87) 2.63 (1.96) 2.77) 2.71) 2.60 (1.86) 2.91) 2.56 (1.62) 1.07 (1.41) 1.91) 1.91) 3.75 2.68) 1.10 (1.29 (1.97) 1.00 (1.94) 1.94 (1.87) 3.32 (1.69) 1.99 (1.97) 2.97 (1.99) 2. & Control Technology Medical Technology Organic Fine Chemistry Chemistry & Polymers Pharmaceuticals & Cosmetics Biotechnology Materials & Metallurgy Agriculture.91) 1.98 (1.75) 2.85) 1.84) 1.86) 2.28 (1.88) 2.87 (1.26 (1.68) 1.61) 1.00) 2.55 (1.99 (1.88) 3.48 (1.67) 1.98) 3.60) 1.80) 1.72 (1.48 (1.43 (1.82) 2.77) 1.78 (1.82) 2.98) Suppliers 1.55 (1.11 (1.54) 2.88) 1.90) 1.40 (1.53 (1.02 (1.86) 2.82) 1.69) 2.28 (1.61 (1.90) 1.83) 2.76) 1.16 (1.53 (1.95) 2.24 (1.46 (2.68 (1.19 (1.73) 3.05) 2.40 (1.79 (1.91 (1.49 (1.56 (1.83) 2.06) 2.05 (1.11 (1.68) 1.60 (1.62) 1.90) 2.95) 2.92 (1. Distribution by “micro” technological class ISI Technological Classes Electrical Devices.94) 3.30 (1.88) 3.81) 1. Measurement.56) 1.87) 3.49) 1.79) 1.59 (1.59 (1.21 (1.41 (1.51) 0.43 (1.59 (1.72) 1.45 (1.90) 2.70 (1.59) 1.72) 1.92) 2.97 (1.32 (1.72) 2.78 (1.67) 1.22) 1.42) 1.76 (1.66) 1.89) 1.76 (1.67) 2.75) 1.99 (1.13 (1.42 (1.64 (1.96 (1.09 (1.52) 3.78) 2.85 (1.88) 2.44 (1.10 (1.90) 1.06 (1.29 (1.58) 2.93) 2.51) 1.60) 1.03) 2.99) 1.72) 1.46) 3.06 (1.56) 3.86 (1.87 (2.88 (2.34) 1.39 (1.86) 2.44 (1.29 (1.83) 2.58) 1.52 (1.35 (2.93) 1.89) 2.95) 1.79) 1.72 (1.74) 1.10 (1.88) 2.85 (1.01) 2.77) 2.30 (1.23 (1.81) 2.90) 1.51) 0. Machinery & Apparatus Transport Nuclear Engineering Space Technology Weapons Consumer Goods & Equipment Civil Eng.95) 3.45 (1.64) 1.92) 1.75) 1..81) 1.42 (1.45) 3.38 (1.01) 2.92) 2.00 (1.90) 1.08 (1.10 (1.34 (1.84 (1.65) 1. & Food Chemistry Chemical & Petrol Industry.06 (1.36 (1.77 (1.82) 1.69) 1.19 (1.73) 1.03 (1. Building & Mining Total Laboratories 1.56 (1.59) 1.87) Note: standard deviations in parenthesis 61 .91 (1.17 (2.69 (1.88) 2.52 (1.75 (1.70) 1.53 (1.61) 1.71 (1.88) 2.91 (1.78) 1.91) 2.55) 1.20 (1. Pumps & Turbines Mechanical Elements Handling & Printing Agricultural & Food Processing.66) 1.79 (1.46 (1.85) 2.29 (1.61 (1.10 (1.88 (1.14 (1.45 (1.79) 1.36 (1.90 (1.80) 1.80) 2.67) 1.12 (1.89 (1.49 (1.07 (1.73) 1.62) 1.88 (1.46 (1.39 (1.96) 1.79 (1.79 (1.09 (1.25 (1.83) 2.75) 1.02 (1.44) 3.90) 3. Textiles & Paper Thermal Processes & Apparatus Environmental Technology Machine Tools Engines.55 (1.85) 2.51 (1.79) 2.76 (1.91 (1.10 (1.77 (1.76 (1.01 (1.30 (1.79 (1.74) 2.92) 3.96) 2.86) 2.39) 1.47 (1.87) 2.82 (1.14 (1.73) 2.21 (1.92) 1.09 (1.78 (1.92 (1.97 (1.69 (1.91 (1.50 (1.87) 3.85 (1.88) 2.81) 1.89) Conferences 1.91 (1.72) Patents 2.64) 1.73) Competitors 2.42 (1.42 (1.93) 1.18 (1.35(1.75) 2.71) 0.00) 2.51 (1.99) 3.63) 1.38 (2.15 (1.87 (1.35 (1.69 (1.45 (1.75) 1.20 (1.73) 1.90) 3.82) 3.85) 1.79) 0.41 (1.97) 3.58 (1.68 (1.76 (1.71) 1.46 (1.67 (1. Electrical Eng.02 (1.98) 1.48 (1.20 (1.77) 2.

followed by the patent literature.59) 2.59) 2.47) 1.39) 1. This might be related to the more basic and general nature of the research performed by these institutions.97 (1.68 (1.95) 1.56 (1.35 (2.14 (1.88) 1.49 (1. Institutions 1.74) 2.51 (1.18) 2.90) 2.61) 3.92) 0.62) Other Governm.91) 0.96) 1.63 (1. medium and small companies.88) 1.84) 2.51 (2.56 (1.95) 3.73) 2. In private research laboratories the ranking is similar to the latter.60 (1.50 (1.94 (1.87) 1.09 (1.58) 2.00 (2.55 (1.86 (1.30 (1.87) Public Research Institutions 3.51 (1.72) 2.24 (1.96 (1.09 (2.76) 1.86) 1. The participation in conferences and meetings ranks third.06 (1.80) Total 1.44 (1. Distribution by type of inventors’ employer Large companies Laboratories Scientific literature Conferences Patents Users Suppliers Competitors 1.58 (1.42 (1. when the inventors belong to public research institutions and universities.75 (1.86) 2.96) 3. the scientific literature and the other research laboratories are the most important sources of knowledge amongst those listed in Table 3.98) 1. The users are the most important source of knowledge for the inventors employed in large.33 looks at the importance of different sources of knowledge for different types of inventors’ employers.15 (1.88) 3.33 Importance of different sources of knowledge.12 (2.26 (1.81) 2. 62 .72) 2. The users.21 (1.13) 2.4 adds two tables on the use of different sources of knowledge when the patents’ applicants are located in one country as compared to the case in which they are located in more than one country.08) 1.80 (1.14 (2.35 (1.82) 2.Table 3.74 (1.63) 2. while the inventors affiliated to “other” government laboratories rank the sources of knowledge similar to those employed by private companies.72) 2.87 (1.25 (1.95 (1.33.39 (1.97) 3.88 (1.08) 1.76) 2.55 (1.80 (1.94 (1.69) Universities 3.79) Others 1.90) 1.25 (1. However.93) 2.90) 2.83) 2.42 (1.89) Small companies 1.87) Note: standard deviations in parenthesis Annex A III.11 (1. Table 3.80 (1.95 (1.86) 1.32) 1.16 (1.31 (1.14 (1.96 (1.83) 2.86) Medium sized companies 1.71 (2. the competitors and the suppliers are at the bottom of the ranking.45 (1.94) 1.77 (1.72) 2.19 (1. non-European origins.37 (1.67) 1.36 (1.66 (1.16 (1.89) 1.01) 2.22) 1.13 (1.27 (1.92) Private Research Institutions 1.91) 1. and when the applicants have European origins vs.97) 3.64 (1.84) 1.01) 1.57) 2.

But we should be aware that this substitution still might have an effect also in the nineties. The Bulletin contains bibliographic data as well as data concerning the legal situation of European patent applications and patents as laid down in Rule 92 EPC.1 The Data Source The ZEW constructed a data set based on the European Patent Bulletin. by country as well as by broad technology areas and the most probable economic sector to use. Also we admit that not all patents with priority year 2001 are included in this data base right now. Variables extracted from Bulletin are the dates of the application process (Priority date. This is done to document the explosion of the use of the European patent system in the last twenty years where the increase in the number of patent application is even stronger as in the increase in the number of patent applications to the US patent and trademark office which induced the academic and policy discussion about the recent surge in patenting. it is clear from aggregated data published by the patent office that the growth rate observed in the nineties no longer prevails in more recent years. However. Through out the analyses we use the following conventions: Some of the increase may be attributed to an substitution of various application to one or more national offices in EU. date of application for examination. There are a publicity lag and administrative lags of creating the data base.10 11 To lay the ground for the discussion on possible change in the (economic) value of patents this chapter documents the recent trends in patenting at the EPO. Hence these descriptive analyses might contribute to sorting out some explanations and highlight some facts which help pointing out direction where reasons for the recent surge in EPO patenting might be found. Hence. Because of these time lags reliable information are available for priority years from 1978 to 2001. 11 See Kortum and Lerner (1999) as an early reference for his still ongoing debate. and other late status information. The observations range from priority years 1977 to 2002. application date at EPO. The bulk of this reason for the increase in EPO patenting is probably occurring throughout the eighties.4 million patent applications were available.4. withdrawals. grant date) as well as information about possible opposition and the opposing company or person. By March 2005 about 1. Applications at the EPO through the PCT applications route are also included. For the purpose of this project we extracted detailed information about the applicants and inventor and calculated some descriptive statistics. we slightly underestimate the growth in patenting in 2001. 4. 10 63 . such as renewal payments. Because of the long period of examination and the physical distances between the PTO patent offices these patents enter the data base with a time lag of about 2 years. revocations. THE EPO PATENT EXPLOSION – SOME DESCRIPTIVE STATISTICS This part of the report examines the development of the number of EPO applications of time.

• • •

The technological field of a patent is defined on the basis of the main IPC code. “Nationality” of a patent is determined by the address of the first inventor.12 We use the priority data as the time stamp because the priority date is closest to the date of invention and hence to the decision to apply for patent protection or not.

Technology areas are defined quite broadly as patents primarily belonging the five areas “Chemical technologies”, “Drugs and Health”, “Electronics, Communication and Electrical Engineering”, “Mechanical technology” and “Others”. This broad classification of technology areas is widely used in the research in patenting (see e.g. Lanjouw and Schankerman 2001, Hall 2004, Kaiser et al. 2005). In order to allow for a more detailed and market based views we implemented the correspondence of patent classification and economic sector classification recent developed by Schmoch et al. (2003) which distinguish 44 different economic sectors based on the NACE classes. The use of this more detailed classification is limited to the analyses of large countries resp. country aggregates. It seems not useful to implement the more detail sector classification when looking at smaller countries or the new member states because the lower number of patent application will immediately the analyses might be biases to a small numbers problem. 4.2 Overall trends in EPO applications

First we start with examining the number of EPO applications overall as well as by the technology class for the large patenting countries resp. different aggregates of EU member states. Here we distinguish EU25, EU15 and the new member states (NMS). Figure 1 documents the rise in the number of patent applications at the EPO. We start in 1985 to avoid that the picture is too much influenced by the initial period of transition from pure national European patents systems to the regional EU patent system. From detailed analyses of major EU countries one can conclude that the first phase of this transition ended by the mid eighties. Figure 4.1 and Figure 4.2 shows that the rise in the use of the EPO system is clearly visible in all country groups not matter whether we determine nationality at the inventor as well as at the applicant level. Even when restricting our view on the nineties and compare the increase to the US patent system the rise in EPO patenting still outperforms the rise in the number of USPTO applications. Both figures also show that EU15 (EU25) is the single most important applicant to the EPO followed by the USA and Japan. However, we should avoid to look at this as an reliable indicator for the international production of technology because the so called home advantage lead to a higher probability of EU inventions to be applied for at the EPO than in the case of the same invention made in the USA or Japan.

12

We also worked with other possible conventions to determine the nationality of patents for counting applications i.e. random selection of one of the inventors, fractional counting, country of the first applicant. However, this does not materially alter our conclusions and hence we opted for the easiest and widespread methods.

64

Both figures show a decline of patenting in the early nineties for EU15 and EU25 aggregates which is mainly due to a slow down in patenting in major EU economies. Also, the number of patent application for Japanese invention is declining in this period. The slow-down is much less visible for the USA. In this period the number of patent applications from the New Member States unstuck dramatically. This is even more visible when looking at patent applicants instead of inventors.

Figure 4.1 Development of EPO Patent Application By Inventor Country

Figure 4.2 Development of EPO Patent Application By Applicant Country

65

In Figures 4.3 to 4.7 the same calculations by technology areas are presented. It is visible that the time series development of the number of patent application by country holds distinct technology specific patterns. So, in the field of “Drugs and Health” the early nineties slow down is less visible than in “Chemicals” or more in “Mechanical Technologies” or “Other technologies” in the case of the EU15, EU25 and the US where in the case of Japan the reduction in the number of patent application is most notable in “Chemistry” and “Health and Drugs”. The break down by technologies also highlights the rise in the importance of “Health and Drugs” as well as “Electronics, etc.” as the most dynamic fields of inventive activity.

Table 4.1 gives some indication on the use of the EPO system at the detailed country level. In order to avoid as small number bias for some smaller countries this league table is based on the sum of patent application of 1996-2002. Not surprising this table shows that within the EU Germany, France, UK and Italy are the larger countries in terms of the number of patent applications. Table 2 also provides a rough picture of technological specialization. When comparing the country shares we note that the ranking of countries might by different even when looking at broad areas of technology. E.g. one can immediately see from the table that Germany is dominating in mechanical technologies and much less stronger in “Drugs and Health”. For the USA the opposite specialization can be found. Similarly, Japan is specialized in “Electronics, etc.” and relatively weaker in “Drugs and Health”. Table 2 also highlights the low level of technological invention in the New Member States.

Figure 4.3 Development of EPO Patent Application by Broad Technology Areas - Chemicals

66

Electronics.5 Development of EPO Patent Application by Broad Technology Areas .Health and Drugs Figure 4. Communication and Electrical Engineering 67 .Figure 4.4 Development of EPO Patent Application by Broad Technology Areas .

Mechanical Technologies Figure 4.7 Development of EPO Patent Application by Broad Technology Areas -Other Technologies 68 .6 Development of EPO Patent Application by Broad Technology Areas .Figure 4.

417 27.604 5.082 0.131 16.481 0.161 40.989 5.058 0.663 21.495 0.711 0.066 0.012 3.111 0.001 0.795 6.367 0.021 0. This is done in Figure 4.001 Other 55.745 0.009 0.196 30.002 0.004 0.380 1.410 8.811 37.227 0.001 Mechanical 58.957 0.064 0.199 0.165 0.004 Total 46.651 0.344 0.002 0.045 6.509 2.312 0.076 15.026 0.093 3.964 0.005 0.222 57.800 5.330 0.895 0.023 0.170 5.932 0.244 0.245 0.677 2.900 0.743 0.006 0.644 46.104 0.402 0.037 2.582 2.688 0.152 0.762 2.170 3. NZ Denmark Israel Spain Other Asia Norway Other Europe New Members China (incl.369 0.135 0.386 1.311 0.966 2.523 0.342 0.031 0.061 0.911 1.237 0.134 0.045 0.267 0.267 2.272 0.435 0.040 2.013 0.220 0.510 42.199 1.014 0.013 0.774 0.966 0.002 Based on the inspection of the trends in overall patenting as well as the geo-political chance we also give the average annual compound growth rates by various periods.422 6.179 0. HK) Latin America Ireland Africa India Hungary Luxemburg Greece Czech. Poland Slovenia Portugal Slovakia Estonia Cyprus Malta Latvia Lithuania Chemicals 43.559 43. Rep.266 0.015 0.003 0.133 0.045 0.004 0.877 2.037 0.097 5. The 69 .284 1.093 0.140 0.443 1.004 0.028 0.395 1.436 0.163 1.058 0.040 0.010 0.104 0.553 23.843 0.069 0.116 9.003 0.411 0.019 0.085 0.009 0.530 0.303 0.929 1.1 Country Shares in Total EPO Application 1996-2002 by Technology Area (%) Country EU25 EU15 USA Germany Japan France Great Britain Italy The Netherlands Switzerland Sweden Canada Finland Belgium South-Korea Austria Australia.432 0.045 17.292 0.100 7.031 0.044 0.053 0.592 1.221 0.189 0.040 0.985 0.675 0.021 1.698 0.010 0.240 0.087 0.607 0.771 2.510 0.346 1.363 2.043 0.002 0.962 14.012 0.786 1.098 0.141 0.384 1.133 0.675 5.207 0.010 0.002 Electronics 40.005 0.145 1.221 0.8.Table 4.094 1.032 0.679 2.014 0.590 0.979 18.001 0.548 2.031 0.242 0.140 0.544 2.963 2.663 0.236 0.396 0.576 1.069 0.007 0.001 0.097 0.365 0.099 0.543 1.758 5.945 16.032 0.028 28.002 Drugs and Health 37.414 22.301 0.920 1.683 1.301 0.294 2.048 0.052 0.602 20.077 0.047 0.355 0.226 0.702 23.030 0.032 0.009 0.028 0.788 4.629 6.347 0.783 1.001 0.832 3.310 0.419 0.229 32.399 0.012 0.841 1.759 0.613 1.085 1.060 0.003 0.304 0.042 0.566 0.003 0.004 0.994 55.595 0.602 0.902 0.031 0.782 0.725 9.094 0.281 0.292 0.080 0.143 0.004 0.

0 0. in the technology area i in country j in the year t. The contribution of each technology area to 70 . The growth rate in EU patenting outperforms US patenting in Europe in the second half of the nineties where as EU growth was much smaller before.8 Contribution to Overall Patent Growth by Country Technology Areas 25.0 5. Figure 4.0 -10.overall trends in patenting come out even more clearly here.0 15. Hence. However. this expression decomposes the actual growth in patenting into a weighted sum of the growth in each area of technology weighted by the share of each technology. The growth rate is calculated as first differences in logs. ∆pitj refer to the change in the number of patent application overall resp.0 EU25 EU15 1988-91 NMS USA 1992-95 JAPAN 1996-99 Others 2000-01 Total 4.3 Contributions by to Patent Explosion by Technology Areas and by Sector In order to gain some more insights to technological shifts we also look at the contribution of areas of technologies as well as the probably sector of technology use to the overall growth in patenting. EU15 as well as other countries.0 Average Annual Growth (%) 20. For all country groups we see a strong decline in the growth of patenting at the EPO in 2000 and 2001. The figure shows an increasing momentum of EPO patenting until the year 2000 in EU25. Starting point for this descriptive analysis is a decomposition of the overall growth rate in line with the following formula ∆ptj = ∑ i j j sit + sit −1 2 ∆pitj Where ∆ptj .0 -5. as mentioned earlier we have to note that this decline might overestimate the real decline in the growth rate of patenting due to the fact that the year 2001 is not covered equally as the early years by the most recent EPO bulletin data due to lags in publication of patent documents via the electronic EPO bulletin data. This momentum is especially notable for the New Member States.0 10. sitj is the share of technology area in total patenting in country j in the year t.0 -15.

”.5 1. The results of this composition can be found in Figures 4. The bulk of the growth in overall patenting can be traced back to the technology area “Electronics. Somewhat surprisingly is the steady decline of “Drugs and Health” as contributing factor in US patenting which is less visible from the charts presented above.0 0. But keep in mind that the large contribution of this field not only results from the significant momentum of patenting but also from the shear size of this area. This is not at all surprising given the technological dynamics in the nineties.5 0. communication.9-4.the overall growth depends on both elements: the size of the technology area as well as the growth rate of patenting in this area. The Figures tell us that the growth contribution differ strongly by areas of technology.5 Growth Contribution (%) 2.12. Figure 4. This increase of the contribution is most notably in the EU15 and NMS whereas it roughly stays at a high level in the USA. The least contribution in case of the EU25 and EU15 results from patenting in “Chemicals” and “Other technologies”.5 -1.0 EU25 EU15 NMS 1988-91 USA 1992-95 1996-99 JAPAN 2000-01 Others Total 71 . etc.0 1.9 Contribution to Overall Patent Growth by Technology Areas – Drugs and Health Drugs and Health 2.0 -0.

0 Growth Contribution (%) 6.0 -3.0 8. Communication.0 -8.0 -4. Electrical Engineering Electronics.0 4.0 7.0 2.Figure 4.0 EU25 EU15 NMS 1988-91 USA 1992-95 1996-99 JAPAN 2000-01 Others Total 72 . Communication.0 Growth Contribution (%) 6.0 5.0 -2.0 -1.0 1.0 4.0 -6.0 EU25 EU15 NMS 1988-91 USA 1992-95 1996-99 JAPAN 2000-01 Others Total Figure 4. Electrical Engineering 8.0 0.0 0.10 Contribution to Overall Patent Growth by Country Technology Areas – Electronics.0 -2.0 3.11 Contribution to Overall Patent Growth by Country Technology Areas – Mechanical Technologies Mechanical Technologies 10.0 2.

The importance of IT is underlined also by the fact that computing and office machines is ranked three over the whole period and its weight is even more increasing in the nineties. The results are reported in Table 4. audio visible electrical equipment. First of all. the technological differences between countries are even more pronounced at the industry level than at the level of broad technological areas. In addition.0 2.0 EU25 EU15 NMS 1988-91 USA 1992-95 1996-99 JAPAN 2000-01 Others Total A deeper insight into the technological dynamic is gained when looking at the much more disaggregated level of potential use industries. automotive industry is especially strong in Europe and Japan whereas its contribution to US patenting is much smaller. We apply the above decomposition formula at the level of industry.5 1. However. medical equipment also plays an important role when looking at the patent explosion.0 -0. In addition. To enhance the readability the contribution of a sector is expressed as per cent of the overall patent growth.0 -1. the automotive industry is a main driver of invention activities are mirrored in EPO patent statistics.0 1.12 Contribution to Overall Patent Growth by Country Technology Areas – Other Technologies Other Technologies 3.2.5 -2. and industrial process control are place in the top quarter of the table.5 -1.Figure 4. The results of the decomposition of overall patent growth by economic sector yield an even more impressive picture of the importance of the information technology for the dynamics of inventive activity in the last 15 years. In addition to telecom and IT.5 0.5 2. In addition. In the case of the USA the leading role is played by the pharmaceutical sector.0 0. Growth Contribution (%) 73 . some other sectors strongly linked to the IT sector like electronic components. Given the small number of patent application of the New Member States this analysis can only be conducted at the EU25 level. communication technology delivers the largest contribution of the growth of patenting in Europe in the last 15 years.

87 1.29 0.45 0.32 1.04 -0.20 0.67 0.80 0.63 0.04 -0.10 0.42 0.08 -0.17 2.19 1.04 10.33 0.19 0.06 12.50 1.30 2.63 2.60 -2.05 2.75 0.66 14.52 -0.19 5.12 -0.05 -0.13 0.58 -3.04 -0.68 1.35 -0.99 1.26 -0.04 -0.18 2.50 11.92 Others 1996-1999 14.38 0.83 0.17 3.01 1.01 0.61 0.32 0.93 1.50 0.06 0.27 1.56 -0.19 2.39 0.86 4.84 0.16 7.30 -0.64 0.12 0.09 0.75 0.79 3.02 -0.18 0.95 0.94 -1.54 0.94 2.33 2.08 -0.16 0.51 5.10 5.18 0.33 3.78 4.49 7.23 0.75 5. consumer goods Electronic components Accumulators.84 -2.55 1.15 -2.12 -3.64 0.48 0.32 USA 1996-1999 16.12 5.07 1.37 0.22 1.12 0.13 0.56 -0.41 -0.01 0.07 0.39 0.44 1992-1995 23.93 7.63 0.65 0.49 0.10 -0.58 7.22 1.53 1.77 0.21 8.61 1.88 2.27 0.17 0.10 0.29 0.40 0.33 0.56 1.90 0.12 2000-2001 14.24 1.75 1.29 -1.72 5.86 4.72 0.12 -0.07 1.58 0.06 4.03 0.34 2.34 5.11 0.46 1.04 -0.45 0.05 0.44 2.85 Total 1996-1999 14.06 1.28 0.34 1.84 1.94 0.02 0.33 18.06 -0.15 9.40 2.79 2.56 -3.09 0.03 0.28 0.70 0.34 0.15 0.20 4.00 -1.68 0.45 -0.64 1.08 0.67 0.17 0.08 -1.23 14.26 0.34 0.16 1992-1995 12.48 5.02 1.99 1.03 2.65 1.27 15.72 2. battery Electrical distribution.11 0.04 3.51 -0.31 0.23 3.92 1.34 0.63 1.36 2.67 2.75 1.79 0.02 0.68 1.64 16.10 0.07 0.01 0.17 0.67 0.26 0.47 -0.51 3.03 1.10 0.54 1.11 -0.54 1996-1999 13.38 0.22 -2.17 4.42 5.31 -0.77 3.96 0.99 -0.56 1.08 3.80 1.03 15.92 1.69 2.65 1.64 -0.08 0.77 -0.51 4.19 -0.67 0.02 -0.26 18.04 0.66 1.71 4.68 1.70 0.68 0.57 1.78 0.74 0.90 1.05 1.59 0.34 5.78 12.66 1.09 0.98 13.05 2.03 0.03 1.02 1.28 3.41 0.02 0.36 5.54 -0.95 5.51 -0.00 1.13 0.03 -0.57 74 .07 -1.38 2.90 2.38 0.72 0.34 1.71 0.71 0.63 17.43 -4.80 2.81 3. generators.31 -0.69 0.80 2.58 3.10 0.62 1.06 1.71 13.11 0.16 0.64 1.Table 4.42 2.65 1.70 2000-2001 12.02 -0.59 1.71 0.83 1.92 0.23 1.86 8.65 2.2 Contribution to Total Patent Growth by Sector (in %) – sorted by size of sector contribution to patent growth EU25 Sector Signal transmission.31 -0.01 0.48 3.05 2.29 2.94 0.68 0.47 -5.31 -2.13 10.31 1. varnishes Basic materials Other chemicals Tobacco products W eapons and ammunition Basic chemicals 1992-1995 25.20 19.45 2.01 0.73 17.43 0.13 0.89 4.39 -2.87 1.89 4.02 0.86 0.05 0.59 10.02 0.58 0.02 0.18 1.32 6.07 2.20 0.42 2.00 0.40 2.10 2000-2001 14.26 6.31 7.56 0.30 0.35 6.22 2.52 0. telecom Pharmaceuticals Office maschinery.63 3.41 0.21 0.91 18. radio.13 8.86 5.95 6. clocks Man-made fibres Paints.00 0.40 0.02 0.42 29.00 5.23 2.06 0.86 3.29 -4.79 4.96 9.11 0.95 9.19 0.17 0.65 0.61 5.68 29.59 1.91 1.46 0.36 2.79 6.23 0.56 1.75 22.35 0.48 -1.20 0.42 1.06 0.49 2.03 -0.12 0.03 2000-2001 15.36 0.04 -0.21 0.09 0.99 1. wire.87 1.02 -0.06 0.73 1.31 0.27 0.87 -0.11 1.45 4.10 1.02 -0.35 1. cable Non-metallic mineral products Electric motors.33 0.07 2.20 0.16 0.55 0.04 -2.07 0.29 0.04 -0.53 0.21 0.24 1.15 -1.53 2.48 2.07 0.66 0.03 0.77 0.09 0.00 0.29 0.16 -0.11 0.36 7.00 -0.40 0.05 3.17 1.42 1.32 1.70 3.41 -4.59 3.59 0.79 1.41 -0.20 0.69 2.61 -7.39 0.93 -1.47 8.84 1.44 -5.90 -0.55 2.20 0.57 0.64 12.29 3.18 0.38 0.70 7.20 12.21 3.05 -0.55 0.12 0.88 1.27 1.43 0. beverages Paper Pesticides.70 0.16 0.40 0.57 3.92 6.35 2. Measuring instruments Energy maschinery Rubber and plastics products Fabricated metal products Non-specific purpose maschinery Domestic appliances Optical instruments Furniture.72 2.68 1.27 10.02 1.16 0.42 -0.22 4.06 2.53 0.22 1.36 -0.32 2.14 2.36 1.28 0.33 4.68 0.09 2.74 0.62 10.12 0.23 2.93 -0.02 -0.49 1.71 0.16 -8.57 8.54 5.23 7.35 2.09 Japan 1996-1999 15.77 0.57 0.09 2.47 2.51 0.45 -0.52 -0. computers Medical equipment Motor vehicles Television.53 2.15 8.93 0.00 -0.11 1.34 0.73 2.13 4.68 1.26 0.32 0.98 15.51 0.91 0.47 0.74 0.18 -0.14 0.25 0.48 -0.10 -0. detergents Machine tools Food.66 0.20 0.81 -0.04 0.44 3.69 1.79 -1.10 0.54 11.41 2.15 1.27 18.28 1.10 0.57 -8.18 1.14 1.98 -0.71 0.57 0.12 3.26 0.63 1.21 0.65 1.27 0.72 0.30 5.83 1.02 0.73 1.90 0.04 4.83 2.52 2.64 1.96 3.68 12.14 1.71 -2.07 -1.60 0.82 -0.18 18.70 0. transformers Other electrical equipment Special purpose maschinery Other transport equipment Soaps.03 2.55 1.55 9.59 1.57 4.12 0.91 0.29 1.73 3.83 10.42 0.22 2.90 1992-1995 15.57 -0.00 1.28 1992-1995 -9.97 0.04 4.77 2.41 9.39 1.02 -7.28 0.66 0.42 -0.89 -0.03 2.24 -1.03 -0.41 2.86 3.25 0.11 1.51 0.83 2.54 1. nuclear fuel Industrial process control equipment Textiles W ood products Leather articles W atches.67 0.95 2000-2001 12.82 0.46 2.05 0.09 0.79 0.92 -2.85 13.07 -0.06 0.60 0.02 0.47 -6.71 0.01 4.42 0.73 0.03 -0.43 3.87 2.34 0.19 2.73 2.06 0. agro-chemicals Lightening equipment Agricultural and forestry maschinery W earing apparel Petroleum products.82 0.23 4. audiovisual electr.15 0.12 0.16 0.21 2.16 2.86 26.33 0.34 3.22 0.46 0.63 3.

US and other countries. and we provided descriptive statistics along with comments about the main patterns that emerged from the data. mainly because of the lack of comparable innovation surveys that are repeated over time. technologies. countries. CONCLUSIONS AND RESEARCH PLAN FOR LOT2 This Report provided detailed data and a comprehensive review of the literature on the economic value of patents. the survey of the literature highlighted the main gaps in the existing literature that would need further investigation either from the theoretical and empirical point of view. We collected the references (scientific papers. type of inventors’ organisation and type of contribution. Some contributions describe the casual link between the changes in the patent premium and R&D in a cross-section perspective for the US. but not vice-versa. Moreover. by summarising the main contributions on the four Themes of our study. and their changes over time) and the investment in innovative activity. Some other studies define the dynamic relationship between patenting or patent value and firm performance. and it does not provide strong policy guidance. firm performance. more information would be needed to evaluate whether patent quality changes significantly over time. 75 . The Report is composed of 5 Sections and 3 Annexes. For each of these Themes we analysed the background literature. most studies on the relationship between R&D and patenting focus on the impact of R&D on patenting activity. For example. Section 2 presented the methodology employed in the survey of the literature and the database that we created on the surveyed publications. Longitudinal studies on this topic are missing. Annex II reports the data on the coverage of the literature by Theme. However. In particular. It is indeed the latter which needs to be analysed in more detail in order to evaluate the intended economic objective of the patent system. and their changes over time. this literature does not employ direct measures of the drivers of the premium afforded by the patent. it is not clear the extent to which the increasing strategic use of patents hinders or stimulates innovation. and employment. economic growth. Moreover. Section 2 also summarised and discussed a representative part of the background literature. Section 1 described the aim and the organisation of the Report. etc. little work has been done on the relationship between the drivers of patent value (such as the characteristics of the litigation environment. and how these changes relate to the overall innovation and growth patterns. However.) on various aspects of the direct and indirect effects of the patent system in Europe. The analysis of the background literature helped us refine the plan for research in Lot 2 of this Project. the empirical evidence is missing for Europe. in particular with reference to the social costs of patent-litigation activities. books. Similarly. The Report moved along 4 Themes. Reports. but they do not distinguish between the impact of patent protection and the impact of innovation. Few studies systematically quantify the social costs of patenting.5.

(B2) Collaboration. and patents that are not used at all – i. (A2) The economic use of patents and the motives for patenting.A few studies have analysed the determinants of licensing and the development of markets for technologies as devices to increase the use of patents. prevention from imitation.e. theoretical and empirical studies on the extent and the determinants of the “unused” patents are scarce or inexistent. and we described the differences among the 6 EU countries. we found out that there is a strong need for further research on specific issues included in the Themes analysed in this Report. Europe and the New Member States). licensing. blocking competitors. systematic evidence on the formation of new firms from the patented inventions in the private sector. licensing. but that received very little attention in the literature. and in some large technological fields: there are technologies like mechanical engineering and electrical engineering in which about half of the EPO patents are classified. blocking competitors. however.e. commercial exploitation. Many patents have multiple inventors. especially for Europe and the US. This Section presented the main patterns that emerge from the data by NUTS2 regions (as classified by Eurostat). As far as the indirect effects of patents are concerned. There is not. We also analysed the private and social costs of the European patents. which suggests that the innovation activity is organised around 76 . applicant organisations. We described the distribution of the monetary value of the PatVal-EU innovations. We defined six possible uses of patents (i. In particular the evidence is weak in some countries and regions (i. and the causal link between such flows of information and firms’ R&D productivity. European regions. Section 3 reported and commented on Tables and Figures of selected descriptive statistics based on the PatVal-EU dataset. All in all. However.e. and reputation). and in Europe or in countries other than the US. and the applicant organisations. technologies and applicant organisations. especially for the US. The 4 Themes analysed in this Report are the following: (A1) The value and social costs of patents. although many contributions studied the relationship between patents and spillovers (mostly by means of patent citations). analysed the extent to which knowledge spillovers are conditioned by patent protection. with only few patents that yield large returns. the “macro” and “micro” technological classes in which the patents are classified. licensing & use. cross-licensing. and technological classes. technological classes and applicant organisations. We reported the share of patents in the PatVal-EU database used to start-up a new firm. These statistics were reported by countries. and we analysed their distribution across countries. a few contributions. spillovers and the sources of knowledge in the innovation process. “sleeping patents”) and six motives for patenting (i. there are some studies on spin-off firms from patented inventions by university scientists. (B1) The creation of new businesses from the patented innovations. It would be crucial to fill this gap to derive policy suggestions on how to increase the use of property rights. country. Finally. and we confirmed that such distribution is skewed. cross-licensing.e. internal use. the European regions.

teams of researchers. The vast majority of these co-inventors belong to the same organisation and are geographically close. Moreover, the most common source of knowledge in the innovation process is the interaction with the customers. University and non-university research is rarely used. These statistics were provided by European countries, regions, “macro” and “micro” technological classes and type of applicants’ organisations. Section 4 reported selected Tables of descriptive statistics on counts of EPO patent applications from 1986 to 2001, and provided trends over time and across technological classes. By showing aggregate data at the level of US, Japan, EU-15, EU-25 and New Member States, Section 4 provided preliminary insights on cross-country comparisons. An increase in the use of the European Patent System is visible in all the technological areas and in nearly most sectors of the economy. However, the momentum of the increase strongly differs by technological area and sector, probably reflecting differences in the technological dynamics. The differences in the momentum of the patent value might help explain the differentials in the dynamics of patenting in Europe. This is suggested by the fact that the dynamics of R&D inputs by sector differ from the dynamics of patenting. Further research is needed to study the relationship between the growth of patents and R&D. The Annexes at the end of the report supplied additional information on: (1) the organisation of the dataset for the survey of the literature; (2) the coverage of the survey of the literature; (3) Tables of descriptive statistics produced by using the PatVal-EU dataset that complement the data presented in Section 3. It is worth stressing that the analysis of the background literature and the empirical work with patent data performed in Lot 1 provided useful suggestions to guide the research in Lot 2. On the one hand, as noted above, they suggested that there are still many open problems and questions to be addressed, and they showed the gaps in the empirical literature. For instance, we found that an important gap in the literature concerns the impact of patenting activity -- and in particular the impact of specific characteristics of patents, applicants and invention processes -- on R&D. This will inspire the empirical models to be developed in Lot 2. On the other hand, the analysis of the data in Section 3 and Section 4 of Lot 1 provided a general base for guiding the empirical investigation in Lot 2, and the search for policy suggestions that are expected to arise from the results of the studies in Lot 2. For example, the descriptive statistics pointed out that there are interesting differences across European countries and regions, technological classes, and types of inventors’ employers (i.e. private firms vs. public organisations; companies vs. universities and other research institutions; large vs. small firms; domestic vs. foreign applicants) concerning the value and the cost of the European patents, their use, the formation of new firms, and the occurrence of spillovers and collaborations among inventors and institutions in the innovation process. More systematic empirical analysis is needed to explore these differences and their causes. More precisely, in order to identify the net effect of different factors on the variables of interest (i.e. the value of patents, the propensity to start a new company, etc.), we need to perform multiple correlation analyses based on specific econometric models. Against this background, we plan to produce a series of scientific studies concerning the advantages and disadvantages of the patent system, and in particular the direct and indirect impact of patents on 77

the economy and society. For each Theme we will develop aggregate econometric contributions and contributions at the level of specific countries, technological classes and inventors’ employers. We will also integrate the datasets used in Lot 1 with additional data required for the specific purpose of the different studies. The remaining part of this Section presents a detailed research plan for Lot 2 in line with the original research project.
(A1) The Economic Value of Patents

The study of the economic value of patents will use several indicators of the importance of patents to proxy for their value, as suggested by the literature in Lot 1. We will use regression-based empirical models to estimate the economic value of patents.
Study A.1.1. The Value of Patents from Indirect Indicators

The contributions under this heading will study the value of patents by using indirect indicators, and they will develop through three empirical studies. The first one will employ indicators that have proven to be correlated with patent value such as the number of forward citations, the number of claims, the patent family size, and to some extent the occurrence of opposition procedures. These indicators will be used to develop an “importance index” of the patent rights. The importance index will then be linked to the monetary value of the patents by using the renewal information contained in the patent database. In so doing, this work will follow the methodology pioneered by Ariel Pakes and Mark Schankerman that we cited in Lot 1. The ensuing estimates of patent values can be assessed for different countries, regions, technological areas, and types of applicants. The second analysis will focus on the private value of patents for the applicants. It will use the methodology proposed by Harhoff, Scherer, and Vopel (2003b, cited in Lot 1) to provide a direct assessment of the value of a patent right to the owner. The analysis can differentiate the applicants in the ZEW-EPO patent database into 3 categories: individual applicant, private enterprise, and public institution. This will be done by analysing the text in the patent document with the names of applicants. We will also take into account the administrative differences among the patent systems in Europe. For example, not all European Universities are allowed to own patents. In countries such as Sweden the Professors will be the applicants or the owner of the patents, whereas in other countries (e.g. Germany) the public institutions such as the Universities and Colleges apply for the patents. Apart from the assessment by applicant types, we will estimate the value of patents by country and technological class, and the results of the two studies will be compared to check for the robustness of the results. The third analysis will look at the value of patents as proxied by the efforts undertaken by the assignees to legally enforce their patent rights. Harhoff and Reitzig (2004, cited in Lot 1) found that European patents that have been opposed and upheld either unchanged or amended are on average more valuable than the average patents. The private interests of the parties either in an opposition procedure or in a litigation case are related to the efforts that the parties undertake as proxied by the costs involved. We will produce case study evidence from data on patent litigations in Germany 78

from 1993 to 1995. The estimated cost of enforcement (probability of litigation and average costs) will indicate the effort that an applicant undertakes to enforce his patent rights.
Study A.1.2. The Value of Patents from Survey Data

Our second study will rely on the measure of the economic value of patents gathered from the PatVal-EU survey, and it will combine this information with data provided by the Regio database on the European regions, and those on firm characteristics drawn from the Who Owns Whom and Amadeus databases. In the PatVal-EU survey we asked the inventor to provide a quantitative estimate of the value of the patent that they contributed to invent by using a scale that goes from zero economic returns to “more than 300 million euros”. We examined the distribution of the patent value in Lot 1. While we should be aware of the potential shortcomings of this measure, we believe that this is a fairly reliable measure of the economic value of patents. Moreover, it is the only available measure we know that bases the estimates of the value of the patent on the correct economic concept, viz. the expected value of its future stream of profits at the moment of grant. In other words, compared to other indicators of the value of patents, which rely on their importance (e.g. citations, claims), not only is this a direct measure, but it is also the only one that provides a monetary estimate of the value. The empirical work will study the determinants of the economic value of the patent. It will use an ordered probit regression where the value intervals are the dependent variable. Our regressors are drawn from the PatVal-EU database and from other datasets to explore the impact of four sets of determinants: i) characteristics of the organisation in which the patent was developed; ii) characteristics of the inventors; iii) characteristics of the patent; iv) characteristics of the location in which the patent was developed. To our knowledge, this is the first attempt that investigates the impact of such a comprehensive set of determinants on the value of patents. It also presents several novelties with respect to previous research on this issue. For example, by using the PatVal-EU data, we will assess the effect of factors that were ignored in previous studies like the inventors’ characteristics (e.g. age, past productivity, educational degree). In addition, it will enable us to understand empirically the relative importance of different factors affecting the value of patents. For example, how important are the technological characteristics of patents in determining its value? That is, is patent value largely determined by the sector or type of technology, or are there differences depending on the individual inventor, the organisation or the location? How important are the inventors’ characteristics vis-àvis the type of applicant organisation? Do more valuable patents depend on “star” inventors, or are they explained mainly by organisational characteristics, like the greater resources provided by the large firms or the more creative atmosphere of the small firms? Interestingly enough, the latter situation suggests that shopping for talents would not be crucial for an organisation, as the proper organisational setting can turn most individuals with suitable characteristics into good inventors, while the opposite is true in the former case. Similarly, our analysis enables us to assess the relative importance of agglomeration economies, spillovers and local factors more generally. 79

We will check the validity of our monetary measure in three ways. by allowing for an impact of the patent system on R&D investment we will shed some light on the ability of the patent system to overcome the externality resulting from limited appropriability of R&D. cited in Lot 1). 80 . we will regress five traditional indicators of patent value on our value intervals. Value of Patents and Relationships with R&D Our analysis of patent value cannot neglect the relationships between patents and R&D. this index captures the combined effect of the underlying indicators. Furthermore. which are intertwined in the study by Bessen and Hunt. In a first step this approach will be applied to EPO applications of firms in selected countries like Germany. We will estimate a patent production function and we will introduce a methodology to separate productivity effects in the invention process from behavioural effects that are also present in the patents – R&D relationship. and an equation that models a firm’s application decision. application year and country dummies as controls. along with technology. In our survey there are 354 French patents whose value question was submitted to both the inventor and to a manager responsible for the development of the patent. we will construct a patent value index using the same methodology and indicators employed by Lanjouw and Schankerman (2004). The Monetary Value of Patents and the Use of Multiple Indicators By using the PatVal-EU data we will provide an assessment of the monetary measure of value against alternative indicators.3. First. If we find that patents have a negative impact on R&D this will give some hint on the social value of the patent system as the latter would fail to provide R&D incentive. rests on the introduction of the invention production equation. Study A.4. and vice versa. The problem is probably not that severe in the case of the smaller firms or non-profit research labs. a third aspect of our patent value measure is that the individual inventors may not know about the value of the patent as much as the managers who are responsible for their development. In addition. To do this we will combine data from the German Mannheim Innovation Survey with EPO data.Study A. We will also try to collect similar data on R&D for other European countries in order to extend the coverage of this study. On comparing the two distributions we will find out whether there is just a slight overestimate of the patent value by the inventors. but it can be more serious in larger firms wherein the organisational distance between the inventor and the managers responsible for their development can be notable. As they note. In particular we will also try to assess the impact of specific characteristics of the invention process drawn from the PatVal survey data on the incentives to invest in R&D. we will not only use simple patent counts as indicators but also try quality-weighed patent counts. and our value intervals on the five indicators. The weighing procedure will be developed during the project. which is the basic rationale for the patent system. We will employ a methodology similar to that of Bessen and Hunt (2004.1.1. The basic rationale for this is that Germany shows an even stronger increase in the patents to R&D ratio than the US. Based on a merge of patent and firm data we will estimate econometric models to determine the relationship between patents and R&D. The identification of both effects. Finally. or if this is a serious issue to deal with. In our second check.

This will enable us to study the extent to which patents are exploited for economic purposes in Europe. the empirical analysis on licensing was mostly centred on historical approaches and case studies in particular 81 . iii) the patent was only used internally by the owner and iv) the patent was not used. An empirical investigation from European Survey Data The last two decades have witnessed a variety of arrangements for the exchange of technologies. licensing has become a necessary means to exploit firm technology and to sustain their competitive advantage. Study A. Small firms are expected to be more likely to license. To our knowledge. In this study we will move one step further by exploring the determinants of these choices. We will also identify the characteristics of the unused patents. iii) internally used and licensed. namely i) the patent was licensed ii) the patent was licensed and used internally by the owner. We expect that the probability to license a patent is positively influenced by the patent scope.(A2) The economic Use of Patents and the Potential for Enhancing It The PatVal-EU dataset enabled us to obtain information on whether the patent in the sample was: i) used internally by the assignee for economic purposes. Especially in technology-based sectors.2. large firms are expected to be more likely to license non core technology. the empirical evidence is still very scattered. mostly due to the limited availability of adequate data. other patent uses. we are interested in estimating the value of the unused patents. when licensing. In particular. The aim of this contribution will be to investigate the effect of different covariates on the probability that a patent is licensed. iv) left unused. The Determinants of Licensing vs. notwithstanding the richness of the theoretical literature on licensing. This can be aggregated at the level of countries or technological areas to provide an estimate of the under exploited value of patents. Moreover. and since there are repeated observations for the most patenting firms. and type of applicant. So far. We will then use our estimates using the PatVal-EU sample to draw inferences for the populations of patents. and the characteristics of the patents employed in the market for technology. This is an important contribution to the existing empirical literature on licensing for several reasons. technological class. this is one of the very first attempts to empirically estimate the extent of the potential economic value of unused patents. We will employ a multinomial logit specification that separates the effects of covariates on different patent fates. by using the estimates of the value of patents produced under the A1 Theme we can estimate the potential value of patents that are left unused. and how they are used. and that it is greater for patents employing more “basic” knowledge.1. These results are consistent with a theoretical background that merges the transaction cost approach and the resource/capability based view of the firm. We will develop empirical models that predict these choices. This study will use the PatVal-EU data to present new empirical evidence on the determinants of the decision of firms to license a patented innovation. Lot 1 produced a series of descriptive statistics on the share of patents in each of these categories by country. ii) licensed. we will control for firm heterogeneity. First. It will also use data provided by Who Owns Whom and Amadeus on firm characteristics. Since our data are both patent and firm level specific.

2. but it can reduce the actual use of the patents. “sleeping” patents). the linkage with science. (B1) New Firm Creation and Employment The PatVal-EU survey asks specifically to the interviewee whether the patent was used to create a new firm. Several recent papers have emphasised how important from a policy perspective is to understand if patents are actually used or left unused (Scotchmer. descriptive statistics of this phenomenon by country.1. the competitive environment (spillovers from competitors or presence of innovation race). the markets for technology. Arora and Ceccagnoli. Hall and Ziedonis. Only a few recent studies use extensive databases (i. and the different motives for not using a patent (i. One of the main concerns in the literature is that the increasing strength and effectiveness of patent protection may increase the propensity to patent. and they are not used by the patenting company. Fosfuri. the motivations to patent. to block competitors vs. The empirical analysis on this issue is scarce.industries or on small sample studies. and 82 . technological class. 2004. it aims at uncovering the determinants of different uses of patents by the firms: internal use. 1998. It will also look at the determinants of the “non use” of patents: strategic (i. 2000. These determinants are: the patent scope. 2004). blocking) and sleeping patents. By using the PatVal-EU data together with data on firms drawn from Who Owns Whom. study. Ziedonis. 1991. Moreover. the availability of complementary assets. 2001. 2004). Country and technological controls will be included. Our paper represents the first extensive analysis conducted on surveyed patent-level data. Mazzoleni and Nelson. This is a key information because we can estimate the different effects of our covariates on different behaviours of firms. Once again. our empirical analysis can detect significant low magnitude effects on the probability of licensing when distinguishing between all the possible outcomes for a firm’s patent. Study A. non-use. especially by dominant firms. and the advantages and drawbacks of the patent system. the size of the firm. Another important advantage of our data is that they allow for the distinction on whether the patent has been only licensed. finally. it was not used. In our study the information on whether the patent has been licensed is available at the level of the single innovation. because they may block potential uses of the inventions for different applications by other actors. the breadth of patent protection.e. we will derive hypotheses on the factors that might explain the outcome of a patent. Anand and Khanna. On the use of patents: a quantitative analysis based on the PatVal-EU data The goal of this work is closely related to the A.e. or it was only used internally. The social risk of unused patents is larger if patents have a broad scope. an important novelty of this work lies on the joint examination of the determinants of the patent uses. By using the existing literature on the determinants of licensing. or. Moreover.e. This allows us to assess more sharply the variables that play a significant major role on the probability of licensing. or it has been both licensed and used internally by the firm. and the cost of projects. Our contribution will cover this gap. by using a large dataset.2.2. Amadeus and Delphion.

As noted in Lot 1. Firm creation and the value of patents By using the PatVal-EU data. However. From the literature and from the relevant trade press we will look for estimates of the weight of the value of knowledge assets to the total value of start-ups or similar firms (possibly by country and sectors). To do so.type of applicant was performed in Lot 1. can then be used to predict the employment potential of these firms. We will assess the contribution of the patents that give rise to new firms in two ways. We will then refine the analysis from the feedbacks coming from the actual empirical implementation of this study. and then infer the value of these firms. organisation. region. By using our estimates of the patent value we can also predict the value of the new firms based on intellectual property. we will use the characteristics of the patents to predict their probability of producing a new firm from the logit regression. Clearly. we will assess the employment potential of these firms. This is a lower bound of the value of the new firms. As principal covariates we will start with the usual set of inventor. The direct effect of the patent system on employment is very difficult to detect empirically. Again. The product between the estimated value of the patent and its predicted probability of giving rise to a new firm will give an estimate of the value of the knowledge asset of the firms created from intellectual property. especially in some countries.1. These statistics revealed that this is not a rare event. (B2) Knowledge Spillovers from Patents The goal of this analysis is to assess the extent of the social benefits of patents produced by knowledge spillovers. from the literature and from the relevant trade press we will try to single out information in order to correlate the value of these firms with the size of their employment (possibly by country. sector. we will begin by running logit regressions with a dummy for whether the patent gave rise to a new firm or not. This will then be used to estimate the total employment of firms created from intellectual property in Europe or by individual countries and technological areas. This contribution will provide important results about the indirect effects of the patent system on employment and competitiveness in different European regions. Second. this analysis is important as it can suggest that while the inherent rationale for patenting implies a reduction of the social ex-post output (as the patents grant 83 . The value of these firms. Study B. we will infer the value of the patents that produced a new firm. We will produce empirical models that predict the creation of new firms from patented innovations. but we think that entrepreneurship is a good way for exploring and assessing indirect effects on employment and regional advantage. along with other characteristics. this kind of analysis requires a great deal of inferences from many sources. We will then estimate the value of these patents from our analysis in Study A1.2. there is so little evidence on this very important issue that such inferences are nonetheless most informative. with breakdowns by country and technological areas. and sector characteristics.1. First. or other dimensions).

Most of them describe the spillovers as merely being “in the air”. This would suggest that: a) the availability of patents does induce their use by others (otherwise why should strategic patents be cited less than others?). To do so. sector and country specificities. and they will enable us to estimate the impact of different motives for patenting on how much the patent is cited. innovative activities cluster geographically because they benefit from localised knowledge spillovers. As discussed in the tender for Lot 1. patent citations have been used by the literature as a measure of spillovers. universities.g. The third step of the research on spillovers will focus on the type of patents that are most used by other patents. An empirical analysis of the effectiveness of different means of knowledge spillovers from patents R&D activities concentrate geographically.). etc. carrier.a monopoly). blocking rivals and prevention from imitation. and it will compare them to the use of patents to transfer knowledge. large firms.1. Study B. the paper will use a non-parametric approach to highlight the average performance of the innovation process in different sub-classes characterised by different “coordination” mechanisms. population density. one limitation of these studies is that they do not explain the sources of such spillovers. Like in production. that too many patents are applied with the sole scope of blocking rivals. It will also suggest the extent to which the knowledge in that patent is used in the research leading to other patents. sectors. this contribution will investigate the mechanisms through which knowledge spillovers arise. We will devote particular attention to the “strategic” motives for patenting. In addition. etc. The analysis under this Theme will provide the basis for a careful study of this issue. viz. The second goal of the paper is to compare the relative effectiveness of these coordination mechanisms to produce valuable innovations. Although there is an extensive literature on the existence of knowledge spillovers and the benefits of locating in the geographical cluster.2. we will test whether the patents that have been applied for strategic reasons are cited less than others. the social value of patents would be enhanced by limiting the number of patents that produce little spillovers. however. presence of scientific institutions. The econometric analysis also considers the case in which the inventors use contemporaneously different mechanisms to transfer knowledge in order to produce a specific patent. Patent uses and patent spillovers. and the characteristics of the inventors (education. 84 . The aim of the study is twofold. number of patents invented in the area. These data will be complemented by the information drawn from Regio. By using the information provided by the PatVal-EU survey on the sources of knowledge used to develop the patents. regions. economies of scale stimulate firms to locate R&D in one geographical area. etc. small vs. their contribution to knowledge spillovers implies that such limitations on social welfare are in fact less severe. Patent citations to the PatVal-EU patents will be used as a measure of the spillovers that arise from the latter. By controlling for many factors. the characteristics of the regions (e. it will estimate the probability to use these mechanisms conditional upon the characteristics of the organisations (e.g. If this is so. Specifically. First. There is concern. it will describe the use of the different mechanisms for transferring knowledge across types of firms.). The PatVal-EU survey identifies the motives for patenting.). Who Owns Whom and Amadeus. and countries.

In short. by means of valuation techniques similar to the ones used in the theory of the firm. to move to a different public research organisation (including another university). the country.b) the social costs of strategic patenting are higher than the mere monopoly on the invention. This is the first quantitative assessment of this phenomenon and it is the basis for a set of econometric models that try to explain how different factors affect the mobility of academics and their choices: to stay.3. With the help of the PatVal-EU data we will discuss the characteristics of European university patents in terms of ownership. Study B.e. or the characteristics of the knowledge base. as they also imply lower spillovers. and more generally their social advantages vis-à-vis their social costs.2. we expect our policy implications to focus on issues such as: (1) 85 . by estimating the effect of spillovers on the present discounted value of the research unit adjusted by the replacement cost of intellectual capital. etc. defining as such the network of inventors that arises from the research done to create new patentable knowledge. Spillovers at the level of the research unit The purpose of this contribution that uses the PatVal-EU data will be to assess the existence of spillovers at the research unit level. If so. to move to the private sector. technological area and type of applicant whenever meaningful. We will construct three different types of spillover pools. and therefore our measure can be interpreted as the present discounted value adjusted by the renewal cost of intellectual capital. we expect the different spillover pools to come up positive and significant. We propose to identify a research unit in a novel way. We will assess empirically the importance of spillovers at the research unit level. the technological sector. Mobility away from academia is a significant phenomenon at least for the sub-sample of university researchers that hold patents from the European Patent Office. country of inventor and mobility. The econometric models will provide evidence on the relationship between the value of the patents and the probability of the inventors to move to a company. policy can meaningfully manoeuvre the extent of the spillovers. C) Policy Implications Under this Heading we will provide a comprehensive discussion of the policy implications for enhancing the value of patents. are young researchers more likely to move than senior researchers?). After correcting for sample selection and endogeneity. Study B. technological class.2. Spillovers through labour mobility from academia to business This study will present the first representative set of information on patenting and spillovers in the Academia in Europe. their use. we will try to understand whether different motives for patenting lead to different degrees of spillovers. From the previous discussion of our empirical studies. following the empirical analyses of the previous headings. We will break down our analysis by country. Then we will focus on the analysis of the university researcher mobility. We will also check for other factors like the characteristics of the inventors (i.2.

g. employment.on firm and subsidiaries characteristics . R&D expenditures) and countries/regions (e. Firm and country/region data provide the necessary covariates for our empirical studies.(available at CERM/LEM) WOW (Who Owns Whom.g. We will integrate the PatVal EU survey data with data from the patent documents or databases (e. (3) Policies for reducing the social costs of patents. We plan. DEP and CIS data for France. 1997) . while the other survey in the other two countries will contribute to develop a better understanding of the patents’ invention process and value in New Member States. oppositions. intensity of technological activities). GDP.g. Denmark and an additional New Member State. We already described in the Project the databases that we will use in our work. DELPHION – for additional patent variables (like number of claims. (2) Policies for enhancing the economic use of patents like licensing. by taking into account that different motives for patenting can give rise to different spillovers). Datasets The empirical analyses that we plan to conduct in Lot 2 require the combination of different datasets possessed by different members of the team. AMADEUS .Policies for encouraging factors that have shown to be important determinants for producing high value patents. markets for technology. renewals).(available at CERM/LEM) Country specific datasets on firm R&D Expenditures (like Mediocredito Centrale for Italy. Hungary and another New Member State (Estonia or Slovenia) with the PatVal-EU survey. however. All our datasets have now been cleaned and organized. citations. Denmark will represent a small Scandinavian country. (6) Policies for a better assessment of the strength of the patent protection by a more careful fine-tuning of social values and social costs of patents (e. population. or else. and with data at the level of firms (e. the Mannheim Innovation Panel for Germany). ISI. We list them again here: • • • • • • • • ZEW-EPO Patent Dataset. number of technological classes used as proxies for patent scope and protection). with the extension of the survey Hungary. and they are ready to be used.on firm characteristics . size. By complementing the patent data with other firm and regional variables we will perform a comprehensive qualitative and quantitative evaluation of the impact of the patent system on the economy and society at different levels.g. to cover Denmark.(available at CERM/LEM) REGIO – on regional characteristics . (4) Policies for enhancing R&D activities (5) Policies for enhancing employment effects from patents like the creation of new firms. technological class. INIPI and OST classifications for the technological fields in which the patents are classified according to the International Patent Classification (IPC) 86 . PatVal-EU Datasets. financial data. applicant.

87 .• The ZEW Search Engine to merge firm and patent data.

European Commission. Working Paper. 45. (2001). “Competition. “Determinants of innovative activity in Canadian manufacturing firms: the role of intellectual property rights”. D. Arora. and Merges. 451-475 Arora. Khanna. Arora. Kogut. 7.” Research Policy.417-442. Management Science. Arora.. what is the question? Revisiting the patent controversies. DG XIII. 467-492. MIT Press. T.611-624.. RAND Journal of Economics. Innovation Strategies of Europe’s Largest Industrial Firms: Results of the PACE Survey for Information Sources. “The Relative Effectiveness of Patents and Secrecy for Appropriation”. 23. Yao. und Pianta.. in: OECD (ed. 30. pp. Vickers. ”The changing technology of technical change: General and abstract knowledge and the division of innovative labor. P.A. A. forthcoming. No.. “Networks of inventors and the role of academia: an exploration of Italian patent data”.J. (1996).103-135. 48. Fosfuri. B. (1996). A. 35 (1). R. “R&D and the Patent Premium”. 127-145. Feldman. Innovation. A. Anand. Arundel. No. A.N. M.905-917. Patents and Technological Strategies. Markets for Technology.. P.. pp. 630–640. Imitation and Growth with Step-by-Step Innovation”.. Lissoni. S. (2000). B. M. 33. E. 1-22.277-295. Audretsch. Technology Analysis & Strategic Management.. “The structure of licensing contracts” Journal of Industrial Economics. A. and firms' incentives for technology licensing”. Working Paper n. (1999). Alcacer. pp 523-532. complementary assets. J..” Journal of Development Economics 50. American Economic Review. Steinmueller.. 68. Research Policy. M. Anton. 3. (2003). “Specialized supply firms. pp.. Andersen. A. A. 9431. “If 'intellectual property rights' is the answer. G. Baldwin. Management Science. “Patent protection. Review of Economic Studies. Van de Paal.157-173. property rights and firm boundaries” Industrial and Corporate Change. A. Cambridge (MA).M. J. M. “Licensing the Market for Technology. Ceccagnoli.”. n.REFERENCES Aghion. “The use of patent databases by European small and medium-sized enterprises”. pp. EIMS Publication. P.. L. pp. “Localization of Knowledge and the Mobility of Engineers in Regional Networks”. M. W. (2004). 13. Arora. A.. (2003).. Arora. “How do I Know what you Know? Patent Examiners and the Generation of Patent Citations” New York University. Cohen. Research Policy. 13. Harris. D. Balconi. A. (2005). pp.). “R&D spillovers and the geography of innovation and production”. (2001). A.. (2004). Arundel. Statistics Canada. A. Howitt. pp. 10.. D.. 122. Sabourin.B. 86. pp. (2004).” Journal of Economic Behavior and Organization 52(2). Arora.. (2003). (1998). pp. pp. 17–57. B. pp. Archibugi. (1995).. Economics of Innovation & New Technology.. “Contracting for Tacit Knowledge: The Provision of Technical Services in Technology Licensing Contracts. pp. F. J. Almeida.. (2000).. 88 . Innovation Surveys and Patents as Technology Indicators: The State of the Art. Soete.. Gambardella. Hanel. pp.233-256. (2004). Gambardella. Arundel. J. A. (1996). (1994). M. NBER Working Papers. Gittleman. “Little patents and big secrets: Managing intellectual property”. Ceccagnoli. (2001). D. A.P. Breschi. and Fosfuri. C. P.

S. 419-23.“Who Deters Entry? Evidence on the Use of Strategic Entry Deterrents”. “Intellectual Property Rights. Cohen. pp.“Patents.. Hunt.M. B. Carnegie Mellon University).M. 3366. Cohen.. The Influence of Public Research on Industrial R&D”. 31. pp. Schankerman. pp. “Patents and Appropriation: Concerns and Evidence”. 89 . (2001).. R&D. 31.“Knowledge Spillovers and Local Innovation Systems: A Critical Survey”.. 30. R. 52. (2004). 2002. Goto.J. (2004). N. 57. patents and the incentives to innovate in Japan and the United States”. B. No. (2001). Branstetter. (2000) “Protecting their intellectual assets: Appropriability conditions and why U. 7522 (revised. pp.Nagata. 3305. n. Cornelli. Cesaroni F.. W.G. Dahlman. 2.. Policy Research Working Paper. Cohen. (1988). pp. (2002). RAND Journal of Economics. L. Review of Economics and Statistics. S. Research Policy. Research Policy. Economie Appliquee.. pp. and Labory S. Strategic Technology Agreements and Market Structure. 78. as mimeo.“Links and Impacts.. Chen.R. G. Microeconometric evidence from the U.. 1169-1184. Journal of Technology Transfer. Breschi. Duysters. American Economic Review.197-213. and Verspagen. Van Reenen.M.. 3.. (2004). 30. I.. Cohen. and Walsh. S. C. (2005).S. manufacturing firms patent (or not)..Bekkers. real options and firm performance”. W. Griliches. R. W. 975-1005.. 12. Blind. Lissoni. D. Bunch. The World Bank.“Do Stronger Intellectual Property Rights Increase International Technology Transfer? Empirical Evidence from U. R. 2. (2002). pp.(1999). Breschi. Cambridge University Press.S. International Handbook of Industrial Policy. H. (2004). F. Foley. N.. Economic Journal. The case of GSM”.. “R&D Cooperation and Spillovers: some empirical evidence from Belgium.“Mobility and Social Networks. J.“R&D spillovers. F. Nelson. pp. 265-280. (2002). 112. Cambridge. Nelson. D. Localised Knowledge Spillovers Revisited”. C.. Breschi.“Patent Renewals and R&D Incentives”. (2001). Cincera. L.“Industry Effects and Appropriability Measures in the Stock Market's Valuation of R&D and Patents”. Lissoni. J. D. F.” NBER Working Paper n. Cockburn. M. An Empirical Look at Software Patents. Working Paper No. Journal of International Economics.R. Cassiman. R. W. Bosworth. Nelson. 03-17/R. pp.. and Japan”. 33. forthcoming.” The American Economic Review 92 (4). R.R.Walsh. R.. pp.. 1141–1161 Bessen.M. foirthcoming. R.“Knowledge and Development. R. evidence from patent data.C. 1583-1598. Firm-Level Panel Data”. F.M. Research Policy. A..P. 5379. Smiley. M. and Veugelers.. “Interrelation between patenting and standardisation strategies: empirical evidence and policy mplications”.71... pp. MA.G. Annales d'Economie et de Statistique. Federal Reserve Bank of Philadelphia. 74. The World Bank.P. in Bianchi P. 48. Z. 509-521. (1997). pp.. No. (2005) “IPRs and market dynamics”. Industrial and Corporate Change. pp. Bloom. (2004). Policy Research Working Paper. No. K. Fisman. Vol. Management Science. and Technological Spillovers at the Firm Level: Some Evidence from Econometric Count Models for Panel Data”. Spatial patterns of innovation. n. Giuri P.“Returns to Intellectual Property in the Pharmaceuticals Sector”. Journal of Applied Econometrics. Branstetter. 53. A Cross-Section Approach”. Thumm.H.1-23. 10.S. (2002). “Patents. A. (1992). S. J. 1349-1367. (2002). pp 97-116. J.“Are knowledge spillovers international or intranational in scope?. (1999). Mahdian. 69-93.

No.'”The geographic sources of foreign subsidiaries 'innovation'”. P. pp.. (2004).“International Technology Diffusion. (2005). RAND Journal of Economics. Cockburn. forthcoming.“Why do some universities generate more start-ups than others?”. (1999). (1990). European Planning Studies 5. S. “Empirical methods to identify and measure research spillovers: a survey”. strategy and policy”. W. No.. No. 514-526.M. 537-570. H. . 3. Lessons from recent economic research. Journal of Finance... Lerner. The Value of European Patents. Boston University. Gans. S. 209-227. pp. Patenting. pp. 28. (2001). Vol. “Technological Activity and Employment in a Panel of UK Firms”. RAND Journal of Economics. Davis. N. 571-586.(2002). Furman. pp. Research Policy.. Gruben. Evidence from a survey of European inventors. (1997).J. 48. “SMEs in Europe. D. Maskus. 23. Journal of Development Economics 48. 249-265. 20-33.. Di Gregorio. (ed) (1997). forthcoming. Fleischer. (1996). Yale University. Denicolo. (1999). “Preemptive Patenting and the Persistence of Monopoly”. Competitiveness. Hsu. Innovation in Japan. Lenox. T. No. S. 16. J. “Patent Statistics as Economic Indicators: A Survey”.T. J. Oxford. Gilbert.C.“Entrepreneurial spin-off enterprises in Goteborg Sweden”. Scottish Journal of Political Economy. Gallini. Vol.T. Shane. 52-63.R. W. Forst. pp. Theory and Measurement”. No. Scotchmer. J. 3.. IP and development. Kortum.“Intellectual property rights.Dahlstrand. Griliches. 40. 44. pp. G. 399-415. 95-119. Brussels. V. Garcia Fontes. 4.. 22. 52.E. 131-154..K. Z. 72. Longland. “The role of intellectual property rights in economic growth”. (2005). C.L. L. (2003).S. pp. pp.. (1996). M. “The economics of patents: Lessons from recent US patent reform”. J. mimeo. Bosworth. innovation and the knowledge-driven society. Journal of Economic Perspectives. RAND Journal of Economics. “When does start-up innovation spur the gale of creative destruction”. Goto. pp. 13.“Innovation. pp. European Commission (2005). 1.. D. (2003). Newberry. M. (2005). Gompers. 33. R. pp. 1. Eurostat (2002).. Greenhalgh. 19962001”. pp. 659-673.World Bank. (2004). Journal of Industrial Economics.L. (2002). European Commission. pp. 260-82. “On the Division of Profit in Sequential Innovation”. Essays on the evaluation of European patents. mimeo. pp. International Economic Review. Deng. 323-350 Green. Dushnitsky.(ed). 101-123. Odagiri. (1992). (1995). (2005).“Patent Races and Optimal Patent Breadth and Length”.. Clarendon Press. Strategic Management Journal.. Fink. Eaton. Gould. M. (2001). D. Kyle. D. (1992). K. PhD Thesis. 26. D. American Economic Review. J. I. “When Do Incumbents Learn from Entrepreneurial Ventures? Corporate Venture Capital and Investing Firm Innovation Rates”. A. S. “Patent Policy and Costly Imitation”. “Public & Private Spillovers: Location and the Productivity of Pharmaceutical Research”. 1661-1707. N. 32. “Entrepreneurial Spawning: Public Corporations and the Formation of New Ventures”. pp. Scharfstein. Stern. D. A. Gallini. Economie Appliquee. Forthcoming. Journal of Economic Literature. Research Policy. M. R. Henderson.. 90 . Economics of Innovation & New Technology.. UPF. and Performance”. Y. S.

S. pp. pp. Trajtenberg. 32 (1). (1999). “R&D. Imitation. NBER Working Paper No. pp. insights and methodological tools”. “Can Patents Deter Innovation? The Anticommons in Biomedical Research”.. (1998). F. “Patentability. Pisano. 280. 76. Haneda. A. Hicks. 511-515. (2005).. “Appropriation of Returns from Technological Assets and the Values of Patents and R&D in Japanese High-Tech Firms”. 51.. Family Size. R. M. pp. (1986). B. 1.H. pp.. 5. K.. D. 577-598.. “The Pharmaceutical Industry and the Revolution in Molecular Biology: Interactions among Scientific.R. forthcoming. International Journal of Industrial Organization. “Intellectual Property Strategy in the Global Cosmetics Industry”. Guellec. Nelson (eds). pp. Jaffe. B. Scherer. and Intellectual Property Rights”. (1993). pp. RAND Journal of Economics. S.. E. Hernan.H. Jaffe. B. pp. mimeo.A. G. D. Hall. Van Pottelsberghe. and Market Value”. “An empirical evaluation of the determinants of research joint venture formation”. 108. D. “Exploring the Patent Explosion”. No.H. Opposition and the Value of Patent Rights”. Countries and Technology Areas Patterns”. an empirical study of patenting in the US semiconductor industry. D. B. pp. (2004). 81.. Kluwer Academic Publishers. M. Quarterly Journal of Economics. 303-321..H. (2002). 984-1001. Breitzman. New York: Cambridge University Press. (1993). Hall. 1979-1995”. Sources of Industrial Leadership: Studies of Seven Industries. A. (2004). 61. Harhoff. (2003b).... D. Scherer.. No. 443480. and Policy”. RAND Journal of Economics.. M. 3. (1999). Harhoff.. Econometrica. A. “The patent paradox revisited.H. 698-701. H. pp. Institutional and Organizational Change”. 6.. Economics of Innovation and New Technology. 91 . M. Hall. (2001). G. Economics. A. patents. and Innovation". F. “Business Method Patents. D. (2002). Granstrand (ed.P. Hall. K. D. Helpman. Science. and Hamilton. A. American Economic Review. “Citations. Heller. “The NBER patent citations data file: lessons. 32. and Vopel. Henderson. 11(2). M. L. Eisenberg. University of California at Berkeley. 22.. Olivastro. pp. 52. Jaffe.B. Hunt. “Technological Opportunity and Spillovers of R & D: Evidence from Firms' Patents. Law and Intellectual Property: Seeking strategies for research and teaching in a developing field. No. M. Vopel. D..C. Research Policy. (2003). 101-128. “Citation Frequency and the Value of Patented Inventions”. Narin. B.. 681-704. and market value revisited: Is there a second (technological opportunity) factor? ” Economics of Innovation and New Technology. Reitzig. Jaffe. T. NBER Working Papers n. Orsenigo. F. R.” Research Policy. Mowery and R. 1343-1363.. Hall. V. (2001) “The changing composition of innovative activity in the US—a portrait based on patent analysis. Ziedonis. (2004).. Review of Economics and Statistics. Harhoff. R. M. Siotis..B. 401-425. Profits. M. and Pakes.“Exploring the Tail of Patented Invention Value" Distribution”... 7. pp. Boston Harhoff. Hall.H. F. Odagiri. “Determinants of Opposition against EPO Patent Grants--The Case of Biotechnology and Pharmaceuticals”. Economics of Innovation and New Technology. Trajtenberg.L. D. (1991). Scherer. P. Hall B.. Industry Structure. (2003). (2004). K. Vopel. „The Value of Patents and Patenting Strategies.. Henderson. “Innovation. WZB Conference Paper. K. Z. B... (1998).H. 8498. Marin.. 101-128. 30 (4). M. “Market value and patent citations”.). in O. 75-89. “Geographic Localization of Knowledge Spillovers as Evidenced by Patent Citations”. R. Harhoff. pp. 1247-1280. R.Griliches. pp.183-202. Innovation. B. 10605. R. Trajtenberg. Journal of Industrial Economics. (2001). Journal of Industrial Economics.

Carnegie Mellon University. Fogarty. D. 34 (1). Nelson. J.B. Putnam. J.The Uses of Patent Renewal and Application Data”.Jaffe. Popp. 783-820.J. No.. 55. 91. Pakes.. pp. pp. European and US Patents by German Companies. Lowe.”. 1. mimeo.. J. 7631. A. D.O. RAND Journal of Economics. J. M. Sokoloff. Jaffe. pp. und Zos. pp. Klevorick. 91. “The Importance of Patent Scope: An Empirical Analysis”. mimeo. Levin. K. RAND Journal of Economics. pp.R. “Patenting in the Shadow of Competitors”. (2000). J. “Patent protection in the shadow of infringement: Simulation estimations of patent value. RAND Journal of Economics. 39-44. 319-333. 65. Philipson.K. (1998). G. Vol. Journal of Law & Economics..O.” Research Policy. Jaffe. 183-205. Patents and R&D: An Econometric Investigation using Applications for German. (1995). A. Lerner. Licht. 29. Marschke. (2001). pp. 25. eds.. Annales d'Economie et de Statistique. pp. and the Firm's Patenting Decision”. (1994). B. 38. Wright. "The Meaning of Patent Citations: Report on the NBER/Case-Western Reserve Survey of Patentees" NBER Working Paper No. Vol. (2002).. 46. pp. pp. 2. “Labor Mobility of Scientists. Lerner. pp. (2004). 240-59. “Does Intellectual Property Protection Spur Technological Change?”.. (1998). 223-46. J. P..A. Brookings Papers on Economic Activity. Lerner. Banks. 2. and Winter. Lerner. pp.. “The dual effects of intellectual property regulations: Within. Technological Diffusion.S. 49/50.B..B.. F. Fogarty. (1998).N. No. (1990). “Evidence from Patents and Patent Citations on the Impact of NASA and other Federal Labs on Commercial Innovation”. R.O.K. 235-64.: National Academies Press. SPRU. Trajtenberg. Lanjouw.R.” in W. J. R. pp. Oxford Economic Papers. pp. Journal of Industrial Economics. (2000). No.. No. “How to Count Patents and Value Intellectual Property. Lockett.D. J. 531-558. 49-50... Journal of Industrial Economics. risk capital and the creation of university spin-out companies”. Annales d'Economie et de Statistique. “The U. 573-603. No. RAND Journal of Economics. 328-360. “The Enforcement of Intellectual Property Rights: A Survey of the Empirical Literature”. (1998).O. patent system in transition: Policy innovation and the innovation process. M. M. (2002). 113-130. “Tilting the Table? The Predatory Use of Preliminary Injunctions. 44. 463-96.. . (2003). 405432. (2003). Lanjouw.S. Merrill. R. No.” The Journal of Law and Economics. Lichtenberg. “Forced out of the closet. “Market Trade in Patents and the Rise of a Class of Specialized Inventors in the 19th-Century United States”. (2001). 2... R. Patents in the Knowledge-Based Economy. Vohora. (1987) `Appropriating the returns from industrial R&D'. Klemperer. M. “Proofs and Prototypes for Sale: The Licensing of University Inventions”. 92 . A. and Schankerman. Thursby.O. (2001). A. Lanjouw.. A.G. 643-672. pp.and between-patent competition in the US pharmaceuticals industry”.. S. Lamoreaux. “Entrepreneurship and Information Asymmetry: Theory and Evidence from the University of California”. R. P. 671-712. (2003) “Enforcement of patent rights in the United States. M. J. 21. Journal of Law and Economics. Cohen and S. Kanwar.C. J. Johnson. Washington. S. N. Kim.C. J. (2005). 45. N. Review of Economic Studies. A. 46. American Economic Review.. Lanjouw. the impact of the American Inventors Protection Act on the timing of patent disclosure”. pp.A. American Economic Review. pp. Evenson. Jensen. J. M. Lanjouw. K. “How Broad Should the Scope of Patent Protection Be?”.S. “Resources. (1996). T.R. 96-112.L. forthcoming. capabilities.

457-476. E. “R&D and Development”. Murray. (1986) “Patents and innovation: An empirical study. “Knowledge Spillovers in Europe: A Patent Citations Analysis”. pp. Discussion Paper No.. D. 104. (2004). Mansfield. (2003). C. Japan and the World Economy. Schankerman M. Merrill. M. S. “The benefits and costs of strong patent protection. E. “Start-ups. 643-659. A. (ed. (2004). (1996).32. Ziedonis. D.. Maurseth. Pakes A. M. Universities: An Assessment of the Effects of the Bayh-Dole Act of 1980”. Scandinavian Journal of Economics. towards an updated picture of the use of patents across industries”. (2003). (1985). M. 32. Wagner. National Research Council (NRC) (2004). (2001). (2004).C. (eds) (2004). J.. Journal of Political Economy. Chicago. Nagaoka. R. Judd. and Licensing with Japanese Company Data”. “The role of academic inventors in entrepreneurial firms. World Bank Policy Research Working Paper n. Levin. Maloney. D. pp. The National Academies Press. A. pp. Verspagen. Nakamura. Mansfield. R. 33. pp. in Griliches.C. 11. Ziedonis. 173-l8l. and the Private Rate of Return to Research Resources”. RAND Journal of Economics. A contribution to the current debate”. PhD Thesis. Growth and Welfare: A Theoretical Treatment of Technological Change. 27. (1999). 91. “On Patents.. (2003). “What Determines Patent Value? Insights from the Semiconductor Industry”. Joint R&D. 93. 273–284. Research Policy. “Patents and Innovation. pp. mimeo.. The value of international patent rights.R. Mansfield. Mazzoleni. 54. (2003).T. Pakes. Press O'Donoghue. R. Reitzig. OECD (2004). Brookings Papers on Economic Activity. K.. B. R. Economics of Innovation & New Technology. E. “A patentability requirement for sequential innovation”. University of Chicago Press. 4. A.Lowe. Washington.C.S. A Patent System for the 21st Century. “Determinants of R&D Boundaries of the Firm: An Empirical Study of Commissioned R&D. (1981). T. Simpson. B.” Management Science. Nordhaus (1969). No. E. A. 99-119. 32. F. “Patents as Options: Some Estimates of the Value of Holding European Patent Stocks”. K. 13. pp. Imitation Costs and Patents: An Empirical Study. (1989). 531-545. Odagiri..N. Lederman. pp. Mowery. Research Policy. NISTEP. Research Policy. 907-918. Patents. Research Policy. Sampat... pp. Pakes. R & D.. (1986). A. M. The Economic Journal. Myers. 2. 654-679. Putnam.B. “Patent Renewal Data”. S.: National Academies Press. Invention. Trends and Policy Challenges”. Maskus. Reitzig.A. H. mimeo. No. A Patent System for the 21st Century. W. Research Gestation Lags. A. “The Rate of Obsolescence of Patents. M. Washington. (1984). 557-574... “Patents and the other IPRs in use”. Yale University.. (1998). pp.I.A. pp. 331-401. OECD Publications. and Productivity.R. Pakes.. Established Firms. 93 . K. R. (2002). 3024. “The private values of 'thickets' and 'fences'. Econometrica. “Impacts of the Japanese Patent System on Productivity Growth”. 30. Z..B. and the Commercialization of University Inventions”. and the Stock Market Rate of Return”.A. sharing the laboratory life”. Hitotsubashi University.. 29.. McDaniel. D. (1998). Carnegie Mellon University. P. Nelson. 13-26. Schwartz. Nelson.F. Microeconomics. M. “The Growth of Patenting and Licensing by U.) R&D. pp.

21. pp. E. pp.M. M. (2000). Harvard Business Review 78(1). (2002). Economic Journal. 114. S. and the Value of Innovation. Kline.. 2nd edn. Management Science. privately published. (1999). 351-364. J. citations & innovations: A window on the knowledge economy. Singh. (2003). “Do Stronger Patents Induce More Innovation? Evidence from the 1988 Japanese Patent Law Reforms”. 1017-1047.. pp.. L. M. Speakman. pp.. Cambridge. 48. “Discovering New Value in Intellectual Property”. Vol. Shane S. Vol. pp. 384. pp. K. D. Scherer M. and Harhoff D. 312-332. 29-41. pp. “Estimates of the Value of Patent Rights in European Countries During the Post-1950 Period”. 33. patent pools. 131-146. Patents and the corporation. (1991). “The return of the Prince of Denmark: A survey on recent developments in the economics of innovation”. 94 . Fleming. J. Schankerman. “Navigating the patent thicket: Cross licenses. (1990).. C. Scotchmer. No. “Novelty and Disclosure in Patent Law”. No. No. und Teece. 96. Economic Journal. et al. Sherry. 179–191. “Technological Opportunities and New Firm Creation”. Vol. (1986). (2004). M. Shapiro. “Standing on the shoulders of giants: cumulative research and the patent law. Management Science. (2001). Schankerman. 77-100. Research Policy 33. 29. S. 205-220. Economic Journal. (2004). R. “Selling University Technology: Patterns from MIT”.M. n. 5. No. eds. Khurana (2003). Schankerman.. 48. 113-491. 93-111. 1. Jaffe. O. Research Policy. F. National Bureau of Economic Research and MIT Press. Rudholm. 19. 1615– 1634. “Collaborative Networks as Determinants of Knowledge Diffusion Patterns”. pp. S. Sorenson. Edward Elgar. 11731190.” Journal of Economic Perspectives. Management Science. Berkeley Technology Law Journal. pp. (2001). 54-66. Stuart. 154-170.G. “Bringing individuals back in: the effects of career experience on new firm founding”. (2002). (2000). Branstetter. Shane. Royalties. S. Shane. (1998). Green. pp. 1. pp. RAND Journal of Economics. C.. RAND Journal of Economics. (2004). A. pp. T. E. N. Pakes. Rogers. 77-107. RAND Journal of Economics. Sena. 32. Academic entrepreneurship. pp. Shane. pp. 1. pp. “How Valuable is Patent Protection? Estimates by Technology Field”. Shapiro. Review of Industrial Organization. (1959). “Entry and the Number of Firms in the Swedish Pharmaceuticals Market”. S. 29. “Patent System Reform: Economic Analysis and Critique”.. Shane. 122137.. (2004). 19. L. Shane. 559-566 Scotchmer. 3. (2000). Scherer.. (2001). Innovation Policy and the Economy. D. “Technology Regimes and New Firm Formation”. “Policy Implications for a World with Skew-Distributed Returns to Innovation”. (2005). S. “Organizational Endowments and the Performance of University Start-ups”. pp.. M.. Sakakibara. 47. Stern. Research Policy. 12. 15. A. 2.”. 47.Rivette.. “Spin-offs from research centers at a research university”. Lerner and S. pp. Boston. “Science and the diffusion of knowledge”. “Patents. 519-543. Steffensen. 113. Management Science. and standard setting. Journal of Business Venturing. Management Science. F653-F654. M.. Industrial and Corporate Change. (2004). (2001). Evolving Patent Rights. forthcoming. J. V. K. S.

R. Verspagen. C.. M. eds. M. Ziedonis. semiconductor industry. 2. (1990). Z. B. 4. B. eds. Research Policy. No. pp. Verspagen.H. Taylor.M. (1994).. Merrill. Waterson. pp. pp. R. Washington. “Is the commercialisation of scientific research affecting the production of public knowledge? Global trends in the output of corporate research articles”..C. 215-235. W. Technological Forecasting and Social Change. Journal of Economic Geography. vol.C. (1999). pp. Cohen and S. (2003). B. No.: National Academies Press. Cambridge: Cambridge University Press.W. and Growth”. Cohen. Trajtenberg.Taylor. 860869. 80. A. Trade. Citations & Innovations: A Window on the Knowledge Economy” (book review). “Trips. No. “Dont Fence Me In: Fragmented Markets for Technology and the Patent Acquisition Strategies of Firms”. 1709-1711. 172-187.. W. “Effects of Research Tool Patents and Licensing on Biomedical Innovation. 50 (6). Silberston. 361-381. 4. Management Science. pp. (2004). RAND Journal of Economics. D. Cohen and S. 35. Merrill. Ziedonis. (2004). 33. Tijssen. International Economic Review. Patents in the Knowledge-Based Economy. Munich School of Management. 709733.J. M. 23-42. (2004). Arora. (1973) The economic impact of the patent system: A study of the British experience. “Patents. 33. R. “Technology Spillovers between Sectors and over Time”. (2004). Wagner. 1.. “The Economics of Product Patents”. Washington. Patents in the Knowledge-Based Economy. “Business Method Patents in Europe and their Strategic Use – Evidence from Franking Device Manufacturers”.S.T. (2004).: National Academies Press. 60. (1990). pp.P. S. (2003) “Patent litigation in the U. American Economic Review.. J. Vol. 95 . mimeo Walsh.. Verspagen. “A Penny for Your Quotes: Patent Citations and the Value of Innovations”. pp. 21. 804-820. pp. Research Policy. Schoenmakers.. “The spatial dimension of patenting by multinational firms in Europe”.” in W.S.” in W.A. D. De Loo I.

Information (i.e. methodology. etc. technological class. results) and the relevance for the research themes of this project (i. country. For each reference we also include a field with the abstract of the contribution (i. the following list provides the details of the classification. etc. Each reference is also classified by research theme. DATASET OF THE LITERATURE Part of this project was devoted to the construction of a database that contains the publications included in survey of the literature. the economic value and the impact of the patent.e. research question. fields in the Access table) on each publication of the reference list ID: Publication identification number Title of the publication Source: Journal/Publisher Volume / Pages Year of publication Author/s: Author 1 Author 2 Author 3 Author 4 Author 5 Type of publication: Journal article Book Book chapter Working paper Report Abstract File availability (availability of the file in pdf format to be included in a CD-Rom): Yes No 96 .e. It is a database in Access format with a main table that lists all the references made to the literature. technological area. This is a very flexible database that will enable a search for publications by theme. To better understand the organisation of the literature dataset. theoretical and empirical content. country. and the implications for the economy and society).ANNEX I.

Semiconductors Communications Electrical engineering Mechanical engineering 97 .New firm creation and employment B2 . we also classified the literature according to the main sub-theme classifies as a key word. Key word – Sub-theme Within each broad theme. Skewness of the Patent Value Distribution. For example.Economic Value A2 . Criteria used for classification of the publications Themes A1 .Economic use B1 . etc.Knowledge spillovers Second Theme Some publications deal with more than one of the four themes of the project.Citation ISI: Number of citations received from the publications collected from the ISI Web of Science database (available since 1990). Theoretical/Empirical Codes: (B) Both theory/empirical (E) Empirical (RB) Review Both theory/empirical (RE) Review Empirical Literature (RT) Review Theoretical Literature (T) Theoretical Technological Areas Multiple Biotechnology Pharmaceutical Bio-pharma Chemicals Computing Electronics Electronic . Litigations. In these cases we specify also the second theme of the publication. within the theme A1 we included sub-themes like Patent Citations.

They are then analysed on a sectoral base.- Instruments Production engineering Other Countries Germany France UK Italy Spain Netherlands Denmark Sweden Hungary US Japan Multiple Multiple (EU) Type of applicants’ organization All Multiple All firms Large firm SMEs Public research institutions (incl. the records are classified in aggregate terms. Universities) No profit research institutions Individual When data had multiple items to be single out. only if this is relevant to the analysis. 98 .

Economic Value 18 8 25 1 3 1 7 1 2 22 88 A2 . many countries or technologies).New firm creation and employment 2 12 1 2 17 B2 Knowledge spillovers 6 5 15 2 1 5 34 Other 1 1 Total 38 20 74 1 8 1 1 1 7 2 2 4 31 190 99 . theoretical/empirical). The rows or columns classified as “multiple” in the Tables below collects the contributions performed on aggregate data or on multiple entries (i. country. It was prepared by using the dataset of the literature that includes 164 key publications on the four themes. technological class. COVERAGE OF THE LITERATURE SURVEY This Annex shows the coverage of the literature we surveyed by theme.e. 26 of these publications deal with two themes.ANNEX II.Economic use 12 6 22 3 1 1 1 4 50 B1 .e. For this reason the total number of observations of the Tables of the coverage of the literature by theme is 190. Themes & Countries Country Multiple Multiple (EU) USA Canada Japan India France Belgium Germany Italy Sweden United Kingdom NA (theory or NA) Total A1 . Each cell shows the frequencies of publications according to the selected criteria. type of inventors’ employer and type of contribution to the literature (i.

Economic Value use 7 24 29 4 1 23 88 7 7 23 2 1 3 2 5 50 B1 .Themes & Technologies Industry Multiple Bio-pharma Biotechnology Pharmaceutical Chemicals Computing Communications Electronic .Semiconductors Instruments Mechanical engineering Other NA (theory or NA) Total A1 .Economic A2 .New firm creation and employment 14 2 1 17 B2 Knowledge spillovers 27 1 1 5 34 Other 1 1 Total 131 3 4 5 1 1 1 7 1 3 2 31 190 Themes & Inventors’ organizations Inventor type Multiple All All firms Large firm SMEs Public research institutions (incl. Universities) Individual NA (theory or NA) Total A1 .Economic Value use 51 1 1 4 1 1 3 1 2 1 22 88 38 2 1 1 3 1 4 50 B1 .New firm creation and employment 1 2 1 2 9 2 17 B2 Knowledge spillovers 4 13 5 3 1 2 1 5 34 Other 1 1 Total 19 44 59 10 6 14 5 33 190 100 .Economic A2 .

Semiconductors Instruments Mechanical engineering Other NA (theory or NA) Total 23 1 2 1 1 1 1 30 17 1 1 19 49 1 2 2 1 5 1 1 62 1 1 8 1 9 1 1 1 1 1 1 6 1 7 6 1 7 1 1 2 3 1 4 117 3 4 5 1 1 1 7 1 2 2 25 169 101 Total India USA Italy .Theme & Type of contribution Theory/Empirical NA (theory or NA) Both theory/empirical Empirical Review Both theory/empirical Review Empirical Literature Review Theoretical Literature Theoretical Total A1 Economic Value 7 49 8 9 1 14 88 A2 Economic use 1 9 29 4 2 1 4 50 B1 .New firm creation and employment 1 15 1 17 B2 Knowledge spillovers 24 3 6 1 34 Other 1 1 Total 2 17 117 15 18 2 19 190 Countries & Technologies Germany United Kingdom Sweden Belgium Canada Multiple Multiple (EU) France Japan Other 25 25 Countries & Technologies Multiple Bio-pharma Biotechnology Pharmaceutical Chemicals Computing Communications Electronic .

Countries & Inventors’ organizations
Germany United Kingdom Sweden Belgium Canada Multiple Multiple (EU) France Japan Other
25 25

Countries & Inventors’ organizations

Multiple All All firms Large firm SMEs Public research institutions (incl. Universities) Individual NA (theory or NA) Total

7 11 10 2 1 31

2 8 3 2 2 2 19

3 11 28 3 3 11 3 62

1 1

3 1 4 8

1 1

1 1

1 1

4 2 1 7

1 1 2

1 1 2

2 1 1 4

15 37 53 9 6 13 4 27 164

Inventors’ organizations & Technologies
Communications Electronic Semiconductors Pharmaceutical

Biotechnology

NA (theory or NA)
25 25

Instruments

Mechanical engineering

Bio-pharma

Computing

Chemicals

Multiple

Other

Inventor type

Multiple All All firms Large firm SMEs Public research institutions (incl. Universities) Individual NA (theory or NA) Total

12 34 39 7 5 12 2 2 113

2 1 3

1 1 1 1 4

4 1 5

1 1

1 1

1 1

4 1 1 6

1 1

2 2

1 1 2

15 37 53 9 6 13 4 27 164

102

Total

Total

India

USA

Italy

ANNEX III. ADDITIONAL DESCRIPTIVE STATISTICS FROM THE PATVAL-EU DATASET
The Tables in this Annex complement the Report by providing additional descriptive statistics on the PatVal-EU data. The first three Tables in this Annex show some basic data on the composition of the PatVal-EU dataset. Specifically, they list the technological classes in which the patents in the sample are classified both at the micro and macro level and, by conditioning on the (micro and macro) technological classes, they provide the share of patents applied by different employers’ organisations. The rest of Annex III collects additional Tables on the four themes presented in the Report, and in so doing it reproduces the structure of the Report as follows: AIII.1. The value and social costs of patents AIII.2. The economic use of patents: - The use of patents - The importance of different motives for patenting AIII.3. The creation of new businesses from the patented inventions AIII.4. Collaborations, spillovers and the sources of knowledge in the invention process: - Collaboration among inventors in the innovations process - Formal and informal collaborations in research - The role of geographical proximity for collaboration - The sources of knowledge.

103

Table A.1 List of Macro and Micro technological classes
Macro Technological Class
Electrical engineering Electrical engineering Electrical engineering Electrical engineering Electrical engineering Instruments Instruments Instruments Instruments Chemistry, Pharmaceuticals Chemistry, Pharmaceuticals Chemistry, Pharmaceuticals Chemistry, Pharmaceuticals Chemistry, Pharmaceuticals Chemistry, Pharmaceuticals Process engineering Process engineering Process engineering Process engineering Process engineering Process engineering Process engineering Mechanical engineering Mechanical engineering Mechanical engineering Mechanical engineering Mechanical engineering Mechanical engineering Mechanical engineering Mechanical engineering

Micro Technological Class
Electrical devices, electrical engineering, electrical energy Audio-visual technology Telecommunications Information technology Semiconductors Optics Analysis, measurement, control technology Medical technology Nuclear engineering Organic fine chemistry Macromolecular chemistry, polymers Pharmaceuticals, cosmetics Biotechnology Agriculture, food chemistry Chemical and petrol industry, basic materials chemistry Materials, metallurgy Chemical engineering Surface technology, coating Materials processing, textiles, paper Environmental technology Handling, printing Agricultural and food processing, machinery and apparatus Thermal processes and apparatus Machine tools Engines, pumps, turbines Mechanical Elements Transport Space technology weapons Consumer goods and equipment Civil engineering, building, mining

Table A. 2 Composition of the sample by macro technological class and type of inventors’ employer
Employer Type
Large companies Electrical Engineering Instruments Chemical & Pharmaceutical tech. Process Engineering Mechanical Engineering Total Medium sized companies Small companies Private Research Institutions Public Research Institutions Universities Other Governm. Institutions Others

Total

79.89% 60.42% 81.16% 64.37% 67.86% 70.58%

5.49% 7.85% 4.88% 12.29% 10.48% 8.81%

9.13% 16.65% 4.94% 17.18% 17.80% 13.75%

0.43% 3.25% 0.61% 0.69% 0.19% 0.76%

1.85% 3.77% 2.56% 2.24% 1.07% 2.05%

2.85% 7.02% 5.67% 2.42% 1.18% 3.22%

0.07% 0.10% 0.12% 0.23% 0.19% 0.16%

0.29% 0.94% 0.06% 0.59% 1.22% 0.67%

100.00% 100.00% 100.00% 100.00% 100.00% 100.00%

Number of observations = 8,809

104

85 4.18 4.36 62.809 105 .56 13.50 75.46 65.12 2. Basic Materials Chemistry Chemical Engineering Surface Technology & Coating Materials Processing.22 64.96 1.00 1.00 0.35 4.72 0.61 7.00 0.00 0.61 17.08 0.27 66.38 12.00 0.00 100.93 6.81 9.00 100.22 0.37 85.00 0.00 100.51 3.00 100.04 23.72 0.27 2.15 10.00 100.30 10.94 47.58 14.15 9.77 45.00 0.14 6.00 100.68 0.02 17.72 5.15 14.00 100.76 0.98 11.00 11.48 0.00 0.34 1.84 0.60 1.35 3.70 0.00 100.06 5.00 0.00 0.35 9. Textiles & Paper Thermal Processes & Apparatus Environmental Technology Machine Tools Engines.52 2.00 100.89 7.86 5.38 11.50 0.44 59.33 6.00 0.58 1.53 6.19 17.00 100.00 0.00 1.67 0.00 0.79 5.27 3.00 100.47 17.35 6.55 1.00 0.00 0.16 56.72 21.73 87.70 0.00 100.00 0.68 17.90 4.58 0.05 61.40 4.89 12.00 0. & Food Chemistry Chemical & Petrol Industry.65 7.00 0.00 100. Electrical Eng.00 100.34 0.48 1.00 Number of observations = 8.36 0.38 2.98 0.06 75.88 76.92 1. Building & Mining Total Large companies Medium sized companies Small companies Private Research Institutions Public Research Institutions Universities Other Governm.00 100.00 100.22 0.47 1.63 0.00 0.56 4.75 11.06 87.54 5.60 1.18 1.30 1.18 15.98 1. & Control Technology Medical Technology Organic Fine Chemistry Macromolecular Chemistry & Polymers Pharmaceuticals & Cosmetics Biotechnology Materials & Metallurgy Agriculture.00 0.07 40.00 0.91 0.51 3.00 100.26 77.00 0.75 0.00 0.00 0.50 14.23 1.31 0.90 73.31 0.50 0.00 1.00 0.00 100.44 70.15 0. Machinery & Apparatus Transport Nuclear Engineering Space Technology Weapons Consumer Goods & Equipment Civil Eng.90 87.10 7.00 0.60 0.89 0.41 77.65 3.27 4.00 100.00 100.12 67. Pumps & Turbines Mechanical Elements Handling & Printing Agricultural & Food Processing.88 22.00 0.00 1.16 66.00 0.91 3.21 0.80 4.27 0.58 7.64 3.15 12.41 0.27 0.76 2.00 2.16 0.22 0.23 6. Measurement.40 24.00 0.65 4.00 100.00 100.00 0.06 2.05 1.77 72.69 14.07 28.64 1.00 0.35 3.00 0.47 5.16 32.00 0.66 8.21 74.08 13.00 100.05 0.53 10.00 100.00 100.47 0.00 0.32 0.51 0.00 0.91 11.00 0.36 2.05 0.00 100.67 82.10 1.67 0.41 7.12 2.48 3.89 3.00 0. Institutions Others Total 78.00 1.34 1.94 4. 3 Composition of the sample by micro technological class and type of inventors’ employer Employer Type ISI Technological Classes Electrical Devices.00 0.49 7.03 3.67 100.78 2.33 0.00 100.49 62.14 30.00 0.Table A. & Electrical Energy Audio-visual Technology Telecommunications Information Technology Semiconductors Optics Analysis..54 0.95 7.94 8.67 0.64 0.25 8.90 1.00 0.19 0.28 66.00 100.35 4.05 0.00 2.24 47.58 1.00 82.71 9.15 0.79 1.53 4.98 0.00 100.94 16.45 7.75 12.01 3.00 0.53 12.87 2.49 1.63 0.71 1.50 10.07 0.50 13.61 0.44 4.00 100.58 0.

American applicants. meaning that the country of the applicant/s is one of the 6 European countries involved in the survey. Table A. It lists only the regions for which the average patent value is lower than the EU6 average. Table A.8 indicates the number of man-months by the country of the applicant.5 looks at the average economic value of European patents conditioned on the country of the applicant.6 shows the average value of patents invented by the NUTS2 regions in which they were invented.A III. Specifically.e. (2) other European countries (“other EU”). Table A. applicants located in countries other than European and American countries. and it is the same as the country of the inventor that we interviewed. If there are multiple applicants located in different countries we classified these patents as applied by foreign countries: (5) Co-assigned Foreign.7 reports the number of man-months needed to develop the innovations by number of countries in which the applicants are located. (4) non EU/US. i.4 shows the average economic value of European patents invented by applicants located in only one country as compared to applicants located in more than one country.e. 106 .1 The value and social costs of patents This section collects five Tables on the economic value and social costs of patents. meaning that the country of the applicant is a European country but it differs from the country of the interviewed inventor. i. or when there are multiple applicants from the same country we differentiate among the following possibilities: (1) “domestic” European countries. When there is only one applicant. while Table A. In so doing it looks at the relationship between the probability of developing important innovations and the setting up of international links among institutions in developing or simply applying for a patent. Table A. (3) USA.

39% 20.20% 9.69% 1.109.88% 17.89% 279 Total 7.88% 17.413) 0.76% 100.The value of patents Table A.98% 16.83% 15.49) 8.00% 8.75% 0.94% 2.07% 0.80% 15.80 (35.54% 11.31% 20.45% 9.422.27% 100.56% 1.211.53% 20.84% 20.52% 3.751.76 (1.74 (1.44% 22.18% 0.66% 21.68% 11.77 (1.88% 17.00% 1.00% 6. The number of observations shown in this Table refers to the number of forward citations and opposed patents.69% 2.92 (1.50% 3.92 (1.60% 9013 Total Average patent value Average number of forward citations Share of opposed patents Number of observations Note: standard deviations in parenthesis. Distribution by applicants’ country Domestic applicant <30k 30k-100k 100k-300k 300k-1m 1m-3m 3m-10m 10m-30m 30m-100m 100m-300m >300m 8. The category “1 country” includes patents with only 1 applicant and patents with more than 1 applicant located in the same country.24% 0.97 (1.36% 4.91 (1.249.59% 23.00% 0.262 (30.079.05% 100.53% 100.86) 10.64% 21.43% 296 Total 7.92 (1.00% 9.60% 9013 Total Average patent value Average number of forward citations Share of opposed patents Number of observations Note: standard deviations in parenthesis.00% 0.39% 4.72) 7. Number of countries in which the applicants are located 1 country <30k 30k-100k 100k-300k 300k-1m 1m-3m 3m-10m 10m-30m 30m-100m 100m-300m >300m 7.43) 0.752.00% 0.64% 8717 More than 1 country 8.61% 17.80) 0.76% 100.486.269.40% 1.76% 0.00% 6.91% 9.55% 21.41 (11.00% 6.79 (30.86 (28.40% 1.00% 6.20% 100. For the average patent value the number of observation is 7. Table A.92) 7.96% 3.66% 335 Non-EU & non-USA applicant 5.76% 0.27) 0.173.77% 100.90% 11.70% 2.26% 39 Non-domestic co-applicant 8.46% 9.12% 21.194) 0.60% 434 USA applicant 4.251 (36. 107 .00% 9.16% 1.17) 0.35% 16.41% 1.94% 0.09 (36.00% 100.00% 6.12% 15.65% 15.77% 0.55% 3. For the average patent value the number of observation is 7.19% 11.61% 15.58% 5.85% 17.80) 8.79) 0.76) 7.81% 2.91 (1.39% 20.94% 2.56% 12.46% 21.02% 3.294) 0.81) 8.91) 0.58% 3.59% 5.81) 8.690.85% 0. 4 The value of European patents. The number of observations shown in this Table refers to the number of forward citations and opposed patents.00% 3.77% 100.359 (30.77% 21.93% 17. 5 The value of European patents.556.359.410.32% 21.59% 20.59% 3.45% 9.13% 17.67% 7926 European applicant 5.49% 25.53% 2.36 (30.76% 11.89% 0.65% 21.80% 15.80) 8.65% 23.

66 (2.85% 14.82) 0.49 (23607.71 (1502.03% 7.05) 2610.94 (26834.99 (20507. The number of observations shown in this Table refers to the number of forward citations and opposed patents.35) 0.94 (11630.97) 3855. For the average patent value the number of observation is 7.37) Average Forward Cits (St dev) 0.72 (1.89% 11.48) 0.20% 10.34 (3076.06 (0.17) 4487.00% 20.76) 1.75 (2.45 (0.21) 0.12) 1.83 (1.00) Hampshire and Isle of Wight Lazio Merseyside Herefordshire.82) 6286.20 (2.35 (30407.42% 9.10) 5630.90% 11.14) 6264.67) 2506.54) 3556.00 (9153.55) 2523.29) 3124.74 (15877.22) 1.18(0.81 (1.13% 13.57 (25370.59) 975.15% 11.26 (8850.63% 12.814) 2384.30% 3.90% 18.26% 11.31) 0.16) 3161.85 (1.89) 0.65 (1.05) 0.29 (10028.20) 1.911683 (1.99 (14345.98) 1.057) 3826.56 (13086.22 (21753.24) 1919.80 (1.52) 2780.97 (5813.10 (29739.63) 5153.Table A.17 (0.33% 3.80) 0.58 (2.57 (10956.27) % Opposed patents 8.94) 1.10) 3100.59 (11915.10) 0.23 (0.28 (1.21 (22511.31 (11463.66) 0.82% 4.90) 0.76% 10.39) 1.64 (9142.16 (1.07) 5481.63 (1.35) 4155.Pas-de-Calais Alsace Picardie Berlin Tübingen Münster Bourgogne Schwaben Detmold Stuttgart Mittelfranken Centre Languedoc-Roussillon Oberpfalz Unterfranken 6350.65% 7.53 (23487.85 (15748.42) 3341.82% 5.23) 1.48 (1.47% 7.51) 0.58) 0.51) 0.00) 0.50% 6. 9013 55 35 38 49 172 71 117 129 86 62 48 78 278 183 114 364 436 65 110 44 217 89 32 52 339 101 311 50 63 40 65 131 61 51 84 44 346 136 65 42 49 58 Total UK IT UK UK IT FR NL NL UK NL FR UK FR NL IT DE FR NL UK UK DE UK DE DE DE DE DE FR FR FR DE DE DE FR DE DE DE DE FR FR DE DE Note: standard deviations in parentheses.34 (1.40 (1.55) 2328.37) 0.26 (2855.90% 9.12 (0. 108 .22 (0.17) 0.60% 1.04% 3.72) 0.32 (0.56 (1.00 (19392.23) 5818.56 (28198.49% 3.67) 1.04) 834.85 (1.46% 5.22 (0.78) 4224. This Table includes the European regions in which the average value of patents is below the EU6 average and the number of observation in each region is ≥30.56) 3024.19 (10264.75) 0.49% 4.71 (9479. 6 The value of European patents across European regions (NUTS2) Country NUTS 2 Average value (st dev) 6358.14% 10.21) 2783.62% 1.19 (0.67 (2.35 (1.41 (22618.84 (19528.38) 0.71% 13.52) 3615. Worcestershire and Warks Emilia-Romagna Midi-pyrénées Noord-Holland Gelderland West midlands Overijssel Pays de la Loire Greater Manchester Rhône-alpes Zuid-Holland Veneto Darmstadt Île de France Utrecht East Anglia Essex Köln Gloucestershire.71% 18.50) 5153. Wiltshire and North Somerset Koblenz Hamburg Oberbayern Arnsberg Düsseldorf Nord .35 (28934.34) 0.37 (9916.50% 9.17) 0.28) 1.92) 4297.08) 5091.84 (1.26) 1.90 (2.92(18704.58) 0.58 (10236.53 (1964.88 (10261.55% 9.23 (0.23% 7.97) 1377.754.76 (2.45% 19.14) 1148.91) 0.22% 4.67 (1.90 (10487.36 (10445.68) 1.21 (13989.42% 10.17) 4735.20% 18.47% 4.00) 4919.93) 4354.27 (0.50) 0.84% 8.07) 2.04) 5407.43) 0.25 (0.07 (13009.97% Number of obs.49 (1.80% 12.79 (2.28) 2361.54) 1.54 (1.09 (2.44) 4249.04 (14316.67% 11.94 (2600.74) 1.64% 7.16 (1.41) 4395.32) 1017.01) 0.

00% Non-domestic co-applicant 6.83% 20.98% 16. Number of countries in which the applicants are located 1 country <1 1-3 4-6 7-12 13-24 25-48 49-72 >72 12.82% 22.86% 2. Table A.95% 7.e.61% 17. Blocking competitors.26% 16.The cost of patents Table A.50% 19.9 and A.71% 18.98% 3.00% More than 1 country 6.22% 15. 109 .00% Total 12.56% 8.43% 22. Specifically.33% 19.22% 100.13% 3.28% 18.71% 2.65% 18.00% USA applicant 13.21% 100.24% 1.12% 3.58% 2.13% 18.86% 8.17% 18.62% 20.51% 19.12% 18.71% 20.51% 18.32% 8.17% 100.86% 100.54% 19.70% 100.14% 9.85% 5. Cross-licensing.07% 15.62% 20.58% 2.12% 3.34% 8.28% 100.91% 1.10 in this section look at the economic use of patents (i.45% 22.2 The economic use of patents Tables A.28% 100.00% Non-EU & nonUSA applicant 11.56% 8.00% Total 12.35% 8.05% 20.9 looks at the use of patents when the applicants are located in only one country as compared to patents whose applicants are located in more than one country.19% 17.66% 2.95% 23. Licensing. They show their distribution conditioned on the country of the applicant.97% 2.10% 100.00% European applicant 13.00% Total AIII.56% 6.23% 1.21% 15.57% 22. 8 The man-months required by the patents’ invention process.10 describes the use of European patents by the country of the applicants.86% 5. Distribution by applicants’ country Domestic applicant <1 1-3 4-6 7-12 13-24 25-48 49-72 >72 12.27% 14.67% 19. Sleeping patents) as defined in Section 3.83% 21. Table A.3.22% 15.66% 100.72% 2. Internal use.00% Total Table A.57% 5.12% 3. 7 The man-months required by the patents’ invention process.00% 15.27% 19.86% 22.56% 2. Licensing & Use.

55) IT 3.46 (1.72% 49.26 (1.53% 2.06 (1.01 (1.00% 100.56% 3.44% Total 100.14 looks at European vs.14 show the importance of six different motives for patenting for the organizations in which the inventors were employed at the time of the invention.00% 100.32 (1.45 (1.14% 18.13% 5.90 (1. Distribution by applicants’ country Internal use Domestic applicant European applicant USA applicant Non-EU & non-USA applicant Non-domestic co-applicant Total 6.11-A.78% 2.79 (1.50% Blocking Competitors 17. non-European origins of the applicants.77) 1.56) 2.65) FR 3.50% 53.00% 100.44) 3.78 (1.45 (1.39 (1. By using a score from 1 to 5.97% 4.60) 3.60) Total 3.33% 2.02% Licensing & Use 4.56) 2. and links this information to the importance of the different motives for patenting.76 (1.58 (1.61 (1. Compared to the “actual use of patents” made by the organisations.86% 17.64) 3. 10 The economic use of European patents. prevention from imitation.02% 3.47 (1.56) Note: standard deviations in parenthesis.56% 21.43% Total 100.07% 4.22) 4.17 (1.75) 1. Table A.89 (1. this part of the survey looked at the motives that led the inventors and their organisations to ask for patent protection. 9 The economic use of European patents.86% 3.63 (1.23 (1. and reputation.34) ES 4.54) 1. Table A.00 (1.85 (1.78 (1. Table A.68 (1.63) 2.93 (1.72) 1.32% 20.61 (1.52 (1. Number of countries in which the applicants are located.The use of patents Table A.80% 42.68% 19.65 (1.09 (1.97% Crosslicensing 2.31) 3.83% 22.28 (1.39) 2.66) 1.00% Number of observations = 7710 Table A.51) 1.75) NL 3.33% 4.12 shows their importance across the micro technological classes used in this Report.40) 2.97% Blocking Competitors 18.38% Licensing 4.20 (1.09 (1.53% 17.35 (1.21% 50.00% 100.67) UK 4.79) 3.77) 3.00% 100.48% 8.62) 3. Table A.65) 1.30% 50. cross-licensing.15 (1.70) 2.00% 100.75% 2.62) 2.11 describes the importance of the six motives for patenting across the countries involved in the survey.50) 2.43% 6.28% 3.28% 6.47) 3.00% 10.41) 3.66 (1. 110 .62% 5.44% 27.75% 24.29) 2.02% Licensing & Use 50. Internal use 1 country More than 1 country Total 50.53% 2.03% 2.80% 55.73) 2. We distinguished among six motives for patenting: commercial exploitation.33) 1.71 (1.50% 5.86) 2.94% 2.00% Number of observations = 7711 The importance of different motives for patenting Tables A.00% 100.79 (1. 11 The importance of different motives for patenting by country DE Commercial exploitation Licensing Cross-licensing Prevention from imitation Blocking patents Reputation 3.69% Sleeping Patents 17.99 (1.64 (1.47) 1.70 (1.69% Sleeping Patents 17.38% 8. blocking patents.53% 18.50% 3.56) 3. Table A.24 (1.38% Crosslicensing 2.45 (1.41) 2.36% 28. licensing.13 looks at the importance of different motives for patenting when the applicants of the patent are located in one country as compared to applicants located in more than one country.51% Licensing 6.83% 15.56% 41.37 (1.80) 3.75) 1.60) 3.

68) 3.42) 1.61 (1.80 (1.06 (1.15) 1.45) 3.58 (1.56) 3. Pumps & Turbines Mechanical Elements Handling & Printing Agricultural & Food Processing.91 (1.49) 1.68 (1.66 (1.63 (1.88 (1.97 (1.62) 1.41 (1.28 (1.50 (1.89 (1. 12 The importance of different motives for patenting by micro technological class ISI Technological Classes Electrical Devices.70) 2.66) 2.57 (1.20 (1.95 (1.57) 3.65) 2.52) 3.92 (1.55) 2.50 (1.66 (1.56 (1.74 (1.55) 1.56) Total Note: standard deviations in parenthesis.35) 3.64) 3.15) 1. Building & Mining Commercial exploitation 3.39) 1.91 (1.52) 3.41 (1.55) 2.56) 2.81) 2.54 (1.37 (1.48) 3.61) 2.33 (1.26) 1.73) 3.74 (1.68 (1.43 (1.87 (1.82 (1.99 (1.97 (1.60) 3.67 (1.10) 1.47) 1.49) 2.51 (1.31) 1.32) 1.77 (1.68) 2.15 (1.88 (1.81 (1.66) 3.61) 3. & Control Technology Medical Technology Organic Fine Chemistry Macromolecular Chemistry & Polymers Pharmaceuticals & Cosmetics Biotechnology Materials & Metallurgy Agriculture.09 (1.64) 2.88 (1.64) 3.64) 2.02 (1.48) 1.50 (1.66) 2.66) 3.98 (1.09 (1.07 (1.69) 3.70) 2.02 (1.65) 3.63) 3.45) 3.67 (1. Measurement.86 (1.89 (1.43) 1.55) 2.27 (1.61 (1.33 (1.96 (1.41 (1.90) 2.48) 3.76) 3.73) 2.22) 1.69) 3.80) 2.20 (1.75 (1.01 (1.61 (1.89 (1.28) 1.16 (1.47 (1.51) 2.22 (2.57) 2.77 (1.44) 3.71) 3.10 (1.59) 3.43 (1.67 (1.69) 2.31) 3.57) 3.95 (1..38) 1.90 (1.58) 4.31) 1.05 (1.63) 3.47) 2.08 (1.63) 2.38 (1.58) 3.67 (1.07 (1.63) 3.40 (1.55) 3.94 (1. & Electrical Energy Audio-visual Technology Telecommunications Information Technology Semiconductors Optics Analysis.30 (1.26) 1.57) 2.00) 4.05 (1.74 (1.62) 2.69 (1.73 (1.29 (1.56) 2.72) 2.86 (1.55) 3.62) 3.89 (1.25 (1.66) 3.05 (1.50 (1. Electrical Eng.69) 3.54 (1.36 (1.63) 3.50) 3.61) 2.67) 2.00 (1.51 (1.05 (1.07 (1.76) 3.16 (1.67) 3.91 (1.25 (1.47) 3.61 (1.26 (1.47 (1.40 (1.95 (1.60) 2.09 (1.54) 2.67) 2.53) 2.11 (1.83 (1.32 (1.66) 2.14 (1.44) 3. 111 .01 (1.73) 3.64 (1.63) 2.74) 3.51) 3.75 (1.81 (1.35) 3.64 (1.49) 3.69) 3.41) 1.75 (1.69) 3.53) 3.70 (1.95 (1.60) Blocking patents 3.06 (1.73 (1.51) 2.59) 2.78 (1.03 (1.31 (1.68) 3.16 (1.08 (1.30) 1.75 (1.05 (1.97 (1.19 (1.47 (1.52) 2.54) Crosslicensing 2.33 (1.63) 1.68) 2.57 (1.66) 3.50) 1.70) 2.92 (1.47) 3.29 (1.45) 3.54) 1.01 (1.29 (1.52) 2.14 (1.81 (1.87 (1.42) 2.47) 2.63 (1.33 (1.70) 3. Basic Materials Chemistry Chemical Engineering Surface Technology & Coating Materials Processing.05 (1.77 (1.86 (1.58) 2.43) 2.84 (1.44 (1.51) 2.88 (1.70) Reputation 2.54) 2.46) 2.78 (1.76) 3.23 (1.65) 3.26 (1.36) 2.63) 2.57) 4.53) 2.06 (1.66 (1.22) 1.78) 3.52) 2.93 (1.02 (1.45) 2. Textiles & Paper Thermal Processes & Apparatus Environmental Technology Machine Tools Engines.58) 2.62) 3.58 (1.38 (1.35) 3.Table A.88) 2.56) Licensing 2.32) 1.79 (1.82) 3.51 (1.54) 1.75) 3. & Food Chemistry Chemical & Petrol Industry.61) 4.64) 3.14 (1.35 (1.68) 2.70) 3.94 (1. Machinery & Apparatus Transport Nuclear Engineering Space Technology Weapons Consumer Goods & Equipment Civil Eng.53 (1.50) 2.79 (1.63) 3.82 (1.75 (1.52) 3.17 (1.31) 1.69) 1.63) 3.41 (1.70) 3.58) 3.63) 2.52) 3.08 (1.80 (1.87 (1.59) 1.48) 2.24) 1.61) 2.51) 2.75) 2.55) 3.58) 2.32) 1.44) 3.54) 3.67 (1.41) 3.70) 1.60) 2.44) Prevention from imitation 3.02 (1.75 (1.35 (1.36 (1.53) 3.61 (1.24) 1.76 (1.68 (1.31) 1.65 (1.76 (1.08 (1.54) 1.84 (1.40) 1.35 (1.58 (1.72) 4.29 (1.18 (1.29 (1.20 (1.17 (1.

76) Non-domestic co-applicant 3.70) 2.06 (1.42) 2.69) 3.3 and A.13 (1.77 (1. 3.71 (1.66) 2.54) 1.85 (1.77 (1.18 (1.56 (1.76 (1.78 (1.06 (1.18 (1.21 (1.59) 1.70) 2.25 (1.87 (1.44) 3. 3.62) 2.56) Table A.71) 2.28 (1.78 (1.43) 3.54) 1.56) 2.76 (1.18 (1.44) 3. Number of countries in which the applicants are located 1 country Commercial exploitation Licensing Cross-licensing Prevention from imitation Blocking patents Reputation Note: standard deviations in parenthesis.59) 2.09 (1.90 (1.75 (1.01 (1.00 (1.4 show the share of patents used to create a new venture by country and nationality of the applicants.93) 3.60) 3. and distinguishes between individual inventors who applied for the patent.06 (1. Table A. 112 .62) 2.1 shows the share of new firms created from the patented inventions.70) 2.93 (1.62) Non-EU & nonUSA applicant 4.31 (1.77 (1.42 (1.77 (1.07 (1.56) A III.04 (1.49) 3.2 looks at the share of patents used to create a new firm when the applicants of the patent are located in one country as compared to applicants located in more than one country.28 (1.64) 3. Figure A.50) 1.58) 2.60) 3.56) More than 1 country 3.25 (1.60) 3.58) 2.78 (1.56) 2.65) 2.78 (1.01 (1.70) 2.68) 3.00 (1.58) Total 3.51 (1. Figure A. 13 The importance of different motives for patenting.26 (1.59) 3.48) 3.79) 3.54) 1.00 (1.06 (1.59) USA applicant 4.46 (1.3 The creation of new businesses from the patented inventions This part of Annex III adds some statistics on the creation of new businesses from the patented inventions.41) 3.62) 2.57) 2.57) 1.03 (1.78 (1.54) 1.Table A.85 (1.23 (1.68) 3.15 shows the share of new firms created from the patented inventions by NUTS2 European region for which this share is below the EU6 average. Figures A.25 (1.51) 2. and applicant organisations (one or more than one).68) 3.15 (1.57) 2. 14 The importance of different motives for patenting by nationality of the applicants Domestic applicant Commercial exploitation Licensing Cross-licensing Prevention from imitation Blocking patents Reputation Note: standard deviations in parenthesis.76 (1.76 (1.90 (1.49) 2.58) Total 3.56 (1.65) 2.96 (1.55) European applicant 3.

04% 2.13% Outer London Overijssel Utrecht Piemonte Tübingen Herefordshire.04% 1.62% 4.Table A.30% 2.82% 1.00% 3.19% 1.07% 4.56% 2.22% 2.64% 3.50% 2.56% 2.96% 1.50% 4.88% 4.00% 0.96% 3.86% 2.11% 4.00% N 7394 41 62 65 195 111 48 73 123 50 101 55 114 182 312 35 73 295 334 39 195 80 45 180 49 49 51 273 55 335 82 84 91 30 47 Total UK NL NL IT DE UK IT DE UK FR IT NL NL DE IT DE DE NL DE DE DE DE DE DE UK DE DE DE DE DE FR DE FR DE Note: This Table includes the European regions in which the share of new firms from patented inventions is below the EU6 average and the number of observation in each region is ≥30.69% 2.10% 0.88% 2.51% 3.84% 4. 113 . Worcestershire and Warks Friuli-Venezia Giulia Mittelfranken Cheshire Île de France Toscana Limburg Zuid-Holland Oberbayern Lazio Schwaben Stuttgart Noord-Brabant Detmold Köln Hannover Oberpfalz Rheinhessen-Pfalz Unterfranken Kent Berlin Düsseldorf Münster Darmstadt Freiburg Rhône-Alpes Arnsberg Alsace Hamburg 4.71% 2.22% 1.49% 1.62% 4.17% 4.74% 2.22% 2. 15 Share of new firms from patented inventions by European regions (NUTS2) Country NUTS2 % New firms 5.83% 1.

13% More than 1 country 2.Figure A.00% 6. Number of countries in which the applicants are located Total 5.24% 0. 2 Share of new firm formation.00% 2.91% Individual inventors 4.00% Figure A.26% 1 country 5. organisations Total 5.00% 10.00% 5.00% 5.91% Multiple organisations 2.00% 20.00% 1.00% 25.00% 15.13% Single organisation 22.00% 3.00% 4.28% 0. 1 Share of new firms created by type of applicant: individual inventors vs.00% 114 .

00% 6.00% 4.26% 0.00% 3.00% 1.63% 5.00% 115 .00% 2.69% 4.00% 7.75% 5.00% 4.27% 6.00% 2.38% 1.72% 3.00% 9. 3 Share of new firms by country of the applicants Total UK NL IT FR ES DE 0.52% Co-assApplGroup NonCo-ass/Ind 5.00% 10.00% 8. 4 Share of new firms by nationality of the applicants Total 5.00% 5.13% Co-assIndepAppl 4.00% 2.00% Figure A.00% 1.00% 1.13% 9.97% 9.Figure A.00% 5.

small firms.63% Patents developed with external coinventors 14. Table A. 16 Affiliation of the inventors involved in the development of the patents.18 conditions on the micro technological class in which the patent is classified.Inventors”). third.00% 100. Finally.4 Collaboration. it describes the sources of knowledge used to develop a patent. Table A. government laboratories. as compared to the share of patents developed by a team of multiple inventors. Table A. Table A.20 gives these shares by micro technological classes.16 shows the share of patents produced by single inventors.22 describes their European and nonEuropean origins.17 describes the affiliation of the inventors involved in the development of a patent.69% 15.78% 43. it highlights the role of geographical proximity in fostering knowledge exchange among the inventors. (ii) applied by multiple applicants all belonging to the same corporate group. Finally. In Tables A. “Internal Co-Inventors” and “External Co. then.19 shows the share of these three categories of patents by macro technological classes.00% 100. Table A.26% 47. medium-sized firms.21 looks at the share of patents in the three categories when the applicants are located in one country as compared to the case in which the applicants are located in more than one country.e. and calculates the share of patents invented by single inventors. and other employers like the inventors themselves or other institutions).05% 37. Number of applicant countries Patents developed with internal coinventors 47. Table A. Spillovers and the Sources of Knowledge in the Invention Process The Tables in this section adds evidence on the importance and the type of collaborative links and knowledge spillovers that arose during the invention process leading to the PatVal-EU patents.36% Total 100.00% Share of Individual inventors’ patents 37. private non-profit research institutions. public research organisations.84% 23. patents applied by multiple applicants belonging to the same corporate group. and patents applied by independent applicants by type of inventors’ employer (i. while Table A. it looks at the collaboration among individual inventors. universities. Consistently with the structure of the Report. it goes into the type of collaborative links that the PatVal-EU inventors established with other partners to develop the innovation. Table A. large firms.23 shows the distribution of patents applied by single applicant.A III. By conditioning on the number of countries where the patent applicant was located.38% 33.Inventors” in each micro technological class. Collaboration among inventors in the innovation process Individual inventors often collaborate in teams of researchers to develop an innovation. either located within the same organization (“Internal Co-Inventors” or different organizations (“External Co.23 we classify the patents as: (i) applied by single applicant or by an individual inventor.19 to A. Table A. (iii) applied by multiple applicants that do not belong to the same corporate group.00% 1 country More than 1 country Total 116 .

77% 8.51% 20.46% 51.73% 14.05% 46.08% 46.64% 42.39% 41.88% 13. 18 Collaboration among inventors.09% 54.00% 100.49% 14.54% 22.19% 39.00% 100.00% 100.23% 38.00% 100. & Food Chemistry Chemical & Petrol Industry.37% Total 100.00% Share of Individual inventors’ patents 37.00% 100.09% 13.79% 37.22% 56.00% 117 .00% 100.45% 14.75% 61.32% 27.01% 47.36% 14.00% 100.38% 55. Distribution by “micro” technological class.18% 34.48% 47.00% Domestic applicant European applicant USA applicant Non-EU & non-USA applicant Non-domestic co-applicant Total Table A.00% 100.77% 51.25% 17.00% 100.30% 16. Textiles & Paper Thermal Processes & Apparatus Environmental Technology Machine Tools Engines.71% 20.50% 49. Pumps & Turbines Mechanical Elements Handling & Printing Agricultural & Food Processing.Table A.00% 100.34% 22. Patents developed with internal coinventors 47.00% 100.52% 38.. 17 Affiliation of the inventors involved in the development of the patents.42% 15.00% 100.18% 73.00% 100.67% 32.82% 36.00% 100.88% 27.00% 100.18% 14.66% 15.00% 49.56% 43.86% 27. foreign applicants.93% 37.02% 38.00% 100.00% 100. Basic Materials Chemistry Chemical Engineering Surface Technology & Coating Materials Processing.92% 49.79% 47.99% 12.73% 70. & Electrical Energy Audio-visual Technology Telecommunications Information Technology Semiconductors Optics Analysis. & Control Technology Medical Technology Organic Fine Chemistry Macromolecular Chemistry & Polymers Pharmaceuticals & Cosmetics Biotechnology Materials & Metallurgy Agriculture.87% 36.00% 100.00% 100.55% 36. Building & Mining Total Patents developed with internal coinventors 47.90% 47.00% 100.01% 41.55% 30.00% 100.78% 44.54% 57.04% 13.00% 100.63% 3.11% 10.62% 10.00% Share of Individual inventors’ patents 40. Measurement.00% 100.11% 34. Machinery & Apparatus Transport Nuclear Engineering Space Technology Weapons Consumer Goods & Equipment Civil Eng.63% Patents developed with external coinventors 12.55% 7.00% 100.63% Patents developed with external coinventors 14.00% 100.81% 28.14% 9.15% 12.00% 100. Electrical Eng.27% 18.31% 15.08% 30.46% 40.82% 52.00% 100.50% 10.79% 24.48% 39.23% 59.04% 38.00% 100.60% 11.59% 15.36% 9. Domestic vs.76% 47.77% 33.76% 18.83% 22.00% 100.11% 39.56% 12.64% 11.21% 14.18% 23.35% 43.00% 100.83% 47.00% 100.03% 40.82% 46.37% Total 100.88% 44.00% 100. ISI Technological Classes Electrical Devices.58% 46.35% 42.26% 59.00% 100.34% 58.00% 100.00% 100.16% 21.18% 54.00% 34.14% 50.56% 37.05% 34.14% 15.41% 52.41% 30.69% 55.

30% 96. 3.00% 100.46% 91.00% 100.62% 6. Distribution by “micro” technological class ISI Technological Classes Electrical Devices.00% 100.00% 2.60% 4.23% 3.20% 91.86% Multiple Appl.74% 2.35% 1.20% 1.61% 94. & Electrical Energy Audio-visual Technology Telecommunications Information Technology Semiconductors Optics Analysis.53% 1.32% 2.00% 93.73% 1. Group 2.00% 100.00% 2. Electrical Eng.33% 91.59% 2.00% 100.Table A.92% 1.22% 9.47% 6.00% 100.30% 11.00% 3.56% 3.86% 2.00% 100.00% 100.92% 12.67% 2. & Food Chemistry Chemical & Petrol Industry.61% 3.55% Multiple Appl.00% 118 .00% 100.05% 4.76% 2.00% 100.19% 2.30% 1.73% 6.03% 94.17% 6.29% 0. Textiles & Paper Thermal Processes & Apparatus Environmental Technology Machine Tools Engines.75% 1.16% 94.47% 0.50% 2.59% 97.87% 86.00% 100.00% 100.91% 4.78% 3.45% 1.76% 94.10% 95. Group 3.00% Electrical Engineering Instruments Chemical & Pharmaceutical tech. Pumps & Turbines Mechanical Elements Handling & Printing Agricultural & Food Processing.12% 93.00% 100.00% 100.71% 95.00% 100. Building & Mining Total Single Applicant 95.00% 100.00% 100.53% 96.48% 3.44% Multiple Appl.28% 95.87% 4.28% 89.00% 100.00% 100.66% Total 100.00% 1.86% 96.53% 94.00% 100. 2. & Control Technology Medical Technology Organic Fine Chemistry Macromolecular Chemistry & Polymers Pharmaceuticals & Cosmetics Biotechnology Materials & Metallurgy Agriculture.00% 100.00% 1.00% 100.00% 100.85% 96.92% 7.58% 80.00% 100.71% 90.37% 91.71% 5.00% 100.00% 0.00% 100.70% 0.02% 3.41% 97.49% 95. Machinery & Apparatus Transport Nuclear Engineering Space Technology Weapons Consumer Goods & Equipment Civil Eng. Distribution by “macro” technological class Single Applicant 93..58% 92.27% 93.73% 94.00% 100.48% 2. Measurement.79% 6.98% 5.15% 0.51% 1.79% 94.00% 100.21% 4.38% 6. Indep.00% 100.00% 100. Process Engineering Mechanical Engineering Total Table A. 20 Number of applicant institutions in a patent.98% 2.65% 3.18% 0.23% 7.94% 0.11% 2.62% 0.86% Multiple Appl.04% 95.00% 100.58% Total 100.83% 4.11% 5. Basic Materials Chemistry Chemical Engineering Surface Technology & Coating Materials Processing.00% 100.30% 3.85% 4.98% 4.42% 3.35% 2.82% 0.00% 100.15% 0.26% 2.28% 92. Indep.13% 93.34% 92.76% 0.86% 88.00% 100.77% 3.00% 100. 19 Number of applicant institutions in a patent.26% 92.

39% 3.86% Multiple Appl.46% 13.00% 0.00% 3.27 look at formal vs. large companies.24 shows the share of collaborative patents developed in each micro technological class.49% 0.78% 3. Table A. while Table A. 2.41% 3.69% 93.53% 2.e. informal collaborations.Table A. Group 0.29% 95.63% 0.76% 95. and Table A.41% 0.56% 34.45% 64.00% 100. Table A.19% 93.00% 100. 23 Number of applicant institutions in a patent. medium sized companies. 3.30% 3.28 to A.25 gives these shares by type of inventors’ employer.26 shows the extent of formal and informal collaborations among institutions when the applicants are located in one country as compared to the case in which the applicants are located in more than one country.00% 100.80% 2.00% 100.00% 1 country More than 1 country Total Table A.00% 100.60% Total 100.39% 99.61% 0. Indep.75% 1.55% Multiple Appl.00% 100. foreign applicants Single Applicant 96.66% 100.00% 100.00% 94.00% 1. Group 3.41% 1.00% 100.00% Large companies Medium sized companies Small companies Private Research Institutions Public Research Institutions Universities Other Governm.24 to A. and indicates the distribution of such collaborations between formal and informal collaborations.58% Total 100. Indep. small companies. Universities.74% 2.00% 100. 2.00% 100.99% 1.79% 93.00% 100. 21 Number of applicant institutions in a patent. Number of applicant countries Single applicant 96. Distribution by type of inventors’ employer Single Applicant 93.00% 63.27 describes the European and non-European origins of the applicants who collaborate to develop a patent.86% Multiple Appl.85% Multiple Appl.55% Multiple Appl.74% 96. 22 Number of applicant institutions in a patent.04% 86.00% 100.56% 33. Domestic vs.30% 2.52% 0. Indep.00% 0.70% 97. Table A.00% 0.28 shows the share of patents invented in collaboration with different external partners when the applicants are located in one country as compared to the case in which the applicants are located in more than one country.30 focus on formal collaborations and classify the patents according to the type of organization to which the external inventors involved in the collaboration are affiliated (i. private research institutions. Tables A.49% 3.70% 91. and “others” as a residual category).70% 0. Table A. Institutions Others Total Formal and informal collaborations in research This section adds empirical evidence on the extent of collaboration among different institutions to develop a patent.55% Multiple Appl.89% 7.00% 100.69% 2. Tables A.29 describes the distribution of collaborations within different type of partners by the European 119 .58% Total 100.00% Domestic applicant European applicant USA applicant Non-EU & non-USA applicant Non-domestic co-applicant Total Table A.44% 1.81% 5.92% 93.00% 100.00% 100. Group 0.

00% 9. Basic Materials Chemistry Chemical Engineering Surface Technology & Coating Materials Processing.69% 23.75% 89.87% 79.97% 23.19% 81.55% 35.00% 80.00% 79.24% 20. & Control Technology Medical Technology Organic Fine Chemistry Macromolecular Chemistry & Polymers Pharmaceuticals & Cosmetics Biotechnology Materials & Metallurgy Agriculture.12% 23.25% 72.67% 30.85% 14.44% 28.63% 83.70% 73.89% 84.00% 19.87% 23.45% 19.77% 15.34% 20.30% 20. 25 Formal vs.35% 22.61% 21.55% 20.94% Number of observations: 8314 120 .28% 40.46% 39. Electrical Eng.13% 29.70% 78. Table A.88% 21.36% 24.57% 76.29% 21.61% 23. Pumps & Turbines Mechanical Elements Handling & Printing Agricultural & Food Processing.91% 77.31% 76.30% 21. Machinery & Apparatus Transport Nuclear Engineering Space Technology Weapons Consumer Goods & Equipment Civil Eng.81% 18.35% 27. Measurement.00% 90.00% 61.21% 18.50% 86.92% 38.87% 23.62% 25.01% 87.73% 23.98% 48.22% 27. Finally.33% 0.07% 17.08% 18. Informal collaboration amongst institutions.12% Number of observations: 8499.49% 83.34% 18.53% 20.67% 100. Building & Mining Total Share of “collaborative” patents 17.53% Share of formal collaborations 76.67% 14. Distribution by “micro” technological class ISI Technological Classes Electrical Devices.18% 23.61% 25.78% 72.21% 0.39% 74.60% 20.68% 18.38% 74.13% 76.95% 27.08% 23.75% 27.82% 16..48% 71.08% 18.79% 100.25% 10.47% 24.43% 24.85% 17.17% 69.95% 5.16% Share of formal collaborations 75.00% 38.09% 24.73% 67. Distribution by type of employer Share of “collaborative” patents Large companies Medium sized companies Small companies Private Research Institutions Public Research Institutions Universities Other Governm.27% 35.09% 22.57% 11.11% 15.27% 32.83% 30.70% 79.50% 13.06% Share of informal collaborations 24.27% 35.52% 28.12% 78. Textiles & Paper Thermal Processes & Apparatus Environmental Technology Machine Tools Engines. & Electrical Energy Audio-visual Technology Telecommunications Information Technology Semiconductors Optics Analysis.35% 29.43% 88. Informal collaboration amongst institutions.65% 72.73% 64. 24 Formal vs. Table A.30% 26.55% 56.14% 17. Institutions Others Total 16.33% 69.39% 76.99% 12.05% 94.05% 72.65% 77.00% 20.72% 60. Table A.88% Share of informal collaborations 23.and non-European origins of the patent applicant.37% 16.51% 16.13% 20. & Food Chemistry Chemical & Petrol Industry.30 gives the same distribution by conditioning on the micro technological class in which the patents are classified.

06% 8. foreign applicants.36% 6.30% 62.88% Share of informal collaborations 23.38% 0. 27 Formal vs.00% 14.45% 84. Large companies Domestic applicant European applicant USA applicant Non-EU & non-USA applicant Non-domestic co-applicant 48.67% 0.74% 1.59% 16.77% 6.08% Affiliation of the “external” co-inventors Small Private Res.67% 0.12% Number of observations: 8498 Table A. Universities companies Institutions 20.85% 20.70% 37.55% 15. Large companies 49.88% 10.00% 5.91% 38.58% “Others” 8.52% 0.89% 6.00% 0. Number of applicant countries.66% 1 country More than 1 country Total Number of observations: 1213 Table A.50% 83.82% 8.47% 49.79% Medium sized companies 7.00% 5.22% 16.36% 8. Affiliation of the co-inventors involved in the development of the patents.77% 16.16% 20.67% 16.15% 23.08% 39.51% 76.50% 16.22% 30.49% 23.33% 2. 29 Formal collaborations. Affiliation of the co-inventors involved in the development of the patents.33% 64.82% 15.64% 20.44% 53.80% 6.53% Share of formal collaborations 76. 28 Formal collaborations.12% 20.52% 50.27% 13. Domestic vs. Share of “collaborative” patents 1 country More than 1 country Total 19.00% 56.00% 7.65% 22. Universities companies Institutions 21.52% Share of formal collaborations 76.67% 9. Share of “collaborative” patents Domestic applicant European applicant USA applicant Non-EU & non-USA applicant Non-domestic co-applicant Total 19.00% 23.58% “Others” 8.06% 15.21% 16. Informal collaboration amongst institutions.66% Total Number of observations: 1213 121 .78% 69.07% 59. Domestic vs.00% 0. Number of applicant countries.00% 7.00% 9.51% 22.00% 11.76% 8.12% Number of observations: 8499 Table A.79% Medium sized companies 8. 26 Formal vs.29% 50.09% 49.41% 84.89% 6.36% 0.88% Share of informal collaborations 23. Informal collaboration amongst institutions.01% 8.Table A.82% 11.00% 76.08% Affiliation of the “external” co-inventors Small Private Res. foreign applicants.

00% 8.53% 23.67% 33. & Control Technology Medical Technology Organic Fine Chemistry Macromolecular Chemistry & Polymers Pharmaceuticals & Cosmetics Biotechnology Materials & Metallurgy Agriculture.14% 16.00% 34.00% 0.11% 62.60% 5.76% 17.81% 6.Table A.53% 4.23% 0.81% 58.00% 8.50% 47.00% 0.34% 16.11% 6.52% 6.00% 28.11% 6.00% 65.00% 100.38% 5.67% 3.00% 7.04% 13.49% 14..89% 23.11% 3.58% 6.90% 9.90% 11.00% 10. & Food Chemistry Chemical & Petrol Industry.98% 5.00% 44.45% 50.00% 2.79% 15.79% 6.49% 7.82% 60.00% 10.29% 18.54% 4.09% 9.53% 10.56% 8.40% 2.00% 36.38% 28.00% 4.00% 42.14% 1.49% 15.12% 23.82% 7.00% 0.50% 12.11% 2.50% 15.53% 6.65% 60.67% 16.71% 12.17% 11.11% 20.89% 8.14% 1.72% 20.85% 4.69% 7.53% 14.37% 53.00% 43.63% 13.00% 17.00% 12.50% 37.54% 7.00% 2.54% 20.79% Affiliation of the “external” co-inventors Medium Small Private Res. Basic Materials Chemistry Chemical Engineering Surface Technology & Coating Materials Processing.84% 10.42% 2. Textiles & Paper Thermal Processes & Apparatus Environmental Technology Machine Tools Engines.23% 17.52% 67.36% 4.08% 42.50% 0.57% 31.57% 4.06% 23.00% 12.65% 21.50% 10.33% 3.52% 3.13% 0.79% 11.00% 0.67% 59.05% 28. Electrical Eng.89% 49.05% 16.24% 8.00% 2.57% 9.44% 34.76% 34.00% 22.58% “Others” 3.00% 6.00% 0.00% 20.00% 42.28% 42.38% 10.33% 5.23% 9.86% 16.00% 28.50% 15.68% 10.53% 20.53% 5.00% 0. Affiliation of the inventors involved in the development of the patents.48% 19.86% 28.38% 2.00% 4.93% 36.41% 0.50% 5.12% 7.56% 15.69% 44.29% 0.65% 23.50% 17.00% 10.26% 9. & Electrical Energy Audio-visual Technology Telecommunications Information Technology Semiconductors Optics Analysis.86% 37.00% 18.53% 24.67% 9.33% 45.86% 42.51% 8.00% 2. Distribution by “micro” technological class in % ISI Technological Classes Electrical Devices.33% 38.86% 40.00% 14.00% 0.00% 7.00% 0.05% 7.77% 19.16% 66.00% 10.62% 62.00% 19. Building & Mining Large companies 57.66% Total 122 .52% 13.57% 19.66% 0.53% 0.88% 38.41% 5. Machinery & Apparatus Transport Nuclear Engineering Space Technology Weapons Consumer Goods & Equipment Civil Eng.68% 0. Measurement.67% 26.40% 31.13% 12.56% 5.00% 11.58% 24. 30 Formal collaborations.18% 15.81% 15. Pumps & Turbines Mechanical Elements Handling & Printing Agricultural & Food Processing.29% 0.52% 24.34% 11.10% 21.14% 42.76% 12.38% 0.00% 12. Universities sized companies Institutions companies 7.

32 shows the importance of these four types of interactions in the European regions in which the importance of close and external interactions for developing innovations is below the EU6 average. By using a scale from 1 to 5 Table A.e. and geographically close.e. more than one hour to reach physically the partner). Table A. Table A. less than one hour to reach physically the partner). (4) interactions with people external to the inventor’s organization.5 we distinguish among 4 types of interactions: (1) interactions with people internal to the inventor’s organization. (2) interactions with people internal to the inventor’s organization. Like in Section 3. and geographically distant (i. and geographically distant. (3) interactions with people external to the inventor’s organization.The role of geographical proximity for collaboration The Tables in this part of Annex III provide descriptive statistics on the role of geographical proximity for fostering collaborations among inventors.32 shows the importance of the four types of interactions in each micro technological class in which the PatValEU patents are classified. and geographically close (i. Table A.34 looks at the importance of the four forms of interaction according to the European and non-European origins of the applicants. 123 .33 describes their importance when the applicants are located in one country as compared to the case in which the applicants are located in more than one country.

48 (2.90 (1.96) 2.29 (1.66 (1.51 (1. 124 .93) 1.05 (1.87) 2.89) 3.66 (1.42 (1.09) 0.13) 0.87) 0.15 (1.70 (1.80 (1.10) 0.83 (1.61) 0.34) 1.10 (1.22) 0.31 (1.73 (1.81 (1.39 (1.31 (1.58 (1.82) 0.45) 0.26 (1.55) 1.79 1.30 (1.10) 1.14 (2.98) Distant & Internal 1.61 (1.09 (1.47) 0.79) 1.22) 0.70 (1.94) 2.95) 2.72 (1.77 (1.00 (1.22) 0.84 (1.63) 0.84) 3.77 (1.56 (0.64 (1.29 (1.28) 0.46 (1.77 (1.17 (1.03 (1.83) 2.67) 1.42 (1.22) 0.05 (2.02 (1.67 (1.00) 3.77) 1. Worcestershire and Warks Friuli-Venezia Giulia Lazio Oberpfalz Essex Lombardia Arnsberg Kent Schwaben Pais Vasco Toscana 3.94) 3.45 (1.33 (1.34 (1.75 (1.47) 0.54 (1.88) 2.72 (1.45) 2.88) Hampshire and Isle of Wight Noord-Holland Gelderland Noord-Brabant Overijssel Gießen Hannover Koblenz Karlsruhe Oberbayern Cheshire Mittelfranken West Yorkshire Freiburg Detmold Münster Köln Cataluña Limburg Hamburg Braunschweig Veneto Darmstadt Piemonte Surrey.21) 1.27) 0.54 (1.56) Close & External 0.95) 3.42 (1.88 (1.89 (1.61 (1.13 (1.57 (1.40) 0.90) 1.91) 3.70) 1.76) 1.80) 3.47 (1.95 (1.70) 2.13) 0.89) Distant & External 1.40) 1.17 (1.06 (2.82 (1.13 (1.76 (1.70) 1.07) 0.75) 0.74) 1.75 (1.83) 2. east and west Sussex Bretagne Greater Manchester Tübingen Stuttgart Rheinhessen-Pfalz Berlin Emilia-Romagna Herefordshire.27 (1.40) 0.19 (1.36) 0.52 (1.42) 0.92) 3.24) 0.57 (1.68) N 8588 52 117 129 337 62 37 86 31 143 333 51 134 57 86 44 60 213 101 114 50 30 97 352 188 90 30 78 122 336 190 64 154 44 69 33 47 41 392 96 45 83 30 50 Total UK NL NL NL NL DE DE DE DE DE UK DE UK DE DE DE DE ES NL DE DE IT DE IT UK FR UK DE DE DE DE IT UK IT IT DE UK IT DE UK DE ES IT Note: This Table includes the European regions in which the importance of Close & External interactions are below the EU6 average and the number of observation in each region is ≥30.97 (1.41 (1.58 (1.38) 0.00 (1.47) 1.25 (1.57) 1.66 (2.30) 1.04 (1.60) 1.62) 2.15) 0.30 (1.80) 1.29 (1.52 (2.03 (1.67 (1.29 (1.68) 2.87 (1.35) 0.76 (1.16 (1.47) 1.03 (1.73 (1.15 (1.83 (1.10) 0.68) 1.61 0.96) 1.76) 1.02) 0.94) 0.13 (1.70 (1.09 (1.09 (1.94) 1.57 (1.41) 0.01) 0.80 (1.98) 1.80) 2.15 (1.81 (1.64) 2.71 (1.40 (1. 31 Importance of the interaction with other people while developing the innovation.61 (1. Distribution by European regions (NUTS2) Country NUTS 2 Close & Internal 3.70) 2.32 (1.81) 1.71) 0.03 (1.68 (1.76) 3.42 (1.96) 1.42) 2.77 (1.40 (1.50 (1.42) 0.19) 1.08 (1.71) 1.88) 0.81 (1.54) 1.88) 1.78) 1.51) 0.75) 3.38) 1.79) 1.50) 1.06) 0.92 (2.61 (1.57 (1.99) 1.03) 2.33) 0.16) 0.07 (1.29 (1.64) 0.34) 0.95) 2.73 (1.87) 1.75 (1.37) 1.82) 3.02 (1.81) 3.97 (1.83 (1.32) 1.59) 2.12) 1.02) 0.35 (1.53 (1.28) 1.44 (0.79) 0.16) 0.33) 0.83 (1.42 (1.77 (1.94) 2.89) 3.33) 0.20 (1.17 (1.48) 1.82 (1.08 (1.85 (1.38 (1.08) 1.87 (1.62 (1.06) 2.65) 1.50) 1.53 (1.31 (0.25 (1.07 (1.40 (1.62) 1.63 (1.74 (1.17 (2.79) 1.63 (1.10) 0.92) 2.61) 1.95) 3.81 (1.68) 1.72 (1.76) 1.87) 2.91) 3.70 (1.55) 1.74 (1.82 (1.15 (1.67 (1.70) 2.40 (1.20 (1.68) 2.10 (2.12 (1.95) 3.58 (1.86) 3.48) 0.40 (1.78) 0.51 (1.85) 1.86 (1.02 (1.14 (1.60) 1.88) 0.62 (1.59) 1.40) 0.70) 1.94) 3.99 (1.89) 2.39) 1.70 (1.69) 2.32 (1.94) 1.95 (2.65) 1.27) 0.00) 3.08 (1.65) 1.21 (1.91) 3.62) 1.06) 0.91 (1.00 (1.80) 0.50) 0.13) 0.Table A.65 (1.11 (1.69) 1.80 (1.49 (1.00 (1.36 (1.37) 0.64 (1.91) 1.43) 1.53 (1.37) 0.27 (1.89) 2.18 (1.28) 0.16 (2.

25 (1.83) 3.87) 2.64) 1.47) 0.58) 1.78) 1.22 (1.88 (1.73 (1.84) 3.51 (1.77) Note: standard deviations in parenthesis.84) 1.49) 0.77) 0.64) 1.00 (1.73) 1.47 (1.50 (1.75 (1.39) 1.59) 1.29) 0.93 (1.64) 1.60 (1.18 (1.17 (1.47 (1.03 (1.38) 0. Measurement.84 (1.02 (1.90 (1. Distribution by “micro” technological class ISI Technological Classes Electrical Devices.06 (1.76 (1. Electrical Eng.69) 0.98) 2.78) 1.76) 1.24 (1.37 (1.16 (1.77) Total Note: standard deviations in parenthesis.19 (1.32 (1.73) 1.88) Distant Internal 1.44) 0.95) 1.99) 3.85 (1.10 (1.23 (1.76) 1.88 (1. & Electrical Energy Audio-visual Technology Telecommunications Information Technology Semiconductors Optics Analysis.81) 1.80) 1.74 (1.41) 0.25 (1.95 (1.78 (1.69) 3.39) 0.93) 2.03 (1.69) 1. Machinery & Apparatus Transport Nuclear Engineering Space Technology Weapons Consumer Goods & Equipment Civil Eng.20) 1.02 (1.77) More than 1 country 3.82) 1.95 (1.95) 2.89) 2.46) 1.28 (1.43) 0.32 (1.74) 2.86 (1.36 (1.42) 1.36 (1.45) Distant External 1.58) 1.00 (1.08 (1.85) 1.88 (1.53) 0.27 (1.73 (1.80 (1.94 (1.32 (1.70) Close External 0. Basic Materials Chemistry Chemical Engineering Surface Technology & Coating Materials Processing.83) 1.42 (1.77 (1.35 (1.31 (1.40 (1.88) 3.02 (1.33 (1.70) 0.73 (1.59) 1.82 (1. 1 country Close Internal Distant Internal Close External Distant External 3.20 (1.73) 1.69) 3..00 (1.55) 1.83 (1.78) 1.70) 2.89) 1.07 (1.88 (1.00 (1.51 (1.97 (1.74) 1.76) 1.87) 1.88) 3.87 (1. 32 Importance of the interaction with other people while developing the innovation.39 (2.93) 2.31 (1.72) 1. Building & Mining Close Internal 3.35) 0.61 (1.71) 1.51) 1.61) 1.55) 1.54) 3.64 (1.64) 1. Table A.32 (1.29 (1.45) 0.77) 1.56 (1.61) 1.61) 1.46 (1.80(1.59) 0.41 (1.78) 1.82) Total 3.65 (1.86 (1.85) 1.84 (1. & Food Chemistry Chemical & Petrol Industry.32) 0.36 (1.84) 3.02 (1.90 (1.28 (1.04 (1.56 (1.87) 1.88 (1.94) 3.35 (1. 33 Importance of the interaction with other people while developing the innovation.53 (1.61) 1.86) 3. & Control Technology Medical Technology Organic Fine Chemistry Macromolecular Chemistry & Polymers Pharmaceuticals & Cosmetics Biotechnology Materials & Metallurgy Agriculture.16 (1.50 (1.04 (1.88) 2.81) 1.39 (1.99) 3.45) 1.34) 0.80) 2.16 (1.46 (1.28 (1.76 (1.89) 2.72) 1.93) 1.70) 1.53) 0.10 (1.66) 1.72) 1.63) 1.56 (1.71) 1.27 (1. Pumps & Turbines Mechanical Elements Handling & Printing Agricultural & Food Processing.45 (1.59) 1.19 (1.31 (1.99 (1.72) 1.00 (1.33 (1.51 (1.94) 1.00) 2.40) 0.34 (1.87) 2.99) 2.73) 3.89) 1.46 (1.36 (1.97 (1.88) 1.67 (1.45) 1. Number of applicant countries.68) 1.06 (1.62) 1.27 (1.89 (1.39 (1.25 (1.30 (1.81 (1.89) 1.45 (1.72) 1.02 (1. 125 .71 (1.87 (1.70) 1.39) 1.44 (1.80) 1.45) 0.56) 0.68) 1.78) 1.47) 1.93) 3.32 (1.37) 0.94) 2.72) 0.16 (1.13 (1.77) 2.89 (1.76) 1.89) 0.66 (1.48 (1.70 (1.87) 2.26 (1.53 (1.67) 1.52) 1.17 (1.Table A.29 (1.54) 0. Textiles & Paper Thermal Processes & Apparatus Environmental Technology Machine Tools Engines.70 (1.

72) 2.90) 2.67 (1.71) 2.88 (1. Table A.71) 1.79) USA applicant 3.89) 1.40 (1.54 (1.86) 0.69 (1.35 (1.73) Non-domestic co-applicant 3. Table A.02 (1.30 shows the importance of these sources of knowledge when the patents’ applicants are located in one country as compared to the case in which they are located in more than one country.31 (1.44) 1.65 (1.98 (1.88) More than 1 country 1.15 (1. Domestic applicant Close Internal Distant Internal Close External Distant External 2.77) Note: standard deviations in parenthesis.81 (1.88) 1.86) 1.83 (1. 1 country Laboratories Scientific literature Conferences Patents Users Suppliers Competitors 1.51 (1.01 (1.56) 1.82 (1.98) 1.99 (1.85) 2.32 (1.66) 0.85) 2.59) 1.83) Total 3. 126 .31 (1.31 describes their importance according to the European and non-European origins of the patents’ applicants.05 (1. foreign applicants.22 (1. They use a scale from 1 to 5 to measure their importance.88 (1.90) 2.35 (1.e.83 (1.78 (1.68) 2.25 (1.77) European applicant 3.60 (1.54 (1.81) Total 1. 35 Importance of different sources of knowledge.89 (1. knowledge developed in university and non-university laboratories.45 (1.88 (1.89) 1.31 (1.74) 2.95) 0.88 (1.70) 0.36 look at the sources of knowledge used to develop the patented innovation (i.93) 0.73) 1.73) 2.40) 1.89 (1.76 (1.35 and Table A. Number of applicant countries.83 (1.97) 1.55 (1. Domestic vs.98) 1.72) 2.80) 1. Table A.Table A.76) Non-EU & non-USA applicant 3.89) 1.60 (1.87) Note: standard deviations in parenthesis.73) 2. suppliers and competitors).54 (1.55 (1.45) 1. the participation in conferences and workshops.45) 1.16 (1. 34 Importance of the interaction with other people while developing the innovation.74) 2. The Sources of Knowledge Table A.19 (1. earlier patents.67 (1.43) 1.23 (1.89) 1. the firm’s users.04 (1. the scientific literature.

55 (1.15 (1.89 (1.80) Total 1.87) Note: standard deviations in parentheses.55 (1.74 (1.79 (1.67 (1.54 (1.77) 3.65 (1.55 (1. Domestic vs.03) 2. 36 Importance of different sources of knowledge.70) 1.33 (1.90) USA applicant 1.75) 2.74) 2.04 (1.67 (1.71) 2.73) 2.71) 2.59 (1. Domestic applicant Laboratories Scientific literature Conferences Patents Users Suppliers Competitors 1.68 (1.98) 1.69 (1.50 (1.71) 1.90) 2.61 (1.93) 2.78) 1.83) 2.77) Non-domestic co-applicant 1.89) 1.32 (1.81 (1.87) European applicant 1. 127 .98) 1.Table A.36 (1.72) 2.90) 2.94) 2.04 (1.60 (1.49 (1.98) 1.88 (1.23 (1.01) 1.88 (1.68) 2.65 (1.65) 2.96) 1.89) 1.79 (1.90) 2.72) 2.11 (1.23 (1.73) 2.87) 1.93 (1.80 (1.97) 1. foreign applicants.18 (1.23 (2.49 (1.77 (1.05 (1.87) 1.58 (1.73) 2.68) 2.00 (2.87 (1.93) 3.73) 1.35 (1.73) 2.85) Non-EU & non-USA applicant 1.

Sign up to vote on this title
UsefulNot useful